Add support for HTTPS stagers

Merge pull request #916 from BishopFox/go/v1.19
Go/v1.19
2026-11-07 17:08:13 -10:00 · 2022-09-28 07:36:27 -07:00 · 2022-09-26 20:04:50 -07:00 · 2022-09-26 18:40:29 -07:00 · 2022-09-26 18:31:46 -07:00 · 2022-09-26 18:00:10 -07:00
863 changed files with 241718 additions and 20598 deletions
@@ -6,57 +6,55 @@ on:
    branches: [master, goreleaser]

 jobs:
-
  linux-windows-build:
    name: Linux & Windows Build
    if: startsWith( github.ref, 'refs/tags/v')
    runs-on: ubuntu-latest
    timeout-minutes: 45
    steps:
+      - name: Go 1.19
+        uses: actions/setup-go@v3
+        with:
+          go-version: "1.19.1"

-    - name: Go 1.18
-      uses: actions/setup-go@v3
-      with:
-        go-version: '1.18.6'
-    
-    - name: OS Packages 
-      run: |
-        sudo apt-get update --fix-missing && sudo apt-get -y install \
-        git build-essential zlib1g zlib1g-dev wget zip unzip \
-        mingw-w64 binutils-mingw-w64 g++-mingw-w64 gcc-multilib
+      - name: OS Packages
+        run: |
+          sudo apt-get update --fix-missing && sudo apt-get -y install \
+          git build-essential zlib1g zlib1g-dev wget zip unzip \
+          mingw-w64 binutils-mingw-w64 g++-mingw-w64 gcc-multilib

-    - name: Check out code into the Go module directory
-      uses: actions/checkout@v3
+      - name: Check out code into the Go module directory
+        uses: actions/checkout@v3

-    - name: Git Fetch Tags
-      run: git fetch --prune --unshallow --tags -f
+      - name: Git Fetch Tags
+        run: git fetch --prune --unshallow --tags -f

-    - name: Go Assets
-      run: ./go-assets.sh
+      - name: Go Assets
+        run: ./go-assets.sh

-    - name: Make
-      env:
-        ARTIFACT_SUFFIX: _linux
-      run: export PATH=/home/runner/go/bin/:$PATH && make linux
-  
-    - name: Linux Artifacts
-      uses: actions/upload-artifact@v3
-      with:
-        name: linux
-        path: ./sliver-*
+      - name: Make
+        env:
+          ARTIFACT_SUFFIX: _linux
+        run: export PATH=/home/runner/go/bin/:$PATH && make linux

-    - name: Make Windows
-      env: 
-        ARTIFACT_SUFFIX: _windows
-      run: |
-        make clean
-        export PATH=/home/runner/go/bin/:$PATH && make windows
+      - name: Linux Artifacts
+        uses: actions/upload-artifact@v3
+        with:
+          name: linux
+          path: ./sliver-*

-    - name: Windows Artifacts
-      uses: actions/upload-artifact@v3
-      with:
-        name: windows
-        path:  ./sliver-*
+      - name: Make Windows
+        env:
+          ARTIFACT_SUFFIX: _windows
+        run: |
+          make clean
+          export PATH=/home/runner/go/bin/:$PATH && make windows
+
+      - name: Windows Artifacts
+        uses: actions/upload-artifact@v3
+        with:
+          name: windows
+          path: ./sliver-*

  macos-build:
    name: MacOS Build
@@ -64,42 +62,40 @@ jobs:
    runs-on: macos-latest
    timeout-minutes: 45
    steps:
+      - name: Go 1.19
+        uses: actions/setup-go@v3
+        with:
+          go-version: "1.19.1"

-    - name: Go 1.18
-      uses: actions/setup-go@v3
-      with:
-        go-version: '1.18.6'
+      - name: Mingw
+        run: brew install mingw-w64

-    - name: Mingw
-      run: brew install mingw-w64
+      - name: Check out code into the Go module directory
+        uses: actions/checkout@v3

-    - name: Check out code into the Go module directory
-      uses: actions/checkout@v3
+      - name: Git Fetch Tags
+        run: git fetch --prune --unshallow --tags -f

-    - name: Git Fetch Tags
-      run: git fetch --prune --unshallow --tags -f
+      - name: Go Assets
+        run: ./go-assets.sh

-    - name: Go Assets
-      run: ./go-assets.sh
+      - name: Make AMD64
+        env:
+          ARTIFACT_SUFFIX: _macos
+        run: make macos

-    - name: Make AMD64
-      env: 
-        ARTIFACT_SUFFIX: _macos
-      run: make macos
+      - name: Make ARM64
+        env:
+          ARTIFACT_SUFFIX: _macos-arm64
+        run: make macos-arm64

-    - name: Make ARM64
-      env: 
-        ARTIFACT_SUFFIX: _macos-arm64
-      run: make macos-arm64
-
-    - name: MacOS Artifacts
-      uses: actions/upload-artifact@v3
-      with:
-        name: macos
-        path: ./sliver-*
+      - name: MacOS Artifacts
+        uses: actions/upload-artifact@v3
+        with:
+          name: macos
+          path: ./sliver-*

  tagged-release:
-
    needs: [linux-windows-build, macos-build]

    name: "Tagged Release"
@@ -108,53 +104,52 @@ jobs:
    timeout-minutes: 45

    steps:
-  
-    - run: mkdir -p ./builds
-    - uses: actions/download-artifact@v3
-      with:
-        path: ./builds
+      - run: mkdir -p ./builds
+      - uses: actions/download-artifact@v3
+        with:
+          path: ./builds

-    - name: Extract Artifacts
-      run: |
-        mkdir -p ./artifacts
-        cp ./builds/linux/* ./artifacts
-        cp ./builds/windows/* ./artifacts
-        cp ./builds/macos/* ./artifacts
+      - name: Extract Artifacts
+        run: |
+          mkdir -p ./artifacts
+          cp ./builds/linux/* ./artifacts
+          cp ./builds/windows/* ./artifacts
+          cp ./builds/macos/* ./artifacts

-    - id: install-secret-key
-      name: GPG Key(s)
-      run: |
-        cat <(echo -e "${{ secrets.SLIVER_GPG }}") | gpg --batch --import
-        gpg --list-secret-keys --keyid-format LONG
+      - id: install-secret-key
+        name: GPG Key(s)
+        run: |
+          cat <(echo -e "${{ secrets.SLIVER_GPG }}") | gpg --batch --import
+          gpg --list-secret-keys --keyid-format LONG

-    # Linux
-    - name: Linux AMD64
-      run: |
-        gpg --default-key 4449039C --output ./artifacts/sliver-server_linux.sig --detach-sig ./artifacts/sliver-server_linux
-        gpg --default-key 4449039C --output ./artifacts/sliver-client_linux.sig --detach-sig ./artifacts/sliver-client_linux
-    
-    # Windows
-    - name: Windows AMD64
-      run: |
-        gpg --default-key 4449039C --output ./artifacts/sliver-server_windows.sig --detach-sig ./artifacts/sliver-server_windows.exe
-        gpg --default-key 4449039C --output ./artifacts/sliver-client_windows.sig --detach-sig ./artifacts/sliver-client_windows.exe
+      # Linux
+      - name: Linux AMD64
+        run: |
+          gpg --default-key 4449039C --output ./artifacts/sliver-server_linux.sig --detach-sig ./artifacts/sliver-server_linux
+          gpg --default-key 4449039C --output ./artifacts/sliver-client_linux.sig --detach-sig ./artifacts/sliver-client_linux

-    # MacOS AMD64
-    - name: MacOS AMD64
-      run: |
-        gpg --default-key 4449039C --output ./artifacts/sliver-server_macos.sig --detach-sig ./artifacts/sliver-server_macos
-        gpg --default-key 4449039C --output ./artifacts/sliver-client_macos.sig --detach-sig ./artifacts/sliver-client_macos
-  
-    # MacOS ARM64
-    - name: Package MacOS ARM64
-      run: |
-        gpg --default-key 4449039C --output ./artifacts/sliver-server_macos-arm64.sig --detach-sig ./artifacts/sliver-server_macos-arm64
-        gpg --default-key 4449039C --output ./artifacts/sliver-client_macos-arm64.sig --detach-sig ./artifacts/sliver-client_macos-arm64
-  
-    - name: "Publish Release"
-      uses: "marvinpinto/action-automatic-releases@latest"
-      with:
-        repo_token: "${{ secrets.GITHUB_TOKEN }}"
-        prerelease: true
-        files: |
-          ./artifacts/*
+      # Windows
+      - name: Windows AMD64
+        run: |
+          gpg --default-key 4449039C --output ./artifacts/sliver-server_windows.sig --detach-sig ./artifacts/sliver-server_windows.exe
+          gpg --default-key 4449039C --output ./artifacts/sliver-client_windows.sig --detach-sig ./artifacts/sliver-client_windows.exe
+
+      # MacOS AMD64
+      - name: MacOS AMD64
+        run: |
+          gpg --default-key 4449039C --output ./artifacts/sliver-server_macos.sig --detach-sig ./artifacts/sliver-server_macos
+          gpg --default-key 4449039C --output ./artifacts/sliver-client_macos.sig --detach-sig ./artifacts/sliver-client_macos
+
+      # MacOS ARM64
+      - name: Package MacOS ARM64
+        run: |
+          gpg --default-key 4449039C --output ./artifacts/sliver-server_macos-arm64.sig --detach-sig ./artifacts/sliver-server_macos-arm64
+          gpg --default-key 4449039C --output ./artifacts/sliver-client_macos-arm64.sig --detach-sig ./artifacts/sliver-client_macos-arm64
+
+      - name: "Publish Release"
+        uses: "marvinpinto/action-automatic-releases@latest"
+        with:
+          repo_token: "${{ secrets.GITHUB_TOKEN }}"
+          prerelease: true
+          files: |
+            ./artifacts/*
@@ -17,10 +17,10 @@ jobs:

    steps:
      # Setup go environment
-      - name: Go 1.18
+      - name: Go 1.19
        uses: actions/setup-go@v3
        with:
-          go-version: "1.18.6"
+          go-version: "1.19.1"

      - name: Checkout repository
        uses: actions/checkout@v3
@@ -1,7 +1,7 @@
 name: Build & Test
 on:
  push:
-    branches: 
+    branches:
      - test/*
  pull_request:
    branches:
@@ -13,70 +13,67 @@ on:
      - ready_for_review

 jobs:
-
  linux-windows-build:
    name: Linux & Windows Build
    runs-on: ubuntu-latest
    timeout-minutes: 90
    steps:
+      - name: Go 1.19
+        uses: actions/setup-go@v3
+        with:
+          go-version: "1.19.1"

-    - name: Go 1.18
-      uses: actions/setup-go@v3
-      with:
-        go-version: '1.18.6'
-    
-    - name: OS Packages 
-      run: |
-        sudo apt-get update --fix-missing && sudo apt-get -y install \
-        git build-essential zlib1g zlib1g-dev wget zip unzip \
-        mingw-w64 binutils-mingw-w64 g++-mingw-w64 gcc-multilib
+      - name: OS Packages
+        run: |
+          sudo apt-get update --fix-missing && sudo apt-get -y install \
+          git build-essential zlib1g zlib1g-dev wget zip unzip \
+          mingw-w64 binutils-mingw-w64 g++-mingw-w64 gcc-multilib

-    - name: Check out code into the Go module directory
-      uses: actions/checkout@v3
+      - name: Check out code into the Go module directory
+        uses: actions/checkout@v3

-    - name: Git Fetch Tags
-      run: git fetch --prune --unshallow --tags -f
+      - name: Git Fetch Tags
+        run: git fetch --prune --unshallow --tags -f

-    - name: Go Assets
-      run: ./go-assets.sh
+      - name: Go Assets
+        run: ./go-assets.sh

-    - name: Make
-      run: make linux
+      - name: Make
+        run: make linux

-    - name: Test
-      run: ./sliver-server unpack --force && ./go-tests.sh
+      - name: Test
+        run: ./sliver-server unpack --force && ./go-tests.sh
+
+      - name: Make Windows
+        run: make windows

-    - name: Make Windows
-      run: make windows
-  
  macos-build:
    name: MacOS Build
    runs-on: macos-latest
    timeout-minutes: 90
    steps:
+      - name: Go 1.19
+        uses: actions/setup-go@v3
+        with:
+          go-version: "1.19.1"

-    - name: Go 1.18
-      uses: actions/setup-go@v3
-      with:
-        go-version: '1.18.6'
+      - name: Mingw
+        run: brew install mingw-w64

-    - name: Mingw
-      run: brew install mingw-w64
+      - name: Check out code into the Go module directory
+        uses: actions/checkout@v3

-    - name: Check out code into the Go module directory
-      uses: actions/checkout@v3
+      - name: Git Fetch Tags
+        run: git fetch --prune --unshallow --tags -f

-    - name: Git Fetch Tags
-      run: git fetch --prune --unshallow --tags -f
+      - name: Go Assets
+        run: ./go-assets.sh

-    - name: Go Assets
-      run: ./go-assets.sh
+      - name: Make
+        run: make macos

-    - name: Make
-      run: make macos
-
-    - name: Test
-      env:
-        SLIVER_CC_32: /usr/local/bin/i686-w64-mingw32-gcc
-        SLIVER_CC_64: /usr/local/bin/x86_64-w64-mingw32-gcc
-      run: ./sliver-server unpack --force && ./go-tests.sh
+      - name: Test
+        env:
+          SLIVER_CC_32: /usr/local/bin/i686-w64-mingw32-gcc
+          SLIVER_CC_64: /usr/local/bin/x86_64-w64-mingw32-gcc
+        run: ./sliver-server unpack --force && ./go-tests.sh
@@ -1,4 +1,4 @@
-FROM golang:1.18.6
+FROM golang:1.19.1

 #
 # IMPORTANT: This Dockerfile is used for testing, I do not recommend deploying
@@ -53,8 +53,9 @@ RUN go install google.golang.org/protobuf/cmd/protoc-gen-go@${PROTOC_GEN_GO_VER}
 # Go assets
 WORKDIR /go/src/github.com/bishopfox/sliver
 ADD . /go/src/github.com/bishopfox/sliver/
-RUN make clean-all
-RUN ./go-assets.sh
+RUN make clean-all \
+    && ./go-assets.sh
+
 RUN make \
    && cp -vv sliver-server /opt/sliver-server \
    && /opt/sliver-server unpack --force 
@@ -20,8 +20,8 @@ set -e

 # Creates the static go asset archives

-GO_VER="1.18.6"
-GARBLE_VER="1.18.0"
+GO_VER="1.19.1"
+GARBLE_VER="1.19.2"
 SGN_VER="0.0.3"

 GO_ARCH_1="amd64"
@@ -53,7 +53,7 @@ require (
 	gorm.io/driver/postgres v1.3.10
 	gorm.io/driver/sqlite v1.3.6
 	gorm.io/gorm v1.23.9
-	gvisor.dev/gvisor v0.0.0-20211020211948-f76a604701b6
+	gvisor.dev/gvisor v0.0.0-20220817001344-846276b3dbc5
 	modernc.org/sqlite v1.16.0
 )

@@ -1,6 +1,6 @@
 module github.com/bishopfox/sliver

-go 1.17
+go 1.18

 require (
 	github.com/Binject/universal v0.0.0-20210304094126-daefaa886313
@@ -10,34 +10,40 @@ require (
 	github.com/lesnuages/go-winio v0.4.19
 	github.com/miekg/dns v1.1.50
 	github.com/moloch--/memmod v0.0.0-20211120144554-8b37cc654945
+	github.com/onsi/ginkgo v1.16.5
+	github.com/onsi/gomega v1.20.2
 	github.com/pquerna/otp v1.3.0
 	github.com/stretchr/testify v1.7.0
 	github.com/things-go/go-socks5 v0.0.3-0.20210722055343-24af464efe43
 	golang.org/x/crypto v0.0.0-20220315160706-3147a52a75dd
-	golang.org/x/net v0.0.0-20220225172249-27dd8689420f
-	golang.org/x/sys v0.0.0-20220330033206-e17cdc41300f
+	golang.org/x/net v0.0.0-20220722155237-a158d28d115b
+	golang.org/x/sys v0.0.0-20220722155257-8c9f86f7a55f
 	golang.zx2c4.com/wireguard v0.0.0-20220316235147-5aff28b14c24
-	google.golang.org/protobuf v1.26.0
-	gvisor.dev/gvisor v0.0.0-20211020211948-f76a604701b6
+	google.golang.org/protobuf v1.28.0
+	gvisor.dev/gvisor v0.0.0-20220817001344-846276b3dbc5
 )

 require (
 	github.com/Binject/debug v0.0.0-20210225042342-c9b8b45728d2 // indirect
 	github.com/BurntSushi/xgb v0.0.0-20210121224620-deaf085860bc // indirect
-	github.com/Microsoft/go-winio v0.5.0 // indirect
+	github.com/Microsoft/go-winio v0.5.1 // indirect
 	github.com/awgh/cppgo v0.0.0-20210224085512-3d24bca8edc0 // indirect
 	github.com/awgh/rawreader v0.0.0-20200626064944-56820a9c6da4 // indirect
 	github.com/boombuler/barcode v1.0.1-0.20190219062509-6c824513bacc // indirect
 	github.com/davecgh/go-spew v1.1.1 // indirect
+	github.com/fsnotify/fsnotify v1.4.9 // indirect
 	github.com/gen2brain/shm v0.0.0-20200228170931-49f9650110c5 // indirect
 	github.com/google/btree v1.0.1 // indirect
+	github.com/google/go-cmp v0.5.8 // indirect
 	github.com/lxn/win v0.0.0-20210218163916-a377121e959e // indirect
+	github.com/nxadm/tail v1.4.8 // indirect
 	github.com/pmezard/go-difflib v1.0.0 // indirect
-	golang.org/x/mod v0.4.2 // indirect
+	golang.org/x/mod v0.5.1 // indirect
 	golang.org/x/text v0.3.7 // indirect
 	golang.org/x/time v0.0.0-20191024005414-555d28b269f0 // indirect
-	golang.org/x/tools v0.1.7 // indirect
+	golang.org/x/tools v0.1.9 // indirect
 	golang.org/x/xerrors v0.0.0-20200804184101-5ec99f83aff1 // indirect
 	golang.zx2c4.com/wintun v0.0.0-20211104114900-415007cec224 // indirect
-	gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c // indirect
+	gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7 // indirect
+	gopkg.in/yaml.v3 v3.0.1 // indirect
 )
@@ -6,6 +6,7 @@
 package netstack

 import (
+	"bytes"
 	"context"
 	"crypto/rand"
 	"encoding/binary"
@@ -23,10 +24,11 @@ import (
 	"golang.zx2c4.com/wireguard/tun"

 	"golang.org/x/net/dns/dnsmessage"
+	"gvisor.dev/gvisor/pkg/bufferv2"
 	"gvisor.dev/gvisor/pkg/tcpip"
 	"gvisor.dev/gvisor/pkg/tcpip/adapters/gonet"
-	"gvisor.dev/gvisor/pkg/tcpip/buffer"
 	"gvisor.dev/gvisor/pkg/tcpip/header"
+	"gvisor.dev/gvisor/pkg/tcpip/link/channel"
 	"gvisor.dev/gvisor/pkg/tcpip/network/ipv4"
 	"gvisor.dev/gvisor/pkg/tcpip/network/ipv6"
 	"gvisor.dev/gvisor/pkg/tcpip/stack"
@@ -37,69 +39,16 @@ import (
 )

 type netTun struct {
+	ep             *channel.Endpoint
 	stack          *stack.Stack
-	dispatcher     stack.NetworkDispatcher
 	events         chan tun.Event
-	incomingPacket chan buffer.VectorisedView
+	incomingPacket chan *bufferv2.View
 	mtu            int
 	dnsServers     []netip.Addr
 	hasV4, hasV6   bool
 }

-type (
-	endpoint netTun
-	Net      netTun
-)
-
-func (e *endpoint) Attach(dispatcher stack.NetworkDispatcher) {
-	e.dispatcher = dispatcher
-}
-
-func (e *endpoint) IsAttached() bool {
-	return e.dispatcher != nil
-}
-
-func (e *endpoint) MTU() uint32 {
-	mtu, err := (*netTun)(e).MTU()
-	if err != nil {
-		panic(err)
-	}
-	return uint32(mtu)
-}
-
-func (*endpoint) Capabilities() stack.LinkEndpointCapabilities {
-	return stack.CapabilityNone
-}
-
-func (*endpoint) MaxHeaderLength() uint16 {
-	return 0
-}
-
-func (*endpoint) LinkAddress() tcpip.LinkAddress {
-	return ""
-}
-
-func (*endpoint) Wait() {}
-
-func (e *endpoint) WritePacket(_ stack.RouteInfo, _ tcpip.NetworkProtocolNumber, pkt *stack.PacketBuffer) tcpip.Error {
-	e.incomingPacket <- buffer.NewVectorisedView(pkt.Size(), pkt.Views())
-	return nil
-}
-
-func (e *endpoint) WritePackets(stack.RouteInfo, stack.PacketBufferList, tcpip.NetworkProtocolNumber) (int, tcpip.Error) {
-	panic("not implemented")
-}
-
-func (e *endpoint) WriteRawPacket(*stack.PacketBuffer) tcpip.Error {
-	panic("not implemented")
-}
-
-func (*endpoint) ARPHardwareType() header.ARPHardwareType {
-	return header.ARPHardwareNone
-}
-
-func (e *endpoint) AddHeader(tcpip.LinkAddress, tcpip.LinkAddress, tcpip.NetworkProtocolNumber, *stack.PacketBuffer) {
-}
+type Net netTun

 func CreateNetTUN(localAddresses, dnsServers []netip.Addr, mtu int) (tun.Device, *Net, error) {
 	opts := stack.Options{
@@ -108,13 +57,15 @@ func CreateNetTUN(localAddresses, dnsServers []netip.Addr, mtu int) (tun.Device,
 		HandleLocal:        true,
 	}
 	dev := &netTun{
+		ep:             channel.New(1024, uint32(mtu), ""),
 		stack:          stack.New(opts),
 		events:         make(chan tun.Event, 10),
-		incomingPacket: make(chan buffer.VectorisedView),
+		incomingPacket: make(chan *bufferv2.View),
 		dnsServers:     dnsServers,
 		mtu:            mtu,
 	}
-	tcpipErr := dev.stack.CreateNIC(1, (*endpoint)(dev))
+	dev.ep.AddNotify(dev)
+	tcpipErr := dev.stack.CreateNIC(1, dev.ep)
 	if tcpipErr != nil {
 		return nil, nil, fmt.Errorf("CreateNIC: %v", tcpipErr)
 	}
@@ -167,6 +118,7 @@ func (tun *netTun) Read(buf []byte, offset int) (int, error) {
 	if !ok {
 		return 0, os.ErrClosed
 	}
+
 	return view.Read(buf[offset:])
 }

@@ -176,17 +128,29 @@ func (tun *netTun) Write(buf []byte, offset int) (int, error) {
 		return 0, nil
 	}

-	pkb := stack.NewPacketBuffer(stack.PacketBufferOptions{Data: buffer.NewVectorisedView(len(packet), []buffer.View{buffer.NewViewFromBytes(packet)})})
+	pkb := stack.NewPacketBuffer(stack.PacketBufferOptions{Payload: bufferv2.MakeWithData(packet)})
 	switch packet[0] >> 4 {
 	case 4:
-		tun.dispatcher.DeliverNetworkPacket("", "", ipv4.ProtocolNumber, pkb)
+		tun.ep.InjectInbound(header.IPv4ProtocolNumber, pkb)
 	case 6:
-		tun.dispatcher.DeliverNetworkPacket("", "", ipv6.ProtocolNumber, pkb)
+		tun.ep.InjectInbound(header.IPv6ProtocolNumber, pkb)
 	}

 	return len(buf), nil
 }

+func (tun *netTun) WriteNotify() {
+	pkt := tun.ep.Read()
+	if pkt == nil {
+		return
+	}
+
+	view := pkt.ToView()
+	pkt.DecRef()
+
+	tun.incomingPacket <- view
+}
+
 func (tun *netTun) Flush() error {
 	return nil
 }
@@ -197,9 +161,9 @@ func (tun *netTun) Close() error {
 	if tun.events != nil {
 		close(tun.events)
 	}
-	if tun.incomingPacket != nil {
-		close(tun.incomingPacket)
-	}
+
+	tun.ep.Close()
+
 	return nil
 }

@@ -434,11 +398,10 @@ func (pc *PingConn) WriteTo(p []byte, addr net.Addr) (n int, err error) {
 		return 0, fmt.Errorf("ping write: mismatched protocols")
 	}

-	buf := buffer.NewViewFromBytes(p)
-	rdr := buf.Reader()
+	buf := bytes.NewReader(p)
 	rfa, _ := convertToFullAddr(netip.AddrPortFrom(na, 0))
 	// won't block, no deadlines
-	n64, tcpipErr := pc.ep.Write(&rdr, tcpip.WriteOptions{
+	n64, tcpipErr := pc.ep.Write(buf, tcpip.WriteOptions{
 		To: &rfa,
 	})
 	if tcpipErr != nil {
@@ -453,8 +416,8 @@ func (pc *PingConn) Write(p []byte) (n int, err error) {
 }

 func (pc *PingConn) ReadFrom(p []byte) (n int, addr net.Addr, err error) {
-	e, notifyCh := waiter.NewChannelEntry(nil)
-	pc.wq.EventRegister(&e, waiter.EventIn)
+	e, notifyCh := waiter.NewChannelEntry(waiter.EventIn)
+	pc.wq.EventRegister(&e)
 	defer pc.wq.EventUnregister(&e)

 	select {
@@ -488,7 +451,7 @@ func (pc *PingConn) SetDeadline(t time.Time) error {
 }

 func (pc *PingConn) SetReadDeadline(t time.Time) error {
-	pc.deadline.Reset(t.Sub(time.Now()))
+	pc.deadline.Reset(time.Until(t))
 	return nil
 }

@@ -0,0 +1,12 @@
+root = true
+
+[*.go]
+indent_style = tab
+indent_size = 4
+insert_final_newline = true
+
+[*.{yml,yaml}]
+indent_style = space
+indent_size = 2
+insert_final_newline = true
+trim_trailing_whitespace = true
@@ -0,0 +1 @@
+go.sum linguist-generated
@@ -0,0 +1,6 @@
+# Setup a Global .gitignore for OS and editor generated files:
+# https://help.github.com/articles/ignoring-files
+# git config --global core.excludesfile ~/.gitignore_global
+
+.vagrant
+*.sublime-project
@@ -0,0 +1,36 @@
+sudo: false
+language: go
+
+go:
+  - "stable"
+  - "1.11.x"
+  - "1.10.x"
+  - "1.9.x"
+
+matrix:
+  include:
+    - go: "stable"
+      env: GOLINT=true
+  allow_failures:
+    - go: tip
+  fast_finish: true
+
+
+before_install:
+  - if [ ! -z "${GOLINT}" ]; then go get -u golang.org/x/lint/golint; fi
+
+script:
+  - go test --race ./...
+
+after_script:
+  - test -z "$(gofmt -s -l -w . | tee /dev/stderr)"
+  - if [ ! -z  "${GOLINT}" ]; then echo running golint; golint --set_exit_status  ./...; else echo skipping golint; fi
+  - go vet ./...
+
+os:
+  - linux
+  - osx
+  - windows
+
+notifications:
+  email: false
@@ -0,0 +1,52 @@
+# Names should be added to this file as
+#	Name or Organization <email address>
+# The email address is not required for organizations.
+
+# You can update this list using the following command:
+#
+#   $ git shortlog -se | awk '{print $2 " " $3 " " $4}'
+
+# Please keep the list sorted.
+
+Aaron L <[email protected]>
+Adrien Bustany <[email protected]>
+Amit Krishnan <[email protected]>
+Anmol Sethi <[email protected]>
+Bjørn Erik Pedersen <[email protected]>
+Bruno Bigras <[email protected]>
+Caleb Spare <[email protected]>
+Case Nelson <[email protected]>
+Chris Howey <[email protected]> <[email protected]>
+Christoffer Buchholz <[email protected]>
+Daniel Wagner-Hall <[email protected]>
+Dave Cheney <[email protected]>
+Evan Phoenix <[email protected]>
+Francisco Souza <[email protected]>
+Hari haran <[email protected]>
+John C Barstow
+Kelvin Fo <[email protected]>
+Ken-ichirou MATSUZAWA <[email protected]>
+Matt Layher <[email protected]>
+Nathan Youngman <[email protected]>
+Nickolai Zeldovich <[email protected]>
+Patrick <[email protected]>
+Paul Hammond <[email protected]>
+Pawel Knap <[email protected]>
+Pieter Droogendijk <[email protected]>
+Pursuit92 <[email protected]>
+Riku Voipio <[email protected]>
+Rob Figueiredo <[email protected]>
+Rodrigo Chiossi <[email protected]>
+Slawek Ligus <[email protected]>
+Soge Zhang <[email protected]>
+Tiffany Jernigan <[email protected]>
+Tilak Sharma <[email protected]>
+Tom Payne <[email protected]>
+Travis Cline <[email protected]>
+Tudor Golubenco <[email protected]>
+Vahe Khachikyan <[email protected]>
+Yukang <[email protected]>
+bronze1man <[email protected]>
+debrando <[email protected]>
+henrikedwards <[email protected]>
+铁哥 <[email protected]>
@@ -0,0 +1,317 @@
+# Changelog
+
+## v1.4.7 / 2018-01-09
+
+* BSD/macOS: Fix possible deadlock on closing the watcher on kqueue (thanks @nhooyr and @glycerine)
+* Tests: Fix missing verb on format string (thanks @rchiossi)
+* Linux: Fix deadlock in Remove (thanks @aarondl)
+* Linux: Watch.Add improvements (avoid race, fix consistency, reduce garbage) (thanks @twpayne)
+* Docs: Moved FAQ into the README (thanks @vahe)
+* Linux: Properly handle inotify's IN_Q_OVERFLOW event (thanks @zeldovich)
+* Docs: replace references to OS X with macOS
+
+## v1.4.2 / 2016-10-10
+
+* Linux: use InotifyInit1 with IN_CLOEXEC to stop leaking a file descriptor to a child process when using fork/exec [#178](https://github.com/fsnotify/fsnotify/pull/178) (thanks @pattyshack)
+
+## v1.4.1 / 2016-10-04
+
+* Fix flaky inotify stress test on Linux [#177](https://github.com/fsnotify/fsnotify/pull/177) (thanks @pattyshack)
+
+## v1.4.0 / 2016-10-01
+
+* add a String() method to Event.Op [#165](https://github.com/fsnotify/fsnotify/pull/165) (thanks @oozie)
+
+## v1.3.1 / 2016-06-28
+
+* Windows: fix for double backslash when watching the root of a drive [#151](https://github.com/fsnotify/fsnotify/issues/151) (thanks @brunoqc)
+
+## v1.3.0 / 2016-04-19
+
+* Support linux/arm64 by [patching](https://go-review.googlesource.com/#/c/21971/) x/sys/unix and switching to to it from syscall (thanks @suihkulokki) [#135](https://github.com/fsnotify/fsnotify/pull/135)
+
+## v1.2.10 / 2016-03-02
+
+* Fix golint errors in windows.go [#121](https://github.com/fsnotify/fsnotify/pull/121) (thanks @tiffanyfj)
+
+## v1.2.9 / 2016-01-13
+
+kqueue: Fix logic for CREATE after REMOVE [#111](https://github.com/fsnotify/fsnotify/pull/111) (thanks @bep)
+
+## v1.2.8 / 2015-12-17
+
+* kqueue: fix race condition in Close [#105](https://github.com/fsnotify/fsnotify/pull/105) (thanks @djui for reporting the issue and @ppknap for writing a failing test)
+* inotify: fix race in test
+* enable race detection for continuous integration (Linux, Mac, Windows)
+
+## v1.2.5 / 2015-10-17
+
+* inotify: use epoll_create1 for arm64 support (requires Linux 2.6.27 or later) [#100](https://github.com/fsnotify/fsnotify/pull/100) (thanks @suihkulokki)
+* inotify: fix path leaks [#73](https://github.com/fsnotify/fsnotify/pull/73) (thanks @chamaken)
+* kqueue: watch for rename events on subdirectories [#83](https://github.com/fsnotify/fsnotify/pull/83) (thanks @guotie)
+* kqueue: avoid infinite loops from symlinks cycles [#101](https://github.com/fsnotify/fsnotify/pull/101) (thanks @illicitonion)
+
+## v1.2.1 / 2015-10-14
+
+* kqueue: don't watch named pipes [#98](https://github.com/fsnotify/fsnotify/pull/98) (thanks @evanphx)
+
+## v1.2.0 / 2015-02-08
+
+* inotify: use epoll to wake up readEvents [#66](https://github.com/fsnotify/fsnotify/pull/66) (thanks @PieterD)
+* inotify: closing watcher should now always shut down goroutine [#63](https://github.com/fsnotify/fsnotify/pull/63) (thanks @PieterD)
+* kqueue: close kqueue after removing watches, fixes [#59](https://github.com/fsnotify/fsnotify/issues/59)
+
+## v1.1.1 / 2015-02-05
+
+* inotify: Retry read on EINTR [#61](https://github.com/fsnotify/fsnotify/issues/61) (thanks @PieterD)
+
+## v1.1.0 / 2014-12-12
+
+* kqueue: rework internals [#43](https://github.com/fsnotify/fsnotify/pull/43)
+    * add low-level functions
+    * only need to store flags on directories
+    * less mutexes [#13](https://github.com/fsnotify/fsnotify/issues/13)
+    * done can be an unbuffered channel
+    * remove calls to os.NewSyscallError
+* More efficient string concatenation for Event.String() [#52](https://github.com/fsnotify/fsnotify/pull/52) (thanks @mdlayher)
+* kqueue: fix regression in  rework causing subdirectories to be watched [#48](https://github.com/fsnotify/fsnotify/issues/48)
+* kqueue: cleanup internal watch before sending remove event [#51](https://github.com/fsnotify/fsnotify/issues/51)
+
+## v1.0.4 / 2014-09-07
+
+* kqueue: add dragonfly to the build tags.
+* Rename source code files, rearrange code so exported APIs are at the top.
+* Add done channel to example code. [#37](https://github.com/fsnotify/fsnotify/pull/37) (thanks @chenyukang)
+
+## v1.0.3 / 2014-08-19
+
+* [Fix] Windows MOVED_TO now translates to Create like on BSD and Linux. [#36](https://github.com/fsnotify/fsnotify/issues/36)
+
+## v1.0.2 / 2014-08-17
+
+* [Fix] Missing create events on macOS. [#14](https://github.com/fsnotify/fsnotify/issues/14) (thanks @zhsso)
+* [Fix] Make ./path and path equivalent. (thanks @zhsso)
+
+## v1.0.0 / 2014-08-15
+
+* [API] Remove AddWatch on Windows, use Add.
+* Improve documentation for exported identifiers. [#30](https://github.com/fsnotify/fsnotify/issues/30)
+* Minor updates based on feedback from golint.
+
+## dev / 2014-07-09
+
+* Moved to [github.com/fsnotify/fsnotify](https://github.com/fsnotify/fsnotify).
+* Use os.NewSyscallError instead of returning errno (thanks @hariharan-uno)
+
+## dev / 2014-07-04
+
+* kqueue: fix incorrect mutex used in Close()
+* Update example to demonstrate usage of Op.
+
+## dev / 2014-06-28
+
+* [API] Don't set the Write Op for attribute notifications [#4](https://github.com/fsnotify/fsnotify/issues/4)
+* Fix for String() method on Event (thanks Alex Brainman)
+* Don't build on Plan 9 or Solaris (thanks @4ad)
+
+## dev / 2014-06-21
+
+* Events channel of type Event rather than *Event.
+* [internal] use syscall constants directly for inotify and kqueue.
+* [internal] kqueue: rename events to kevents and fileEvent to event.
+
+## dev / 2014-06-19
+
+* Go 1.3+ required on Windows (uses syscall.ERROR_MORE_DATA internally).
+* [internal] remove cookie from Event struct (unused).
+* [internal] Event struct has the same definition across every OS.
+* [internal] remove internal watch and removeWatch methods.
+
+## dev / 2014-06-12
+
+* [API] Renamed Watch() to Add() and RemoveWatch() to Remove().
+* [API] Pluralized channel names: Events and Errors.
+* [API] Renamed FileEvent struct to Event.
+* [API] Op constants replace methods like IsCreate().
+
+## dev / 2014-06-12
+
+* Fix data race on kevent buffer (thanks @tilaks) [#98](https://github.com/howeyc/fsnotify/pull/98)
+
+## dev / 2014-05-23
+
+* [API] Remove current implementation of WatchFlags.
+    * current implementation doesn't take advantage of OS for efficiency
+    * provides little benefit over filtering events as they are received, but has  extra bookkeeping and mutexes
+    * no tests for the current implementation
+    * not fully implemented on Windows [#93](https://github.com/howeyc/fsnotify/issues/93#issuecomment-39285195)
+
+## v0.9.3 / 2014-12-31
+
+* kqueue: cleanup internal watch before sending remove event [#51](https://github.com/fsnotify/fsnotify/issues/51)
+
+## v0.9.2 / 2014-08-17
+
+* [Backport] Fix missing create events on macOS. [#14](https://github.com/fsnotify/fsnotify/issues/14) (thanks @zhsso)
+
+## v0.9.1 / 2014-06-12
+
+* Fix data race on kevent buffer (thanks @tilaks) [#98](https://github.com/howeyc/fsnotify/pull/98)
+
+## v0.9.0 / 2014-01-17
+
+* IsAttrib() for events that only concern a file's metadata [#79][] (thanks @abustany)
+* [Fix] kqueue: fix deadlock [#77][] (thanks @cespare)
+* [NOTICE] Development has moved to `code.google.com/p/go.exp/fsnotify` in preparation for inclusion in the Go standard library.
+
+## v0.8.12 / 2013-11-13
+
+* [API] Remove FD_SET and friends from Linux adapter
+
+## v0.8.11 / 2013-11-02
+
+* [Doc] Add Changelog [#72][] (thanks @nathany)
+* [Doc] Spotlight and double modify events on macOS [#62][] (reported by @paulhammond)
+
+## v0.8.10 / 2013-10-19
+
+* [Fix] kqueue: remove file watches when parent directory is removed [#71][] (reported by @mdwhatcott)
+* [Fix] kqueue: race between Close and readEvents [#70][] (reported by @bernerdschaefer)
+* [Doc] specify OS-specific limits in README (thanks @debrando)
+
+## v0.8.9 / 2013-09-08
+
+* [Doc] Contributing (thanks @nathany)
+* [Doc] update package path in example code [#63][] (thanks @paulhammond)
+* [Doc] GoCI badge in README (Linux only) [#60][]
+* [Doc] Cross-platform testing with Vagrant  [#59][] (thanks @nathany)
+
+## v0.8.8 / 2013-06-17
+
+* [Fix] Windows: handle `ERROR_MORE_DATA` on Windows [#49][] (thanks @jbowtie)
+
+## v0.8.7 / 2013-06-03
+
+* [API] Make syscall flags internal
+* [Fix] inotify: ignore event changes
+* [Fix] race in symlink test [#45][] (reported by @srid)
+* [Fix] tests on Windows
+* lower case error messages
+
+## v0.8.6 / 2013-05-23
+
+* kqueue: Use EVT_ONLY flag on Darwin
+* [Doc] Update README with full example
+
+## v0.8.5 / 2013-05-09
+
+* [Fix] inotify: allow monitoring of "broken" symlinks (thanks @tsg)
+
+## v0.8.4 / 2013-04-07
+
+* [Fix] kqueue: watch all file events [#40][] (thanks @ChrisBuchholz)
+
+## v0.8.3 / 2013-03-13
+
+* [Fix] inoitfy/kqueue memory leak [#36][] (reported by @nbkolchin)
+* [Fix] kqueue: use fsnFlags for watching a directory [#33][] (reported by @nbkolchin)
+
+## v0.8.2 / 2013-02-07
+
+* [Doc] add Authors
+* [Fix] fix data races for map access [#29][] (thanks @fsouza)
+
+## v0.8.1 / 2013-01-09
+
+* [Fix] Windows path separators
+* [Doc] BSD License
+
+## v0.8.0 / 2012-11-09
+
+* kqueue: directory watching improvements (thanks @vmirage)
+* inotify: add `IN_MOVED_TO` [#25][] (requested by @cpisto)
+* [Fix] kqueue: deleting watched directory [#24][] (reported by @jakerr)
+
+## v0.7.4 / 2012-10-09
+
+* [Fix] inotify: fixes from https://codereview.appspot.com/5418045/ (ugorji)
+* [Fix] kqueue: preserve watch flags when watching for delete [#21][] (reported by @robfig)
+* [Fix] kqueue: watch the directory even if it isn't a new watch (thanks @robfig)
+* [Fix] kqueue: modify after recreation of file
+
+## v0.7.3 / 2012-09-27
+
+* [Fix] kqueue: watch with an existing folder inside the watched folder (thanks @vmirage)
+* [Fix] kqueue: no longer get duplicate CREATE events
+
+## v0.7.2 / 2012-09-01
+
+* kqueue: events for created directories
+
+## v0.7.1 / 2012-07-14
+
+* [Fix] for renaming files
+
+## v0.7.0 / 2012-07-02
+
+* [Feature] FSNotify flags
+* [Fix] inotify: Added file name back to event path
+
+## v0.6.0 / 2012-06-06
+
+* kqueue: watch files after directory created (thanks @tmc)
+
+## v0.5.1 / 2012-05-22
+
+* [Fix] inotify: remove all watches before Close()
+
+## v0.5.0 / 2012-05-03
+
+* [API] kqueue: return errors during watch instead of sending over channel
+* kqueue: match symlink behavior on Linux
+* inotify: add `DELETE_SELF` (requested by @taralx)
+* [Fix] kqueue: handle EINTR (reported by @robfig)
+* [Doc] Godoc example [#1][] (thanks @davecheney)
+
+## v0.4.0 / 2012-03-30
+
+* Go 1 released: build with go tool
+* [Feature] Windows support using winfsnotify
+* Windows does not have attribute change notifications
+* Roll attribute notifications into IsModify
+
+## v0.3.0 / 2012-02-19
+
+* kqueue: add files when watch directory
+
+## v0.2.0 / 2011-12-30
+
+* update to latest Go weekly code
+
+## v0.1.0 / 2011-10-19
+
+* kqueue: add watch on file creation to match inotify
+* kqueue: create file event
+* inotify: ignore `IN_IGNORED` events
+* event String()
+* linux: common FileEvent functions
+* initial commit
+
+[#79]: https://github.com/howeyc/fsnotify/pull/79
+[#77]: https://github.com/howeyc/fsnotify/pull/77
+[#72]: https://github.com/howeyc/fsnotify/issues/72
+[#71]: https://github.com/howeyc/fsnotify/issues/71
+[#70]: https://github.com/howeyc/fsnotify/issues/70
+[#63]: https://github.com/howeyc/fsnotify/issues/63
+[#62]: https://github.com/howeyc/fsnotify/issues/62
+[#60]: https://github.com/howeyc/fsnotify/issues/60
+[#59]: https://github.com/howeyc/fsnotify/issues/59
+[#49]: https://github.com/howeyc/fsnotify/issues/49
+[#45]: https://github.com/howeyc/fsnotify/issues/45
+[#40]: https://github.com/howeyc/fsnotify/issues/40
+[#36]: https://github.com/howeyc/fsnotify/issues/36
+[#33]: https://github.com/howeyc/fsnotify/issues/33
+[#29]: https://github.com/howeyc/fsnotify/issues/29
+[#25]: https://github.com/howeyc/fsnotify/issues/25
+[#24]: https://github.com/howeyc/fsnotify/issues/24
+[#21]: https://github.com/howeyc/fsnotify/issues/21
@@ -0,0 +1,77 @@
+# Contributing
+
+## Issues
+
+* Request features and report bugs using the [GitHub Issue Tracker](https://github.com/fsnotify/fsnotify/issues).
+* Please indicate the platform you are using fsnotify on.
+* A code example to reproduce the problem is appreciated.
+
+## Pull Requests
+
+### Contributor License Agreement
+
+fsnotify is derived from code in the [golang.org/x/exp](https://godoc.org/golang.org/x/exp) package and it may be included [in the standard library](https://github.com/fsnotify/fsnotify/issues/1) in the future. Therefore fsnotify carries the same [LICENSE](https://github.com/fsnotify/fsnotify/blob/master/LICENSE) as Go. Contributors retain their copyright, so you need to fill out a short form before we can accept your contribution: [Google Individual Contributor License Agreement](https://developers.google.com/open-source/cla/individual).
+
+Please indicate that you have signed the CLA in your pull request.
+
+### How fsnotify is Developed
+
+* Development is done on feature branches.
+* Tests are run on BSD, Linux, macOS and Windows.
+* Pull requests are reviewed and [applied to master][am] using [hub][].
+  * Maintainers may modify or squash commits rather than asking contributors to.
+* To issue a new release, the maintainers will:
+  * Update the CHANGELOG
+  * Tag a version, which will become available through gopkg.in.
+ 
+### How to Fork
+
+For smooth sailing, always use the original import path. Installing with `go get` makes this easy. 
+
+1. Install from GitHub (`go get -u github.com/fsnotify/fsnotify`)
+2. Create your feature branch (`git checkout -b my-new-feature`)
+3. Ensure everything works and the tests pass (see below)
+4. Commit your changes (`git commit -am 'Add some feature'`)
+
+Contribute upstream:
+
+1. Fork fsnotify on GitHub
+2. Add your remote (`git remote add fork [email protected]:mycompany/repo.git`)
+3. Push to the branch (`git push fork my-new-feature`)
+4. Create a new Pull Request on GitHub
+
+This workflow is [thoroughly explained by Katrina Owen](https://splice.com/blog/contributing-open-source-git-repositories-go/).
+
+### Testing
+
+fsnotify uses build tags to compile different code on Linux, BSD, macOS, and Windows.
+
+Before doing a pull request, please do your best to test your changes on multiple platforms, and list which platforms you were able/unable to test on.
+
+To aid in cross-platform testing there is a Vagrantfile for Linux and BSD.
+
+* Install [Vagrant](http://www.vagrantup.com/) and [VirtualBox](https://www.virtualbox.org/)
+* Setup [Vagrant Gopher](https://github.com/nathany/vagrant-gopher) in your `src` folder.
+* Run `vagrant up` from the project folder. You can also setup just one box with `vagrant up linux` or `vagrant up bsd` (note: the BSD box doesn't support Windows hosts at this time, and NFS may prompt for your host OS password)
+* Once setup, you can run the test suite on a given OS with a single command `vagrant ssh linux -c 'cd fsnotify/fsnotify; go test'`.
+* When you're done, you will want to halt or destroy the Vagrant boxes.
+
+Notice: fsnotify file system events won't trigger in shared folders. The tests get around this limitation by using the /tmp directory.
+
+Right now there is no equivalent solution for Windows and macOS, but there are Windows VMs [freely available from Microsoft](http://www.modern.ie/en-us/virtualization-tools#downloads).
+
+### Maintainers
+
+Help maintaining fsnotify is welcome. To be a maintainer:
+
+* Submit a pull request and sign the CLA as above.
+* You must be able to run the test suite on Mac, Windows, Linux and BSD.
+
+To keep master clean, the fsnotify project uses the "apply mail" workflow outlined in Nathaniel Talbott's post ["Merge pull request" Considered Harmful][am]. This requires installing [hub][].
+
+All code changes should be internal pull requests.
+
+Releases are tagged using [Semantic Versioning](http://semver.org/).
+
+[hub]: https://github.com/github/hub
+[am]: http://blog.spreedly.com/2014/06/24/merge-pull-request-considered-harmful/#.VGa5yZPF_Zs
@@ -0,0 +1,28 @@
+Copyright (c) 2012 The Go Authors. All rights reserved.
+Copyright (c) 2012-2019 fsnotify Authors. All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+   * Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+   * Redistributions in binary form must reproduce the above
+copyright notice, this list of conditions and the following disclaimer
+in the documentation and/or other materials provided with the
+distribution.
+   * Neither the name of Google Inc. nor the names of its
+contributors may be used to endorse or promote products derived from
+this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
@@ -0,0 +1,130 @@
+# File system notifications for Go
+
+[![GoDoc](https://godoc.org/github.com/fsnotify/fsnotify?status.svg)](https://godoc.org/github.com/fsnotify/fsnotify) [![Go Report Card](https://goreportcard.com/badge/github.com/fsnotify/fsnotify)](https://goreportcard.com/report/github.com/fsnotify/fsnotify)
+
+fsnotify utilizes [golang.org/x/sys](https://godoc.org/golang.org/x/sys) rather than `syscall` from the standard library. Ensure you have the latest version installed by running:
+
+```console
+go get -u golang.org/x/sys/...
+```
+
+Cross platform: Windows, Linux, BSD and macOS.
+
+| Adapter               | OS                               | Status                                                                                                                          |
+| --------------------- | -------------------------------- | ------------------------------------------------------------------------------------------------------------------------------- |
+| inotify               | Linux 2.6.27 or later, Android\* | Supported [![Build Status](https://travis-ci.org/fsnotify/fsnotify.svg?branch=master)](https://travis-ci.org/fsnotify/fsnotify) |
+| kqueue                | BSD, macOS, iOS\*                | Supported [![Build Status](https://travis-ci.org/fsnotify/fsnotify.svg?branch=master)](https://travis-ci.org/fsnotify/fsnotify) |
+| ReadDirectoryChangesW | Windows                          | Supported [![Build Status](https://travis-ci.org/fsnotify/fsnotify.svg?branch=master)](https://travis-ci.org/fsnotify/fsnotify) |
+| FSEvents              | macOS                            | [Planned](https://github.com/fsnotify/fsnotify/issues/11)                                                                       |
+| FEN                   | Solaris 11                       | [In Progress](https://github.com/fsnotify/fsnotify/issues/12)                                                                   |
+| fanotify              | Linux 2.6.37+                    | [Planned](https://github.com/fsnotify/fsnotify/issues/114)                                                                      |
+| USN Journals          | Windows                          | [Maybe](https://github.com/fsnotify/fsnotify/issues/53)                                                                         |
+| Polling               | *All*                            | [Maybe](https://github.com/fsnotify/fsnotify/issues/9)                                                                          |
+
+\* Android and iOS are untested.
+
+Please see [the documentation](https://godoc.org/github.com/fsnotify/fsnotify) and consult the [FAQ](#faq) for usage information.
+
+## API stability
+
+fsnotify is a fork of [howeyc/fsnotify](https://godoc.org/github.com/howeyc/fsnotify) with a new API as of v1.0. The API is based on [this design document](http://goo.gl/MrYxyA). 
+
+All [releases](https://github.com/fsnotify/fsnotify/releases) are tagged based on [Semantic Versioning](http://semver.org/). Further API changes are [planned](https://github.com/fsnotify/fsnotify/milestones), and will be tagged with a new major revision number.
+
+Go 1.6 supports dependencies located in the `vendor/` folder. Unless you are creating a library, it is recommended that you copy fsnotify into `vendor/github.com/fsnotify/fsnotify` within your project, and likewise for `golang.org/x/sys`.
+
+## Usage
+
+```go
+package main
+
+import (
+	"log"
+
+	"github.com/fsnotify/fsnotify"
+)
+
+func main() {
+	watcher, err := fsnotify.NewWatcher()
+	if err != nil {
+		log.Fatal(err)
+	}
+	defer watcher.Close()
+
+	done := make(chan bool)
+	go func() {
+		for {
+			select {
+			case event, ok := <-watcher.Events:
+				if !ok {
+					return
+				}
+				log.Println("event:", event)
+				if event.Op&fsnotify.Write == fsnotify.Write {
+					log.Println("modified file:", event.Name)
+				}
+			case err, ok := <-watcher.Errors:
+				if !ok {
+					return
+				}
+				log.Println("error:", err)
+			}
+		}
+	}()
+
+	err = watcher.Add("/tmp/foo")
+	if err != nil {
+		log.Fatal(err)
+	}
+	<-done
+}
+```
+
+## Contributing
+
+Please refer to [CONTRIBUTING][] before opening an issue or pull request.
+
+## Example
+
+See [example_test.go](https://github.com/fsnotify/fsnotify/blob/master/example_test.go).
+
+## FAQ
+
+**When a file is moved to another directory is it still being watched?**
+
+No (it shouldn't be, unless you are watching where it was moved to).
+
+**When I watch a directory, are all subdirectories watched as well?**
+
+No, you must add watches for any directory you want to watch (a recursive watcher is on the roadmap [#18][]).
+
+**Do I have to watch the Error and Event channels in a separate goroutine?**
+
+As of now, yes. Looking into making this single-thread friendly (see [howeyc #7][#7])
+
+**Why am I receiving multiple events for the same file on OS X?**
+
+Spotlight indexing on OS X can result in multiple events (see [howeyc #62][#62]). A temporary workaround is to add your folder(s) to the *Spotlight Privacy settings* until we have a native FSEvents implementation (see [#11][]).
+
+**How many files can be watched at once?**
+
+There are OS-specific limits as to how many watches can be created:
+* Linux: /proc/sys/fs/inotify/max_user_watches contains the limit, reaching this limit results in a "no space left on device" error.
+* BSD / OSX: sysctl variables "kern.maxfiles" and "kern.maxfilesperproc", reaching these limits results in a "too many open files" error.
+
+**Why don't notifications work with NFS filesystems or filesystem in userspace (FUSE)?**
+
+fsnotify requires support from underlying OS to work. The current NFS protocol does not provide network level support for file notifications.
+
+[#62]: https://github.com/howeyc/fsnotify/issues/62
+[#18]: https://github.com/fsnotify/fsnotify/issues/18
+[#11]: https://github.com/fsnotify/fsnotify/issues/11
+[#7]: https://github.com/howeyc/fsnotify/issues/7
+
+[contributing]: https://github.com/fsnotify/fsnotify/blob/master/CONTRIBUTING.md
+
+## Related Projects
+
+* [notify](https://github.com/rjeczalik/notify)
+* [fsevents](https://github.com/fsnotify/fsevents)
+
@@ -0,0 +1,37 @@
+// Copyright 2010 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build solaris
+
+package fsnotify
+
+import (
+	"errors"
+)
+
+// Watcher watches a set of files, delivering events to a channel.
+type Watcher struct {
+	Events chan Event
+	Errors chan error
+}
+
+// NewWatcher establishes a new watcher with the underlying OS and begins waiting for events.
+func NewWatcher() (*Watcher, error) {
+	return nil, errors.New("FEN based watcher not yet supported for fsnotify\n")
+}
+
+// Close removes all watches and closes the events channel.
+func (w *Watcher) Close() error {
+	return nil
+}
+
+// Add starts watching the named file or directory (non-recursively).
+func (w *Watcher) Add(name string) error {
+	return nil
+}
+
+// Remove stops watching the the named file or directory (non-recursively).
+func (w *Watcher) Remove(name string) error {
+	return nil
+}
@@ -0,0 +1,68 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build !plan9
+
+// Package fsnotify provides a platform-independent interface for file system notifications.
+package fsnotify
+
+import (
+	"bytes"
+	"errors"
+	"fmt"
+)
+
+// Event represents a single file system notification.
+type Event struct {
+	Name string // Relative path to the file or directory.
+	Op   Op     // File operation that triggered the event.
+}
+
+// Op describes a set of file operations.
+type Op uint32
+
+// These are the generalized file operations that can trigger a notification.
+const (
+	Create Op = 1 << iota
+	Write
+	Remove
+	Rename
+	Chmod
+)
+
+func (op Op) String() string {
+	// Use a buffer for efficient string concatenation
+	var buffer bytes.Buffer
+
+	if op&Create == Create {
+		buffer.WriteString("|CREATE")
+	}
+	if op&Remove == Remove {
+		buffer.WriteString("|REMOVE")
+	}
+	if op&Write == Write {
+		buffer.WriteString("|WRITE")
+	}
+	if op&Rename == Rename {
+		buffer.WriteString("|RENAME")
+	}
+	if op&Chmod == Chmod {
+		buffer.WriteString("|CHMOD")
+	}
+	if buffer.Len() == 0 {
+		return ""
+	}
+	return buffer.String()[1:] // Strip leading pipe
+}
+
+// String returns a string representation of the event in the form
+// "file: REMOVE|WRITE|..."
+func (e Event) String() string {
+	return fmt.Sprintf("%q: %s", e.Name, e.Op.String())
+}
+
+// Common errors that can be reported by a watcher
+var (
+	ErrEventOverflow = errors.New("fsnotify queue overflow")
+)
@@ -0,0 +1,337 @@
+// Copyright 2010 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build linux
+
+package fsnotify
+
+import (
+	"errors"
+	"fmt"
+	"io"
+	"os"
+	"path/filepath"
+	"strings"
+	"sync"
+	"unsafe"
+
+	"golang.org/x/sys/unix"
+)
+
+// Watcher watches a set of files, delivering events to a channel.
+type Watcher struct {
+	Events   chan Event
+	Errors   chan error
+	mu       sync.Mutex // Map access
+	fd       int
+	poller   *fdPoller
+	watches  map[string]*watch // Map of inotify watches (key: path)
+	paths    map[int]string    // Map of watched paths (key: watch descriptor)
+	done     chan struct{}     // Channel for sending a "quit message" to the reader goroutine
+	doneResp chan struct{}     // Channel to respond to Close
+}
+
+// NewWatcher establishes a new watcher with the underlying OS and begins waiting for events.
+func NewWatcher() (*Watcher, error) {
+	// Create inotify fd
+	fd, errno := unix.InotifyInit1(unix.IN_CLOEXEC)
+	if fd == -1 {
+		return nil, errno
+	}
+	// Create epoll
+	poller, err := newFdPoller(fd)
+	if err != nil {
+		unix.Close(fd)
+		return nil, err
+	}
+	w := &Watcher{
+		fd:       fd,
+		poller:   poller,
+		watches:  make(map[string]*watch),
+		paths:    make(map[int]string),
+		Events:   make(chan Event),
+		Errors:   make(chan error),
+		done:     make(chan struct{}),
+		doneResp: make(chan struct{}),
+	}
+
+	go w.readEvents()
+	return w, nil
+}
+
+func (w *Watcher) isClosed() bool {
+	select {
+	case <-w.done:
+		return true
+	default:
+		return false
+	}
+}
+
+// Close removes all watches and closes the events channel.
+func (w *Watcher) Close() error {
+	if w.isClosed() {
+		return nil
+	}
+
+	// Send 'close' signal to goroutine, and set the Watcher to closed.
+	close(w.done)
+
+	// Wake up goroutine
+	w.poller.wake()
+
+	// Wait for goroutine to close
+	<-w.doneResp
+
+	return nil
+}
+
+// Add starts watching the named file or directory (non-recursively).
+func (w *Watcher) Add(name string) error {
+	name = filepath.Clean(name)
+	if w.isClosed() {
+		return errors.New("inotify instance already closed")
+	}
+
+	const agnosticEvents = unix.IN_MOVED_TO | unix.IN_MOVED_FROM |
+		unix.IN_CREATE | unix.IN_ATTRIB | unix.IN_MODIFY |
+		unix.IN_MOVE_SELF | unix.IN_DELETE | unix.IN_DELETE_SELF
+
+	var flags uint32 = agnosticEvents
+
+	w.mu.Lock()
+	defer w.mu.Unlock()
+	watchEntry := w.watches[name]
+	if watchEntry != nil {
+		flags |= watchEntry.flags | unix.IN_MASK_ADD
+	}
+	wd, errno := unix.InotifyAddWatch(w.fd, name, flags)
+	if wd == -1 {
+		return errno
+	}
+
+	if watchEntry == nil {
+		w.watches[name] = &watch{wd: uint32(wd), flags: flags}
+		w.paths[wd] = name
+	} else {
+		watchEntry.wd = uint32(wd)
+		watchEntry.flags = flags
+	}
+
+	return nil
+}
+
+// Remove stops watching the named file or directory (non-recursively).
+func (w *Watcher) Remove(name string) error {
+	name = filepath.Clean(name)
+
+	// Fetch the watch.
+	w.mu.Lock()
+	defer w.mu.Unlock()
+	watch, ok := w.watches[name]
+
+	// Remove it from inotify.
+	if !ok {
+		return fmt.Errorf("can't remove non-existent inotify watch for: %s", name)
+	}
+
+	// We successfully removed the watch if InotifyRmWatch doesn't return an
+	// error, we need to clean up our internal state to ensure it matches
+	// inotify's kernel state.
+	delete(w.paths, int(watch.wd))
+	delete(w.watches, name)
+
+	// inotify_rm_watch will return EINVAL if the file has been deleted;
+	// the inotify will already have been removed.
+	// watches and pathes are deleted in ignoreLinux() implicitly and asynchronously
+	// by calling inotify_rm_watch() below. e.g. readEvents() goroutine receives IN_IGNORE
+	// so that EINVAL means that the wd is being rm_watch()ed or its file removed
+	// by another thread and we have not received IN_IGNORE event.
+	success, errno := unix.InotifyRmWatch(w.fd, watch.wd)
+	if success == -1 {
+		// TODO: Perhaps it's not helpful to return an error here in every case.
+		// the only two possible errors are:
+		// EBADF, which happens when w.fd is not a valid file descriptor of any kind.
+		// EINVAL, which is when fd is not an inotify descriptor or wd is not a valid watch descriptor.
+		// Watch descriptors are invalidated when they are removed explicitly or implicitly;
+		// explicitly by inotify_rm_watch, implicitly when the file they are watching is deleted.
+		return errno
+	}
+
+	return nil
+}
+
+type watch struct {
+	wd    uint32 // Watch descriptor (as returned by the inotify_add_watch() syscall)
+	flags uint32 // inotify flags of this watch (see inotify(7) for the list of valid flags)
+}
+
+// readEvents reads from the inotify file descriptor, converts the
+// received events into Event objects and sends them via the Events channel
+func (w *Watcher) readEvents() {
+	var (
+		buf   [unix.SizeofInotifyEvent * 4096]byte // Buffer for a maximum of 4096 raw events
+		n     int                                  // Number of bytes read with read()
+		errno error                                // Syscall errno
+		ok    bool                                 // For poller.wait
+	)
+
+	defer close(w.doneResp)
+	defer close(w.Errors)
+	defer close(w.Events)
+	defer unix.Close(w.fd)
+	defer w.poller.close()
+
+	for {
+		// See if we have been closed.
+		if w.isClosed() {
+			return
+		}
+
+		ok, errno = w.poller.wait()
+		if errno != nil {
+			select {
+			case w.Errors <- errno:
+			case <-w.done:
+				return
+			}
+			continue
+		}
+
+		if !ok {
+			continue
+		}
+
+		n, errno = unix.Read(w.fd, buf[:])
+		// If a signal interrupted execution, see if we've been asked to close, and try again.
+		// http://man7.org/linux/man-pages/man7/signal.7.html :
+		// "Before Linux 3.8, reads from an inotify(7) file descriptor were not restartable"
+		if errno == unix.EINTR {
+			continue
+		}
+
+		// unix.Read might have been woken up by Close. If so, we're done.
+		if w.isClosed() {
+			return
+		}
+
+		if n < unix.SizeofInotifyEvent {
+			var err error
+			if n == 0 {
+				// If EOF is received. This should really never happen.
+				err = io.EOF
+			} else if n < 0 {
+				// If an error occurred while reading.
+				err = errno
+			} else {
+				// Read was too short.
+				err = errors.New("notify: short read in readEvents()")
+			}
+			select {
+			case w.Errors <- err:
+			case <-w.done:
+				return
+			}
+			continue
+		}
+
+		var offset uint32
+		// We don't know how many events we just read into the buffer
+		// While the offset points to at least one whole event...
+		for offset <= uint32(n-unix.SizeofInotifyEvent) {
+			// Point "raw" to the event in the buffer
+			raw := (*unix.InotifyEvent)(unsafe.Pointer(&buf[offset]))
+
+			mask := uint32(raw.Mask)
+			nameLen := uint32(raw.Len)
+
+			if mask&unix.IN_Q_OVERFLOW != 0 {
+				select {
+				case w.Errors <- ErrEventOverflow:
+				case <-w.done:
+					return
+				}
+			}
+
+			// If the event happened to the watched directory or the watched file, the kernel
+			// doesn't append the filename to the event, but we would like to always fill the
+			// the "Name" field with a valid filename. We retrieve the path of the watch from
+			// the "paths" map.
+			w.mu.Lock()
+			name, ok := w.paths[int(raw.Wd)]
+			// IN_DELETE_SELF occurs when the file/directory being watched is removed.
+			// This is a sign to clean up the maps, otherwise we are no longer in sync
+			// with the inotify kernel state which has already deleted the watch
+			// automatically.
+			if ok && mask&unix.IN_DELETE_SELF == unix.IN_DELETE_SELF {
+				delete(w.paths, int(raw.Wd))
+				delete(w.watches, name)
+			}
+			w.mu.Unlock()
+
+			if nameLen > 0 {
+				// Point "bytes" at the first byte of the filename
+				bytes := (*[unix.PathMax]byte)(unsafe.Pointer(&buf[offset+unix.SizeofInotifyEvent]))
+				// The filename is padded with NULL bytes. TrimRight() gets rid of those.
+				name += "/" + strings.TrimRight(string(bytes[0:nameLen]), "\000")
+			}
+
+			event := newEvent(name, mask)
+
+			// Send the events that are not ignored on the events channel
+			if !event.ignoreLinux(mask) {
+				select {
+				case w.Events <- event:
+				case <-w.done:
+					return
+				}
+			}
+
+			// Move to the next event in the buffer
+			offset += unix.SizeofInotifyEvent + nameLen
+		}
+	}
+}
+
+// Certain types of events can be "ignored" and not sent over the Events
+// channel. Such as events marked ignore by the kernel, or MODIFY events
+// against files that do not exist.
+func (e *Event) ignoreLinux(mask uint32) bool {
+	// Ignore anything the inotify API says to ignore
+	if mask&unix.IN_IGNORED == unix.IN_IGNORED {
+		return true
+	}
+
+	// If the event is not a DELETE or RENAME, the file must exist.
+	// Otherwise the event is ignored.
+	// *Note*: this was put in place because it was seen that a MODIFY
+	// event was sent after the DELETE. This ignores that MODIFY and
+	// assumes a DELETE will come or has come if the file doesn't exist.
+	if !(e.Op&Remove == Remove || e.Op&Rename == Rename) {
+		_, statErr := os.Lstat(e.Name)
+		return os.IsNotExist(statErr)
+	}
+	return false
+}
+
+// newEvent returns an platform-independent Event based on an inotify mask.
+func newEvent(name string, mask uint32) Event {
+	e := Event{Name: name}
+	if mask&unix.IN_CREATE == unix.IN_CREATE || mask&unix.IN_MOVED_TO == unix.IN_MOVED_TO {
+		e.Op |= Create
+	}
+	if mask&unix.IN_DELETE_SELF == unix.IN_DELETE_SELF || mask&unix.IN_DELETE == unix.IN_DELETE {
+		e.Op |= Remove
+	}
+	if mask&unix.IN_MODIFY == unix.IN_MODIFY {
+		e.Op |= Write
+	}
+	if mask&unix.IN_MOVE_SELF == unix.IN_MOVE_SELF || mask&unix.IN_MOVED_FROM == unix.IN_MOVED_FROM {
+		e.Op |= Rename
+	}
+	if mask&unix.IN_ATTRIB == unix.IN_ATTRIB {
+		e.Op |= Chmod
+	}
+	return e
+}
@@ -0,0 +1,187 @@
+// Copyright 2015 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build linux
+
+package fsnotify
+
+import (
+	"errors"
+
+	"golang.org/x/sys/unix"
+)
+
+type fdPoller struct {
+	fd   int    // File descriptor (as returned by the inotify_init() syscall)
+	epfd int    // Epoll file descriptor
+	pipe [2]int // Pipe for waking up
+}
+
+func emptyPoller(fd int) *fdPoller {
+	poller := new(fdPoller)
+	poller.fd = fd
+	poller.epfd = -1
+	poller.pipe[0] = -1
+	poller.pipe[1] = -1
+	return poller
+}
+
+// Create a new inotify poller.
+// This creates an inotify handler, and an epoll handler.
+func newFdPoller(fd int) (*fdPoller, error) {
+	var errno error
+	poller := emptyPoller(fd)
+	defer func() {
+		if errno != nil {
+			poller.close()
+		}
+	}()
+	poller.fd = fd
+
+	// Create epoll fd
+	poller.epfd, errno = unix.EpollCreate1(unix.EPOLL_CLOEXEC)
+	if poller.epfd == -1 {
+		return nil, errno
+	}
+	// Create pipe; pipe[0] is the read end, pipe[1] the write end.
+	errno = unix.Pipe2(poller.pipe[:], unix.O_NONBLOCK|unix.O_CLOEXEC)
+	if errno != nil {
+		return nil, errno
+	}
+
+	// Register inotify fd with epoll
+	event := unix.EpollEvent{
+		Fd:     int32(poller.fd),
+		Events: unix.EPOLLIN,
+	}
+	errno = unix.EpollCtl(poller.epfd, unix.EPOLL_CTL_ADD, poller.fd, &event)
+	if errno != nil {
+		return nil, errno
+	}
+
+	// Register pipe fd with epoll
+	event = unix.EpollEvent{
+		Fd:     int32(poller.pipe[0]),
+		Events: unix.EPOLLIN,
+	}
+	errno = unix.EpollCtl(poller.epfd, unix.EPOLL_CTL_ADD, poller.pipe[0], &event)
+	if errno != nil {
+		return nil, errno
+	}
+
+	return poller, nil
+}
+
+// Wait using epoll.
+// Returns true if something is ready to be read,
+// false if there is not.
+func (poller *fdPoller) wait() (bool, error) {
+	// 3 possible events per fd, and 2 fds, makes a maximum of 6 events.
+	// I don't know whether epoll_wait returns the number of events returned,
+	// or the total number of events ready.
+	// I decided to catch both by making the buffer one larger than the maximum.
+	events := make([]unix.EpollEvent, 7)
+	for {
+		n, errno := unix.EpollWait(poller.epfd, events, -1)
+		if n == -1 {
+			if errno == unix.EINTR {
+				continue
+			}
+			return false, errno
+		}
+		if n == 0 {
+			// If there are no events, try again.
+			continue
+		}
+		if n > 6 {
+			// This should never happen. More events were returned than should be possible.
+			return false, errors.New("epoll_wait returned more events than I know what to do with")
+		}
+		ready := events[:n]
+		epollhup := false
+		epollerr := false
+		epollin := false
+		for _, event := range ready {
+			if event.Fd == int32(poller.fd) {
+				if event.Events&unix.EPOLLHUP != 0 {
+					// This should not happen, but if it does, treat it as a wakeup.
+					epollhup = true
+				}
+				if event.Events&unix.EPOLLERR != 0 {
+					// If an error is waiting on the file descriptor, we should pretend
+					// something is ready to read, and let unix.Read pick up the error.
+					epollerr = true
+				}
+				if event.Events&unix.EPOLLIN != 0 {
+					// There is data to read.
+					epollin = true
+				}
+			}
+			if event.Fd == int32(poller.pipe[0]) {
+				if event.Events&unix.EPOLLHUP != 0 {
+					// Write pipe descriptor was closed, by us. This means we're closing down the
+					// watcher, and we should wake up.
+				}
+				if event.Events&unix.EPOLLERR != 0 {
+					// If an error is waiting on the pipe file descriptor.
+					// This is an absolute mystery, and should never ever happen.
+					return false, errors.New("Error on the pipe descriptor.")
+				}
+				if event.Events&unix.EPOLLIN != 0 {
+					// This is a regular wakeup, so we have to clear the buffer.
+					err := poller.clearWake()
+					if err != nil {
+						return false, err
+					}
+				}
+			}
+		}
+
+		if epollhup || epollerr || epollin {
+			return true, nil
+		}
+		return false, nil
+	}
+}
+
+// Close the write end of the poller.
+func (poller *fdPoller) wake() error {
+	buf := make([]byte, 1)
+	n, errno := unix.Write(poller.pipe[1], buf)
+	if n == -1 {
+		if errno == unix.EAGAIN {
+			// Buffer is full, poller will wake.
+			return nil
+		}
+		return errno
+	}
+	return nil
+}
+
+func (poller *fdPoller) clearWake() error {
+	// You have to be woken up a LOT in order to get to 100!
+	buf := make([]byte, 100)
+	n, errno := unix.Read(poller.pipe[0], buf)
+	if n == -1 {
+		if errno == unix.EAGAIN {
+			// Buffer is empty, someone else cleared our wake.
+			return nil
+		}
+		return errno
+	}
+	return nil
+}
+
+// Close all poller file descriptors, but not the one passed to it.
+func (poller *fdPoller) close() {
+	if poller.pipe[1] != -1 {
+		unix.Close(poller.pipe[1])
+	}
+	if poller.pipe[0] != -1 {
+		unix.Close(poller.pipe[0])
+	}
+	if poller.epfd != -1 {
+		unix.Close(poller.epfd)
+	}
+}
@@ -0,0 +1,521 @@
+// Copyright 2010 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build freebsd openbsd netbsd dragonfly darwin
+
+package fsnotify
+
+import (
+	"errors"
+	"fmt"
+	"io/ioutil"
+	"os"
+	"path/filepath"
+	"sync"
+	"time"
+
+	"golang.org/x/sys/unix"
+)
+
+// Watcher watches a set of files, delivering events to a channel.
+type Watcher struct {
+	Events chan Event
+	Errors chan error
+	done   chan struct{} // Channel for sending a "quit message" to the reader goroutine
+
+	kq int // File descriptor (as returned by the kqueue() syscall).
+
+	mu              sync.Mutex        // Protects access to watcher data
+	watches         map[string]int    // Map of watched file descriptors (key: path).
+	externalWatches map[string]bool   // Map of watches added by user of the library.
+	dirFlags        map[string]uint32 // Map of watched directories to fflags used in kqueue.
+	paths           map[int]pathInfo  // Map file descriptors to path names for processing kqueue events.
+	fileExists      map[string]bool   // Keep track of if we know this file exists (to stop duplicate create events).
+	isClosed        bool              // Set to true when Close() is first called
+}
+
+type pathInfo struct {
+	name  string
+	isDir bool
+}
+
+// NewWatcher establishes a new watcher with the underlying OS and begins waiting for events.
+func NewWatcher() (*Watcher, error) {
+	kq, err := kqueue()
+	if err != nil {
+		return nil, err
+	}
+
+	w := &Watcher{
+		kq:              kq,
+		watches:         make(map[string]int),
+		dirFlags:        make(map[string]uint32),
+		paths:           make(map[int]pathInfo),
+		fileExists:      make(map[string]bool),
+		externalWatches: make(map[string]bool),
+		Events:          make(chan Event),
+		Errors:          make(chan error),
+		done:            make(chan struct{}),
+	}
+
+	go w.readEvents()
+	return w, nil
+}
+
+// Close removes all watches and closes the events channel.
+func (w *Watcher) Close() error {
+	w.mu.Lock()
+	if w.isClosed {
+		w.mu.Unlock()
+		return nil
+	}
+	w.isClosed = true
+
+	// copy paths to remove while locked
+	var pathsToRemove = make([]string, 0, len(w.watches))
+	for name := range w.watches {
+		pathsToRemove = append(pathsToRemove, name)
+	}
+	w.mu.Unlock()
+	// unlock before calling Remove, which also locks
+
+	for _, name := range pathsToRemove {
+		w.Remove(name)
+	}
+
+	// send a "quit" message to the reader goroutine
+	close(w.done)
+
+	return nil
+}
+
+// Add starts watching the named file or directory (non-recursively).
+func (w *Watcher) Add(name string) error {
+	w.mu.Lock()
+	w.externalWatches[name] = true
+	w.mu.Unlock()
+	_, err := w.addWatch(name, noteAllEvents)
+	return err
+}
+
+// Remove stops watching the the named file or directory (non-recursively).
+func (w *Watcher) Remove(name string) error {
+	name = filepath.Clean(name)
+	w.mu.Lock()
+	watchfd, ok := w.watches[name]
+	w.mu.Unlock()
+	if !ok {
+		return fmt.Errorf("can't remove non-existent kevent watch for: %s", name)
+	}
+
+	const registerRemove = unix.EV_DELETE
+	if err := register(w.kq, []int{watchfd}, registerRemove, 0); err != nil {
+		return err
+	}
+
+	unix.Close(watchfd)
+
+	w.mu.Lock()
+	isDir := w.paths[watchfd].isDir
+	delete(w.watches, name)
+	delete(w.paths, watchfd)
+	delete(w.dirFlags, name)
+	w.mu.Unlock()
+
+	// Find all watched paths that are in this directory that are not external.
+	if isDir {
+		var pathsToRemove []string
+		w.mu.Lock()
+		for _, path := range w.paths {
+			wdir, _ := filepath.Split(path.name)
+			if filepath.Clean(wdir) == name {
+				if !w.externalWatches[path.name] {
+					pathsToRemove = append(pathsToRemove, path.name)
+				}
+			}
+		}
+		w.mu.Unlock()
+		for _, name := range pathsToRemove {
+			// Since these are internal, not much sense in propagating error
+			// to the user, as that will just confuse them with an error about
+			// a path they did not explicitly watch themselves.
+			w.Remove(name)
+		}
+	}
+
+	return nil
+}
+
+// Watch all events (except NOTE_EXTEND, NOTE_LINK, NOTE_REVOKE)
+const noteAllEvents = unix.NOTE_DELETE | unix.NOTE_WRITE | unix.NOTE_ATTRIB | unix.NOTE_RENAME
+
+// keventWaitTime to block on each read from kevent
+var keventWaitTime = durationToTimespec(100 * time.Millisecond)
+
+// addWatch adds name to the watched file set.
+// The flags are interpreted as described in kevent(2).
+// Returns the real path to the file which was added, if any, which may be different from the one passed in the case of symlinks.
+func (w *Watcher) addWatch(name string, flags uint32) (string, error) {
+	var isDir bool
+	// Make ./name and name equivalent
+	name = filepath.Clean(name)
+
+	w.mu.Lock()
+	if w.isClosed {
+		w.mu.Unlock()
+		return "", errors.New("kevent instance already closed")
+	}
+	watchfd, alreadyWatching := w.watches[name]
+	// We already have a watch, but we can still override flags.
+	if alreadyWatching {
+		isDir = w.paths[watchfd].isDir
+	}
+	w.mu.Unlock()
+
+	if !alreadyWatching {
+		fi, err := os.Lstat(name)
+		if err != nil {
+			return "", err
+		}
+
+		// Don't watch sockets.
+		if fi.Mode()&os.ModeSocket == os.ModeSocket {
+			return "", nil
+		}
+
+		// Don't watch named pipes.
+		if fi.Mode()&os.ModeNamedPipe == os.ModeNamedPipe {
+			return "", nil
+		}
+
+		// Follow Symlinks
+		// Unfortunately, Linux can add bogus symlinks to watch list without
+		// issue, and Windows can't do symlinks period (AFAIK). To  maintain
+		// consistency, we will act like everything is fine. There will simply
+		// be no file events for broken symlinks.
+		// Hence the returns of nil on errors.
+		if fi.Mode()&os.ModeSymlink == os.ModeSymlink {
+			name, err = filepath.EvalSymlinks(name)
+			if err != nil {
+				return "", nil
+			}
+
+			w.mu.Lock()
+			_, alreadyWatching = w.watches[name]
+			w.mu.Unlock()
+
+			if alreadyWatching {
+				return name, nil
+			}
+
+			fi, err = os.Lstat(name)
+			if err != nil {
+				return "", nil
+			}
+		}
+
+		watchfd, err = unix.Open(name, openMode, 0700)
+		if watchfd == -1 {
+			return "", err
+		}
+
+		isDir = fi.IsDir()
+	}
+
+	const registerAdd = unix.EV_ADD | unix.EV_CLEAR | unix.EV_ENABLE
+	if err := register(w.kq, []int{watchfd}, registerAdd, flags); err != nil {
+		unix.Close(watchfd)
+		return "", err
+	}
+
+	if !alreadyWatching {
+		w.mu.Lock()
+		w.watches[name] = watchfd
+		w.paths[watchfd] = pathInfo{name: name, isDir: isDir}
+		w.mu.Unlock()
+	}
+
+	if isDir {
+		// Watch the directory if it has not been watched before,
+		// or if it was watched before, but perhaps only a NOTE_DELETE (watchDirectoryFiles)
+		w.mu.Lock()
+
+		watchDir := (flags&unix.NOTE_WRITE) == unix.NOTE_WRITE &&
+			(!alreadyWatching || (w.dirFlags[name]&unix.NOTE_WRITE) != unix.NOTE_WRITE)
+		// Store flags so this watch can be updated later
+		w.dirFlags[name] = flags
+		w.mu.Unlock()
+
+		if watchDir {
+			if err := w.watchDirectoryFiles(name); err != nil {
+				return "", err
+			}
+		}
+	}
+	return name, nil
+}
+
+// readEvents reads from kqueue and converts the received kevents into
+// Event values that it sends down the Events channel.
+func (w *Watcher) readEvents() {
+	eventBuffer := make([]unix.Kevent_t, 10)
+
+loop:
+	for {
+		// See if there is a message on the "done" channel
+		select {
+		case <-w.done:
+			break loop
+		default:
+		}
+
+		// Get new events
+		kevents, err := read(w.kq, eventBuffer, &keventWaitTime)
+		// EINTR is okay, the syscall was interrupted before timeout expired.
+		if err != nil && err != unix.EINTR {
+			select {
+			case w.Errors <- err:
+			case <-w.done:
+				break loop
+			}
+			continue
+		}
+
+		// Flush the events we received to the Events channel
+		for len(kevents) > 0 {
+			kevent := &kevents[0]
+			watchfd := int(kevent.Ident)
+			mask := uint32(kevent.Fflags)
+			w.mu.Lock()
+			path := w.paths[watchfd]
+			w.mu.Unlock()
+			event := newEvent(path.name, mask)
+
+			if path.isDir && !(event.Op&Remove == Remove) {
+				// Double check to make sure the directory exists. This can happen when
+				// we do a rm -fr on a recursively watched folders and we receive a
+				// modification event first but the folder has been deleted and later
+				// receive the delete event
+				if _, err := os.Lstat(event.Name); os.IsNotExist(err) {
+					// mark is as delete event
+					event.Op |= Remove
+				}
+			}
+
+			if event.Op&Rename == Rename || event.Op&Remove == Remove {
+				w.Remove(event.Name)
+				w.mu.Lock()
+				delete(w.fileExists, event.Name)
+				w.mu.Unlock()
+			}
+
+			if path.isDir && event.Op&Write == Write && !(event.Op&Remove == Remove) {
+				w.sendDirectoryChangeEvents(event.Name)
+			} else {
+				// Send the event on the Events channel.
+				select {
+				case w.Events <- event:
+				case <-w.done:
+					break loop
+				}
+			}
+
+			if event.Op&Remove == Remove {
+				// Look for a file that may have overwritten this.
+				// For example, mv f1 f2 will delete f2, then create f2.
+				if path.isDir {
+					fileDir := filepath.Clean(event.Name)
+					w.mu.Lock()
+					_, found := w.watches[fileDir]
+					w.mu.Unlock()
+					if found {
+						// make sure the directory exists before we watch for changes. When we
+						// do a recursive watch and perform rm -fr, the parent directory might
+						// have gone missing, ignore the missing directory and let the
+						// upcoming delete event remove the watch from the parent directory.
+						if _, err := os.Lstat(fileDir); err == nil {
+							w.sendDirectoryChangeEvents(fileDir)
+						}
+					}
+				} else {
+					filePath := filepath.Clean(event.Name)
+					if fileInfo, err := os.Lstat(filePath); err == nil {
+						w.sendFileCreatedEventIfNew(filePath, fileInfo)
+					}
+				}
+			}
+
+			// Move to next event
+			kevents = kevents[1:]
+		}
+	}
+
+	// cleanup
+	err := unix.Close(w.kq)
+	if err != nil {
+		// only way the previous loop breaks is if w.done was closed so we need to async send to w.Errors.
+		select {
+		case w.Errors <- err:
+		default:
+		}
+	}
+	close(w.Events)
+	close(w.Errors)
+}
+
+// newEvent returns an platform-independent Event based on kqueue Fflags.
+func newEvent(name string, mask uint32) Event {
+	e := Event{Name: name}
+	if mask&unix.NOTE_DELETE == unix.NOTE_DELETE {
+		e.Op |= Remove
+	}
+	if mask&unix.NOTE_WRITE == unix.NOTE_WRITE {
+		e.Op |= Write
+	}
+	if mask&unix.NOTE_RENAME == unix.NOTE_RENAME {
+		e.Op |= Rename
+	}
+	if mask&unix.NOTE_ATTRIB == unix.NOTE_ATTRIB {
+		e.Op |= Chmod
+	}
+	return e
+}
+
+func newCreateEvent(name string) Event {
+	return Event{Name: name, Op: Create}
+}
+
+// watchDirectoryFiles to mimic inotify when adding a watch on a directory
+func (w *Watcher) watchDirectoryFiles(dirPath string) error {
+	// Get all files
+	files, err := ioutil.ReadDir(dirPath)
+	if err != nil {
+		return err
+	}
+
+	for _, fileInfo := range files {
+		filePath := filepath.Join(dirPath, fileInfo.Name())
+		filePath, err = w.internalWatch(filePath, fileInfo)
+		if err != nil {
+			return err
+		}
+
+		w.mu.Lock()
+		w.fileExists[filePath] = true
+		w.mu.Unlock()
+	}
+
+	return nil
+}
+
+// sendDirectoryEvents searches the directory for newly created files
+// and sends them over the event channel. This functionality is to have
+// the BSD version of fsnotify match Linux inotify which provides a
+// create event for files created in a watched directory.
+func (w *Watcher) sendDirectoryChangeEvents(dirPath string) {
+	// Get all files
+	files, err := ioutil.ReadDir(dirPath)
+	if err != nil {
+		select {
+		case w.Errors <- err:
+		case <-w.done:
+			return
+		}
+	}
+
+	// Search for new files
+	for _, fileInfo := range files {
+		filePath := filepath.Join(dirPath, fileInfo.Name())
+		err := w.sendFileCreatedEventIfNew(filePath, fileInfo)
+
+		if err != nil {
+			return
+		}
+	}
+}
+
+// sendFileCreatedEvent sends a create event if the file isn't already being tracked.
+func (w *Watcher) sendFileCreatedEventIfNew(filePath string, fileInfo os.FileInfo) (err error) {
+	w.mu.Lock()
+	_, doesExist := w.fileExists[filePath]
+	w.mu.Unlock()
+	if !doesExist {
+		// Send create event
+		select {
+		case w.Events <- newCreateEvent(filePath):
+		case <-w.done:
+			return
+		}
+	}
+
+	// like watchDirectoryFiles (but without doing another ReadDir)
+	filePath, err = w.internalWatch(filePath, fileInfo)
+	if err != nil {
+		return err
+	}
+
+	w.mu.Lock()
+	w.fileExists[filePath] = true
+	w.mu.Unlock()
+
+	return nil
+}
+
+func (w *Watcher) internalWatch(name string, fileInfo os.FileInfo) (string, error) {
+	if fileInfo.IsDir() {
+		// mimic Linux providing delete events for subdirectories
+		// but preserve the flags used if currently watching subdirectory
+		w.mu.Lock()
+		flags := w.dirFlags[name]
+		w.mu.Unlock()
+
+		flags |= unix.NOTE_DELETE | unix.NOTE_RENAME
+		return w.addWatch(name, flags)
+	}
+
+	// watch file to mimic Linux inotify
+	return w.addWatch(name, noteAllEvents)
+}
+
+// kqueue creates a new kernel event queue and returns a descriptor.
+func kqueue() (kq int, err error) {
+	kq, err = unix.Kqueue()
+	if kq == -1 {
+		return kq, err
+	}
+	return kq, nil
+}
+
+// register events with the queue
+func register(kq int, fds []int, flags int, fflags uint32) error {
+	changes := make([]unix.Kevent_t, len(fds))
+
+	for i, fd := range fds {
+		// SetKevent converts int to the platform-specific types:
+		unix.SetKevent(&changes[i], fd, unix.EVFILT_VNODE, flags)
+		changes[i].Fflags = fflags
+	}
+
+	// register the events
+	success, err := unix.Kevent(kq, changes, nil, nil)
+	if success == -1 {
+		return err
+	}
+	return nil
+}
+
+// read retrieves pending events, or waits until an event occurs.
+// A timeout of nil blocks indefinitely, while 0 polls the queue.
+func read(kq int, events []unix.Kevent_t, timeout *unix.Timespec) ([]unix.Kevent_t, error) {
+	n, err := unix.Kevent(kq, nil, events, timeout)
+	if err != nil {
+		return nil, err
+	}
+	return events[0:n], nil
+}
+
+// durationToTimespec prepares a timeout value
+func durationToTimespec(d time.Duration) unix.Timespec {
+	return unix.NsecToTimespec(d.Nanoseconds())
+}
@@ -0,0 +1,11 @@
+// Copyright 2013 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build freebsd openbsd netbsd dragonfly
+
+package fsnotify
+
+import "golang.org/x/sys/unix"
+
+const openMode = unix.O_NONBLOCK | unix.O_RDONLY | unix.O_CLOEXEC
@@ -0,0 +1,12 @@
+// Copyright 2013 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build darwin
+
+package fsnotify
+
+import "golang.org/x/sys/unix"
+
+// note: this constant is not defined on BSD
+const openMode = unix.O_EVTONLY | unix.O_CLOEXEC
@@ -0,0 +1,561 @@
+// Copyright 2011 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build windows
+
+package fsnotify
+
+import (
+	"errors"
+	"fmt"
+	"os"
+	"path/filepath"
+	"runtime"
+	"sync"
+	"syscall"
+	"unsafe"
+)
+
+// Watcher watches a set of files, delivering events to a channel.
+type Watcher struct {
+	Events   chan Event
+	Errors   chan error
+	isClosed bool           // Set to true when Close() is first called
+	mu       sync.Mutex     // Map access
+	port     syscall.Handle // Handle to completion port
+	watches  watchMap       // Map of watches (key: i-number)
+	input    chan *input    // Inputs to the reader are sent on this channel
+	quit     chan chan<- error
+}
+
+// NewWatcher establishes a new watcher with the underlying OS and begins waiting for events.
+func NewWatcher() (*Watcher, error) {
+	port, e := syscall.CreateIoCompletionPort(syscall.InvalidHandle, 0, 0, 0)
+	if e != nil {
+		return nil, os.NewSyscallError("CreateIoCompletionPort", e)
+	}
+	w := &Watcher{
+		port:    port,
+		watches: make(watchMap),
+		input:   make(chan *input, 1),
+		Events:  make(chan Event, 50),
+		Errors:  make(chan error),
+		quit:    make(chan chan<- error, 1),
+	}
+	go w.readEvents()
+	return w, nil
+}
+
+// Close removes all watches and closes the events channel.
+func (w *Watcher) Close() error {
+	if w.isClosed {
+		return nil
+	}
+	w.isClosed = true
+
+	// Send "quit" message to the reader goroutine
+	ch := make(chan error)
+	w.quit <- ch
+	if err := w.wakeupReader(); err != nil {
+		return err
+	}
+	return <-ch
+}
+
+// Add starts watching the named file or directory (non-recursively).
+func (w *Watcher) Add(name string) error {
+	if w.isClosed {
+		return errors.New("watcher already closed")
+	}
+	in := &input{
+		op:    opAddWatch,
+		path:  filepath.Clean(name),
+		flags: sysFSALLEVENTS,
+		reply: make(chan error),
+	}
+	w.input <- in
+	if err := w.wakeupReader(); err != nil {
+		return err
+	}
+	return <-in.reply
+}
+
+// Remove stops watching the the named file or directory (non-recursively).
+func (w *Watcher) Remove(name string) error {
+	in := &input{
+		op:    opRemoveWatch,
+		path:  filepath.Clean(name),
+		reply: make(chan error),
+	}
+	w.input <- in
+	if err := w.wakeupReader(); err != nil {
+		return err
+	}
+	return <-in.reply
+}
+
+const (
+	// Options for AddWatch
+	sysFSONESHOT = 0x80000000
+	sysFSONLYDIR = 0x1000000
+
+	// Events
+	sysFSACCESS     = 0x1
+	sysFSALLEVENTS  = 0xfff
+	sysFSATTRIB     = 0x4
+	sysFSCLOSE      = 0x18
+	sysFSCREATE     = 0x100
+	sysFSDELETE     = 0x200
+	sysFSDELETESELF = 0x400
+	sysFSMODIFY     = 0x2
+	sysFSMOVE       = 0xc0
+	sysFSMOVEDFROM  = 0x40
+	sysFSMOVEDTO    = 0x80
+	sysFSMOVESELF   = 0x800
+
+	// Special events
+	sysFSIGNORED   = 0x8000
+	sysFSQOVERFLOW = 0x4000
+)
+
+func newEvent(name string, mask uint32) Event {
+	e := Event{Name: name}
+	if mask&sysFSCREATE == sysFSCREATE || mask&sysFSMOVEDTO == sysFSMOVEDTO {
+		e.Op |= Create
+	}
+	if mask&sysFSDELETE == sysFSDELETE || mask&sysFSDELETESELF == sysFSDELETESELF {
+		e.Op |= Remove
+	}
+	if mask&sysFSMODIFY == sysFSMODIFY {
+		e.Op |= Write
+	}
+	if mask&sysFSMOVE == sysFSMOVE || mask&sysFSMOVESELF == sysFSMOVESELF || mask&sysFSMOVEDFROM == sysFSMOVEDFROM {
+		e.Op |= Rename
+	}
+	if mask&sysFSATTRIB == sysFSATTRIB {
+		e.Op |= Chmod
+	}
+	return e
+}
+
+const (
+	opAddWatch = iota
+	opRemoveWatch
+)
+
+const (
+	provisional uint64 = 1 << (32 + iota)
+)
+
+type input struct {
+	op    int
+	path  string
+	flags uint32
+	reply chan error
+}
+
+type inode struct {
+	handle syscall.Handle
+	volume uint32
+	index  uint64
+}
+
+type watch struct {
+	ov     syscall.Overlapped
+	ino    *inode            // i-number
+	path   string            // Directory path
+	mask   uint64            // Directory itself is being watched with these notify flags
+	names  map[string]uint64 // Map of names being watched and their notify flags
+	rename string            // Remembers the old name while renaming a file
+	buf    [4096]byte
+}
+
+type indexMap map[uint64]*watch
+type watchMap map[uint32]indexMap
+
+func (w *Watcher) wakeupReader() error {
+	e := syscall.PostQueuedCompletionStatus(w.port, 0, 0, nil)
+	if e != nil {
+		return os.NewSyscallError("PostQueuedCompletionStatus", e)
+	}
+	return nil
+}
+
+func getDir(pathname string) (dir string, err error) {
+	attr, e := syscall.GetFileAttributes(syscall.StringToUTF16Ptr(pathname))
+	if e != nil {
+		return "", os.NewSyscallError("GetFileAttributes", e)
+	}
+	if attr&syscall.FILE_ATTRIBUTE_DIRECTORY != 0 {
+		dir = pathname
+	} else {
+		dir, _ = filepath.Split(pathname)
+		dir = filepath.Clean(dir)
+	}
+	return
+}
+
+func getIno(path string) (ino *inode, err error) {
+	h, e := syscall.CreateFile(syscall.StringToUTF16Ptr(path),
+		syscall.FILE_LIST_DIRECTORY,
+		syscall.FILE_SHARE_READ|syscall.FILE_SHARE_WRITE|syscall.FILE_SHARE_DELETE,
+		nil, syscall.OPEN_EXISTING,
+		syscall.FILE_FLAG_BACKUP_SEMANTICS|syscall.FILE_FLAG_OVERLAPPED, 0)
+	if e != nil {
+		return nil, os.NewSyscallError("CreateFile", e)
+	}
+	var fi syscall.ByHandleFileInformation
+	if e = syscall.GetFileInformationByHandle(h, &fi); e != nil {
+		syscall.CloseHandle(h)
+		return nil, os.NewSyscallError("GetFileInformationByHandle", e)
+	}
+	ino = &inode{
+		handle: h,
+		volume: fi.VolumeSerialNumber,
+		index:  uint64(fi.FileIndexHigh)<<32 | uint64(fi.FileIndexLow),
+	}
+	return ino, nil
+}
+
+// Must run within the I/O thread.
+func (m watchMap) get(ino *inode) *watch {
+	if i := m[ino.volume]; i != nil {
+		return i[ino.index]
+	}
+	return nil
+}
+
+// Must run within the I/O thread.
+func (m watchMap) set(ino *inode, watch *watch) {
+	i := m[ino.volume]
+	if i == nil {
+		i = make(indexMap)
+		m[ino.volume] = i
+	}
+	i[ino.index] = watch
+}
+
+// Must run within the I/O thread.
+func (w *Watcher) addWatch(pathname string, flags uint64) error {
+	dir, err := getDir(pathname)
+	if err != nil {
+		return err
+	}
+	if flags&sysFSONLYDIR != 0 && pathname != dir {
+		return nil
+	}
+	ino, err := getIno(dir)
+	if err != nil {
+		return err
+	}
+	w.mu.Lock()
+	watchEntry := w.watches.get(ino)
+	w.mu.Unlock()
+	if watchEntry == nil {
+		if _, e := syscall.CreateIoCompletionPort(ino.handle, w.port, 0, 0); e != nil {
+			syscall.CloseHandle(ino.handle)
+			return os.NewSyscallError("CreateIoCompletionPort", e)
+		}
+		watchEntry = &watch{
+			ino:   ino,
+			path:  dir,
+			names: make(map[string]uint64),
+		}
+		w.mu.Lock()
+		w.watches.set(ino, watchEntry)
+		w.mu.Unlock()
+		flags |= provisional
+	} else {
+		syscall.CloseHandle(ino.handle)
+	}
+	if pathname == dir {
+		watchEntry.mask |= flags
+	} else {
+		watchEntry.names[filepath.Base(pathname)] |= flags
+	}
+	if err = w.startRead(watchEntry); err != nil {
+		return err
+	}
+	if pathname == dir {
+		watchEntry.mask &= ^provisional
+	} else {
+		watchEntry.names[filepath.Base(pathname)] &= ^provisional
+	}
+	return nil
+}
+
+// Must run within the I/O thread.
+func (w *Watcher) remWatch(pathname string) error {
+	dir, err := getDir(pathname)
+	if err != nil {
+		return err
+	}
+	ino, err := getIno(dir)
+	if err != nil {
+		return err
+	}
+	w.mu.Lock()
+	watch := w.watches.get(ino)
+	w.mu.Unlock()
+	if watch == nil {
+		return fmt.Errorf("can't remove non-existent watch for: %s", pathname)
+	}
+	if pathname == dir {
+		w.sendEvent(watch.path, watch.mask&sysFSIGNORED)
+		watch.mask = 0
+	} else {
+		name := filepath.Base(pathname)
+		w.sendEvent(filepath.Join(watch.path, name), watch.names[name]&sysFSIGNORED)
+		delete(watch.names, name)
+	}
+	return w.startRead(watch)
+}
+
+// Must run within the I/O thread.
+func (w *Watcher) deleteWatch(watch *watch) {
+	for name, mask := range watch.names {
+		if mask&provisional == 0 {
+			w.sendEvent(filepath.Join(watch.path, name), mask&sysFSIGNORED)
+		}
+		delete(watch.names, name)
+	}
+	if watch.mask != 0 {
+		if watch.mask&provisional == 0 {
+			w.sendEvent(watch.path, watch.mask&sysFSIGNORED)
+		}
+		watch.mask = 0
+	}
+}
+
+// Must run within the I/O thread.
+func (w *Watcher) startRead(watch *watch) error {
+	if e := syscall.CancelIo(watch.ino.handle); e != nil {
+		w.Errors <- os.NewSyscallError("CancelIo", e)
+		w.deleteWatch(watch)
+	}
+	mask := toWindowsFlags(watch.mask)
+	for _, m := range watch.names {
+		mask |= toWindowsFlags(m)
+	}
+	if mask == 0 {
+		if e := syscall.CloseHandle(watch.ino.handle); e != nil {
+			w.Errors <- os.NewSyscallError("CloseHandle", e)
+		}
+		w.mu.Lock()
+		delete(w.watches[watch.ino.volume], watch.ino.index)
+		w.mu.Unlock()
+		return nil
+	}
+	e := syscall.ReadDirectoryChanges(watch.ino.handle, &watch.buf[0],
+		uint32(unsafe.Sizeof(watch.buf)), false, mask, nil, &watch.ov, 0)
+	if e != nil {
+		err := os.NewSyscallError("ReadDirectoryChanges", e)
+		if e == syscall.ERROR_ACCESS_DENIED && watch.mask&provisional == 0 {
+			// Watched directory was probably removed
+			if w.sendEvent(watch.path, watch.mask&sysFSDELETESELF) {
+				if watch.mask&sysFSONESHOT != 0 {
+					watch.mask = 0
+				}
+			}
+			err = nil
+		}
+		w.deleteWatch(watch)
+		w.startRead(watch)
+		return err
+	}
+	return nil
+}
+
+// readEvents reads from the I/O completion port, converts the
+// received events into Event objects and sends them via the Events channel.
+// Entry point to the I/O thread.
+func (w *Watcher) readEvents() {
+	var (
+		n, key uint32
+		ov     *syscall.Overlapped
+	)
+	runtime.LockOSThread()
+
+	for {
+		e := syscall.GetQueuedCompletionStatus(w.port, &n, &key, &ov, syscall.INFINITE)
+		watch := (*watch)(unsafe.Pointer(ov))
+
+		if watch == nil {
+			select {
+			case ch := <-w.quit:
+				w.mu.Lock()
+				var indexes []indexMap
+				for _, index := range w.watches {
+					indexes = append(indexes, index)
+				}
+				w.mu.Unlock()
+				for _, index := range indexes {
+					for _, watch := range index {
+						w.deleteWatch(watch)
+						w.startRead(watch)
+					}
+				}
+				var err error
+				if e := syscall.CloseHandle(w.port); e != nil {
+					err = os.NewSyscallError("CloseHandle", e)
+				}
+				close(w.Events)
+				close(w.Errors)
+				ch <- err
+				return
+			case in := <-w.input:
+				switch in.op {
+				case opAddWatch:
+					in.reply <- w.addWatch(in.path, uint64(in.flags))
+				case opRemoveWatch:
+					in.reply <- w.remWatch(in.path)
+				}
+			default:
+			}
+			continue
+		}
+
+		switch e {
+		case syscall.ERROR_MORE_DATA:
+			if watch == nil {
+				w.Errors <- errors.New("ERROR_MORE_DATA has unexpectedly null lpOverlapped buffer")
+			} else {
+				// The i/o succeeded but the buffer is full.
+				// In theory we should be building up a full packet.
+				// In practice we can get away with just carrying on.
+				n = uint32(unsafe.Sizeof(watch.buf))
+			}
+		case syscall.ERROR_ACCESS_DENIED:
+			// Watched directory was probably removed
+			w.sendEvent(watch.path, watch.mask&sysFSDELETESELF)
+			w.deleteWatch(watch)
+			w.startRead(watch)
+			continue
+		case syscall.ERROR_OPERATION_ABORTED:
+			// CancelIo was called on this handle
+			continue
+		default:
+			w.Errors <- os.NewSyscallError("GetQueuedCompletionPort", e)
+			continue
+		case nil:
+		}
+
+		var offset uint32
+		for {
+			if n == 0 {
+				w.Events <- newEvent("", sysFSQOVERFLOW)
+				w.Errors <- errors.New("short read in readEvents()")
+				break
+			}
+
+			// Point "raw" to the event in the buffer
+			raw := (*syscall.FileNotifyInformation)(unsafe.Pointer(&watch.buf[offset]))
+			buf := (*[syscall.MAX_PATH]uint16)(unsafe.Pointer(&raw.FileName))
+			name := syscall.UTF16ToString(buf[:raw.FileNameLength/2])
+			fullname := filepath.Join(watch.path, name)
+
+			var mask uint64
+			switch raw.Action {
+			case syscall.FILE_ACTION_REMOVED:
+				mask = sysFSDELETESELF
+			case syscall.FILE_ACTION_MODIFIED:
+				mask = sysFSMODIFY
+			case syscall.FILE_ACTION_RENAMED_OLD_NAME:
+				watch.rename = name
+			case syscall.FILE_ACTION_RENAMED_NEW_NAME:
+				if watch.names[watch.rename] != 0 {
+					watch.names[name] |= watch.names[watch.rename]
+					delete(watch.names, watch.rename)
+					mask = sysFSMOVESELF
+				}
+			}
+
+			sendNameEvent := func() {
+				if w.sendEvent(fullname, watch.names[name]&mask) {
+					if watch.names[name]&sysFSONESHOT != 0 {
+						delete(watch.names, name)
+					}
+				}
+			}
+			if raw.Action != syscall.FILE_ACTION_RENAMED_NEW_NAME {
+				sendNameEvent()
+			}
+			if raw.Action == syscall.FILE_ACTION_REMOVED {
+				w.sendEvent(fullname, watch.names[name]&sysFSIGNORED)
+				delete(watch.names, name)
+			}
+			if w.sendEvent(fullname, watch.mask&toFSnotifyFlags(raw.Action)) {
+				if watch.mask&sysFSONESHOT != 0 {
+					watch.mask = 0
+				}
+			}
+			if raw.Action == syscall.FILE_ACTION_RENAMED_NEW_NAME {
+				fullname = filepath.Join(watch.path, watch.rename)
+				sendNameEvent()
+			}
+
+			// Move to the next event in the buffer
+			if raw.NextEntryOffset == 0 {
+				break
+			}
+			offset += raw.NextEntryOffset
+
+			// Error!
+			if offset >= n {
+				w.Errors <- errors.New("Windows system assumed buffer larger than it is, events have likely been missed.")
+				break
+			}
+		}
+
+		if err := w.startRead(watch); err != nil {
+			w.Errors <- err
+		}
+	}
+}
+
+func (w *Watcher) sendEvent(name string, mask uint64) bool {
+	if mask == 0 {
+		return false
+	}
+	event := newEvent(name, uint32(mask))
+	select {
+	case ch := <-w.quit:
+		w.quit <- ch
+	case w.Events <- event:
+	}
+	return true
+}
+
+func toWindowsFlags(mask uint64) uint32 {
+	var m uint32
+	if mask&sysFSACCESS != 0 {
+		m |= syscall.FILE_NOTIFY_CHANGE_LAST_ACCESS
+	}
+	if mask&sysFSMODIFY != 0 {
+		m |= syscall.FILE_NOTIFY_CHANGE_LAST_WRITE
+	}
+	if mask&sysFSATTRIB != 0 {
+		m |= syscall.FILE_NOTIFY_CHANGE_ATTRIBUTES
+	}
+	if mask&(sysFSMOVE|sysFSCREATE|sysFSDELETE) != 0 {
+		m |= syscall.FILE_NOTIFY_CHANGE_FILE_NAME | syscall.FILE_NOTIFY_CHANGE_DIR_NAME
+	}
+	return m
+}
+
+func toFSnotifyFlags(action uint32) uint64 {
+	switch action {
+	case syscall.FILE_ACTION_ADDED:
+		return sysFSCREATE
+	case syscall.FILE_ACTION_REMOVED:
+		return sysFSDELETE
+	case syscall.FILE_ACTION_MODIFIED:
+		return sysFSMODIFY
+	case syscall.FILE_ACTION_RENAMED_OLD_NAME:
+		return sysFSMOVEDFROM
+	case syscall.FILE_ACTION_RENAMED_NEW_NAME:
+		return sysFSMOVEDTO
+	}
+	return 0
+}
@@ -0,0 +1,27 @@
+Copyright (c) 2017 The Go Authors. All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+   * Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+   * Redistributions in binary form must reproduce the above
+copyright notice, this list of conditions and the following disclaimer
+in the documentation and/or other materials provided with the
+distribution.
+   * Neither the name of Google Inc. nor the names of its
+contributors may be used to endorse or promote products derived from
+this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
@@ -0,0 +1,667 @@
+// Copyright 2017, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Package cmp determines equality of values.
+//
+// This package is intended to be a more powerful and safer alternative to
+// reflect.DeepEqual for comparing whether two values are semantically equal.
+// It is intended to only be used in tests, as performance is not a goal and
+// it may panic if it cannot compare the values. Its propensity towards
+// panicking means that its unsuitable for production environments where a
+// spurious panic may be fatal.
+//
+// The primary features of cmp are:
+//
+// • When the default behavior of equality does not suit the needs of the test,
+// custom equality functions can override the equality operation.
+// For example, an equality function may report floats as equal so long as they
+// are within some tolerance of each other.
+//
+// • Types that have an Equal method may use that method to determine equality.
+// This allows package authors to determine the equality operation for the types
+// that they define.
+//
+// • If no custom equality functions are used and no Equal method is defined,
+// equality is determined by recursively comparing the primitive kinds on both
+// values, much like reflect.DeepEqual. Unlike reflect.DeepEqual, unexported
+// fields are not compared by default; they result in panics unless suppressed
+// by using an Ignore option (see cmpopts.IgnoreUnexported) or explicitly
+// compared using the Exporter option.
+package cmp
+
+import (
+	"fmt"
+	"reflect"
+	"strings"
+
+	"github.com/google/go-cmp/cmp/internal/diff"
+	"github.com/google/go-cmp/cmp/internal/function"
+	"github.com/google/go-cmp/cmp/internal/value"
+)
+
+// TODO(≥go1.18): Use any instead of interface{}.
+
+// Equal reports whether x and y are equal by recursively applying the
+// following rules in the given order to x and y and all of their sub-values:
+//
+// • Let S be the set of all Ignore, Transformer, and Comparer options that
+// remain after applying all path filters, value filters, and type filters.
+// If at least one Ignore exists in S, then the comparison is ignored.
+// If the number of Transformer and Comparer options in S is greater than one,
+// then Equal panics because it is ambiguous which option to use.
+// If S contains a single Transformer, then use that to transform the current
+// values and recursively call Equal on the output values.
+// If S contains a single Comparer, then use that to compare the current values.
+// Otherwise, evaluation proceeds to the next rule.
+//
+// • If the values have an Equal method of the form "(T) Equal(T) bool" or
+// "(T) Equal(I) bool" where T is assignable to I, then use the result of
+// x.Equal(y) even if x or y is nil. Otherwise, no such method exists and
+// evaluation proceeds to the next rule.
+//
+// • Lastly, try to compare x and y based on their basic kinds.
+// Simple kinds like booleans, integers, floats, complex numbers, strings, and
+// channels are compared using the equivalent of the == operator in Go.
+// Functions are only equal if they are both nil, otherwise they are unequal.
+//
+// Structs are equal if recursively calling Equal on all fields report equal.
+// If a struct contains unexported fields, Equal panics unless an Ignore option
+// (e.g., cmpopts.IgnoreUnexported) ignores that field or the Exporter option
+// explicitly permits comparing the unexported field.
+//
+// Slices are equal if they are both nil or both non-nil, where recursively
+// calling Equal on all non-ignored slice or array elements report equal.
+// Empty non-nil slices and nil slices are not equal; to equate empty slices,
+// consider using cmpopts.EquateEmpty.
+//
+// Maps are equal if they are both nil or both non-nil, where recursively
+// calling Equal on all non-ignored map entries report equal.
+// Map keys are equal according to the == operator.
+// To use custom comparisons for map keys, consider using cmpopts.SortMaps.
+// Empty non-nil maps and nil maps are not equal; to equate empty maps,
+// consider using cmpopts.EquateEmpty.
+//
+// Pointers and interfaces are equal if they are both nil or both non-nil,
+// where they have the same underlying concrete type and recursively
+// calling Equal on the underlying values reports equal.
+//
+// Before recursing into a pointer, slice element, or map, the current path
+// is checked to detect whether the address has already been visited.
+// If there is a cycle, then the pointed at values are considered equal
+// only if both addresses were previously visited in the same path step.
+func Equal(x, y interface{}, opts ...Option) bool {
+	s := newState(opts)
+	s.compareAny(rootStep(x, y))
+	return s.result.Equal()
+}
+
+// Diff returns a human-readable report of the differences between two values:
+// y - x. It returns an empty string if and only if Equal returns true for the
+// same input values and options.
+//
+// The output is displayed as a literal in pseudo-Go syntax.
+// At the start of each line, a "-" prefix indicates an element removed from x,
+// a "+" prefix to indicates an element added from y, and the lack of a prefix
+// indicates an element common to both x and y. If possible, the output
+// uses fmt.Stringer.String or error.Error methods to produce more humanly
+// readable outputs. In such cases, the string is prefixed with either an
+// 's' or 'e' character, respectively, to indicate that the method was called.
+//
+// Do not depend on this output being stable. If you need the ability to
+// programmatically interpret the difference, consider using a custom Reporter.
+func Diff(x, y interface{}, opts ...Option) string {
+	s := newState(opts)
+
+	// Optimization: If there are no other reporters, we can optimize for the
+	// common case where the result is equal (and thus no reported difference).
+	// This avoids the expensive construction of a difference tree.
+	if len(s.reporters) == 0 {
+		s.compareAny(rootStep(x, y))
+		if s.result.Equal() {
+			return ""
+		}
+		s.result = diff.Result{} // Reset results
+	}
+
+	r := new(defaultReporter)
+	s.reporters = append(s.reporters, reporter{r})
+	s.compareAny(rootStep(x, y))
+	d := r.String()
+	if (d == "") != s.result.Equal() {
+		panic("inconsistent difference and equality results")
+	}
+	return d
+}
+
+// rootStep constructs the first path step. If x and y have differing types,
+// then they are stored within an empty interface type.
+func rootStep(x, y interface{}) PathStep {
+	vx := reflect.ValueOf(x)
+	vy := reflect.ValueOf(y)
+
+	// If the inputs are different types, auto-wrap them in an empty interface
+	// so that they have the same parent type.
+	var t reflect.Type
+	if !vx.IsValid() || !vy.IsValid() || vx.Type() != vy.Type() {
+		t = reflect.TypeOf((*interface{})(nil)).Elem()
+		if vx.IsValid() {
+			vvx := reflect.New(t).Elem()
+			vvx.Set(vx)
+			vx = vvx
+		}
+		if vy.IsValid() {
+			vvy := reflect.New(t).Elem()
+			vvy.Set(vy)
+			vy = vvy
+		}
+	} else {
+		t = vx.Type()
+	}
+
+	return &pathStep{t, vx, vy}
+}
+
+type state struct {
+	// These fields represent the "comparison state".
+	// Calling statelessCompare must not result in observable changes to these.
+	result    diff.Result // The current result of comparison
+	curPath   Path        // The current path in the value tree
+	curPtrs   pointerPath // The current set of visited pointers
+	reporters []reporter  // Optional reporters
+
+	// recChecker checks for infinite cycles applying the same set of
+	// transformers upon the output of itself.
+	recChecker recChecker
+
+	// dynChecker triggers pseudo-random checks for option correctness.
+	// It is safe for statelessCompare to mutate this value.
+	dynChecker dynChecker
+
+	// These fields, once set by processOption, will not change.
+	exporters []exporter // List of exporters for structs with unexported fields
+	opts      Options    // List of all fundamental and filter options
+}
+
+func newState(opts []Option) *state {
+	// Always ensure a validator option exists to validate the inputs.
+	s := &state{opts: Options{validator{}}}
+	s.curPtrs.Init()
+	s.processOption(Options(opts))
+	return s
+}
+
+func (s *state) processOption(opt Option) {
+	switch opt := opt.(type) {
+	case nil:
+	case Options:
+		for _, o := range opt {
+			s.processOption(o)
+		}
+	case coreOption:
+		type filtered interface {
+			isFiltered() bool
+		}
+		if fopt, ok := opt.(filtered); ok && !fopt.isFiltered() {
+			panic(fmt.Sprintf("cannot use an unfiltered option: %v", opt))
+		}
+		s.opts = append(s.opts, opt)
+	case exporter:
+		s.exporters = append(s.exporters, opt)
+	case reporter:
+		s.reporters = append(s.reporters, opt)
+	default:
+		panic(fmt.Sprintf("unknown option %T", opt))
+	}
+}
+
+// statelessCompare compares two values and returns the result.
+// This function is stateless in that it does not alter the current result,
+// or output to any registered reporters.
+func (s *state) statelessCompare(step PathStep) diff.Result {
+	// We do not save and restore curPath and curPtrs because all of the
+	// compareX methods should properly push and pop from them.
+	// It is an implementation bug if the contents of the paths differ from
+	// when calling this function to when returning from it.
+
+	oldResult, oldReporters := s.result, s.reporters
+	s.result = diff.Result{} // Reset result
+	s.reporters = nil        // Remove reporters to avoid spurious printouts
+	s.compareAny(step)
+	res := s.result
+	s.result, s.reporters = oldResult, oldReporters
+	return res
+}
+
+func (s *state) compareAny(step PathStep) {
+	// Update the path stack.
+	s.curPath.push(step)
+	defer s.curPath.pop()
+	for _, r := range s.reporters {
+		r.PushStep(step)
+		defer r.PopStep()
+	}
+	s.recChecker.Check(s.curPath)
+
+	// Cycle-detection for slice elements (see NOTE in compareSlice).
+	t := step.Type()
+	vx, vy := step.Values()
+	if si, ok := step.(SliceIndex); ok && si.isSlice && vx.IsValid() && vy.IsValid() {
+		px, py := vx.Addr(), vy.Addr()
+		if eq, visited := s.curPtrs.Push(px, py); visited {
+			s.report(eq, reportByCycle)
+			return
+		}
+		defer s.curPtrs.Pop(px, py)
+	}
+
+	// Rule 1: Check whether an option applies on this node in the value tree.
+	if s.tryOptions(t, vx, vy) {
+		return
+	}
+
+	// Rule 2: Check whether the type has a valid Equal method.
+	if s.tryMethod(t, vx, vy) {
+		return
+	}
+
+	// Rule 3: Compare based on the underlying kind.
+	switch t.Kind() {
+	case reflect.Bool:
+		s.report(vx.Bool() == vy.Bool(), 0)
+	case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
+		s.report(vx.Int() == vy.Int(), 0)
+	case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr:
+		s.report(vx.Uint() == vy.Uint(), 0)
+	case reflect.Float32, reflect.Float64:
+		s.report(vx.Float() == vy.Float(), 0)
+	case reflect.Complex64, reflect.Complex128:
+		s.report(vx.Complex() == vy.Complex(), 0)
+	case reflect.String:
+		s.report(vx.String() == vy.String(), 0)
+	case reflect.Chan, reflect.UnsafePointer:
+		s.report(vx.Pointer() == vy.Pointer(), 0)
+	case reflect.Func:
+		s.report(vx.IsNil() && vy.IsNil(), 0)
+	case reflect.Struct:
+		s.compareStruct(t, vx, vy)
+	case reflect.Slice, reflect.Array:
+		s.compareSlice(t, vx, vy)
+	case reflect.Map:
+		s.compareMap(t, vx, vy)
+	case reflect.Ptr:
+		s.comparePtr(t, vx, vy)
+	case reflect.Interface:
+		s.compareInterface(t, vx, vy)
+	default:
+		panic(fmt.Sprintf("%v kind not handled", t.Kind()))
+	}
+}
+
+func (s *state) tryOptions(t reflect.Type, vx, vy reflect.Value) bool {
+	// Evaluate all filters and apply the remaining options.
+	if opt := s.opts.filter(s, t, vx, vy); opt != nil {
+		opt.apply(s, vx, vy)
+		return true
+	}
+	return false
+}
+
+func (s *state) tryMethod(t reflect.Type, vx, vy reflect.Value) bool {
+	// Check if this type even has an Equal method.
+	m, ok := t.MethodByName("Equal")
+	if !ok || !function.IsType(m.Type, function.EqualAssignable) {
+		return false
+	}
+
+	eq := s.callTTBFunc(m.Func, vx, vy)
+	s.report(eq, reportByMethod)
+	return true
+}
+
+func (s *state) callTRFunc(f, v reflect.Value, step Transform) reflect.Value {
+	if !s.dynChecker.Next() {
+		return f.Call([]reflect.Value{v})[0]
+	}
+
+	// Run the function twice and ensure that we get the same results back.
+	// We run in goroutines so that the race detector (if enabled) can detect
+	// unsafe mutations to the input.
+	c := make(chan reflect.Value)
+	go detectRaces(c, f, v)
+	got := <-c
+	want := f.Call([]reflect.Value{v})[0]
+	if step.vx, step.vy = got, want; !s.statelessCompare(step).Equal() {
+		// To avoid false-positives with non-reflexive equality operations,
+		// we sanity check whether a value is equal to itself.
+		if step.vx, step.vy = want, want; !s.statelessCompare(step).Equal() {
+			return want
+		}
+		panic(fmt.Sprintf("non-deterministic function detected: %s", function.NameOf(f)))
+	}
+	return want
+}
+
+func (s *state) callTTBFunc(f, x, y reflect.Value) bool {
+	if !s.dynChecker.Next() {
+		return f.Call([]reflect.Value{x, y})[0].Bool()
+	}
+
+	// Swapping the input arguments is sufficient to check that
+	// f is symmetric and deterministic.
+	// We run in goroutines so that the race detector (if enabled) can detect
+	// unsafe mutations to the input.
+	c := make(chan reflect.Value)
+	go detectRaces(c, f, y, x)
+	got := <-c
+	want := f.Call([]reflect.Value{x, y})[0].Bool()
+	if !got.IsValid() || got.Bool() != want {
+		panic(fmt.Sprintf("non-deterministic or non-symmetric function detected: %s", function.NameOf(f)))
+	}
+	return want
+}
+
+func detectRaces(c chan<- reflect.Value, f reflect.Value, vs ...reflect.Value) {
+	var ret reflect.Value
+	defer func() {
+		recover() // Ignore panics, let the other call to f panic instead
+		c <- ret
+	}()
+	ret = f.Call(vs)[0]
+}
+
+func (s *state) compareStruct(t reflect.Type, vx, vy reflect.Value) {
+	var addr bool
+	var vax, vay reflect.Value // Addressable versions of vx and vy
+
+	var mayForce, mayForceInit bool
+	step := StructField{&structField{}}
+	for i := 0; i < t.NumField(); i++ {
+		step.typ = t.Field(i).Type
+		step.vx = vx.Field(i)
+		step.vy = vy.Field(i)
+		step.name = t.Field(i).Name
+		step.idx = i
+		step.unexported = !isExported(step.name)
+		if step.unexported {
+			if step.name == "_" {
+				continue
+			}
+			// Defer checking of unexported fields until later to give an
+			// Ignore a chance to ignore the field.
+			if !vax.IsValid() || !vay.IsValid() {
+				// For retrieveUnexportedField to work, the parent struct must
+				// be addressable. Create a new copy of the values if
+				// necessary to make them addressable.
+				addr = vx.CanAddr() || vy.CanAddr()
+				vax = makeAddressable(vx)
+				vay = makeAddressable(vy)
+			}
+			if !mayForceInit {
+				for _, xf := range s.exporters {
+					mayForce = mayForce || xf(t)
+				}
+				mayForceInit = true
+			}
+			step.mayForce = mayForce
+			step.paddr = addr
+			step.pvx = vax
+			step.pvy = vay
+			step.field = t.Field(i)
+		}
+		s.compareAny(step)
+	}
+}
+
+func (s *state) compareSlice(t reflect.Type, vx, vy reflect.Value) {
+	isSlice := t.Kind() == reflect.Slice
+	if isSlice && (vx.IsNil() || vy.IsNil()) {
+		s.report(vx.IsNil() && vy.IsNil(), 0)
+		return
+	}
+
+	// NOTE: It is incorrect to call curPtrs.Push on the slice header pointer
+	// since slices represents a list of pointers, rather than a single pointer.
+	// The pointer checking logic must be handled on a per-element basis
+	// in compareAny.
+	//
+	// A slice header (see reflect.SliceHeader) in Go is a tuple of a starting
+	// pointer P, a length N, and a capacity C. Supposing each slice element has
+	// a memory size of M, then the slice is equivalent to the list of pointers:
+	//	[P+i*M for i in range(N)]
+	//
+	// For example, v[:0] and v[:1] are slices with the same starting pointer,
+	// but they are clearly different values. Using the slice pointer alone
+	// violates the assumption that equal pointers implies equal values.
+
+	step := SliceIndex{&sliceIndex{pathStep: pathStep{typ: t.Elem()}, isSlice: isSlice}}
+	withIndexes := func(ix, iy int) SliceIndex {
+		if ix >= 0 {
+			step.vx, step.xkey = vx.Index(ix), ix
+		} else {
+			step.vx, step.xkey = reflect.Value{}, -1
+		}
+		if iy >= 0 {
+			step.vy, step.ykey = vy.Index(iy), iy
+		} else {
+			step.vy, step.ykey = reflect.Value{}, -1
+		}
+		return step
+	}
+
+	// Ignore options are able to ignore missing elements in a slice.
+	// However, detecting these reliably requires an optimal differencing
+	// algorithm, for which diff.Difference is not.
+	//
+	// Instead, we first iterate through both slices to detect which elements
+	// would be ignored if standing alone. The index of non-discarded elements
+	// are stored in a separate slice, which diffing is then performed on.
+	var indexesX, indexesY []int
+	var ignoredX, ignoredY []bool
+	for ix := 0; ix < vx.Len(); ix++ {
+		ignored := s.statelessCompare(withIndexes(ix, -1)).NumDiff == 0
+		if !ignored {
+			indexesX = append(indexesX, ix)
+		}
+		ignoredX = append(ignoredX, ignored)
+	}
+	for iy := 0; iy < vy.Len(); iy++ {
+		ignored := s.statelessCompare(withIndexes(-1, iy)).NumDiff == 0
+		if !ignored {
+			indexesY = append(indexesY, iy)
+		}
+		ignoredY = append(ignoredY, ignored)
+	}
+
+	// Compute an edit-script for slices vx and vy (excluding ignored elements).
+	edits := diff.Difference(len(indexesX), len(indexesY), func(ix, iy int) diff.Result {
+		return s.statelessCompare(withIndexes(indexesX[ix], indexesY[iy]))
+	})
+
+	// Replay the ignore-scripts and the edit-script.
+	var ix, iy int
+	for ix < vx.Len() || iy < vy.Len() {
+		var e diff.EditType
+		switch {
+		case ix < len(ignoredX) && ignoredX[ix]:
+			e = diff.UniqueX
+		case iy < len(ignoredY) && ignoredY[iy]:
+			e = diff.UniqueY
+		default:
+			e, edits = edits[0], edits[1:]
+		}
+		switch e {
+		case diff.UniqueX:
+			s.compareAny(withIndexes(ix, -1))
+			ix++
+		case diff.UniqueY:
+			s.compareAny(withIndexes(-1, iy))
+			iy++
+		default:
+			s.compareAny(withIndexes(ix, iy))
+			ix++
+			iy++
+		}
+	}
+}
+
+func (s *state) compareMap(t reflect.Type, vx, vy reflect.Value) {
+	if vx.IsNil() || vy.IsNil() {
+		s.report(vx.IsNil() && vy.IsNil(), 0)
+		return
+	}
+
+	// Cycle-detection for maps.
+	if eq, visited := s.curPtrs.Push(vx, vy); visited {
+		s.report(eq, reportByCycle)
+		return
+	}
+	defer s.curPtrs.Pop(vx, vy)
+
+	// We combine and sort the two map keys so that we can perform the
+	// comparisons in a deterministic order.
+	step := MapIndex{&mapIndex{pathStep: pathStep{typ: t.Elem()}}}
+	for _, k := range value.SortKeys(append(vx.MapKeys(), vy.MapKeys()...)) {
+		step.vx = vx.MapIndex(k)
+		step.vy = vy.MapIndex(k)
+		step.key = k
+		if !step.vx.IsValid() && !step.vy.IsValid() {
+			// It is possible for both vx and vy to be invalid if the
+			// key contained a NaN value in it.
+			//
+			// Even with the ability to retrieve NaN keys in Go 1.12,
+			// there still isn't a sensible way to compare the values since
+			// a NaN key may map to multiple unordered values.
+			// The most reasonable way to compare NaNs would be to compare the
+			// set of values. However, this is impossible to do efficiently
+			// since set equality is provably an O(n^2) operation given only
+			// an Equal function. If we had a Less function or Hash function,
+			// this could be done in O(n*log(n)) or O(n), respectively.
+			//
+			// Rather than adding complex logic to deal with NaNs, make it
+			// the user's responsibility to compare such obscure maps.
+			const help = "consider providing a Comparer to compare the map"
+			panic(fmt.Sprintf("%#v has map key with NaNs\n%s", s.curPath, help))
+		}
+		s.compareAny(step)
+	}
+}
+
+func (s *state) comparePtr(t reflect.Type, vx, vy reflect.Value) {
+	if vx.IsNil() || vy.IsNil() {
+		s.report(vx.IsNil() && vy.IsNil(), 0)
+		return
+	}
+
+	// Cycle-detection for pointers.
+	if eq, visited := s.curPtrs.Push(vx, vy); visited {
+		s.report(eq, reportByCycle)
+		return
+	}
+	defer s.curPtrs.Pop(vx, vy)
+
+	vx, vy = vx.Elem(), vy.Elem()
+	s.compareAny(Indirect{&indirect{pathStep{t.Elem(), vx, vy}}})
+}
+
+func (s *state) compareInterface(t reflect.Type, vx, vy reflect.Value) {
+	if vx.IsNil() || vy.IsNil() {
+		s.report(vx.IsNil() && vy.IsNil(), 0)
+		return
+	}
+	vx, vy = vx.Elem(), vy.Elem()
+	if vx.Type() != vy.Type() {
+		s.report(false, 0)
+		return
+	}
+	s.compareAny(TypeAssertion{&typeAssertion{pathStep{vx.Type(), vx, vy}}})
+}
+
+func (s *state) report(eq bool, rf resultFlags) {
+	if rf&reportByIgnore == 0 {
+		if eq {
+			s.result.NumSame++
+			rf |= reportEqual
+		} else {
+			s.result.NumDiff++
+			rf |= reportUnequal
+		}
+	}
+	for _, r := range s.reporters {
+		r.Report(Result{flags: rf})
+	}
+}
+
+// recChecker tracks the state needed to periodically perform checks that
+// user provided transformers are not stuck in an infinitely recursive cycle.
+type recChecker struct{ next int }
+
+// Check scans the Path for any recursive transformers and panics when any
+// recursive transformers are detected. Note that the presence of a
+// recursive Transformer does not necessarily imply an infinite cycle.
+// As such, this check only activates after some minimal number of path steps.
+func (rc *recChecker) Check(p Path) {
+	const minLen = 1 << 16
+	if rc.next == 0 {
+		rc.next = minLen
+	}
+	if len(p) < rc.next {
+		return
+	}
+	rc.next <<= 1
+
+	// Check whether the same transformer has appeared at least twice.
+	var ss []string
+	m := map[Option]int{}
+	for _, ps := range p {
+		if t, ok := ps.(Transform); ok {
+			t := t.Option()
+			if m[t] == 1 { // Transformer was used exactly once before
+				tf := t.(*transformer).fnc.Type()
+				ss = append(ss, fmt.Sprintf("%v: %v => %v", t, tf.In(0), tf.Out(0)))
+			}
+			m[t]++
+		}
+	}
+	if len(ss) > 0 {
+		const warning = "recursive set of Transformers detected"
+		const help = "consider using cmpopts.AcyclicTransformer"
+		set := strings.Join(ss, "\n\t")
+		panic(fmt.Sprintf("%s:\n\t%s\n%s", warning, set, help))
+	}
+}
+
+// dynChecker tracks the state needed to periodically perform checks that
+// user provided functions are symmetric and deterministic.
+// The zero value is safe for immediate use.
+type dynChecker struct{ curr, next int }
+
+// Next increments the state and reports whether a check should be performed.
+//
+// Checks occur every Nth function call, where N is a triangular number:
+//	0 1 3 6 10 15 21 28 36 45 55 66 78 91 105 120 136 153 171 190 ...
+// See https://en.wikipedia.org/wiki/Triangular_number
+//
+// This sequence ensures that the cost of checks drops significantly as
+// the number of functions calls grows larger.
+func (dc *dynChecker) Next() bool {
+	ok := dc.curr == dc.next
+	if ok {
+		dc.curr = 0
+		dc.next++
+	}
+	dc.curr++
+	return ok
+}
+
+// makeAddressable returns a value that is always addressable.
+// It returns the input verbatim if it is already addressable,
+// otherwise it creates a new value and returns an addressable copy.
+func makeAddressable(v reflect.Value) reflect.Value {
+	if v.CanAddr() {
+		return v
+	}
+	vc := reflect.New(v.Type()).Elem()
+	vc.Set(v)
+	return vc
+}
@@ -0,0 +1,16 @@
+// Copyright 2017, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build purego
+// +build purego
+
+package cmp
+
+import "reflect"
+
+const supportExporters = false
+
+func retrieveUnexportedField(reflect.Value, reflect.StructField, bool) reflect.Value {
+	panic("no support for forcibly accessing unexported fields")
+}
@@ -0,0 +1,36 @@
+// Copyright 2017, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build !purego
+// +build !purego
+
+package cmp
+
+import (
+	"reflect"
+	"unsafe"
+)
+
+const supportExporters = true
+
+// retrieveUnexportedField uses unsafe to forcibly retrieve any field from
+// a struct such that the value has read-write permissions.
+//
+// The parent struct, v, must be addressable, while f must be a StructField
+// describing the field to retrieve. If addr is false,
+// then the returned value will be shallowed copied to be non-addressable.
+func retrieveUnexportedField(v reflect.Value, f reflect.StructField, addr bool) reflect.Value {
+	ve := reflect.NewAt(f.Type, unsafe.Pointer(uintptr(unsafe.Pointer(v.UnsafeAddr()))+f.Offset)).Elem()
+	if !addr {
+		// A field is addressable if and only if the struct is addressable.
+		// If the original parent value was not addressable, shallow copy the
+		// value to make it non-addressable to avoid leaking an implementation
+		// detail of how forcibly exporting a field works.
+		if ve.Kind() == reflect.Interface && ve.IsNil() {
+			return reflect.Zero(f.Type)
+		}
+		return reflect.ValueOf(ve.Interface()).Convert(f.Type)
+	}
+	return ve
+}
@@ -0,0 +1,18 @@
+// Copyright 2017, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build !cmp_debug
+// +build !cmp_debug
+
+package diff
+
+var debug debugger
+
+type debugger struct{}
+
+func (debugger) Begin(_, _ int, f EqualFunc, _, _ *EditScript) EqualFunc {
+	return f
+}
+func (debugger) Update() {}
+func (debugger) Finish() {}
@@ -0,0 +1,123 @@
+// Copyright 2017, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build cmp_debug
+// +build cmp_debug
+
+package diff
+
+import (
+	"fmt"
+	"strings"
+	"sync"
+	"time"
+)
+
+// The algorithm can be seen running in real-time by enabling debugging:
+//	go test -tags=cmp_debug -v
+//
+// Example output:
+//	=== RUN   TestDifference/#34
+//	┌───────────────────────────────┐
+//	│ \ · · · · · · · · · · · · · · │
+//	│ · # · · · · · · · · · · · · · │
+//	│ · \ · · · · · · · · · · · · · │
+//	│ · · \ · · · · · · · · · · · · │
+//	│ · · · X # · · · · · · · · · · │
+//	│ · · · # \ · · · · · · · · · · │
+//	│ · · · · · # # · · · · · · · · │
+//	│ · · · · · # \ · · · · · · · · │
+//	│ · · · · · · · \ · · · · · · · │
+//	│ · · · · · · · · \ · · · · · · │
+//	│ · · · · · · · · · \ · · · · · │
+//	│ · · · · · · · · · · \ · · # · │
+//	│ · · · · · · · · · · · \ # # · │
+//	│ · · · · · · · · · · · # # # · │
+//	│ · · · · · · · · · · # # # # · │
+//	│ · · · · · · · · · # # # # # · │
+//	│ · · · · · · · · · · · · · · \ │
+//	└───────────────────────────────┘
+//	[.Y..M.XY......YXYXY.|]
+//
+// The grid represents the edit-graph where the horizontal axis represents
+// list X and the vertical axis represents list Y. The start of the two lists
+// is the top-left, while the ends are the bottom-right. The '·' represents
+// an unexplored node in the graph. The '\' indicates that the two symbols
+// from list X and Y are equal. The 'X' indicates that two symbols are similar
+// (but not exactly equal) to each other. The '#' indicates that the two symbols
+// are different (and not similar). The algorithm traverses this graph trying to
+// make the paths starting in the top-left and the bottom-right connect.
+//
+// The series of '.', 'X', 'Y', and 'M' characters at the bottom represents
+// the currently established path from the forward and reverse searches,
+// separated by a '|' character.
+
+const (
+	updateDelay  = 100 * time.Millisecond
+	finishDelay  = 500 * time.Millisecond
+	ansiTerminal = true // ANSI escape codes used to move terminal cursor
+)
+
+var debug debugger
+
+type debugger struct {
+	sync.Mutex
+	p1, p2           EditScript
+	fwdPath, revPath *EditScript
+	grid             []byte
+	lines            int
+}
+
+func (dbg *debugger) Begin(nx, ny int, f EqualFunc, p1, p2 *EditScript) EqualFunc {
+	dbg.Lock()
+	dbg.fwdPath, dbg.revPath = p1, p2
+	top := "┌─" + strings.Repeat("──", nx) + "┐\n"
+	row := "│ " + strings.Repeat("· ", nx) + "│\n"
+	btm := "└─" + strings.Repeat("──", nx) + "┘\n"
+	dbg.grid = []byte(top + strings.Repeat(row, ny) + btm)
+	dbg.lines = strings.Count(dbg.String(), "\n")
+	fmt.Print(dbg)
+
+	// Wrap the EqualFunc so that we can intercept each result.
+	return func(ix, iy int) (r Result) {
+		cell := dbg.grid[len(top)+iy*len(row):][len("│ ")+len("· ")*ix:][:len("·")]
+		for i := range cell {
+			cell[i] = 0 // Zero out the multiple bytes of UTF-8 middle-dot
+		}
+		switch r = f(ix, iy); {
+		case r.Equal():
+			cell[0] = '\\'
+		case r.Similar():
+			cell[0] = 'X'
+		default:
+			cell[0] = '#'
+		}
+		return
+	}
+}
+
+func (dbg *debugger) Update() {
+	dbg.print(updateDelay)
+}
+
+func (dbg *debugger) Finish() {
+	dbg.print(finishDelay)
+	dbg.Unlock()
+}
+
+func (dbg *debugger) String() string {
+	dbg.p1, dbg.p2 = *dbg.fwdPath, dbg.p2[:0]
+	for i := len(*dbg.revPath) - 1; i >= 0; i-- {
+		dbg.p2 = append(dbg.p2, (*dbg.revPath)[i])
+	}
+	return fmt.Sprintf("%s[%v|%v]\n\n", dbg.grid, dbg.p1, dbg.p2)
+}
+
+func (dbg *debugger) print(d time.Duration) {
+	if ansiTerminal {
+		fmt.Printf("\x1b[%dA", dbg.lines) // Reset terminal cursor
+	}
+	fmt.Print(dbg)
+	time.Sleep(d)
+}
@@ -0,0 +1,398 @@
+// Copyright 2017, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Package diff implements an algorithm for producing edit-scripts.
+// The edit-script is a sequence of operations needed to transform one list
+// of symbols into another (or vice-versa). The edits allowed are insertions,
+// deletions, and modifications. The summation of all edits is called the
+// Levenshtein distance as this problem is well-known in computer science.
+//
+// This package prioritizes performance over accuracy. That is, the run time
+// is more important than obtaining a minimal Levenshtein distance.
+package diff
+
+import (
+	"math/rand"
+	"time"
+
+	"github.com/google/go-cmp/cmp/internal/flags"
+)
+
+// EditType represents a single operation within an edit-script.
+type EditType uint8
+
+const (
+	// Identity indicates that a symbol pair is identical in both list X and Y.
+	Identity EditType = iota
+	// UniqueX indicates that a symbol only exists in X and not Y.
+	UniqueX
+	// UniqueY indicates that a symbol only exists in Y and not X.
+	UniqueY
+	// Modified indicates that a symbol pair is a modification of each other.
+	Modified
+)
+
+// EditScript represents the series of differences between two lists.
+type EditScript []EditType
+
+// String returns a human-readable string representing the edit-script where
+// Identity, UniqueX, UniqueY, and Modified are represented by the
+// '.', 'X', 'Y', and 'M' characters, respectively.
+func (es EditScript) String() string {
+	b := make([]byte, len(es))
+	for i, e := range es {
+		switch e {
+		case Identity:
+			b[i] = '.'
+		case UniqueX:
+			b[i] = 'X'
+		case UniqueY:
+			b[i] = 'Y'
+		case Modified:
+			b[i] = 'M'
+		default:
+			panic("invalid edit-type")
+		}
+	}
+	return string(b)
+}
+
+// stats returns a histogram of the number of each type of edit operation.
+func (es EditScript) stats() (s struct{ NI, NX, NY, NM int }) {
+	for _, e := range es {
+		switch e {
+		case Identity:
+			s.NI++
+		case UniqueX:
+			s.NX++
+		case UniqueY:
+			s.NY++
+		case Modified:
+			s.NM++
+		default:
+			panic("invalid edit-type")
+		}
+	}
+	return
+}
+
+// Dist is the Levenshtein distance and is guaranteed to be 0 if and only if
+// lists X and Y are equal.
+func (es EditScript) Dist() int { return len(es) - es.stats().NI }
+
+// LenX is the length of the X list.
+func (es EditScript) LenX() int { return len(es) - es.stats().NY }
+
+// LenY is the length of the Y list.
+func (es EditScript) LenY() int { return len(es) - es.stats().NX }
+
+// EqualFunc reports whether the symbols at indexes ix and iy are equal.
+// When called by Difference, the index is guaranteed to be within nx and ny.
+type EqualFunc func(ix int, iy int) Result
+
+// Result is the result of comparison.
+// NumSame is the number of sub-elements that are equal.
+// NumDiff is the number of sub-elements that are not equal.
+type Result struct{ NumSame, NumDiff int }
+
+// BoolResult returns a Result that is either Equal or not Equal.
+func BoolResult(b bool) Result {
+	if b {
+		return Result{NumSame: 1} // Equal, Similar
+	} else {
+		return Result{NumDiff: 2} // Not Equal, not Similar
+	}
+}
+
+// Equal indicates whether the symbols are equal. Two symbols are equal
+// if and only if NumDiff == 0. If Equal, then they are also Similar.
+func (r Result) Equal() bool { return r.NumDiff == 0 }
+
+// Similar indicates whether two symbols are similar and may be represented
+// by using the Modified type. As a special case, we consider binary comparisons
+// (i.e., those that return Result{1, 0} or Result{0, 1}) to be similar.
+//
+// The exact ratio of NumSame to NumDiff to determine similarity may change.
+func (r Result) Similar() bool {
+	// Use NumSame+1 to offset NumSame so that binary comparisons are similar.
+	return r.NumSame+1 >= r.NumDiff
+}
+
+var randBool = rand.New(rand.NewSource(time.Now().Unix())).Intn(2) == 0
+
+// Difference reports whether two lists of lengths nx and ny are equal
+// given the definition of equality provided as f.
+//
+// This function returns an edit-script, which is a sequence of operations
+// needed to convert one list into the other. The following invariants for
+// the edit-script are maintained:
+//	• eq == (es.Dist()==0)
+//	• nx == es.LenX()
+//	• ny == es.LenY()
+//
+// This algorithm is not guaranteed to be an optimal solution (i.e., one that
+// produces an edit-script with a minimal Levenshtein distance). This algorithm
+// favors performance over optimality. The exact output is not guaranteed to
+// be stable and may change over time.
+func Difference(nx, ny int, f EqualFunc) (es EditScript) {
+	// This algorithm is based on traversing what is known as an "edit-graph".
+	// See Figure 1 from "An O(ND) Difference Algorithm and Its Variations"
+	// by Eugene W. Myers. Since D can be as large as N itself, this is
+	// effectively O(N^2). Unlike the algorithm from that paper, we are not
+	// interested in the optimal path, but at least some "decent" path.
+	//
+	// For example, let X and Y be lists of symbols:
+	//	X = [A B C A B B A]
+	//	Y = [C B A B A C]
+	//
+	// The edit-graph can be drawn as the following:
+	//	   A B C A B B A
+	//	  ┌─────────────┐
+	//	C │_|_|\|_|_|_|_│ 0
+	//	B │_|\|_|_|\|\|_│ 1
+	//	A │\|_|_|\|_|_|\│ 2
+	//	B │_|\|_|_|\|\|_│ 3
+	//	A │\|_|_|\|_|_|\│ 4
+	//	C │ | |\| | | | │ 5
+	//	  └─────────────┘ 6
+	//	   0 1 2 3 4 5 6 7
+	//
+	// List X is written along the horizontal axis, while list Y is written
+	// along the vertical axis. At any point on this grid, if the symbol in
+	// list X matches the corresponding symbol in list Y, then a '\' is drawn.
+	// The goal of any minimal edit-script algorithm is to find a path from the
+	// top-left corner to the bottom-right corner, while traveling through the
+	// fewest horizontal or vertical edges.
+	// A horizontal edge is equivalent to inserting a symbol from list X.
+	// A vertical edge is equivalent to inserting a symbol from list Y.
+	// A diagonal edge is equivalent to a matching symbol between both X and Y.
+
+	// Invariants:
+	//	• 0 ≤ fwdPath.X ≤ (fwdFrontier.X, revFrontier.X) ≤ revPath.X ≤ nx
+	//	• 0 ≤ fwdPath.Y ≤ (fwdFrontier.Y, revFrontier.Y) ≤ revPath.Y ≤ ny
+	//
+	// In general:
+	//	• fwdFrontier.X < revFrontier.X
+	//	• fwdFrontier.Y < revFrontier.Y
+	// Unless, it is time for the algorithm to terminate.
+	fwdPath := path{+1, point{0, 0}, make(EditScript, 0, (nx+ny)/2)}
+	revPath := path{-1, point{nx, ny}, make(EditScript, 0)}
+	fwdFrontier := fwdPath.point // Forward search frontier
+	revFrontier := revPath.point // Reverse search frontier
+
+	// Search budget bounds the cost of searching for better paths.
+	// The longest sequence of non-matching symbols that can be tolerated is
+	// approximately the square-root of the search budget.
+	searchBudget := 4 * (nx + ny) // O(n)
+
+	// Running the tests with the "cmp_debug" build tag prints a visualization
+	// of the algorithm running in real-time. This is educational for
+	// understanding how the algorithm works. See debug_enable.go.
+	f = debug.Begin(nx, ny, f, &fwdPath.es, &revPath.es)
+
+	// The algorithm below is a greedy, meet-in-the-middle algorithm for
+	// computing sub-optimal edit-scripts between two lists.
+	//
+	// The algorithm is approximately as follows:
+	//	• Searching for differences switches back-and-forth between
+	//	a search that starts at the beginning (the top-left corner), and
+	//	a search that starts at the end (the bottom-right corner). The goal of
+	//	the search is connect with the search from the opposite corner.
+	//	• As we search, we build a path in a greedy manner, where the first
+	//	match seen is added to the path (this is sub-optimal, but provides a
+	//	decent result in practice). When matches are found, we try the next pair
+	//	of symbols in the lists and follow all matches as far as possible.
+	//	• When searching for matches, we search along a diagonal going through
+	//	through the "frontier" point. If no matches are found, we advance the
+	//	frontier towards the opposite corner.
+	//	• This algorithm terminates when either the X coordinates or the
+	//	Y coordinates of the forward and reverse frontier points ever intersect.
+
+	// This algorithm is correct even if searching only in the forward direction
+	// or in the reverse direction. We do both because it is commonly observed
+	// that two lists commonly differ because elements were added to the front
+	// or end of the other list.
+	//
+	// Non-deterministically start with either the forward or reverse direction
+	// to introduce some deliberate instability so that we have the flexibility
+	// to change this algorithm in the future.
+	if flags.Deterministic || randBool {
+		goto forwardSearch
+	} else {
+		goto reverseSearch
+	}
+
+forwardSearch:
+	{
+		// Forward search from the beginning.
+		if fwdFrontier.X >= revFrontier.X || fwdFrontier.Y >= revFrontier.Y || searchBudget == 0 {
+			goto finishSearch
+		}
+		for stop1, stop2, i := false, false, 0; !(stop1 && stop2) && searchBudget > 0; i++ {
+			// Search in a diagonal pattern for a match.
+			z := zigzag(i)
+			p := point{fwdFrontier.X + z, fwdFrontier.Y - z}
+			switch {
+			case p.X >= revPath.X || p.Y < fwdPath.Y:
+				stop1 = true // Hit top-right corner
+			case p.Y >= revPath.Y || p.X < fwdPath.X:
+				stop2 = true // Hit bottom-left corner
+			case f(p.X, p.Y).Equal():
+				// Match found, so connect the path to this point.
+				fwdPath.connect(p, f)
+				fwdPath.append(Identity)
+				// Follow sequence of matches as far as possible.
+				for fwdPath.X < revPath.X && fwdPath.Y < revPath.Y {
+					if !f(fwdPath.X, fwdPath.Y).Equal() {
+						break
+					}
+					fwdPath.append(Identity)
+				}
+				fwdFrontier = fwdPath.point
+				stop1, stop2 = true, true
+			default:
+				searchBudget-- // Match not found
+			}
+			debug.Update()
+		}
+		// Advance the frontier towards reverse point.
+		if revPath.X-fwdFrontier.X >= revPath.Y-fwdFrontier.Y {
+			fwdFrontier.X++
+		} else {
+			fwdFrontier.Y++
+		}
+		goto reverseSearch
+	}
+
+reverseSearch:
+	{
+		// Reverse search from the end.
+		if fwdFrontier.X >= revFrontier.X || fwdFrontier.Y >= revFrontier.Y || searchBudget == 0 {
+			goto finishSearch
+		}
+		for stop1, stop2, i := false, false, 0; !(stop1 && stop2) && searchBudget > 0; i++ {
+			// Search in a diagonal pattern for a match.
+			z := zigzag(i)
+			p := point{revFrontier.X - z, revFrontier.Y + z}
+			switch {
+			case fwdPath.X >= p.X || revPath.Y < p.Y:
+				stop1 = true // Hit bottom-left corner
+			case fwdPath.Y >= p.Y || revPath.X < p.X:
+				stop2 = true // Hit top-right corner
+			case f(p.X-1, p.Y-1).Equal():
+				// Match found, so connect the path to this point.
+				revPath.connect(p, f)
+				revPath.append(Identity)
+				// Follow sequence of matches as far as possible.
+				for fwdPath.X < revPath.X && fwdPath.Y < revPath.Y {
+					if !f(revPath.X-1, revPath.Y-1).Equal() {
+						break
+					}
+					revPath.append(Identity)
+				}
+				revFrontier = revPath.point
+				stop1, stop2 = true, true
+			default:
+				searchBudget-- // Match not found
+			}
+			debug.Update()
+		}
+		// Advance the frontier towards forward point.
+		if revFrontier.X-fwdPath.X >= revFrontier.Y-fwdPath.Y {
+			revFrontier.X--
+		} else {
+			revFrontier.Y--
+		}
+		goto forwardSearch
+	}
+
+finishSearch:
+	// Join the forward and reverse paths and then append the reverse path.
+	fwdPath.connect(revPath.point, f)
+	for i := len(revPath.es) - 1; i >= 0; i-- {
+		t := revPath.es[i]
+		revPath.es = revPath.es[:i]
+		fwdPath.append(t)
+	}
+	debug.Finish()
+	return fwdPath.es
+}
+
+type path struct {
+	dir   int // +1 if forward, -1 if reverse
+	point     // Leading point of the EditScript path
+	es    EditScript
+}
+
+// connect appends any necessary Identity, Modified, UniqueX, or UniqueY types
+// to the edit-script to connect p.point to dst.
+func (p *path) connect(dst point, f EqualFunc) {
+	if p.dir > 0 {
+		// Connect in forward direction.
+		for dst.X > p.X && dst.Y > p.Y {
+			switch r := f(p.X, p.Y); {
+			case r.Equal():
+				p.append(Identity)
+			case r.Similar():
+				p.append(Modified)
+			case dst.X-p.X >= dst.Y-p.Y:
+				p.append(UniqueX)
+			default:
+				p.append(UniqueY)
+			}
+		}
+		for dst.X > p.X {
+			p.append(UniqueX)
+		}
+		for dst.Y > p.Y {
+			p.append(UniqueY)
+		}
+	} else {
+		// Connect in reverse direction.
+		for p.X > dst.X && p.Y > dst.Y {
+			switch r := f(p.X-1, p.Y-1); {
+			case r.Equal():
+				p.append(Identity)
+			case r.Similar():
+				p.append(Modified)
+			case p.Y-dst.Y >= p.X-dst.X:
+				p.append(UniqueY)
+			default:
+				p.append(UniqueX)
+			}
+		}
+		for p.X > dst.X {
+			p.append(UniqueX)
+		}
+		for p.Y > dst.Y {
+			p.append(UniqueY)
+		}
+	}
+}
+
+func (p *path) append(t EditType) {
+	p.es = append(p.es, t)
+	switch t {
+	case Identity, Modified:
+		p.add(p.dir, p.dir)
+	case UniqueX:
+		p.add(p.dir, 0)
+	case UniqueY:
+		p.add(0, p.dir)
+	}
+	debug.Update()
+}
+
+type point struct{ X, Y int }
+
+func (p *point) add(dx, dy int) { p.X += dx; p.Y += dy }
+
+// zigzag maps a consecutive sequence of integers to a zig-zag sequence.
+//	[0 1 2 3 4 5 ...] => [0 -1 +1 -2 +2 ...]
+func zigzag(x int) int {
+	if x&1 != 0 {
+		x = ^x
+	}
+	return x >> 1
+}
@@ -0,0 +1,9 @@
+// Copyright 2019, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package flags
+
+// Deterministic controls whether the output of Diff should be deterministic.
+// This is only used for testing.
+var Deterministic bool
@@ -0,0 +1,99 @@
+// Copyright 2017, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Package function provides functionality for identifying function types.
+package function
+
+import (
+	"reflect"
+	"regexp"
+	"runtime"
+	"strings"
+)
+
+type funcType int
+
+const (
+	_ funcType = iota
+
+	tbFunc  // func(T) bool
+	ttbFunc // func(T, T) bool
+	trbFunc // func(T, R) bool
+	tibFunc // func(T, I) bool
+	trFunc  // func(T) R
+
+	Equal             = ttbFunc // func(T, T) bool
+	EqualAssignable   = tibFunc // func(T, I) bool; encapsulates func(T, T) bool
+	Transformer       = trFunc  // func(T) R
+	ValueFilter       = ttbFunc // func(T, T) bool
+	Less              = ttbFunc // func(T, T) bool
+	ValuePredicate    = tbFunc  // func(T) bool
+	KeyValuePredicate = trbFunc // func(T, R) bool
+)
+
+var boolType = reflect.TypeOf(true)
+
+// IsType reports whether the reflect.Type is of the specified function type.
+func IsType(t reflect.Type, ft funcType) bool {
+	if t == nil || t.Kind() != reflect.Func || t.IsVariadic() {
+		return false
+	}
+	ni, no := t.NumIn(), t.NumOut()
+	switch ft {
+	case tbFunc: // func(T) bool
+		if ni == 1 && no == 1 && t.Out(0) == boolType {
+			return true
+		}
+	case ttbFunc: // func(T, T) bool
+		if ni == 2 && no == 1 && t.In(0) == t.In(1) && t.Out(0) == boolType {
+			return true
+		}
+	case trbFunc: // func(T, R) bool
+		if ni == 2 && no == 1 && t.Out(0) == boolType {
+			return true
+		}
+	case tibFunc: // func(T, I) bool
+		if ni == 2 && no == 1 && t.In(0).AssignableTo(t.In(1)) && t.Out(0) == boolType {
+			return true
+		}
+	case trFunc: // func(T) R
+		if ni == 1 && no == 1 {
+			return true
+		}
+	}
+	return false
+}
+
+var lastIdentRx = regexp.MustCompile(`[_\p{L}][_\p{L}\p{N}]*$`)
+
+// NameOf returns the name of the function value.
+func NameOf(v reflect.Value) string {
+	fnc := runtime.FuncForPC(v.Pointer())
+	if fnc == nil {
+		return "<unknown>"
+	}
+	fullName := fnc.Name() // e.g., "long/path/name/mypkg.(*MyType).(long/path/name/mypkg.myMethod)-fm"
+
+	// Method closures have a "-fm" suffix.
+	fullName = strings.TrimSuffix(fullName, "-fm")
+
+	var name string
+	for len(fullName) > 0 {
+		inParen := strings.HasSuffix(fullName, ")")
+		fullName = strings.TrimSuffix(fullName, ")")
+
+		s := lastIdentRx.FindString(fullName)
+		if s == "" {
+			break
+		}
+		name = s + "." + name
+		fullName = strings.TrimSuffix(fullName, s)
+
+		if i := strings.LastIndexByte(fullName, '('); inParen && i >= 0 {
+			fullName = fullName[:i]
+		}
+		fullName = strings.TrimSuffix(fullName, ".")
+	}
+	return strings.TrimSuffix(name, ".")
+}
@@ -0,0 +1,164 @@
+// Copyright 2020, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package value
+
+import (
+	"reflect"
+	"strconv"
+)
+
+var anyType = reflect.TypeOf((*interface{})(nil)).Elem()
+
+// TypeString is nearly identical to reflect.Type.String,
+// but has an additional option to specify that full type names be used.
+func TypeString(t reflect.Type, qualified bool) string {
+	return string(appendTypeName(nil, t, qualified, false))
+}
+
+func appendTypeName(b []byte, t reflect.Type, qualified, elideFunc bool) []byte {
+	// BUG: Go reflection provides no way to disambiguate two named types
+	// of the same name and within the same package,
+	// but declared within the namespace of different functions.
+
+	// Use the "any" alias instead of "interface{}" for better readability.
+	if t == anyType {
+		return append(b, "any"...)
+	}
+
+	// Named type.
+	if t.Name() != "" {
+		if qualified && t.PkgPath() != "" {
+			b = append(b, '"')
+			b = append(b, t.PkgPath()...)
+			b = append(b, '"')
+			b = append(b, '.')
+			b = append(b, t.Name()...)
+		} else {
+			b = append(b, t.String()...)
+		}
+		return b
+	}
+
+	// Unnamed type.
+	switch k := t.Kind(); k {
+	case reflect.Bool, reflect.String, reflect.UnsafePointer,
+		reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64,
+		reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr,
+		reflect.Float32, reflect.Float64, reflect.Complex64, reflect.Complex128:
+		b = append(b, k.String()...)
+	case reflect.Chan:
+		if t.ChanDir() == reflect.RecvDir {
+			b = append(b, "<-"...)
+		}
+		b = append(b, "chan"...)
+		if t.ChanDir() == reflect.SendDir {
+			b = append(b, "<-"...)
+		}
+		b = append(b, ' ')
+		b = appendTypeName(b, t.Elem(), qualified, false)
+	case reflect.Func:
+		if !elideFunc {
+			b = append(b, "func"...)
+		}
+		b = append(b, '(')
+		for i := 0; i < t.NumIn(); i++ {
+			if i > 0 {
+				b = append(b, ", "...)
+			}
+			if i == t.NumIn()-1 && t.IsVariadic() {
+				b = append(b, "..."...)
+				b = appendTypeName(b, t.In(i).Elem(), qualified, false)
+			} else {
+				b = appendTypeName(b, t.In(i), qualified, false)
+			}
+		}
+		b = append(b, ')')
+		switch t.NumOut() {
+		case 0:
+			// Do nothing
+		case 1:
+			b = append(b, ' ')
+			b = appendTypeName(b, t.Out(0), qualified, false)
+		default:
+			b = append(b, " ("...)
+			for i := 0; i < t.NumOut(); i++ {
+				if i > 0 {
+					b = append(b, ", "...)
+				}
+				b = appendTypeName(b, t.Out(i), qualified, false)
+			}
+			b = append(b, ')')
+		}
+	case reflect.Struct:
+		b = append(b, "struct{ "...)
+		for i := 0; i < t.NumField(); i++ {
+			if i > 0 {
+				b = append(b, "; "...)
+			}
+			sf := t.Field(i)
+			if !sf.Anonymous {
+				if qualified && sf.PkgPath != "" {
+					b = append(b, '"')
+					b = append(b, sf.PkgPath...)
+					b = append(b, '"')
+					b = append(b, '.')
+				}
+				b = append(b, sf.Name...)
+				b = append(b, ' ')
+			}
+			b = appendTypeName(b, sf.Type, qualified, false)
+			if sf.Tag != "" {
+				b = append(b, ' ')
+				b = strconv.AppendQuote(b, string(sf.Tag))
+			}
+		}
+		if b[len(b)-1] == ' ' {
+			b = b[:len(b)-1]
+		} else {
+			b = append(b, ' ')
+		}
+		b = append(b, '}')
+	case reflect.Slice, reflect.Array:
+		b = append(b, '[')
+		if k == reflect.Array {
+			b = strconv.AppendUint(b, uint64(t.Len()), 10)
+		}
+		b = append(b, ']')
+		b = appendTypeName(b, t.Elem(), qualified, false)
+	case reflect.Map:
+		b = append(b, "map["...)
+		b = appendTypeName(b, t.Key(), qualified, false)
+		b = append(b, ']')
+		b = appendTypeName(b, t.Elem(), qualified, false)
+	case reflect.Ptr:
+		b = append(b, '*')
+		b = appendTypeName(b, t.Elem(), qualified, false)
+	case reflect.Interface:
+		b = append(b, "interface{ "...)
+		for i := 0; i < t.NumMethod(); i++ {
+			if i > 0 {
+				b = append(b, "; "...)
+			}
+			m := t.Method(i)
+			if qualified && m.PkgPath != "" {
+				b = append(b, '"')
+				b = append(b, m.PkgPath...)
+				b = append(b, '"')
+				b = append(b, '.')
+			}
+			b = append(b, m.Name...)
+			b = appendTypeName(b, m.Type, qualified, true)
+		}
+		if b[len(b)-1] == ' ' {
+			b = b[:len(b)-1]
+		} else {
+			b = append(b, ' ')
+		}
+		b = append(b, '}')
+	default:
+		panic("invalid kind: " + k.String())
+	}
+	return b
+}
@@ -0,0 +1,34 @@
+// Copyright 2018, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build purego
+// +build purego
+
+package value
+
+import "reflect"
+
+// Pointer is an opaque typed pointer and is guaranteed to be comparable.
+type Pointer struct {
+	p uintptr
+	t reflect.Type
+}
+
+// PointerOf returns a Pointer from v, which must be a
+// reflect.Ptr, reflect.Slice, or reflect.Map.
+func PointerOf(v reflect.Value) Pointer {
+	// NOTE: Storing a pointer as an uintptr is technically incorrect as it
+	// assumes that the GC implementation does not use a moving collector.
+	return Pointer{v.Pointer(), v.Type()}
+}
+
+// IsNil reports whether the pointer is nil.
+func (p Pointer) IsNil() bool {
+	return p.p == 0
+}
+
+// Uintptr returns the pointer as a uintptr.
+func (p Pointer) Uintptr() uintptr {
+	return p.p
+}
@@ -0,0 +1,37 @@
+// Copyright 2018, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build !purego
+// +build !purego
+
+package value
+
+import (
+	"reflect"
+	"unsafe"
+)
+
+// Pointer is an opaque typed pointer and is guaranteed to be comparable.
+type Pointer struct {
+	p unsafe.Pointer
+	t reflect.Type
+}
+
+// PointerOf returns a Pointer from v, which must be a
+// reflect.Ptr, reflect.Slice, or reflect.Map.
+func PointerOf(v reflect.Value) Pointer {
+	// The proper representation of a pointer is unsafe.Pointer,
+	// which is necessary if the GC ever uses a moving collector.
+	return Pointer{unsafe.Pointer(v.Pointer()), v.Type()}
+}
+
+// IsNil reports whether the pointer is nil.
+func (p Pointer) IsNil() bool {
+	return p.p == nil
+}
+
+// Uintptr returns the pointer as a uintptr.
+func (p Pointer) Uintptr() uintptr {
+	return uintptr(p.p)
+}
@@ -0,0 +1,106 @@
+// Copyright 2017, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package value
+
+import (
+	"fmt"
+	"math"
+	"reflect"
+	"sort"
+)
+
+// SortKeys sorts a list of map keys, deduplicating keys if necessary.
+// The type of each value must be comparable.
+func SortKeys(vs []reflect.Value) []reflect.Value {
+	if len(vs) == 0 {
+		return vs
+	}
+
+	// Sort the map keys.
+	sort.SliceStable(vs, func(i, j int) bool { return isLess(vs[i], vs[j]) })
+
+	// Deduplicate keys (fails for NaNs).
+	vs2 := vs[:1]
+	for _, v := range vs[1:] {
+		if isLess(vs2[len(vs2)-1], v) {
+			vs2 = append(vs2, v)
+		}
+	}
+	return vs2
+}
+
+// isLess is a generic function for sorting arbitrary map keys.
+// The inputs must be of the same type and must be comparable.
+func isLess(x, y reflect.Value) bool {
+	switch x.Type().Kind() {
+	case reflect.Bool:
+		return !x.Bool() && y.Bool()
+	case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
+		return x.Int() < y.Int()
+	case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr:
+		return x.Uint() < y.Uint()
+	case reflect.Float32, reflect.Float64:
+		// NOTE: This does not sort -0 as less than +0
+		// since Go maps treat -0 and +0 as equal keys.
+		fx, fy := x.Float(), y.Float()
+		return fx < fy || math.IsNaN(fx) && !math.IsNaN(fy)
+	case reflect.Complex64, reflect.Complex128:
+		cx, cy := x.Complex(), y.Complex()
+		rx, ix, ry, iy := real(cx), imag(cx), real(cy), imag(cy)
+		if rx == ry || (math.IsNaN(rx) && math.IsNaN(ry)) {
+			return ix < iy || math.IsNaN(ix) && !math.IsNaN(iy)
+		}
+		return rx < ry || math.IsNaN(rx) && !math.IsNaN(ry)
+	case reflect.Ptr, reflect.UnsafePointer, reflect.Chan:
+		return x.Pointer() < y.Pointer()
+	case reflect.String:
+		return x.String() < y.String()
+	case reflect.Array:
+		for i := 0; i < x.Len(); i++ {
+			if isLess(x.Index(i), y.Index(i)) {
+				return true
+			}
+			if isLess(y.Index(i), x.Index(i)) {
+				return false
+			}
+		}
+		return false
+	case reflect.Struct:
+		for i := 0; i < x.NumField(); i++ {
+			if isLess(x.Field(i), y.Field(i)) {
+				return true
+			}
+			if isLess(y.Field(i), x.Field(i)) {
+				return false
+			}
+		}
+		return false
+	case reflect.Interface:
+		vx, vy := x.Elem(), y.Elem()
+		if !vx.IsValid() || !vy.IsValid() {
+			return !vx.IsValid() && vy.IsValid()
+		}
+		tx, ty := vx.Type(), vy.Type()
+		if tx == ty {
+			return isLess(x.Elem(), y.Elem())
+		}
+		if tx.Kind() != ty.Kind() {
+			return vx.Kind() < vy.Kind()
+		}
+		if tx.String() != ty.String() {
+			return tx.String() < ty.String()
+		}
+		if tx.PkgPath() != ty.PkgPath() {
+			return tx.PkgPath() < ty.PkgPath()
+		}
+		// This can happen in rare situations, so we fallback to just comparing
+		// the unique pointer for a reflect.Type. This guarantees deterministic
+		// ordering within a program, but it is obviously not stable.
+		return reflect.ValueOf(vx.Type()).Pointer() < reflect.ValueOf(vy.Type()).Pointer()
+	default:
+		// Must be Func, Map, or Slice; which are not comparable.
+		panic(fmt.Sprintf("%T is not comparable", x.Type()))
+	}
+}
@@ -0,0 +1,48 @@
+// Copyright 2017, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package value
+
+import (
+	"math"
+	"reflect"
+)
+
+// IsZero reports whether v is the zero value.
+// This does not rely on Interface and so can be used on unexported fields.
+func IsZero(v reflect.Value) bool {
+	switch v.Kind() {
+	case reflect.Bool:
+		return v.Bool() == false
+	case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
+		return v.Int() == 0
+	case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr:
+		return v.Uint() == 0
+	case reflect.Float32, reflect.Float64:
+		return math.Float64bits(v.Float()) == 0
+	case reflect.Complex64, reflect.Complex128:
+		return math.Float64bits(real(v.Complex())) == 0 && math.Float64bits(imag(v.Complex())) == 0
+	case reflect.String:
+		return v.String() == ""
+	case reflect.UnsafePointer:
+		return v.Pointer() == 0
+	case reflect.Chan, reflect.Func, reflect.Interface, reflect.Ptr, reflect.Map, reflect.Slice:
+		return v.IsNil()
+	case reflect.Array:
+		for i := 0; i < v.Len(); i++ {
+			if !IsZero(v.Index(i)) {
+				return false
+			}
+		}
+		return true
+	case reflect.Struct:
+		for i := 0; i < v.NumField(); i++ {
+			if !IsZero(v.Field(i)) {
+				return false
+			}
+		}
+		return true
+	}
+	return false
+}
@@ -0,0 +1,552 @@
+// Copyright 2017, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package cmp
+
+import (
+	"fmt"
+	"reflect"
+	"regexp"
+	"strings"
+
+	"github.com/google/go-cmp/cmp/internal/function"
+)
+
+// Option configures for specific behavior of Equal and Diff. In particular,
+// the fundamental Option functions (Ignore, Transformer, and Comparer),
+// configure how equality is determined.
+//
+// The fundamental options may be composed with filters (FilterPath and
+// FilterValues) to control the scope over which they are applied.
+//
+// The cmp/cmpopts package provides helper functions for creating options that
+// may be used with Equal and Diff.
+type Option interface {
+	// filter applies all filters and returns the option that remains.
+	// Each option may only read s.curPath and call s.callTTBFunc.
+	//
+	// An Options is returned only if multiple comparers or transformers
+	// can apply simultaneously and will only contain values of those types
+	// or sub-Options containing values of those types.
+	filter(s *state, t reflect.Type, vx, vy reflect.Value) applicableOption
+}
+
+// applicableOption represents the following types:
+//	Fundamental: ignore | validator | *comparer | *transformer
+//	Grouping:    Options
+type applicableOption interface {
+	Option
+
+	// apply executes the option, which may mutate s or panic.
+	apply(s *state, vx, vy reflect.Value)
+}
+
+// coreOption represents the following types:
+//	Fundamental: ignore | validator | *comparer | *transformer
+//	Filters:     *pathFilter | *valuesFilter
+type coreOption interface {
+	Option
+	isCore()
+}
+
+type core struct{}
+
+func (core) isCore() {}
+
+// Options is a list of Option values that also satisfies the Option interface.
+// Helper comparison packages may return an Options value when packing multiple
+// Option values into a single Option. When this package processes an Options,
+// it will be implicitly expanded into a flat list.
+//
+// Applying a filter on an Options is equivalent to applying that same filter
+// on all individual options held within.
+type Options []Option
+
+func (opts Options) filter(s *state, t reflect.Type, vx, vy reflect.Value) (out applicableOption) {
+	for _, opt := range opts {
+		switch opt := opt.filter(s, t, vx, vy); opt.(type) {
+		case ignore:
+			return ignore{} // Only ignore can short-circuit evaluation
+		case validator:
+			out = validator{} // Takes precedence over comparer or transformer
+		case *comparer, *transformer, Options:
+			switch out.(type) {
+			case nil:
+				out = opt
+			case validator:
+				// Keep validator
+			case *comparer, *transformer, Options:
+				out = Options{out, opt} // Conflicting comparers or transformers
+			}
+		}
+	}
+	return out
+}
+
+func (opts Options) apply(s *state, _, _ reflect.Value) {
+	const warning = "ambiguous set of applicable options"
+	const help = "consider using filters to ensure at most one Comparer or Transformer may apply"
+	var ss []string
+	for _, opt := range flattenOptions(nil, opts) {
+		ss = append(ss, fmt.Sprint(opt))
+	}
+	set := strings.Join(ss, "\n\t")
+	panic(fmt.Sprintf("%s at %#v:\n\t%s\n%s", warning, s.curPath, set, help))
+}
+
+func (opts Options) String() string {
+	var ss []string
+	for _, opt := range opts {
+		ss = append(ss, fmt.Sprint(opt))
+	}
+	return fmt.Sprintf("Options{%s}", strings.Join(ss, ", "))
+}
+
+// FilterPath returns a new Option where opt is only evaluated if filter f
+// returns true for the current Path in the value tree.
+//
+// This filter is called even if a slice element or map entry is missing and
+// provides an opportunity to ignore such cases. The filter function must be
+// symmetric such that the filter result is identical regardless of whether the
+// missing value is from x or y.
+//
+// The option passed in may be an Ignore, Transformer, Comparer, Options, or
+// a previously filtered Option.
+func FilterPath(f func(Path) bool, opt Option) Option {
+	if f == nil {
+		panic("invalid path filter function")
+	}
+	if opt := normalizeOption(opt); opt != nil {
+		return &pathFilter{fnc: f, opt: opt}
+	}
+	return nil
+}
+
+type pathFilter struct {
+	core
+	fnc func(Path) bool
+	opt Option
+}
+
+func (f pathFilter) filter(s *state, t reflect.Type, vx, vy reflect.Value) applicableOption {
+	if f.fnc(s.curPath) {
+		return f.opt.filter(s, t, vx, vy)
+	}
+	return nil
+}
+
+func (f pathFilter) String() string {
+	return fmt.Sprintf("FilterPath(%s, %v)", function.NameOf(reflect.ValueOf(f.fnc)), f.opt)
+}
+
+// FilterValues returns a new Option where opt is only evaluated if filter f,
+// which is a function of the form "func(T, T) bool", returns true for the
+// current pair of values being compared. If either value is invalid or
+// the type of the values is not assignable to T, then this filter implicitly
+// returns false.
+//
+// The filter function must be
+// symmetric (i.e., agnostic to the order of the inputs) and
+// deterministic (i.e., produces the same result when given the same inputs).
+// If T is an interface, it is possible that f is called with two values with
+// different concrete types that both implement T.
+//
+// The option passed in may be an Ignore, Transformer, Comparer, Options, or
+// a previously filtered Option.
+func FilterValues(f interface{}, opt Option) Option {
+	v := reflect.ValueOf(f)
+	if !function.IsType(v.Type(), function.ValueFilter) || v.IsNil() {
+		panic(fmt.Sprintf("invalid values filter function: %T", f))
+	}
+	if opt := normalizeOption(opt); opt != nil {
+		vf := &valuesFilter{fnc: v, opt: opt}
+		if ti := v.Type().In(0); ti.Kind() != reflect.Interface || ti.NumMethod() > 0 {
+			vf.typ = ti
+		}
+		return vf
+	}
+	return nil
+}
+
+type valuesFilter struct {
+	core
+	typ reflect.Type  // T
+	fnc reflect.Value // func(T, T) bool
+	opt Option
+}
+
+func (f valuesFilter) filter(s *state, t reflect.Type, vx, vy reflect.Value) applicableOption {
+	if !vx.IsValid() || !vx.CanInterface() || !vy.IsValid() || !vy.CanInterface() {
+		return nil
+	}
+	if (f.typ == nil || t.AssignableTo(f.typ)) && s.callTTBFunc(f.fnc, vx, vy) {
+		return f.opt.filter(s, t, vx, vy)
+	}
+	return nil
+}
+
+func (f valuesFilter) String() string {
+	return fmt.Sprintf("FilterValues(%s, %v)", function.NameOf(f.fnc), f.opt)
+}
+
+// Ignore is an Option that causes all comparisons to be ignored.
+// This value is intended to be combined with FilterPath or FilterValues.
+// It is an error to pass an unfiltered Ignore option to Equal.
+func Ignore() Option { return ignore{} }
+
+type ignore struct{ core }
+
+func (ignore) isFiltered() bool                                                     { return false }
+func (ignore) filter(_ *state, _ reflect.Type, _, _ reflect.Value) applicableOption { return ignore{} }
+func (ignore) apply(s *state, _, _ reflect.Value)                                   { s.report(true, reportByIgnore) }
+func (ignore) String() string                                                       { return "Ignore()" }
+
+// validator is a sentinel Option type to indicate that some options could not
+// be evaluated due to unexported fields, missing slice elements, or
+// missing map entries. Both values are validator only for unexported fields.
+type validator struct{ core }
+
+func (validator) filter(_ *state, _ reflect.Type, vx, vy reflect.Value) applicableOption {
+	if !vx.IsValid() || !vy.IsValid() {
+		return validator{}
+	}
+	if !vx.CanInterface() || !vy.CanInterface() {
+		return validator{}
+	}
+	return nil
+}
+func (validator) apply(s *state, vx, vy reflect.Value) {
+	// Implies missing slice element or map entry.
+	if !vx.IsValid() || !vy.IsValid() {
+		s.report(vx.IsValid() == vy.IsValid(), 0)
+		return
+	}
+
+	// Unable to Interface implies unexported field without visibility access.
+	if !vx.CanInterface() || !vy.CanInterface() {
+		help := "consider using a custom Comparer; if you control the implementation of type, you can also consider using an Exporter, AllowUnexported, or cmpopts.IgnoreUnexported"
+		var name string
+		if t := s.curPath.Index(-2).Type(); t.Name() != "" {
+			// Named type with unexported fields.
+			name = fmt.Sprintf("%q.%v", t.PkgPath(), t.Name()) // e.g., "path/to/package".MyType
+			if _, ok := reflect.New(t).Interface().(error); ok {
+				help = "consider using cmpopts.EquateErrors to compare error values"
+			}
+		} else {
+			// Unnamed type with unexported fields. Derive PkgPath from field.
+			var pkgPath string
+			for i := 0; i < t.NumField() && pkgPath == ""; i++ {
+				pkgPath = t.Field(i).PkgPath
+			}
+			name = fmt.Sprintf("%q.(%v)", pkgPath, t.String()) // e.g., "path/to/package".(struct { a int })
+		}
+		panic(fmt.Sprintf("cannot handle unexported field at %#v:\n\t%v\n%s", s.curPath, name, help))
+	}
+
+	panic("not reachable")
+}
+
+// identRx represents a valid identifier according to the Go specification.
+const identRx = `[_\p{L}][_\p{L}\p{N}]*`
+
+var identsRx = regexp.MustCompile(`^` + identRx + `(\.` + identRx + `)*$`)
+
+// Transformer returns an Option that applies a transformation function that
+// converts values of a certain type into that of another.
+//
+// The transformer f must be a function "func(T) R" that converts values of
+// type T to those of type R and is implicitly filtered to input values
+// assignable to T. The transformer must not mutate T in any way.
+//
+// To help prevent some cases of infinite recursive cycles applying the
+// same transform to the output of itself (e.g., in the case where the
+// input and output types are the same), an implicit filter is added such that
+// a transformer is applicable only if that exact transformer is not already
+// in the tail of the Path since the last non-Transform step.
+// For situations where the implicit filter is still insufficient,
+// consider using cmpopts.AcyclicTransformer, which adds a filter
+// to prevent the transformer from being recursively applied upon itself.
+//
+// The name is a user provided label that is used as the Transform.Name in the
+// transformation PathStep (and eventually shown in the Diff output).
+// The name must be a valid identifier or qualified identifier in Go syntax.
+// If empty, an arbitrary name is used.
+func Transformer(name string, f interface{}) Option {
+	v := reflect.ValueOf(f)
+	if !function.IsType(v.Type(), function.Transformer) || v.IsNil() {
+		panic(fmt.Sprintf("invalid transformer function: %T", f))
+	}
+	if name == "" {
+		name = function.NameOf(v)
+		if !identsRx.MatchString(name) {
+			name = "λ" // Lambda-symbol as placeholder name
+		}
+	} else if !identsRx.MatchString(name) {
+		panic(fmt.Sprintf("invalid name: %q", name))
+	}
+	tr := &transformer{name: name, fnc: reflect.ValueOf(f)}
+	if ti := v.Type().In(0); ti.Kind() != reflect.Interface || ti.NumMethod() > 0 {
+		tr.typ = ti
+	}
+	return tr
+}
+
+type transformer struct {
+	core
+	name string
+	typ  reflect.Type  // T
+	fnc  reflect.Value // func(T) R
+}
+
+func (tr *transformer) isFiltered() bool { return tr.typ != nil }
+
+func (tr *transformer) filter(s *state, t reflect.Type, _, _ reflect.Value) applicableOption {
+	for i := len(s.curPath) - 1; i >= 0; i-- {
+		if t, ok := s.curPath[i].(Transform); !ok {
+			break // Hit most recent non-Transform step
+		} else if tr == t.trans {
+			return nil // Cannot directly use same Transform
+		}
+	}
+	if tr.typ == nil || t.AssignableTo(tr.typ) {
+		return tr
+	}
+	return nil
+}
+
+func (tr *transformer) apply(s *state, vx, vy reflect.Value) {
+	step := Transform{&transform{pathStep{typ: tr.fnc.Type().Out(0)}, tr}}
+	vvx := s.callTRFunc(tr.fnc, vx, step)
+	vvy := s.callTRFunc(tr.fnc, vy, step)
+	step.vx, step.vy = vvx, vvy
+	s.compareAny(step)
+}
+
+func (tr transformer) String() string {
+	return fmt.Sprintf("Transformer(%s, %s)", tr.name, function.NameOf(tr.fnc))
+}
+
+// Comparer returns an Option that determines whether two values are equal
+// to each other.
+//
+// The comparer f must be a function "func(T, T) bool" and is implicitly
+// filtered to input values assignable to T. If T is an interface, it is
+// possible that f is called with two values of different concrete types that
+// both implement T.
+//
+// The equality function must be:
+//	• Symmetric: equal(x, y) == equal(y, x)
+//	• Deterministic: equal(x, y) == equal(x, y)
+//	• Pure: equal(x, y) does not modify x or y
+func Comparer(f interface{}) Option {
+	v := reflect.ValueOf(f)
+	if !function.IsType(v.Type(), function.Equal) || v.IsNil() {
+		panic(fmt.Sprintf("invalid comparer function: %T", f))
+	}
+	cm := &comparer{fnc: v}
+	if ti := v.Type().In(0); ti.Kind() != reflect.Interface || ti.NumMethod() > 0 {
+		cm.typ = ti
+	}
+	return cm
+}
+
+type comparer struct {
+	core
+	typ reflect.Type  // T
+	fnc reflect.Value // func(T, T) bool
+}
+
+func (cm *comparer) isFiltered() bool { return cm.typ != nil }
+
+func (cm *comparer) filter(_ *state, t reflect.Type, _, _ reflect.Value) applicableOption {
+	if cm.typ == nil || t.AssignableTo(cm.typ) {
+		return cm
+	}
+	return nil
+}
+
+func (cm *comparer) apply(s *state, vx, vy reflect.Value) {
+	eq := s.callTTBFunc(cm.fnc, vx, vy)
+	s.report(eq, reportByFunc)
+}
+
+func (cm comparer) String() string {
+	return fmt.Sprintf("Comparer(%s)", function.NameOf(cm.fnc))
+}
+
+// Exporter returns an Option that specifies whether Equal is allowed to
+// introspect into the unexported fields of certain struct types.
+//
+// Users of this option must understand that comparing on unexported fields
+// from external packages is not safe since changes in the internal
+// implementation of some external package may cause the result of Equal
+// to unexpectedly change. However, it may be valid to use this option on types
+// defined in an internal package where the semantic meaning of an unexported
+// field is in the control of the user.
+//
+// In many cases, a custom Comparer should be used instead that defines
+// equality as a function of the public API of a type rather than the underlying
+// unexported implementation.
+//
+// For example, the reflect.Type documentation defines equality to be determined
+// by the == operator on the interface (essentially performing a shallow pointer
+// comparison) and most attempts to compare *regexp.Regexp types are interested
+// in only checking that the regular expression strings are equal.
+// Both of these are accomplished using Comparers:
+//
+//	Comparer(func(x, y reflect.Type) bool { return x == y })
+//	Comparer(func(x, y *regexp.Regexp) bool { return x.String() == y.String() })
+//
+// In other cases, the cmpopts.IgnoreUnexported option can be used to ignore
+// all unexported fields on specified struct types.
+func Exporter(f func(reflect.Type) bool) Option {
+	if !supportExporters {
+		panic("Exporter is not supported on purego builds")
+	}
+	return exporter(f)
+}
+
+type exporter func(reflect.Type) bool
+
+func (exporter) filter(_ *state, _ reflect.Type, _, _ reflect.Value) applicableOption {
+	panic("not implemented")
+}
+
+// AllowUnexported returns an Options that allows Equal to forcibly introspect
+// unexported fields of the specified struct types.
+//
+// See Exporter for the proper use of this option.
+func AllowUnexported(types ...interface{}) Option {
+	m := make(map[reflect.Type]bool)
+	for _, typ := range types {
+		t := reflect.TypeOf(typ)
+		if t.Kind() != reflect.Struct {
+			panic(fmt.Sprintf("invalid struct type: %T", typ))
+		}
+		m[t] = true
+	}
+	return exporter(func(t reflect.Type) bool { return m[t] })
+}
+
+// Result represents the comparison result for a single node and
+// is provided by cmp when calling Result (see Reporter).
+type Result struct {
+	_     [0]func() // Make Result incomparable
+	flags resultFlags
+}
+
+// Equal reports whether the node was determined to be equal or not.
+// As a special case, ignored nodes are considered equal.
+func (r Result) Equal() bool {
+	return r.flags&(reportEqual|reportByIgnore) != 0
+}
+
+// ByIgnore reports whether the node is equal because it was ignored.
+// This never reports true if Equal reports false.
+func (r Result) ByIgnore() bool {
+	return r.flags&reportByIgnore != 0
+}
+
+// ByMethod reports whether the Equal method determined equality.
+func (r Result) ByMethod() bool {
+	return r.flags&reportByMethod != 0
+}
+
+// ByFunc reports whether a Comparer function determined equality.
+func (r Result) ByFunc() bool {
+	return r.flags&reportByFunc != 0
+}
+
+// ByCycle reports whether a reference cycle was detected.
+func (r Result) ByCycle() bool {
+	return r.flags&reportByCycle != 0
+}
+
+type resultFlags uint
+
+const (
+	_ resultFlags = (1 << iota) / 2
+
+	reportEqual
+	reportUnequal
+	reportByIgnore
+	reportByMethod
+	reportByFunc
+	reportByCycle
+)
+
+// Reporter is an Option that can be passed to Equal. When Equal traverses
+// the value trees, it calls PushStep as it descends into each node in the
+// tree and PopStep as it ascend out of the node. The leaves of the tree are
+// either compared (determined to be equal or not equal) or ignored and reported
+// as such by calling the Report method.
+func Reporter(r interface {
+	// PushStep is called when a tree-traversal operation is performed.
+	// The PathStep itself is only valid until the step is popped.
+	// The PathStep.Values are valid for the duration of the entire traversal
+	// and must not be mutated.
+	//
+	// Equal always calls PushStep at the start to provide an operation-less
+	// PathStep used to report the root values.
+	//
+	// Within a slice, the exact set of inserted, removed, or modified elements
+	// is unspecified and may change in future implementations.
+	// The entries of a map are iterated through in an unspecified order.
+	PushStep(PathStep)
+
+	// Report is called exactly once on leaf nodes to report whether the
+	// comparison identified the node as equal, unequal, or ignored.
+	// A leaf node is one that is immediately preceded by and followed by
+	// a pair of PushStep and PopStep calls.
+	Report(Result)
+
+	// PopStep ascends back up the value tree.
+	// There is always a matching pop call for every push call.
+	PopStep()
+}) Option {
+	return reporter{r}
+}
+
+type reporter struct{ reporterIface }
+type reporterIface interface {
+	PushStep(PathStep)
+	Report(Result)
+	PopStep()
+}
+
+func (reporter) filter(_ *state, _ reflect.Type, _, _ reflect.Value) applicableOption {
+	panic("not implemented")
+}
+
+// normalizeOption normalizes the input options such that all Options groups
+// are flattened and groups with a single element are reduced to that element.
+// Only coreOptions and Options containing coreOptions are allowed.
+func normalizeOption(src Option) Option {
+	switch opts := flattenOptions(nil, Options{src}); len(opts) {
+	case 0:
+		return nil
+	case 1:
+		return opts[0]
+	default:
+		return opts
+	}
+}
+
+// flattenOptions copies all options in src to dst as a flat list.
+// Only coreOptions and Options containing coreOptions are allowed.
+func flattenOptions(dst, src Options) Options {
+	for _, opt := range src {
+		switch opt := opt.(type) {
+		case nil:
+			continue
+		case Options:
+			dst = flattenOptions(dst, opt)
+		case coreOption:
+			dst = append(dst, opt)
+		default:
+			panic(fmt.Sprintf("invalid option type: %T", opt))
+		}
+	}
+	return dst
+}
@@ -0,0 +1,378 @@
+// Copyright 2017, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package cmp
+
+import (
+	"fmt"
+	"reflect"
+	"strings"
+	"unicode"
+	"unicode/utf8"
+
+	"github.com/google/go-cmp/cmp/internal/value"
+)
+
+// Path is a list of PathSteps describing the sequence of operations to get
+// from some root type to the current position in the value tree.
+// The first Path element is always an operation-less PathStep that exists
+// simply to identify the initial type.
+//
+// When traversing structs with embedded structs, the embedded struct will
+// always be accessed as a field before traversing the fields of the
+// embedded struct themselves. That is, an exported field from the
+// embedded struct will never be accessed directly from the parent struct.
+type Path []PathStep
+
+// PathStep is a union-type for specific operations to traverse
+// a value's tree structure. Users of this package never need to implement
+// these types as values of this type will be returned by this package.
+//
+// Implementations of this interface are
+// StructField, SliceIndex, MapIndex, Indirect, TypeAssertion, and Transform.
+type PathStep interface {
+	String() string
+
+	// Type is the resulting type after performing the path step.
+	Type() reflect.Type
+
+	// Values is the resulting values after performing the path step.
+	// The type of each valid value is guaranteed to be identical to Type.
+	//
+	// In some cases, one or both may be invalid or have restrictions:
+	//	• For StructField, both are not interface-able if the current field
+	//	is unexported and the struct type is not explicitly permitted by
+	//	an Exporter to traverse unexported fields.
+	//	• For SliceIndex, one may be invalid if an element is missing from
+	//	either the x or y slice.
+	//	• For MapIndex, one may be invalid if an entry is missing from
+	//	either the x or y map.
+	//
+	// The provided values must not be mutated.
+	Values() (vx, vy reflect.Value)
+}
+
+var (
+	_ PathStep = StructField{}
+	_ PathStep = SliceIndex{}
+	_ PathStep = MapIndex{}
+	_ PathStep = Indirect{}
+	_ PathStep = TypeAssertion{}
+	_ PathStep = Transform{}
+)
+
+func (pa *Path) push(s PathStep) {
+	*pa = append(*pa, s)
+}
+
+func (pa *Path) pop() {
+	*pa = (*pa)[:len(*pa)-1]
+}
+
+// Last returns the last PathStep in the Path.
+// If the path is empty, this returns a non-nil PathStep that reports a nil Type.
+func (pa Path) Last() PathStep {
+	return pa.Index(-1)
+}
+
+// Index returns the ith step in the Path and supports negative indexing.
+// A negative index starts counting from the tail of the Path such that -1
+// refers to the last step, -2 refers to the second-to-last step, and so on.
+// If index is invalid, this returns a non-nil PathStep that reports a nil Type.
+func (pa Path) Index(i int) PathStep {
+	if i < 0 {
+		i = len(pa) + i
+	}
+	if i < 0 || i >= len(pa) {
+		return pathStep{}
+	}
+	return pa[i]
+}
+
+// String returns the simplified path to a node.
+// The simplified path only contains struct field accesses.
+//
+// For example:
+//	MyMap.MySlices.MyField
+func (pa Path) String() string {
+	var ss []string
+	for _, s := range pa {
+		if _, ok := s.(StructField); ok {
+			ss = append(ss, s.String())
+		}
+	}
+	return strings.TrimPrefix(strings.Join(ss, ""), ".")
+}
+
+// GoString returns the path to a specific node using Go syntax.
+//
+// For example:
+//	(*root.MyMap["key"].(*mypkg.MyStruct).MySlices)[2][3].MyField
+func (pa Path) GoString() string {
+	var ssPre, ssPost []string
+	var numIndirect int
+	for i, s := range pa {
+		var nextStep PathStep
+		if i+1 < len(pa) {
+			nextStep = pa[i+1]
+		}
+		switch s := s.(type) {
+		case Indirect:
+			numIndirect++
+			pPre, pPost := "(", ")"
+			switch nextStep.(type) {
+			case Indirect:
+				continue // Next step is indirection, so let them batch up
+			case StructField:
+				numIndirect-- // Automatic indirection on struct fields
+			case nil:
+				pPre, pPost = "", "" // Last step; no need for parenthesis
+			}
+			if numIndirect > 0 {
+				ssPre = append(ssPre, pPre+strings.Repeat("*", numIndirect))
+				ssPost = append(ssPost, pPost)
+			}
+			numIndirect = 0
+			continue
+		case Transform:
+			ssPre = append(ssPre, s.trans.name+"(")
+			ssPost = append(ssPost, ")")
+			continue
+		}
+		ssPost = append(ssPost, s.String())
+	}
+	for i, j := 0, len(ssPre)-1; i < j; i, j = i+1, j-1 {
+		ssPre[i], ssPre[j] = ssPre[j], ssPre[i]
+	}
+	return strings.Join(ssPre, "") + strings.Join(ssPost, "")
+}
+
+type pathStep struct {
+	typ    reflect.Type
+	vx, vy reflect.Value
+}
+
+func (ps pathStep) Type() reflect.Type             { return ps.typ }
+func (ps pathStep) Values() (vx, vy reflect.Value) { return ps.vx, ps.vy }
+func (ps pathStep) String() string {
+	if ps.typ == nil {
+		return "<nil>"
+	}
+	s := ps.typ.String()
+	if s == "" || strings.ContainsAny(s, "{}\n") {
+		return "root" // Type too simple or complex to print
+	}
+	return fmt.Sprintf("{%s}", s)
+}
+
+// StructField represents a struct field access on a field called Name.
+type StructField struct{ *structField }
+type structField struct {
+	pathStep
+	name string
+	idx  int
+
+	// These fields are used for forcibly accessing an unexported field.
+	// pvx, pvy, and field are only valid if unexported is true.
+	unexported bool
+	mayForce   bool                // Forcibly allow visibility
+	paddr      bool                // Was parent addressable?
+	pvx, pvy   reflect.Value       // Parent values (always addressable)
+	field      reflect.StructField // Field information
+}
+
+func (sf StructField) Type() reflect.Type { return sf.typ }
+func (sf StructField) Values() (vx, vy reflect.Value) {
+	if !sf.unexported {
+		return sf.vx, sf.vy // CanInterface reports true
+	}
+
+	// Forcibly obtain read-write access to an unexported struct field.
+	if sf.mayForce {
+		vx = retrieveUnexportedField(sf.pvx, sf.field, sf.paddr)
+		vy = retrieveUnexportedField(sf.pvy, sf.field, sf.paddr)
+		return vx, vy // CanInterface reports true
+	}
+	return sf.vx, sf.vy // CanInterface reports false
+}
+func (sf StructField) String() string { return fmt.Sprintf(".%s", sf.name) }
+
+// Name is the field name.
+func (sf StructField) Name() string { return sf.name }
+
+// Index is the index of the field in the parent struct type.
+// See reflect.Type.Field.
+func (sf StructField) Index() int { return sf.idx }
+
+// SliceIndex is an index operation on a slice or array at some index Key.
+type SliceIndex struct{ *sliceIndex }
+type sliceIndex struct {
+	pathStep
+	xkey, ykey int
+	isSlice    bool // False for reflect.Array
+}
+
+func (si SliceIndex) Type() reflect.Type             { return si.typ }
+func (si SliceIndex) Values() (vx, vy reflect.Value) { return si.vx, si.vy }
+func (si SliceIndex) String() string {
+	switch {
+	case si.xkey == si.ykey:
+		return fmt.Sprintf("[%d]", si.xkey)
+	case si.ykey == -1:
+		// [5->?] means "I don't know where X[5] went"
+		return fmt.Sprintf("[%d->?]", si.xkey)
+	case si.xkey == -1:
+		// [?->3] means "I don't know where Y[3] came from"
+		return fmt.Sprintf("[?->%d]", si.ykey)
+	default:
+		// [5->3] means "X[5] moved to Y[3]"
+		return fmt.Sprintf("[%d->%d]", si.xkey, si.ykey)
+	}
+}
+
+// Key is the index key; it may return -1 if in a split state
+func (si SliceIndex) Key() int {
+	if si.xkey != si.ykey {
+		return -1
+	}
+	return si.xkey
+}
+
+// SplitKeys are the indexes for indexing into slices in the
+// x and y values, respectively. These indexes may differ due to the
+// insertion or removal of an element in one of the slices, causing
+// all of the indexes to be shifted. If an index is -1, then that
+// indicates that the element does not exist in the associated slice.
+//
+// Key is guaranteed to return -1 if and only if the indexes returned
+// by SplitKeys are not the same. SplitKeys will never return -1 for
+// both indexes.
+func (si SliceIndex) SplitKeys() (ix, iy int) { return si.xkey, si.ykey }
+
+// MapIndex is an index operation on a map at some index Key.
+type MapIndex struct{ *mapIndex }
+type mapIndex struct {
+	pathStep
+	key reflect.Value
+}
+
+func (mi MapIndex) Type() reflect.Type             { return mi.typ }
+func (mi MapIndex) Values() (vx, vy reflect.Value) { return mi.vx, mi.vy }
+func (mi MapIndex) String() string                 { return fmt.Sprintf("[%#v]", mi.key) }
+
+// Key is the value of the map key.
+func (mi MapIndex) Key() reflect.Value { return mi.key }
+
+// Indirect represents pointer indirection on the parent type.
+type Indirect struct{ *indirect }
+type indirect struct {
+	pathStep
+}
+
+func (in Indirect) Type() reflect.Type             { return in.typ }
+func (in Indirect) Values() (vx, vy reflect.Value) { return in.vx, in.vy }
+func (in Indirect) String() string                 { return "*" }
+
+// TypeAssertion represents a type assertion on an interface.
+type TypeAssertion struct{ *typeAssertion }
+type typeAssertion struct {
+	pathStep
+}
+
+func (ta TypeAssertion) Type() reflect.Type             { return ta.typ }
+func (ta TypeAssertion) Values() (vx, vy reflect.Value) { return ta.vx, ta.vy }
+func (ta TypeAssertion) String() string                 { return fmt.Sprintf(".(%v)", ta.typ) }
+
+// Transform is a transformation from the parent type to the current type.
+type Transform struct{ *transform }
+type transform struct {
+	pathStep
+	trans *transformer
+}
+
+func (tf Transform) Type() reflect.Type             { return tf.typ }
+func (tf Transform) Values() (vx, vy reflect.Value) { return tf.vx, tf.vy }
+func (tf Transform) String() string                 { return fmt.Sprintf("%s()", tf.trans.name) }
+
+// Name is the name of the Transformer.
+func (tf Transform) Name() string { return tf.trans.name }
+
+// Func is the function pointer to the transformer function.
+func (tf Transform) Func() reflect.Value { return tf.trans.fnc }
+
+// Option returns the originally constructed Transformer option.
+// The == operator can be used to detect the exact option used.
+func (tf Transform) Option() Option { return tf.trans }
+
+// pointerPath represents a dual-stack of pointers encountered when
+// recursively traversing the x and y values. This data structure supports
+// detection of cycles and determining whether the cycles are equal.
+// In Go, cycles can occur via pointers, slices, and maps.
+//
+// The pointerPath uses a map to represent a stack; where descension into a
+// pointer pushes the address onto the stack, and ascension from a pointer
+// pops the address from the stack. Thus, when traversing into a pointer from
+// reflect.Ptr, reflect.Slice element, or reflect.Map, we can detect cycles
+// by checking whether the pointer has already been visited. The cycle detection
+// uses a separate stack for the x and y values.
+//
+// If a cycle is detected we need to determine whether the two pointers
+// should be considered equal. The definition of equality chosen by Equal
+// requires two graphs to have the same structure. To determine this, both the
+// x and y values must have a cycle where the previous pointers were also
+// encountered together as a pair.
+//
+// Semantically, this is equivalent to augmenting Indirect, SliceIndex, and
+// MapIndex with pointer information for the x and y values.
+// Suppose px and py are two pointers to compare, we then search the
+// Path for whether px was ever encountered in the Path history of x, and
+// similarly so with py. If either side has a cycle, the comparison is only
+// equal if both px and py have a cycle resulting from the same PathStep.
+//
+// Using a map as a stack is more performant as we can perform cycle detection
+// in O(1) instead of O(N) where N is len(Path).
+type pointerPath struct {
+	// mx is keyed by x pointers, where the value is the associated y pointer.
+	mx map[value.Pointer]value.Pointer
+	// my is keyed by y pointers, where the value is the associated x pointer.
+	my map[value.Pointer]value.Pointer
+}
+
+func (p *pointerPath) Init() {
+	p.mx = make(map[value.Pointer]value.Pointer)
+	p.my = make(map[value.Pointer]value.Pointer)
+}
+
+// Push indicates intent to descend into pointers vx and vy where
+// visited reports whether either has been seen before. If visited before,
+// equal reports whether both pointers were encountered together.
+// Pop must be called if and only if the pointers were never visited.
+//
+// The pointers vx and vy must be a reflect.Ptr, reflect.Slice, or reflect.Map
+// and be non-nil.
+func (p pointerPath) Push(vx, vy reflect.Value) (equal, visited bool) {
+	px := value.PointerOf(vx)
+	py := value.PointerOf(vy)
+	_, ok1 := p.mx[px]
+	_, ok2 := p.my[py]
+	if ok1 || ok2 {
+		equal = p.mx[px] == py && p.my[py] == px // Pointers paired together
+		return equal, true
+	}
+	p.mx[px] = py
+	p.my[py] = px
+	return false, false
+}
+
+// Pop ascends from pointers vx and vy.
+func (p pointerPath) Pop(vx, vy reflect.Value) {
+	delete(p.mx, value.PointerOf(vx))
+	delete(p.my, value.PointerOf(vy))
+}
+
+// isExported reports whether the identifier is exported.
+func isExported(id string) bool {
+	r, _ := utf8.DecodeRuneInString(id)
+	return unicode.IsUpper(r)
+}
@@ -0,0 +1,54 @@
+// Copyright 2017, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package cmp
+
+// defaultReporter implements the reporter interface.
+//
+// As Equal serially calls the PushStep, Report, and PopStep methods, the
+// defaultReporter constructs a tree-based representation of the compared value
+// and the result of each comparison (see valueNode).
+//
+// When the String method is called, the FormatDiff method transforms the
+// valueNode tree into a textNode tree, which is a tree-based representation
+// of the textual output (see textNode).
+//
+// Lastly, the textNode.String method produces the final report as a string.
+type defaultReporter struct {
+	root *valueNode
+	curr *valueNode
+}
+
+func (r *defaultReporter) PushStep(ps PathStep) {
+	r.curr = r.curr.PushStep(ps)
+	if r.root == nil {
+		r.root = r.curr
+	}
+}
+func (r *defaultReporter) Report(rs Result) {
+	r.curr.Report(rs)
+}
+func (r *defaultReporter) PopStep() {
+	r.curr = r.curr.PopStep()
+}
+
+// String provides a full report of the differences detected as a structured
+// literal in pseudo-Go syntax. String may only be called after the entire tree
+// has been traversed.
+func (r *defaultReporter) String() string {
+	assert(r.root != nil && r.curr == nil)
+	if r.root.NumDiff == 0 {
+		return ""
+	}
+	ptrs := new(pointerReferences)
+	text := formatOptions{}.FormatDiff(r.root, ptrs)
+	resolveReferences(text)
+	return text.String()
+}
+
+func assert(ok bool) {
+	if !ok {
+		panic("assertion failure")
+	}
+}
@@ -0,0 +1,435 @@
+// Copyright 2019, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package cmp
+
+import (
+	"fmt"
+	"reflect"
+
+	"github.com/google/go-cmp/cmp/internal/value"
+)
+
+// numContextRecords is the number of surrounding equal records to print.
+const numContextRecords = 2
+
+type diffMode byte
+
+const (
+	diffUnknown   diffMode = 0
+	diffIdentical diffMode = ' '
+	diffRemoved   diffMode = '-'
+	diffInserted  diffMode = '+'
+)
+
+type typeMode int
+
+const (
+	// emitType always prints the type.
+	emitType typeMode = iota
+	// elideType never prints the type.
+	elideType
+	// autoType prints the type only for composite kinds
+	// (i.e., structs, slices, arrays, and maps).
+	autoType
+)
+
+type formatOptions struct {
+	// DiffMode controls the output mode of FormatDiff.
+	//
+	// If diffUnknown,   then produce a diff of the x and y values.
+	// If diffIdentical, then emit values as if they were equal.
+	// If diffRemoved,   then only emit x values (ignoring y values).
+	// If diffInserted,  then only emit y values (ignoring x values).
+	DiffMode diffMode
+
+	// TypeMode controls whether to print the type for the current node.
+	//
+	// As a general rule of thumb, we always print the type of the next node
+	// after an interface, and always elide the type of the next node after
+	// a slice or map node.
+	TypeMode typeMode
+
+	// formatValueOptions are options specific to printing reflect.Values.
+	formatValueOptions
+}
+
+func (opts formatOptions) WithDiffMode(d diffMode) formatOptions {
+	opts.DiffMode = d
+	return opts
+}
+func (opts formatOptions) WithTypeMode(t typeMode) formatOptions {
+	opts.TypeMode = t
+	return opts
+}
+func (opts formatOptions) WithVerbosity(level int) formatOptions {
+	opts.VerbosityLevel = level
+	opts.LimitVerbosity = true
+	return opts
+}
+func (opts formatOptions) verbosity() uint {
+	switch {
+	case opts.VerbosityLevel < 0:
+		return 0
+	case opts.VerbosityLevel > 16:
+		return 16 // some reasonable maximum to avoid shift overflow
+	default:
+		return uint(opts.VerbosityLevel)
+	}
+}
+
+const maxVerbosityPreset = 6
+
+// verbosityPreset modifies the verbosity settings given an index
+// between 0 and maxVerbosityPreset, inclusive.
+func verbosityPreset(opts formatOptions, i int) formatOptions {
+	opts.VerbosityLevel = int(opts.verbosity()) + 2*i
+	if i > 0 {
+		opts.AvoidStringer = true
+	}
+	if i >= maxVerbosityPreset {
+		opts.PrintAddresses = true
+		opts.QualifiedNames = true
+	}
+	return opts
+}
+
+// FormatDiff converts a valueNode tree into a textNode tree, where the later
+// is a textual representation of the differences detected in the former.
+func (opts formatOptions) FormatDiff(v *valueNode, ptrs *pointerReferences) (out textNode) {
+	if opts.DiffMode == diffIdentical {
+		opts = opts.WithVerbosity(1)
+	} else if opts.verbosity() < 3 {
+		opts = opts.WithVerbosity(3)
+	}
+
+	// Check whether we have specialized formatting for this node.
+	// This is not necessary, but helpful for producing more readable outputs.
+	if opts.CanFormatDiffSlice(v) {
+		return opts.FormatDiffSlice(v)
+	}
+
+	var parentKind reflect.Kind
+	if v.parent != nil && v.parent.TransformerName == "" {
+		parentKind = v.parent.Type.Kind()
+	}
+
+	// For leaf nodes, format the value based on the reflect.Values alone.
+	// As a special case, treat equal []byte as a leaf nodes.
+	isBytes := v.Type.Kind() == reflect.Slice && v.Type.Elem() == reflect.TypeOf(byte(0))
+	isEqualBytes := isBytes && v.NumDiff+v.NumIgnored+v.NumTransformed == 0
+	if v.MaxDepth == 0 || isEqualBytes {
+		switch opts.DiffMode {
+		case diffUnknown, diffIdentical:
+			// Format Equal.
+			if v.NumDiff == 0 {
+				outx := opts.FormatValue(v.ValueX, parentKind, ptrs)
+				outy := opts.FormatValue(v.ValueY, parentKind, ptrs)
+				if v.NumIgnored > 0 && v.NumSame == 0 {
+					return textEllipsis
+				} else if outx.Len() < outy.Len() {
+					return outx
+				} else {
+					return outy
+				}
+			}
+
+			// Format unequal.
+			assert(opts.DiffMode == diffUnknown)
+			var list textList
+			outx := opts.WithTypeMode(elideType).FormatValue(v.ValueX, parentKind, ptrs)
+			outy := opts.WithTypeMode(elideType).FormatValue(v.ValueY, parentKind, ptrs)
+			for i := 0; i <= maxVerbosityPreset && outx != nil && outy != nil && outx.Equal(outy); i++ {
+				opts2 := verbosityPreset(opts, i).WithTypeMode(elideType)
+				outx = opts2.FormatValue(v.ValueX, parentKind, ptrs)
+				outy = opts2.FormatValue(v.ValueY, parentKind, ptrs)
+			}
+			if outx != nil {
+				list = append(list, textRecord{Diff: '-', Value: outx})
+			}
+			if outy != nil {
+				list = append(list, textRecord{Diff: '+', Value: outy})
+			}
+			return opts.WithTypeMode(emitType).FormatType(v.Type, list)
+		case diffRemoved:
+			return opts.FormatValue(v.ValueX, parentKind, ptrs)
+		case diffInserted:
+			return opts.FormatValue(v.ValueY, parentKind, ptrs)
+		default:
+			panic("invalid diff mode")
+		}
+	}
+
+	// Register slice element to support cycle detection.
+	if parentKind == reflect.Slice {
+		ptrRefs := ptrs.PushPair(v.ValueX, v.ValueY, opts.DiffMode, true)
+		defer ptrs.Pop()
+		defer func() { out = wrapTrunkReferences(ptrRefs, out) }()
+	}
+
+	// Descend into the child value node.
+	if v.TransformerName != "" {
+		out := opts.WithTypeMode(emitType).FormatDiff(v.Value, ptrs)
+		out = &textWrap{Prefix: "Inverse(" + v.TransformerName + ", ", Value: out, Suffix: ")"}
+		return opts.FormatType(v.Type, out)
+	} else {
+		switch k := v.Type.Kind(); k {
+		case reflect.Struct, reflect.Array, reflect.Slice:
+			out = opts.formatDiffList(v.Records, k, ptrs)
+			out = opts.FormatType(v.Type, out)
+		case reflect.Map:
+			// Register map to support cycle detection.
+			ptrRefs := ptrs.PushPair(v.ValueX, v.ValueY, opts.DiffMode, false)
+			defer ptrs.Pop()
+
+			out = opts.formatDiffList(v.Records, k, ptrs)
+			out = wrapTrunkReferences(ptrRefs, out)
+			out = opts.FormatType(v.Type, out)
+		case reflect.Ptr:
+			// Register pointer to support cycle detection.
+			ptrRefs := ptrs.PushPair(v.ValueX, v.ValueY, opts.DiffMode, false)
+			defer ptrs.Pop()
+
+			out = opts.FormatDiff(v.Value, ptrs)
+			out = wrapTrunkReferences(ptrRefs, out)
+			out = &textWrap{Prefix: "&", Value: out}
+		case reflect.Interface:
+			out = opts.WithTypeMode(emitType).FormatDiff(v.Value, ptrs)
+		default:
+			panic(fmt.Sprintf("%v cannot have children", k))
+		}
+		return out
+	}
+}
+
+func (opts formatOptions) formatDiffList(recs []reportRecord, k reflect.Kind, ptrs *pointerReferences) textNode {
+	// Derive record name based on the data structure kind.
+	var name string
+	var formatKey func(reflect.Value) string
+	switch k {
+	case reflect.Struct:
+		name = "field"
+		opts = opts.WithTypeMode(autoType)
+		formatKey = func(v reflect.Value) string { return v.String() }
+	case reflect.Slice, reflect.Array:
+		name = "element"
+		opts = opts.WithTypeMode(elideType)
+		formatKey = func(reflect.Value) string { return "" }
+	case reflect.Map:
+		name = "entry"
+		opts = opts.WithTypeMode(elideType)
+		formatKey = func(v reflect.Value) string { return formatMapKey(v, false, ptrs) }
+	}
+
+	maxLen := -1
+	if opts.LimitVerbosity {
+		if opts.DiffMode == diffIdentical {
+			maxLen = ((1 << opts.verbosity()) >> 1) << 2 // 0, 4, 8, 16, 32, etc...
+		} else {
+			maxLen = (1 << opts.verbosity()) << 1 // 2, 4, 8, 16, 32, 64, etc...
+		}
+		opts.VerbosityLevel--
+	}
+
+	// Handle unification.
+	switch opts.DiffMode {
+	case diffIdentical, diffRemoved, diffInserted:
+		var list textList
+		var deferredEllipsis bool // Add final "..." to indicate records were dropped
+		for _, r := range recs {
+			if len(list) == maxLen {
+				deferredEllipsis = true
+				break
+			}
+
+			// Elide struct fields that are zero value.
+			if k == reflect.Struct {
+				var isZero bool
+				switch opts.DiffMode {
+				case diffIdentical:
+					isZero = value.IsZero(r.Value.ValueX) || value.IsZero(r.Value.ValueY)
+				case diffRemoved:
+					isZero = value.IsZero(r.Value.ValueX)
+				case diffInserted:
+					isZero = value.IsZero(r.Value.ValueY)
+				}
+				if isZero {
+					continue
+				}
+			}
+			// Elide ignored nodes.
+			if r.Value.NumIgnored > 0 && r.Value.NumSame+r.Value.NumDiff == 0 {
+				deferredEllipsis = !(k == reflect.Slice || k == reflect.Array)
+				if !deferredEllipsis {
+					list.AppendEllipsis(diffStats{})
+				}
+				continue
+			}
+			if out := opts.FormatDiff(r.Value, ptrs); out != nil {
+				list = append(list, textRecord{Key: formatKey(r.Key), Value: out})
+			}
+		}
+		if deferredEllipsis {
+			list.AppendEllipsis(diffStats{})
+		}
+		return &textWrap{Prefix: "{", Value: list, Suffix: "}"}
+	case diffUnknown:
+	default:
+		panic("invalid diff mode")
+	}
+
+	// Handle differencing.
+	var numDiffs int
+	var list textList
+	var keys []reflect.Value // invariant: len(list) == len(keys)
+	groups := coalesceAdjacentRecords(name, recs)
+	maxGroup := diffStats{Name: name}
+	for i, ds := range groups {
+		if maxLen >= 0 && numDiffs >= maxLen {
+			maxGroup = maxGroup.Append(ds)
+			continue
+		}
+
+		// Handle equal records.
+		if ds.NumDiff() == 0 {
+			// Compute the number of leading and trailing records to print.
+			var numLo, numHi int
+			numEqual := ds.NumIgnored + ds.NumIdentical
+			for numLo < numContextRecords && numLo+numHi < numEqual && i != 0 {
+				if r := recs[numLo].Value; r.NumIgnored > 0 && r.NumSame+r.NumDiff == 0 {
+					break
+				}
+				numLo++
+			}
+			for numHi < numContextRecords && numLo+numHi < numEqual && i != len(groups)-1 {
+				if r := recs[numEqual-numHi-1].Value; r.NumIgnored > 0 && r.NumSame+r.NumDiff == 0 {
+					break
+				}
+				numHi++
+			}
+			if numEqual-(numLo+numHi) == 1 && ds.NumIgnored == 0 {
+				numHi++ // Avoid pointless coalescing of a single equal record
+			}
+
+			// Format the equal values.
+			for _, r := range recs[:numLo] {
+				out := opts.WithDiffMode(diffIdentical).FormatDiff(r.Value, ptrs)
+				list = append(list, textRecord{Key: formatKey(r.Key), Value: out})
+				keys = append(keys, r.Key)
+			}
+			if numEqual > numLo+numHi {
+				ds.NumIdentical -= numLo + numHi
+				list.AppendEllipsis(ds)
+				for len(keys) < len(list) {
+					keys = append(keys, reflect.Value{})
+				}
+			}
+			for _, r := range recs[numEqual-numHi : numEqual] {
+				out := opts.WithDiffMode(diffIdentical).FormatDiff(r.Value, ptrs)
+				list = append(list, textRecord{Key: formatKey(r.Key), Value: out})
+				keys = append(keys, r.Key)
+			}
+			recs = recs[numEqual:]
+			continue
+		}
+
+		// Handle unequal records.
+		for _, r := range recs[:ds.NumDiff()] {
+			switch {
+			case opts.CanFormatDiffSlice(r.Value):
+				out := opts.FormatDiffSlice(r.Value)
+				list = append(list, textRecord{Key: formatKey(r.Key), Value: out})
+				keys = append(keys, r.Key)
+			case r.Value.NumChildren == r.Value.MaxDepth:
+				outx := opts.WithDiffMode(diffRemoved).FormatDiff(r.Value, ptrs)
+				outy := opts.WithDiffMode(diffInserted).FormatDiff(r.Value, ptrs)
+				for i := 0; i <= maxVerbosityPreset && outx != nil && outy != nil && outx.Equal(outy); i++ {
+					opts2 := verbosityPreset(opts, i)
+					outx = opts2.WithDiffMode(diffRemoved).FormatDiff(r.Value, ptrs)
+					outy = opts2.WithDiffMode(diffInserted).FormatDiff(r.Value, ptrs)
+				}
+				if outx != nil {
+					list = append(list, textRecord{Diff: diffRemoved, Key: formatKey(r.Key), Value: outx})
+					keys = append(keys, r.Key)
+				}
+				if outy != nil {
+					list = append(list, textRecord{Diff: diffInserted, Key: formatKey(r.Key), Value: outy})
+					keys = append(keys, r.Key)
+				}
+			default:
+				out := opts.FormatDiff(r.Value, ptrs)
+				list = append(list, textRecord{Key: formatKey(r.Key), Value: out})
+				keys = append(keys, r.Key)
+			}
+		}
+		recs = recs[ds.NumDiff():]
+		numDiffs += ds.NumDiff()
+	}
+	if maxGroup.IsZero() {
+		assert(len(recs) == 0)
+	} else {
+		list.AppendEllipsis(maxGroup)
+		for len(keys) < len(list) {
+			keys = append(keys, reflect.Value{})
+		}
+	}
+	assert(len(list) == len(keys))
+
+	// For maps, the default formatting logic uses fmt.Stringer which may
+	// produce ambiguous output. Avoid calling String to disambiguate.
+	if k == reflect.Map {
+		var ambiguous bool
+		seenKeys := map[string]reflect.Value{}
+		for i, currKey := range keys {
+			if currKey.IsValid() {
+				strKey := list[i].Key
+				prevKey, seen := seenKeys[strKey]
+				if seen && prevKey.CanInterface() && currKey.CanInterface() {
+					ambiguous = prevKey.Interface() != currKey.Interface()
+					if ambiguous {
+						break
+					}
+				}
+				seenKeys[strKey] = currKey
+			}
+		}
+		if ambiguous {
+			for i, k := range keys {
+				if k.IsValid() {
+					list[i].Key = formatMapKey(k, true, ptrs)
+				}
+			}
+		}
+	}
+
+	return &textWrap{Prefix: "{", Value: list, Suffix: "}"}
+}
+
+// coalesceAdjacentRecords coalesces the list of records into groups of
+// adjacent equal, or unequal counts.
+func coalesceAdjacentRecords(name string, recs []reportRecord) (groups []diffStats) {
+	var prevCase int // Arbitrary index into which case last occurred
+	lastStats := func(i int) *diffStats {
+		if prevCase != i {
+			groups = append(groups, diffStats{Name: name})
+			prevCase = i
+		}
+		return &groups[len(groups)-1]
+	}
+	for _, r := range recs {
+		switch rv := r.Value; {
+		case rv.NumIgnored > 0 && rv.NumSame+rv.NumDiff == 0:
+			lastStats(1).NumIgnored++
+		case rv.NumDiff == 0:
+			lastStats(1).NumIdentical++
+		case rv.NumDiff > 0 && !rv.ValueY.IsValid():
+			lastStats(2).NumRemoved++
+		case rv.NumDiff > 0 && !rv.ValueX.IsValid():
+			lastStats(2).NumInserted++
+		default:
+			lastStats(2).NumModified++
+		}
+	}
+	return groups
+}
@@ -0,0 +1,264 @@
+// Copyright 2020, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package cmp
+
+import (
+	"fmt"
+	"reflect"
+	"strings"
+
+	"github.com/google/go-cmp/cmp/internal/flags"
+	"github.com/google/go-cmp/cmp/internal/value"
+)
+
+const (
+	pointerDelimPrefix = "⟪"
+	pointerDelimSuffix = "⟫"
+)
+
+// formatPointer prints the address of the pointer.
+func formatPointer(p value.Pointer, withDelims bool) string {
+	v := p.Uintptr()
+	if flags.Deterministic {
+		v = 0xdeadf00f // Only used for stable testing purposes
+	}
+	if withDelims {
+		return pointerDelimPrefix + formatHex(uint64(v)) + pointerDelimSuffix
+	}
+	return formatHex(uint64(v))
+}
+
+// pointerReferences is a stack of pointers visited so far.
+type pointerReferences [][2]value.Pointer
+
+func (ps *pointerReferences) PushPair(vx, vy reflect.Value, d diffMode, deref bool) (pp [2]value.Pointer) {
+	if deref && vx.IsValid() {
+		vx = vx.Addr()
+	}
+	if deref && vy.IsValid() {
+		vy = vy.Addr()
+	}
+	switch d {
+	case diffUnknown, diffIdentical:
+		pp = [2]value.Pointer{value.PointerOf(vx), value.PointerOf(vy)}
+	case diffRemoved:
+		pp = [2]value.Pointer{value.PointerOf(vx), value.Pointer{}}
+	case diffInserted:
+		pp = [2]value.Pointer{value.Pointer{}, value.PointerOf(vy)}
+	}
+	*ps = append(*ps, pp)
+	return pp
+}
+
+func (ps *pointerReferences) Push(v reflect.Value) (p value.Pointer, seen bool) {
+	p = value.PointerOf(v)
+	for _, pp := range *ps {
+		if p == pp[0] || p == pp[1] {
+			return p, true
+		}
+	}
+	*ps = append(*ps, [2]value.Pointer{p, p})
+	return p, false
+}
+
+func (ps *pointerReferences) Pop() {
+	*ps = (*ps)[:len(*ps)-1]
+}
+
+// trunkReferences is metadata for a textNode indicating that the sub-tree
+// represents the value for either pointer in a pair of references.
+type trunkReferences struct{ pp [2]value.Pointer }
+
+// trunkReference is metadata for a textNode indicating that the sub-tree
+// represents the value for the given pointer reference.
+type trunkReference struct{ p value.Pointer }
+
+// leafReference is metadata for a textNode indicating that the value is
+// truncated as it refers to another part of the tree (i.e., a trunk).
+type leafReference struct{ p value.Pointer }
+
+func wrapTrunkReferences(pp [2]value.Pointer, s textNode) textNode {
+	switch {
+	case pp[0].IsNil():
+		return &textWrap{Value: s, Metadata: trunkReference{pp[1]}}
+	case pp[1].IsNil():
+		return &textWrap{Value: s, Metadata: trunkReference{pp[0]}}
+	case pp[0] == pp[1]:
+		return &textWrap{Value: s, Metadata: trunkReference{pp[0]}}
+	default:
+		return &textWrap{Value: s, Metadata: trunkReferences{pp}}
+	}
+}
+func wrapTrunkReference(p value.Pointer, printAddress bool, s textNode) textNode {
+	var prefix string
+	if printAddress {
+		prefix = formatPointer(p, true)
+	}
+	return &textWrap{Prefix: prefix, Value: s, Metadata: trunkReference{p}}
+}
+func makeLeafReference(p value.Pointer, printAddress bool) textNode {
+	out := &textWrap{Prefix: "(", Value: textEllipsis, Suffix: ")"}
+	var prefix string
+	if printAddress {
+		prefix = formatPointer(p, true)
+	}
+	return &textWrap{Prefix: prefix, Value: out, Metadata: leafReference{p}}
+}
+
+// resolveReferences walks the textNode tree searching for any leaf reference
+// metadata and resolves each against the corresponding trunk references.
+// Since pointer addresses in memory are not particularly readable to the user,
+// it replaces each pointer value with an arbitrary and unique reference ID.
+func resolveReferences(s textNode) {
+	var walkNodes func(textNode, func(textNode))
+	walkNodes = func(s textNode, f func(textNode)) {
+		f(s)
+		switch s := s.(type) {
+		case *textWrap:
+			walkNodes(s.Value, f)
+		case textList:
+			for _, r := range s {
+				walkNodes(r.Value, f)
+			}
+		}
+	}
+
+	// Collect all trunks and leaves with reference metadata.
+	var trunks, leaves []*textWrap
+	walkNodes(s, func(s textNode) {
+		if s, ok := s.(*textWrap); ok {
+			switch s.Metadata.(type) {
+			case leafReference:
+				leaves = append(leaves, s)
+			case trunkReference, trunkReferences:
+				trunks = append(trunks, s)
+			}
+		}
+	})
+
+	// No leaf references to resolve.
+	if len(leaves) == 0 {
+		return
+	}
+
+	// Collect the set of all leaf references to resolve.
+	leafPtrs := make(map[value.Pointer]bool)
+	for _, leaf := range leaves {
+		leafPtrs[leaf.Metadata.(leafReference).p] = true
+	}
+
+	// Collect the set of trunk pointers that are always paired together.
+	// This allows us to assign a single ID to both pointers for brevity.
+	// If a pointer in a pair ever occurs by itself or as a different pair,
+	// then the pair is broken.
+	pairedTrunkPtrs := make(map[value.Pointer]value.Pointer)
+	unpair := func(p value.Pointer) {
+		if !pairedTrunkPtrs[p].IsNil() {
+			pairedTrunkPtrs[pairedTrunkPtrs[p]] = value.Pointer{} // invalidate other half
+		}
+		pairedTrunkPtrs[p] = value.Pointer{} // invalidate this half
+	}
+	for _, trunk := range trunks {
+		switch p := trunk.Metadata.(type) {
+		case trunkReference:
+			unpair(p.p) // standalone pointer cannot be part of a pair
+		case trunkReferences:
+			p0, ok0 := pairedTrunkPtrs[p.pp[0]]
+			p1, ok1 := pairedTrunkPtrs[p.pp[1]]
+			switch {
+			case !ok0 && !ok1:
+				// Register the newly seen pair.
+				pairedTrunkPtrs[p.pp[0]] = p.pp[1]
+				pairedTrunkPtrs[p.pp[1]] = p.pp[0]
+			case ok0 && ok1 && p0 == p.pp[1] && p1 == p.pp[0]:
+				// Exact pair already seen; do nothing.
+			default:
+				// Pair conflicts with some other pair; break all pairs.
+				unpair(p.pp[0])
+				unpair(p.pp[1])
+			}
+		}
+	}
+
+	// Correlate each pointer referenced by leaves to a unique identifier,
+	// and print the IDs for each trunk that matches those pointers.
+	var nextID uint
+	ptrIDs := make(map[value.Pointer]uint)
+	newID := func() uint {
+		id := nextID
+		nextID++
+		return id
+	}
+	for _, trunk := range trunks {
+		switch p := trunk.Metadata.(type) {
+		case trunkReference:
+			if print := leafPtrs[p.p]; print {
+				id, ok := ptrIDs[p.p]
+				if !ok {
+					id = newID()
+					ptrIDs[p.p] = id
+				}
+				trunk.Prefix = updateReferencePrefix(trunk.Prefix, formatReference(id))
+			}
+		case trunkReferences:
+			print0 := leafPtrs[p.pp[0]]
+			print1 := leafPtrs[p.pp[1]]
+			if print0 || print1 {
+				id0, ok0 := ptrIDs[p.pp[0]]
+				id1, ok1 := ptrIDs[p.pp[1]]
+				isPair := pairedTrunkPtrs[p.pp[0]] == p.pp[1] && pairedTrunkPtrs[p.pp[1]] == p.pp[0]
+				if isPair {
+					var id uint
+					assert(ok0 == ok1) // must be seen together or not at all
+					if ok0 {
+						assert(id0 == id1) // must have the same ID
+						id = id0
+					} else {
+						id = newID()
+						ptrIDs[p.pp[0]] = id
+						ptrIDs[p.pp[1]] = id
+					}
+					trunk.Prefix = updateReferencePrefix(trunk.Prefix, formatReference(id))
+				} else {
+					if print0 && !ok0 {
+						id0 = newID()
+						ptrIDs[p.pp[0]] = id0
+					}
+					if print1 && !ok1 {
+						id1 = newID()
+						ptrIDs[p.pp[1]] = id1
+					}
+					switch {
+					case print0 && print1:
+						trunk.Prefix = updateReferencePrefix(trunk.Prefix, formatReference(id0)+","+formatReference(id1))
+					case print0:
+						trunk.Prefix = updateReferencePrefix(trunk.Prefix, formatReference(id0))
+					case print1:
+						trunk.Prefix = updateReferencePrefix(trunk.Prefix, formatReference(id1))
+					}
+				}
+			}
+		}
+	}
+
+	// Update all leaf references with the unique identifier.
+	for _, leaf := range leaves {
+		if id, ok := ptrIDs[leaf.Metadata.(leafReference).p]; ok {
+			leaf.Prefix = updateReferencePrefix(leaf.Prefix, formatReference(id))
+		}
+	}
+}
+
+func formatReference(id uint) string {
+	return fmt.Sprintf("ref#%d", id)
+}
+
+func updateReferencePrefix(prefix, ref string) string {
+	if prefix == "" {
+		return pointerDelimPrefix + ref + pointerDelimSuffix
+	}
+	suffix := strings.TrimPrefix(prefix, pointerDelimPrefix)
+	return pointerDelimPrefix + ref + ": " + suffix
+}
@@ -0,0 +1,407 @@
+// Copyright 2019, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package cmp
+
+import (
+	"bytes"
+	"fmt"
+	"reflect"
+	"strconv"
+	"strings"
+	"unicode"
+	"unicode/utf8"
+
+	"github.com/google/go-cmp/cmp/internal/value"
+)
+
+type formatValueOptions struct {
+	// AvoidStringer controls whether to avoid calling custom stringer
+	// methods like error.Error or fmt.Stringer.String.
+	AvoidStringer bool
+
+	// PrintAddresses controls whether to print the address of all pointers,
+	// slice elements, and maps.
+	PrintAddresses bool
+
+	// QualifiedNames controls whether FormatType uses the fully qualified name
+	// (including the full package path as opposed to just the package name).
+	QualifiedNames bool
+
+	// VerbosityLevel controls the amount of output to produce.
+	// A higher value produces more output. A value of zero or lower produces
+	// no output (represented using an ellipsis).
+	// If LimitVerbosity is false, then the level is treated as infinite.
+	VerbosityLevel int
+
+	// LimitVerbosity specifies that formatting should respect VerbosityLevel.
+	LimitVerbosity bool
+}
+
+// FormatType prints the type as if it were wrapping s.
+// This may return s as-is depending on the current type and TypeMode mode.
+func (opts formatOptions) FormatType(t reflect.Type, s textNode) textNode {
+	// Check whether to emit the type or not.
+	switch opts.TypeMode {
+	case autoType:
+		switch t.Kind() {
+		case reflect.Struct, reflect.Slice, reflect.Array, reflect.Map:
+			if s.Equal(textNil) {
+				return s
+			}
+		default:
+			return s
+		}
+		if opts.DiffMode == diffIdentical {
+			return s // elide type for identical nodes
+		}
+	case elideType:
+		return s
+	}
+
+	// Determine the type label, applying special handling for unnamed types.
+	typeName := value.TypeString(t, opts.QualifiedNames)
+	if t.Name() == "" {
+		// According to Go grammar, certain type literals contain symbols that
+		// do not strongly bind to the next lexicographical token (e.g., *T).
+		switch t.Kind() {
+		case reflect.Chan, reflect.Func, reflect.Ptr:
+			typeName = "(" + typeName + ")"
+		}
+	}
+	return &textWrap{Prefix: typeName, Value: wrapParens(s)}
+}
+
+// wrapParens wraps s with a set of parenthesis, but avoids it if the
+// wrapped node itself is already surrounded by a pair of parenthesis or braces.
+// It handles unwrapping one level of pointer-reference nodes.
+func wrapParens(s textNode) textNode {
+	var refNode *textWrap
+	if s2, ok := s.(*textWrap); ok {
+		// Unwrap a single pointer reference node.
+		switch s2.Metadata.(type) {
+		case leafReference, trunkReference, trunkReferences:
+			refNode = s2
+			if s3, ok := refNode.Value.(*textWrap); ok {
+				s2 = s3
+			}
+		}
+
+		// Already has delimiters that make parenthesis unnecessary.
+		hasParens := strings.HasPrefix(s2.Prefix, "(") && strings.HasSuffix(s2.Suffix, ")")
+		hasBraces := strings.HasPrefix(s2.Prefix, "{") && strings.HasSuffix(s2.Suffix, "}")
+		if hasParens || hasBraces {
+			return s
+		}
+	}
+	if refNode != nil {
+		refNode.Value = &textWrap{Prefix: "(", Value: refNode.Value, Suffix: ")"}
+		return s
+	}
+	return &textWrap{Prefix: "(", Value: s, Suffix: ")"}
+}
+
+// FormatValue prints the reflect.Value, taking extra care to avoid descending
+// into pointers already in ptrs. As pointers are visited, ptrs is also updated.
+func (opts formatOptions) FormatValue(v reflect.Value, parentKind reflect.Kind, ptrs *pointerReferences) (out textNode) {
+	if !v.IsValid() {
+		return nil
+	}
+	t := v.Type()
+
+	// Check slice element for cycles.
+	if parentKind == reflect.Slice {
+		ptrRef, visited := ptrs.Push(v.Addr())
+		if visited {
+			return makeLeafReference(ptrRef, false)
+		}
+		defer ptrs.Pop()
+		defer func() { out = wrapTrunkReference(ptrRef, false, out) }()
+	}
+
+	// Check whether there is an Error or String method to call.
+	if !opts.AvoidStringer && v.CanInterface() {
+		// Avoid calling Error or String methods on nil receivers since many
+		// implementations crash when doing so.
+		if (t.Kind() != reflect.Ptr && t.Kind() != reflect.Interface) || !v.IsNil() {
+			var prefix, strVal string
+			func() {
+				// Swallow and ignore any panics from String or Error.
+				defer func() { recover() }()
+				switch v := v.Interface().(type) {
+				case error:
+					strVal = v.Error()
+					prefix = "e"
+				case fmt.Stringer:
+					strVal = v.String()
+					prefix = "s"
+				}
+			}()
+			if prefix != "" {
+				return opts.formatString(prefix, strVal)
+			}
+		}
+	}
+
+	// Check whether to explicitly wrap the result with the type.
+	var skipType bool
+	defer func() {
+		if !skipType {
+			out = opts.FormatType(t, out)
+		}
+	}()
+
+	switch t.Kind() {
+	case reflect.Bool:
+		return textLine(fmt.Sprint(v.Bool()))
+	case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
+		return textLine(fmt.Sprint(v.Int()))
+	case reflect.Uint, reflect.Uint16, reflect.Uint32, reflect.Uint64:
+		return textLine(fmt.Sprint(v.Uint()))
+	case reflect.Uint8:
+		if parentKind == reflect.Slice || parentKind == reflect.Array {
+			return textLine(formatHex(v.Uint()))
+		}
+		return textLine(fmt.Sprint(v.Uint()))
+	case reflect.Uintptr:
+		return textLine(formatHex(v.Uint()))
+	case reflect.Float32, reflect.Float64:
+		return textLine(fmt.Sprint(v.Float()))
+	case reflect.Complex64, reflect.Complex128:
+		return textLine(fmt.Sprint(v.Complex()))
+	case reflect.String:
+		return opts.formatString("", v.String())
+	case reflect.UnsafePointer, reflect.Chan, reflect.Func:
+		return textLine(formatPointer(value.PointerOf(v), true))
+	case reflect.Struct:
+		var list textList
+		v := makeAddressable(v) // needed for retrieveUnexportedField
+		maxLen := v.NumField()
+		if opts.LimitVerbosity {
+			maxLen = ((1 << opts.verbosity()) >> 1) << 2 // 0, 4, 8, 16, 32, etc...
+			opts.VerbosityLevel--
+		}
+		for i := 0; i < v.NumField(); i++ {
+			vv := v.Field(i)
+			if value.IsZero(vv) {
+				continue // Elide fields with zero values
+			}
+			if len(list) == maxLen {
+				list.AppendEllipsis(diffStats{})
+				break
+			}
+			sf := t.Field(i)
+			if supportExporters && !isExported(sf.Name) {
+				vv = retrieveUnexportedField(v, sf, true)
+			}
+			s := opts.WithTypeMode(autoType).FormatValue(vv, t.Kind(), ptrs)
+			list = append(list, textRecord{Key: sf.Name, Value: s})
+		}
+		return &textWrap{Prefix: "{", Value: list, Suffix: "}"}
+	case reflect.Slice:
+		if v.IsNil() {
+			return textNil
+		}
+
+		// Check whether this is a []byte of text data.
+		if t.Elem() == reflect.TypeOf(byte(0)) {
+			b := v.Bytes()
+			isPrintSpace := func(r rune) bool { return unicode.IsPrint(r) || unicode.IsSpace(r) }
+			if len(b) > 0 && utf8.Valid(b) && len(bytes.TrimFunc(b, isPrintSpace)) == 0 {
+				out = opts.formatString("", string(b))
+				skipType = true
+				return opts.FormatType(t, out)
+			}
+		}
+
+		fallthrough
+	case reflect.Array:
+		maxLen := v.Len()
+		if opts.LimitVerbosity {
+			maxLen = ((1 << opts.verbosity()) >> 1) << 2 // 0, 4, 8, 16, 32, etc...
+			opts.VerbosityLevel--
+		}
+		var list textList
+		for i := 0; i < v.Len(); i++ {
+			if len(list) == maxLen {
+				list.AppendEllipsis(diffStats{})
+				break
+			}
+			s := opts.WithTypeMode(elideType).FormatValue(v.Index(i), t.Kind(), ptrs)
+			list = append(list, textRecord{Value: s})
+		}
+
+		out = &textWrap{Prefix: "{", Value: list, Suffix: "}"}
+		if t.Kind() == reflect.Slice && opts.PrintAddresses {
+			header := fmt.Sprintf("ptr:%v, len:%d, cap:%d", formatPointer(value.PointerOf(v), false), v.Len(), v.Cap())
+			out = &textWrap{Prefix: pointerDelimPrefix + header + pointerDelimSuffix, Value: out}
+		}
+		return out
+	case reflect.Map:
+		if v.IsNil() {
+			return textNil
+		}
+
+		// Check pointer for cycles.
+		ptrRef, visited := ptrs.Push(v)
+		if visited {
+			return makeLeafReference(ptrRef, opts.PrintAddresses)
+		}
+		defer ptrs.Pop()
+
+		maxLen := v.Len()
+		if opts.LimitVerbosity {
+			maxLen = ((1 << opts.verbosity()) >> 1) << 2 // 0, 4, 8, 16, 32, etc...
+			opts.VerbosityLevel--
+		}
+		var list textList
+		for _, k := range value.SortKeys(v.MapKeys()) {
+			if len(list) == maxLen {
+				list.AppendEllipsis(diffStats{})
+				break
+			}
+			sk := formatMapKey(k, false, ptrs)
+			sv := opts.WithTypeMode(elideType).FormatValue(v.MapIndex(k), t.Kind(), ptrs)
+			list = append(list, textRecord{Key: sk, Value: sv})
+		}
+
+		out = &textWrap{Prefix: "{", Value: list, Suffix: "}"}
+		out = wrapTrunkReference(ptrRef, opts.PrintAddresses, out)
+		return out
+	case reflect.Ptr:
+		if v.IsNil() {
+			return textNil
+		}
+
+		// Check pointer for cycles.
+		ptrRef, visited := ptrs.Push(v)
+		if visited {
+			out = makeLeafReference(ptrRef, opts.PrintAddresses)
+			return &textWrap{Prefix: "&", Value: out}
+		}
+		defer ptrs.Pop()
+
+		// Skip the name only if this is an unnamed pointer type.
+		// Otherwise taking the address of a value does not reproduce
+		// the named pointer type.
+		if v.Type().Name() == "" {
+			skipType = true // Let the underlying value print the type instead
+		}
+		out = opts.FormatValue(v.Elem(), t.Kind(), ptrs)
+		out = wrapTrunkReference(ptrRef, opts.PrintAddresses, out)
+		out = &textWrap{Prefix: "&", Value: out}
+		return out
+	case reflect.Interface:
+		if v.IsNil() {
+			return textNil
+		}
+		// Interfaces accept different concrete types,
+		// so configure the underlying value to explicitly print the type.
+		return opts.WithTypeMode(emitType).FormatValue(v.Elem(), t.Kind(), ptrs)
+	default:
+		panic(fmt.Sprintf("%v kind not handled", v.Kind()))
+	}
+}
+
+func (opts formatOptions) formatString(prefix, s string) textNode {
+	maxLen := len(s)
+	maxLines := strings.Count(s, "\n") + 1
+	if opts.LimitVerbosity {
+		maxLen = (1 << opts.verbosity()) << 5   // 32, 64, 128, 256, etc...
+		maxLines = (1 << opts.verbosity()) << 2 //  4, 8, 16, 32, 64, etc...
+	}
+
+	// For multiline strings, use the triple-quote syntax,
+	// but only use it when printing removed or inserted nodes since
+	// we only want the extra verbosity for those cases.
+	lines := strings.Split(strings.TrimSuffix(s, "\n"), "\n")
+	isTripleQuoted := len(lines) >= 4 && (opts.DiffMode == '-' || opts.DiffMode == '+')
+	for i := 0; i < len(lines) && isTripleQuoted; i++ {
+		lines[i] = strings.TrimPrefix(strings.TrimSuffix(lines[i], "\r"), "\r") // trim leading/trailing carriage returns for legacy Windows endline support
+		isPrintable := func(r rune) bool {
+			return unicode.IsPrint(r) || r == '\t' // specially treat tab as printable
+		}
+		line := lines[i]
+		isTripleQuoted = !strings.HasPrefix(strings.TrimPrefix(line, prefix), `"""`) && !strings.HasPrefix(line, "...") && strings.TrimFunc(line, isPrintable) == "" && len(line) <= maxLen
+	}
+	if isTripleQuoted {
+		var list textList
+		list = append(list, textRecord{Diff: opts.DiffMode, Value: textLine(prefix + `"""`), ElideComma: true})
+		for i, line := range lines {
+			if numElided := len(lines) - i; i == maxLines-1 && numElided > 1 {
+				comment := commentString(fmt.Sprintf("%d elided lines", numElided))
+				list = append(list, textRecord{Diff: opts.DiffMode, Value: textEllipsis, ElideComma: true, Comment: comment})
+				break
+			}
+			list = append(list, textRecord{Diff: opts.DiffMode, Value: textLine(line), ElideComma: true})
+		}
+		list = append(list, textRecord{Diff: opts.DiffMode, Value: textLine(prefix + `"""`), ElideComma: true})
+		return &textWrap{Prefix: "(", Value: list, Suffix: ")"}
+	}
+
+	// Format the string as a single-line quoted string.
+	if len(s) > maxLen+len(textEllipsis) {
+		return textLine(prefix + formatString(s[:maxLen]) + string(textEllipsis))
+	}
+	return textLine(prefix + formatString(s))
+}
+
+// formatMapKey formats v as if it were a map key.
+// The result is guaranteed to be a single line.
+func formatMapKey(v reflect.Value, disambiguate bool, ptrs *pointerReferences) string {
+	var opts formatOptions
+	opts.DiffMode = diffIdentical
+	opts.TypeMode = elideType
+	opts.PrintAddresses = disambiguate
+	opts.AvoidStringer = disambiguate
+	opts.QualifiedNames = disambiguate
+	opts.VerbosityLevel = maxVerbosityPreset
+	opts.LimitVerbosity = true
+	s := opts.FormatValue(v, reflect.Map, ptrs).String()
+	return strings.TrimSpace(s)
+}
+
+// formatString prints s as a double-quoted or backtick-quoted string.
+func formatString(s string) string {
+	// Use quoted string if it the same length as a raw string literal.
+	// Otherwise, attempt to use the raw string form.
+	qs := strconv.Quote(s)
+	if len(qs) == 1+len(s)+1 {
+		return qs
+	}
+
+	// Disallow newlines to ensure output is a single line.
+	// Only allow printable runes for readability purposes.
+	rawInvalid := func(r rune) bool {
+		return r == '`' || r == '\n' || !(unicode.IsPrint(r) || r == '\t')
+	}
+	if utf8.ValidString(s) && strings.IndexFunc(s, rawInvalid) < 0 {
+		return "`" + s + "`"
+	}
+	return qs
+}
+
+// formatHex prints u as a hexadecimal integer in Go notation.
+func formatHex(u uint64) string {
+	var f string
+	switch {
+	case u <= 0xff:
+		f = "0x%02x"
+	case u <= 0xffff:
+		f = "0x%04x"
+	case u <= 0xffffff:
+		f = "0x%06x"
+	case u <= 0xffffffff:
+		f = "0x%08x"
+	case u <= 0xffffffffff:
+		f = "0x%010x"
+	case u <= 0xffffffffffff:
+		f = "0x%012x"
+	case u <= 0xffffffffffffff:
+		f = "0x%014x"
+	case u <= 0xffffffffffffffff:
+		f = "0x%016x"
+	}
+	return fmt.Sprintf(f, u)
+}
@@ -0,0 +1,613 @@
+// Copyright 2019, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package cmp
+
+import (
+	"bytes"
+	"fmt"
+	"math"
+	"reflect"
+	"strconv"
+	"strings"
+	"unicode"
+	"unicode/utf8"
+
+	"github.com/google/go-cmp/cmp/internal/diff"
+)
+
+// CanFormatDiffSlice reports whether we support custom formatting for nodes
+// that are slices of primitive kinds or strings.
+func (opts formatOptions) CanFormatDiffSlice(v *valueNode) bool {
+	switch {
+	case opts.DiffMode != diffUnknown:
+		return false // Must be formatting in diff mode
+	case v.NumDiff == 0:
+		return false // No differences detected
+	case !v.ValueX.IsValid() || !v.ValueY.IsValid():
+		return false // Both values must be valid
+	case v.NumIgnored > 0:
+		return false // Some ignore option was used
+	case v.NumTransformed > 0:
+		return false // Some transform option was used
+	case v.NumCompared > 1:
+		return false // More than one comparison was used
+	case v.NumCompared == 1 && v.Type.Name() != "":
+		// The need for cmp to check applicability of options on every element
+		// in a slice is a significant performance detriment for large []byte.
+		// The workaround is to specify Comparer(bytes.Equal),
+		// which enables cmp to compare []byte more efficiently.
+		// If they differ, we still want to provide batched diffing.
+		// The logic disallows named types since they tend to have their own
+		// String method, with nicer formatting than what this provides.
+		return false
+	}
+
+	// Check whether this is an interface with the same concrete types.
+	t := v.Type
+	vx, vy := v.ValueX, v.ValueY
+	if t.Kind() == reflect.Interface && !vx.IsNil() && !vy.IsNil() && vx.Elem().Type() == vy.Elem().Type() {
+		vx, vy = vx.Elem(), vy.Elem()
+		t = vx.Type()
+	}
+
+	// Check whether we provide specialized diffing for this type.
+	switch t.Kind() {
+	case reflect.String:
+	case reflect.Array, reflect.Slice:
+		// Only slices of primitive types have specialized handling.
+		switch t.Elem().Kind() {
+		case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64,
+			reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr,
+			reflect.Bool, reflect.Float32, reflect.Float64, reflect.Complex64, reflect.Complex128:
+		default:
+			return false
+		}
+
+		// Both slice values have to be non-empty.
+		if t.Kind() == reflect.Slice && (vx.Len() == 0 || vy.Len() == 0) {
+			return false
+		}
+
+		// If a sufficient number of elements already differ,
+		// use specialized formatting even if length requirement is not met.
+		if v.NumDiff > v.NumSame {
+			return true
+		}
+	default:
+		return false
+	}
+
+	// Use specialized string diffing for longer slices or strings.
+	const minLength = 32
+	return vx.Len() >= minLength && vy.Len() >= minLength
+}
+
+// FormatDiffSlice prints a diff for the slices (or strings) represented by v.
+// This provides custom-tailored logic to make printing of differences in
+// textual strings and slices of primitive kinds more readable.
+func (opts formatOptions) FormatDiffSlice(v *valueNode) textNode {
+	assert(opts.DiffMode == diffUnknown)
+	t, vx, vy := v.Type, v.ValueX, v.ValueY
+	if t.Kind() == reflect.Interface {
+		vx, vy = vx.Elem(), vy.Elem()
+		t = vx.Type()
+		opts = opts.WithTypeMode(emitType)
+	}
+
+	// Auto-detect the type of the data.
+	var sx, sy string
+	var ssx, ssy []string
+	var isString, isMostlyText, isPureLinedText, isBinary bool
+	switch {
+	case t.Kind() == reflect.String:
+		sx, sy = vx.String(), vy.String()
+		isString = true
+	case t.Kind() == reflect.Slice && t.Elem() == reflect.TypeOf(byte(0)):
+		sx, sy = string(vx.Bytes()), string(vy.Bytes())
+		isString = true
+	case t.Kind() == reflect.Array:
+		// Arrays need to be addressable for slice operations to work.
+		vx2, vy2 := reflect.New(t).Elem(), reflect.New(t).Elem()
+		vx2.Set(vx)
+		vy2.Set(vy)
+		vx, vy = vx2, vy2
+	}
+	if isString {
+		var numTotalRunes, numValidRunes, numLines, lastLineIdx, maxLineLen int
+		for i, r := range sx + sy {
+			numTotalRunes++
+			if (unicode.IsPrint(r) || unicode.IsSpace(r)) && r != utf8.RuneError {
+				numValidRunes++
+			}
+			if r == '\n' {
+				if maxLineLen < i-lastLineIdx {
+					maxLineLen = i - lastLineIdx
+				}
+				lastLineIdx = i + 1
+				numLines++
+			}
+		}
+		isPureText := numValidRunes == numTotalRunes
+		isMostlyText = float64(numValidRunes) > math.Floor(0.90*float64(numTotalRunes))
+		isPureLinedText = isPureText && numLines >= 4 && maxLineLen <= 1024
+		isBinary = !isMostlyText
+
+		// Avoid diffing by lines if it produces a significantly more complex
+		// edit script than diffing by bytes.
+		if isPureLinedText {
+			ssx = strings.Split(sx, "\n")
+			ssy = strings.Split(sy, "\n")
+			esLines := diff.Difference(len(ssx), len(ssy), func(ix, iy int) diff.Result {
+				return diff.BoolResult(ssx[ix] == ssy[iy])
+			})
+			esBytes := diff.Difference(len(sx), len(sy), func(ix, iy int) diff.Result {
+				return diff.BoolResult(sx[ix] == sy[iy])
+			})
+			efficiencyLines := float64(esLines.Dist()) / float64(len(esLines))
+			efficiencyBytes := float64(esBytes.Dist()) / float64(len(esBytes))
+			isPureLinedText = efficiencyLines < 4*efficiencyBytes
+		}
+	}
+
+	// Format the string into printable records.
+	var list textList
+	var delim string
+	switch {
+	// If the text appears to be multi-lined text,
+	// then perform differencing across individual lines.
+	case isPureLinedText:
+		list = opts.formatDiffSlice(
+			reflect.ValueOf(ssx), reflect.ValueOf(ssy), 1, "line",
+			func(v reflect.Value, d diffMode) textRecord {
+				s := formatString(v.Index(0).String())
+				return textRecord{Diff: d, Value: textLine(s)}
+			},
+		)
+		delim = "\n"
+
+		// If possible, use a custom triple-quote (""") syntax for printing
+		// differences in a string literal. This format is more readable,
+		// but has edge-cases where differences are visually indistinguishable.
+		// This format is avoided under the following conditions:
+		//	• A line starts with `"""`
+		//	• A line starts with "..."
+		//	• A line contains non-printable characters
+		//	• Adjacent different lines differ only by whitespace
+		//
+		// For example:
+		//		"""
+		//		... // 3 identical lines
+		//		foo
+		//		bar
+		//	-	baz
+		//	+	BAZ
+		//		"""
+		isTripleQuoted := true
+		prevRemoveLines := map[string]bool{}
+		prevInsertLines := map[string]bool{}
+		var list2 textList
+		list2 = append(list2, textRecord{Value: textLine(`"""`), ElideComma: true})
+		for _, r := range list {
+			if !r.Value.Equal(textEllipsis) {
+				line, _ := strconv.Unquote(string(r.Value.(textLine)))
+				line = strings.TrimPrefix(strings.TrimSuffix(line, "\r"), "\r") // trim leading/trailing carriage returns for legacy Windows endline support
+				normLine := strings.Map(func(r rune) rune {
+					if unicode.IsSpace(r) {
+						return -1 // drop whitespace to avoid visually indistinguishable output
+					}
+					return r
+				}, line)
+				isPrintable := func(r rune) bool {
+					return unicode.IsPrint(r) || r == '\t' // specially treat tab as printable
+				}
+				isTripleQuoted = !strings.HasPrefix(line, `"""`) && !strings.HasPrefix(line, "...") && strings.TrimFunc(line, isPrintable) == ""
+				switch r.Diff {
+				case diffRemoved:
+					isTripleQuoted = isTripleQuoted && !prevInsertLines[normLine]
+					prevRemoveLines[normLine] = true
+				case diffInserted:
+					isTripleQuoted = isTripleQuoted && !prevRemoveLines[normLine]
+					prevInsertLines[normLine] = true
+				}
+				if !isTripleQuoted {
+					break
+				}
+				r.Value = textLine(line)
+				r.ElideComma = true
+			}
+			if !(r.Diff == diffRemoved || r.Diff == diffInserted) { // start a new non-adjacent difference group
+				prevRemoveLines = map[string]bool{}
+				prevInsertLines = map[string]bool{}
+			}
+			list2 = append(list2, r)
+		}
+		if r := list2[len(list2)-1]; r.Diff == diffIdentical && len(r.Value.(textLine)) == 0 {
+			list2 = list2[:len(list2)-1] // elide single empty line at the end
+		}
+		list2 = append(list2, textRecord{Value: textLine(`"""`), ElideComma: true})
+		if isTripleQuoted {
+			var out textNode = &textWrap{Prefix: "(", Value: list2, Suffix: ")"}
+			switch t.Kind() {
+			case reflect.String:
+				if t != reflect.TypeOf(string("")) {
+					out = opts.FormatType(t, out)
+				}
+			case reflect.Slice:
+				// Always emit type for slices since the triple-quote syntax
+				// looks like a string (not a slice).
+				opts = opts.WithTypeMode(emitType)
+				out = opts.FormatType(t, out)
+			}
+			return out
+		}
+
+	// If the text appears to be single-lined text,
+	// then perform differencing in approximately fixed-sized chunks.
+	// The output is printed as quoted strings.
+	case isMostlyText:
+		list = opts.formatDiffSlice(
+			reflect.ValueOf(sx), reflect.ValueOf(sy), 64, "byte",
+			func(v reflect.Value, d diffMode) textRecord {
+				s := formatString(v.String())
+				return textRecord{Diff: d, Value: textLine(s)}
+			},
+		)
+
+	// If the text appears to be binary data,
+	// then perform differencing in approximately fixed-sized chunks.
+	// The output is inspired by hexdump.
+	case isBinary:
+		list = opts.formatDiffSlice(
+			reflect.ValueOf(sx), reflect.ValueOf(sy), 16, "byte",
+			func(v reflect.Value, d diffMode) textRecord {
+				var ss []string
+				for i := 0; i < v.Len(); i++ {
+					ss = append(ss, formatHex(v.Index(i).Uint()))
+				}
+				s := strings.Join(ss, ", ")
+				comment := commentString(fmt.Sprintf("%c|%v|", d, formatASCII(v.String())))
+				return textRecord{Diff: d, Value: textLine(s), Comment: comment}
+			},
+		)
+
+	// For all other slices of primitive types,
+	// then perform differencing in approximately fixed-sized chunks.
+	// The size of each chunk depends on the width of the element kind.
+	default:
+		var chunkSize int
+		if t.Elem().Kind() == reflect.Bool {
+			chunkSize = 16
+		} else {
+			switch t.Elem().Bits() {
+			case 8:
+				chunkSize = 16
+			case 16:
+				chunkSize = 12
+			case 32:
+				chunkSize = 8
+			default:
+				chunkSize = 8
+			}
+		}
+		list = opts.formatDiffSlice(
+			vx, vy, chunkSize, t.Elem().Kind().String(),
+			func(v reflect.Value, d diffMode) textRecord {
+				var ss []string
+				for i := 0; i < v.Len(); i++ {
+					switch t.Elem().Kind() {
+					case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
+						ss = append(ss, fmt.Sprint(v.Index(i).Int()))
+					case reflect.Uint, reflect.Uint16, reflect.Uint32, reflect.Uint64:
+						ss = append(ss, fmt.Sprint(v.Index(i).Uint()))
+					case reflect.Uint8, reflect.Uintptr:
+						ss = append(ss, formatHex(v.Index(i).Uint()))
+					case reflect.Bool, reflect.Float32, reflect.Float64, reflect.Complex64, reflect.Complex128:
+						ss = append(ss, fmt.Sprint(v.Index(i).Interface()))
+					}
+				}
+				s := strings.Join(ss, ", ")
+				return textRecord{Diff: d, Value: textLine(s)}
+			},
+		)
+	}
+
+	// Wrap the output with appropriate type information.
+	var out textNode = &textWrap{Prefix: "{", Value: list, Suffix: "}"}
+	if !isMostlyText {
+		// The "{...}" byte-sequence literal is not valid Go syntax for strings.
+		// Emit the type for extra clarity (e.g. "string{...}").
+		if t.Kind() == reflect.String {
+			opts = opts.WithTypeMode(emitType)
+		}
+		return opts.FormatType(t, out)
+	}
+	switch t.Kind() {
+	case reflect.String:
+		out = &textWrap{Prefix: "strings.Join(", Value: out, Suffix: fmt.Sprintf(", %q)", delim)}
+		if t != reflect.TypeOf(string("")) {
+			out = opts.FormatType(t, out)
+		}
+	case reflect.Slice:
+		out = &textWrap{Prefix: "bytes.Join(", Value: out, Suffix: fmt.Sprintf(", %q)", delim)}
+		if t != reflect.TypeOf([]byte(nil)) {
+			out = opts.FormatType(t, out)
+		}
+	}
+	return out
+}
+
+// formatASCII formats s as an ASCII string.
+// This is useful for printing binary strings in a semi-legible way.
+func formatASCII(s string) string {
+	b := bytes.Repeat([]byte{'.'}, len(s))
+	for i := 0; i < len(s); i++ {
+		if ' ' <= s[i] && s[i] <= '~' {
+			b[i] = s[i]
+		}
+	}
+	return string(b)
+}
+
+func (opts formatOptions) formatDiffSlice(
+	vx, vy reflect.Value, chunkSize int, name string,
+	makeRec func(reflect.Value, diffMode) textRecord,
+) (list textList) {
+	eq := func(ix, iy int) bool {
+		return vx.Index(ix).Interface() == vy.Index(iy).Interface()
+	}
+	es := diff.Difference(vx.Len(), vy.Len(), func(ix, iy int) diff.Result {
+		return diff.BoolResult(eq(ix, iy))
+	})
+
+	appendChunks := func(v reflect.Value, d diffMode) int {
+		n0 := v.Len()
+		for v.Len() > 0 {
+			n := chunkSize
+			if n > v.Len() {
+				n = v.Len()
+			}
+			list = append(list, makeRec(v.Slice(0, n), d))
+			v = v.Slice(n, v.Len())
+		}
+		return n0 - v.Len()
+	}
+
+	var numDiffs int
+	maxLen := -1
+	if opts.LimitVerbosity {
+		maxLen = (1 << opts.verbosity()) << 2 // 4, 8, 16, 32, 64, etc...
+		opts.VerbosityLevel--
+	}
+
+	groups := coalesceAdjacentEdits(name, es)
+	groups = coalesceInterveningIdentical(groups, chunkSize/4)
+	groups = cleanupSurroundingIdentical(groups, eq)
+	maxGroup := diffStats{Name: name}
+	for i, ds := range groups {
+		if maxLen >= 0 && numDiffs >= maxLen {
+			maxGroup = maxGroup.Append(ds)
+			continue
+		}
+
+		// Print equal.
+		if ds.NumDiff() == 0 {
+			// Compute the number of leading and trailing equal bytes to print.
+			var numLo, numHi int
+			numEqual := ds.NumIgnored + ds.NumIdentical
+			for numLo < chunkSize*numContextRecords && numLo+numHi < numEqual && i != 0 {
+				numLo++
+			}
+			for numHi < chunkSize*numContextRecords && numLo+numHi < numEqual && i != len(groups)-1 {
+				numHi++
+			}
+			if numEqual-(numLo+numHi) <= chunkSize && ds.NumIgnored == 0 {
+				numHi = numEqual - numLo // Avoid pointless coalescing of single equal row
+			}
+
+			// Print the equal bytes.
+			appendChunks(vx.Slice(0, numLo), diffIdentical)
+			if numEqual > numLo+numHi {
+				ds.NumIdentical -= numLo + numHi
+				list.AppendEllipsis(ds)
+			}
+			appendChunks(vx.Slice(numEqual-numHi, numEqual), diffIdentical)
+			vx = vx.Slice(numEqual, vx.Len())
+			vy = vy.Slice(numEqual, vy.Len())
+			continue
+		}
+
+		// Print unequal.
+		len0 := len(list)
+		nx := appendChunks(vx.Slice(0, ds.NumIdentical+ds.NumRemoved+ds.NumModified), diffRemoved)
+		vx = vx.Slice(nx, vx.Len())
+		ny := appendChunks(vy.Slice(0, ds.NumIdentical+ds.NumInserted+ds.NumModified), diffInserted)
+		vy = vy.Slice(ny, vy.Len())
+		numDiffs += len(list) - len0
+	}
+	if maxGroup.IsZero() {
+		assert(vx.Len() == 0 && vy.Len() == 0)
+	} else {
+		list.AppendEllipsis(maxGroup)
+	}
+	return list
+}
+
+// coalesceAdjacentEdits coalesces the list of edits into groups of adjacent
+// equal or unequal counts.
+//
+// Example:
+//
+//	Input:  "..XXY...Y"
+//	Output: [
+//		{NumIdentical: 2},
+//		{NumRemoved: 2, NumInserted 1},
+//		{NumIdentical: 3},
+//		{NumInserted: 1},
+//	]
+//
+func coalesceAdjacentEdits(name string, es diff.EditScript) (groups []diffStats) {
+	var prevMode byte
+	lastStats := func(mode byte) *diffStats {
+		if prevMode != mode {
+			groups = append(groups, diffStats{Name: name})
+			prevMode = mode
+		}
+		return &groups[len(groups)-1]
+	}
+	for _, e := range es {
+		switch e {
+		case diff.Identity:
+			lastStats('=').NumIdentical++
+		case diff.UniqueX:
+			lastStats('!').NumRemoved++
+		case diff.UniqueY:
+			lastStats('!').NumInserted++
+		case diff.Modified:
+			lastStats('!').NumModified++
+		}
+	}
+	return groups
+}
+
+// coalesceInterveningIdentical coalesces sufficiently short (<= windowSize)
+// equal groups into adjacent unequal groups that currently result in a
+// dual inserted/removed printout. This acts as a high-pass filter to smooth
+// out high-frequency changes within the windowSize.
+//
+// Example:
+//
+//	WindowSize: 16,
+//	Input: [
+//		{NumIdentical: 61},              // group 0
+//		{NumRemoved: 3, NumInserted: 1}, // group 1
+//		{NumIdentical: 6},               // ├── coalesce
+//		{NumInserted: 2},                // ├── coalesce
+//		{NumIdentical: 1},               // ├── coalesce
+//		{NumRemoved: 9},                 // └── coalesce
+//		{NumIdentical: 64},              // group 2
+//		{NumRemoved: 3, NumInserted: 1}, // group 3
+//		{NumIdentical: 6},               // ├── coalesce
+//		{NumInserted: 2},                // ├── coalesce
+//		{NumIdentical: 1},               // ├── coalesce
+//		{NumRemoved: 7},                 // ├── coalesce
+//		{NumIdentical: 1},               // ├── coalesce
+//		{NumRemoved: 2},                 // └── coalesce
+//		{NumIdentical: 63},              // group 4
+//	]
+//	Output: [
+//		{NumIdentical: 61},
+//		{NumIdentical: 7, NumRemoved: 12, NumInserted: 3},
+//		{NumIdentical: 64},
+//		{NumIdentical: 8, NumRemoved: 12, NumInserted: 3},
+//		{NumIdentical: 63},
+//	]
+//
+func coalesceInterveningIdentical(groups []diffStats, windowSize int) []diffStats {
+	groups, groupsOrig := groups[:0], groups
+	for i, ds := range groupsOrig {
+		if len(groups) >= 2 && ds.NumDiff() > 0 {
+			prev := &groups[len(groups)-2] // Unequal group
+			curr := &groups[len(groups)-1] // Equal group
+			next := &groupsOrig[i]         // Unequal group
+			hadX, hadY := prev.NumRemoved > 0, prev.NumInserted > 0
+			hasX, hasY := next.NumRemoved > 0, next.NumInserted > 0
+			if ((hadX || hasX) && (hadY || hasY)) && curr.NumIdentical <= windowSize {
+				*prev = prev.Append(*curr).Append(*next)
+				groups = groups[:len(groups)-1] // Truncate off equal group
+				continue
+			}
+		}
+		groups = append(groups, ds)
+	}
+	return groups
+}
+
+// cleanupSurroundingIdentical scans through all unequal groups, and
+// moves any leading sequence of equal elements to the preceding equal group and
+// moves and trailing sequence of equal elements to the succeeding equal group.
+//
+// This is necessary since coalesceInterveningIdentical may coalesce edit groups
+// together such that leading/trailing spans of equal elements becomes possible.
+// Note that this can occur even with an optimal diffing algorithm.
+//
+// Example:
+//
+//	Input: [
+//		{NumIdentical: 61},
+//		{NumIdentical: 1 , NumRemoved: 11, NumInserted: 2}, // assume 3 leading identical elements
+//		{NumIdentical: 67},
+//		{NumIdentical: 7, NumRemoved: 12, NumInserted: 3},  // assume 10 trailing identical elements
+//		{NumIdentical: 54},
+//	]
+//	Output: [
+//		{NumIdentical: 64}, // incremented by 3
+//		{NumRemoved: 9},
+//		{NumIdentical: 67},
+//		{NumRemoved: 9},
+//		{NumIdentical: 64}, // incremented by 10
+//	]
+//
+func cleanupSurroundingIdentical(groups []diffStats, eq func(i, j int) bool) []diffStats {
+	var ix, iy int // indexes into sequence x and y
+	for i, ds := range groups {
+		// Handle equal group.
+		if ds.NumDiff() == 0 {
+			ix += ds.NumIdentical
+			iy += ds.NumIdentical
+			continue
+		}
+
+		// Handle unequal group.
+		nx := ds.NumIdentical + ds.NumRemoved + ds.NumModified
+		ny := ds.NumIdentical + ds.NumInserted + ds.NumModified
+		var numLeadingIdentical, numTrailingIdentical int
+		for j := 0; j < nx && j < ny && eq(ix+j, iy+j); j++ {
+			numLeadingIdentical++
+		}
+		for j := 0; j < nx && j < ny && eq(ix+nx-1-j, iy+ny-1-j); j++ {
+			numTrailingIdentical++
+		}
+		if numIdentical := numLeadingIdentical + numTrailingIdentical; numIdentical > 0 {
+			if numLeadingIdentical > 0 {
+				// Remove leading identical span from this group and
+				// insert it into the preceding group.
+				if i-1 >= 0 {
+					groups[i-1].NumIdentical += numLeadingIdentical
+				} else {
+					// No preceding group exists, so prepend a new group,
+					// but do so after we finish iterating over all groups.
+					defer func() {
+						groups = append([]diffStats{{Name: groups[0].Name, NumIdentical: numLeadingIdentical}}, groups...)
+					}()
+				}
+				// Increment indexes since the preceding group would have handled this.
+				ix += numLeadingIdentical
+				iy += numLeadingIdentical
+			}
+			if numTrailingIdentical > 0 {
+				// Remove trailing identical span from this group and
+				// insert it into the succeeding group.
+				if i+1 < len(groups) {
+					groups[i+1].NumIdentical += numTrailingIdentical
+				} else {
+					// No succeeding group exists, so append a new group,
+					// but do so after we finish iterating over all groups.
+					defer func() {
+						groups = append(groups, diffStats{Name: groups[len(groups)-1].Name, NumIdentical: numTrailingIdentical})
+					}()
+				}
+				// Do not increment indexes since the succeeding group will handle this.
+			}
+
+			// Update this group since some identical elements were removed.
+			nx -= numIdentical
+			ny -= numIdentical
+			groups[i] = diffStats{Name: ds.Name, NumRemoved: nx, NumInserted: ny}
+		}
+		ix += nx
+		iy += ny
+	}
+	return groups
+}
@@ -0,0 +1,431 @@
+// Copyright 2019, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package cmp
+
+import (
+	"bytes"
+	"fmt"
+	"math/rand"
+	"strings"
+	"time"
+	"unicode/utf8"
+
+	"github.com/google/go-cmp/cmp/internal/flags"
+)
+
+var randBool = rand.New(rand.NewSource(time.Now().Unix())).Intn(2) == 0
+
+const maxColumnLength = 80
+
+type indentMode int
+
+func (n indentMode) appendIndent(b []byte, d diffMode) []byte {
+	// The output of Diff is documented as being unstable to provide future
+	// flexibility in changing the output for more humanly readable reports.
+	// This logic intentionally introduces instability to the exact output
+	// so that users can detect accidental reliance on stability early on,
+	// rather than much later when an actual change to the format occurs.
+	if flags.Deterministic || randBool {
+		// Use regular spaces (U+0020).
+		switch d {
+		case diffUnknown, diffIdentical:
+			b = append(b, "  "...)
+		case diffRemoved:
+			b = append(b, "- "...)
+		case diffInserted:
+			b = append(b, "+ "...)
+		}
+	} else {
+		// Use non-breaking spaces (U+00a0).
+		switch d {
+		case diffUnknown, diffIdentical:
+			b = append(b, "  "...)
+		case diffRemoved:
+			b = append(b, "- "...)
+		case diffInserted:
+			b = append(b, "+ "...)
+		}
+	}
+	return repeatCount(n).appendChar(b, '\t')
+}
+
+type repeatCount int
+
+func (n repeatCount) appendChar(b []byte, c byte) []byte {
+	for ; n > 0; n-- {
+		b = append(b, c)
+	}
+	return b
+}
+
+// textNode is a simplified tree-based representation of structured text.
+// Possible node types are textWrap, textList, or textLine.
+type textNode interface {
+	// Len reports the length in bytes of a single-line version of the tree.
+	// Nested textRecord.Diff and textRecord.Comment fields are ignored.
+	Len() int
+	// Equal reports whether the two trees are structurally identical.
+	// Nested textRecord.Diff and textRecord.Comment fields are compared.
+	Equal(textNode) bool
+	// String returns the string representation of the text tree.
+	// It is not guaranteed that len(x.String()) == x.Len(),
+	// nor that x.String() == y.String() implies that x.Equal(y).
+	String() string
+
+	// formatCompactTo formats the contents of the tree as a single-line string
+	// to the provided buffer. Any nested textRecord.Diff and textRecord.Comment
+	// fields are ignored.
+	//
+	// However, not all nodes in the tree should be collapsed as a single-line.
+	// If a node can be collapsed as a single-line, it is replaced by a textLine
+	// node. Since the top-level node cannot replace itself, this also returns
+	// the current node itself.
+	//
+	// This does not mutate the receiver.
+	formatCompactTo([]byte, diffMode) ([]byte, textNode)
+	// formatExpandedTo formats the contents of the tree as a multi-line string
+	// to the provided buffer. In order for column alignment to operate well,
+	// formatCompactTo must be called before calling formatExpandedTo.
+	formatExpandedTo([]byte, diffMode, indentMode) []byte
+}
+
+// textWrap is a wrapper that concatenates a prefix and/or a suffix
+// to the underlying node.
+type textWrap struct {
+	Prefix   string      // e.g., "bytes.Buffer{"
+	Value    textNode    // textWrap | textList | textLine
+	Suffix   string      // e.g., "}"
+	Metadata interface{} // arbitrary metadata; has no effect on formatting
+}
+
+func (s *textWrap) Len() int {
+	return len(s.Prefix) + s.Value.Len() + len(s.Suffix)
+}
+func (s1 *textWrap) Equal(s2 textNode) bool {
+	if s2, ok := s2.(*textWrap); ok {
+		return s1.Prefix == s2.Prefix && s1.Value.Equal(s2.Value) && s1.Suffix == s2.Suffix
+	}
+	return false
+}
+func (s *textWrap) String() string {
+	var d diffMode
+	var n indentMode
+	_, s2 := s.formatCompactTo(nil, d)
+	b := n.appendIndent(nil, d)      // Leading indent
+	b = s2.formatExpandedTo(b, d, n) // Main body
+	b = append(b, '\n')              // Trailing newline
+	return string(b)
+}
+func (s *textWrap) formatCompactTo(b []byte, d diffMode) ([]byte, textNode) {
+	n0 := len(b) // Original buffer length
+	b = append(b, s.Prefix...)
+	b, s.Value = s.Value.formatCompactTo(b, d)
+	b = append(b, s.Suffix...)
+	if _, ok := s.Value.(textLine); ok {
+		return b, textLine(b[n0:])
+	}
+	return b, s
+}
+func (s *textWrap) formatExpandedTo(b []byte, d diffMode, n indentMode) []byte {
+	b = append(b, s.Prefix...)
+	b = s.Value.formatExpandedTo(b, d, n)
+	b = append(b, s.Suffix...)
+	return b
+}
+
+// textList is a comma-separated list of textWrap or textLine nodes.
+// The list may be formatted as multi-lines or single-line at the discretion
+// of the textList.formatCompactTo method.
+type textList []textRecord
+type textRecord struct {
+	Diff       diffMode     // e.g., 0 or '-' or '+'
+	Key        string       // e.g., "MyField"
+	Value      textNode     // textWrap | textLine
+	ElideComma bool         // avoid trailing comma
+	Comment    fmt.Stringer // e.g., "6 identical fields"
+}
+
+// AppendEllipsis appends a new ellipsis node to the list if none already
+// exists at the end. If cs is non-zero it coalesces the statistics with the
+// previous diffStats.
+func (s *textList) AppendEllipsis(ds diffStats) {
+	hasStats := !ds.IsZero()
+	if len(*s) == 0 || !(*s)[len(*s)-1].Value.Equal(textEllipsis) {
+		if hasStats {
+			*s = append(*s, textRecord{Value: textEllipsis, ElideComma: true, Comment: ds})
+		} else {
+			*s = append(*s, textRecord{Value: textEllipsis, ElideComma: true})
+		}
+		return
+	}
+	if hasStats {
+		(*s)[len(*s)-1].Comment = (*s)[len(*s)-1].Comment.(diffStats).Append(ds)
+	}
+}
+
+func (s textList) Len() (n int) {
+	for i, r := range s {
+		n += len(r.Key)
+		if r.Key != "" {
+			n += len(": ")
+		}
+		n += r.Value.Len()
+		if i < len(s)-1 {
+			n += len(", ")
+		}
+	}
+	return n
+}
+
+func (s1 textList) Equal(s2 textNode) bool {
+	if s2, ok := s2.(textList); ok {
+		if len(s1) != len(s2) {
+			return false
+		}
+		for i := range s1 {
+			r1, r2 := s1[i], s2[i]
+			if !(r1.Diff == r2.Diff && r1.Key == r2.Key && r1.Value.Equal(r2.Value) && r1.Comment == r2.Comment) {
+				return false
+			}
+		}
+		return true
+	}
+	return false
+}
+
+func (s textList) String() string {
+	return (&textWrap{Prefix: "{", Value: s, Suffix: "}"}).String()
+}
+
+func (s textList) formatCompactTo(b []byte, d diffMode) ([]byte, textNode) {
+	s = append(textList(nil), s...) // Avoid mutating original
+
+	// Determine whether we can collapse this list as a single line.
+	n0 := len(b) // Original buffer length
+	var multiLine bool
+	for i, r := range s {
+		if r.Diff == diffInserted || r.Diff == diffRemoved {
+			multiLine = true
+		}
+		b = append(b, r.Key...)
+		if r.Key != "" {
+			b = append(b, ": "...)
+		}
+		b, s[i].Value = r.Value.formatCompactTo(b, d|r.Diff)
+		if _, ok := s[i].Value.(textLine); !ok {
+			multiLine = true
+		}
+		if r.Comment != nil {
+			multiLine = true
+		}
+		if i < len(s)-1 {
+			b = append(b, ", "...)
+		}
+	}
+	// Force multi-lined output when printing a removed/inserted node that
+	// is sufficiently long.
+	if (d == diffInserted || d == diffRemoved) && len(b[n0:]) > maxColumnLength {
+		multiLine = true
+	}
+	if !multiLine {
+		return b, textLine(b[n0:])
+	}
+	return b, s
+}
+
+func (s textList) formatExpandedTo(b []byte, d diffMode, n indentMode) []byte {
+	alignKeyLens := s.alignLens(
+		func(r textRecord) bool {
+			_, isLine := r.Value.(textLine)
+			return r.Key == "" || !isLine
+		},
+		func(r textRecord) int { return utf8.RuneCountInString(r.Key) },
+	)
+	alignValueLens := s.alignLens(
+		func(r textRecord) bool {
+			_, isLine := r.Value.(textLine)
+			return !isLine || r.Value.Equal(textEllipsis) || r.Comment == nil
+		},
+		func(r textRecord) int { return utf8.RuneCount(r.Value.(textLine)) },
+	)
+
+	// Format lists of simple lists in a batched form.
+	// If the list is sequence of only textLine values,
+	// then batch multiple values on a single line.
+	var isSimple bool
+	for _, r := range s {
+		_, isLine := r.Value.(textLine)
+		isSimple = r.Diff == 0 && r.Key == "" && isLine && r.Comment == nil
+		if !isSimple {
+			break
+		}
+	}
+	if isSimple {
+		n++
+		var batch []byte
+		emitBatch := func() {
+			if len(batch) > 0 {
+				b = n.appendIndent(append(b, '\n'), d)
+				b = append(b, bytes.TrimRight(batch, " ")...)
+				batch = batch[:0]
+			}
+		}
+		for _, r := range s {
+			line := r.Value.(textLine)
+			if len(batch)+len(line)+len(", ") > maxColumnLength {
+				emitBatch()
+			}
+			batch = append(batch, line...)
+			batch = append(batch, ", "...)
+		}
+		emitBatch()
+		n--
+		return n.appendIndent(append(b, '\n'), d)
+	}
+
+	// Format the list as a multi-lined output.
+	n++
+	for i, r := range s {
+		b = n.appendIndent(append(b, '\n'), d|r.Diff)
+		if r.Key != "" {
+			b = append(b, r.Key+": "...)
+		}
+		b = alignKeyLens[i].appendChar(b, ' ')
+
+		b = r.Value.formatExpandedTo(b, d|r.Diff, n)
+		if !r.ElideComma {
+			b = append(b, ',')
+		}
+		b = alignValueLens[i].appendChar(b, ' ')
+
+		if r.Comment != nil {
+			b = append(b, " // "+r.Comment.String()...)
+		}
+	}
+	n--
+
+	return n.appendIndent(append(b, '\n'), d)
+}
+
+func (s textList) alignLens(
+	skipFunc func(textRecord) bool,
+	lenFunc func(textRecord) int,
+) []repeatCount {
+	var startIdx, endIdx, maxLen int
+	lens := make([]repeatCount, len(s))
+	for i, r := range s {
+		if skipFunc(r) {
+			for j := startIdx; j < endIdx && j < len(s); j++ {
+				lens[j] = repeatCount(maxLen - lenFunc(s[j]))
+			}
+			startIdx, endIdx, maxLen = i+1, i+1, 0
+		} else {
+			if maxLen < lenFunc(r) {
+				maxLen = lenFunc(r)
+			}
+			endIdx = i + 1
+		}
+	}
+	for j := startIdx; j < endIdx && j < len(s); j++ {
+		lens[j] = repeatCount(maxLen - lenFunc(s[j]))
+	}
+	return lens
+}
+
+// textLine is a single-line segment of text and is always a leaf node
+// in the textNode tree.
+type textLine []byte
+
+var (
+	textNil      = textLine("nil")
+	textEllipsis = textLine("...")
+)
+
+func (s textLine) Len() int {
+	return len(s)
+}
+func (s1 textLine) Equal(s2 textNode) bool {
+	if s2, ok := s2.(textLine); ok {
+		return bytes.Equal([]byte(s1), []byte(s2))
+	}
+	return false
+}
+func (s textLine) String() string {
+	return string(s)
+}
+func (s textLine) formatCompactTo(b []byte, d diffMode) ([]byte, textNode) {
+	return append(b, s...), s
+}
+func (s textLine) formatExpandedTo(b []byte, _ diffMode, _ indentMode) []byte {
+	return append(b, s...)
+}
+
+type diffStats struct {
+	Name         string
+	NumIgnored   int
+	NumIdentical int
+	NumRemoved   int
+	NumInserted  int
+	NumModified  int
+}
+
+func (s diffStats) IsZero() bool {
+	s.Name = ""
+	return s == diffStats{}
+}
+
+func (s diffStats) NumDiff() int {
+	return s.NumRemoved + s.NumInserted + s.NumModified
+}
+
+func (s diffStats) Append(ds diffStats) diffStats {
+	assert(s.Name == ds.Name)
+	s.NumIgnored += ds.NumIgnored
+	s.NumIdentical += ds.NumIdentical
+	s.NumRemoved += ds.NumRemoved
+	s.NumInserted += ds.NumInserted
+	s.NumModified += ds.NumModified
+	return s
+}
+
+// String prints a humanly-readable summary of coalesced records.
+//
+// Example:
+//	diffStats{Name: "Field", NumIgnored: 5}.String() => "5 ignored fields"
+func (s diffStats) String() string {
+	var ss []string
+	var sum int
+	labels := [...]string{"ignored", "identical", "removed", "inserted", "modified"}
+	counts := [...]int{s.NumIgnored, s.NumIdentical, s.NumRemoved, s.NumInserted, s.NumModified}
+	for i, n := range counts {
+		if n > 0 {
+			ss = append(ss, fmt.Sprintf("%d %v", n, labels[i]))
+		}
+		sum += n
+	}
+
+	// Pluralize the name (adjusting for some obscure English grammar rules).
+	name := s.Name
+	if sum > 1 {
+		name += "s"
+		if strings.HasSuffix(name, "ys") {
+			name = name[:len(name)-2] + "ies" // e.g., "entrys" => "entries"
+		}
+	}
+
+	// Format the list according to English grammar (with Oxford comma).
+	switch n := len(ss); n {
+	case 0:
+		return ""
+	case 1, 2:
+		return strings.Join(ss, " and ") + " " + name
+	default:
+		return strings.Join(ss[:n-1], ", ") + ", and " + ss[n-1] + " " + name
+	}
+}
+
+type commentString string
+
+func (s commentString) String() string { return string(s) }
@@ -0,0 +1,121 @@
+// Copyright 2019, The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package cmp
+
+import "reflect"
+
+// valueNode represents a single node within a report, which is a
+// structured representation of the value tree, containing information
+// regarding which nodes are equal or not.
+type valueNode struct {
+	parent *valueNode
+
+	Type   reflect.Type
+	ValueX reflect.Value
+	ValueY reflect.Value
+
+	// NumSame is the number of leaf nodes that are equal.
+	// All descendants are equal only if NumDiff is 0.
+	NumSame int
+	// NumDiff is the number of leaf nodes that are not equal.
+	NumDiff int
+	// NumIgnored is the number of leaf nodes that are ignored.
+	NumIgnored int
+	// NumCompared is the number of leaf nodes that were compared
+	// using an Equal method or Comparer function.
+	NumCompared int
+	// NumTransformed is the number of non-leaf nodes that were transformed.
+	NumTransformed int
+	// NumChildren is the number of transitive descendants of this node.
+	// This counts from zero; thus, leaf nodes have no descendants.
+	NumChildren int
+	// MaxDepth is the maximum depth of the tree. This counts from zero;
+	// thus, leaf nodes have a depth of zero.
+	MaxDepth int
+
+	// Records is a list of struct fields, slice elements, or map entries.
+	Records []reportRecord // If populated, implies Value is not populated
+
+	// Value is the result of a transformation, pointer indirect, of
+	// type assertion.
+	Value *valueNode // If populated, implies Records is not populated
+
+	// TransformerName is the name of the transformer.
+	TransformerName string // If non-empty, implies Value is populated
+}
+type reportRecord struct {
+	Key   reflect.Value // Invalid for slice element
+	Value *valueNode
+}
+
+func (parent *valueNode) PushStep(ps PathStep) (child *valueNode) {
+	vx, vy := ps.Values()
+	child = &valueNode{parent: parent, Type: ps.Type(), ValueX: vx, ValueY: vy}
+	switch s := ps.(type) {
+	case StructField:
+		assert(parent.Value == nil)
+		parent.Records = append(parent.Records, reportRecord{Key: reflect.ValueOf(s.Name()), Value: child})
+	case SliceIndex:
+		assert(parent.Value == nil)
+		parent.Records = append(parent.Records, reportRecord{Value: child})
+	case MapIndex:
+		assert(parent.Value == nil)
+		parent.Records = append(parent.Records, reportRecord{Key: s.Key(), Value: child})
+	case Indirect:
+		assert(parent.Value == nil && parent.Records == nil)
+		parent.Value = child
+	case TypeAssertion:
+		assert(parent.Value == nil && parent.Records == nil)
+		parent.Value = child
+	case Transform:
+		assert(parent.Value == nil && parent.Records == nil)
+		parent.Value = child
+		parent.TransformerName = s.Name()
+		parent.NumTransformed++
+	default:
+		assert(parent == nil) // Must be the root step
+	}
+	return child
+}
+
+func (r *valueNode) Report(rs Result) {
+	assert(r.MaxDepth == 0) // May only be called on leaf nodes
+
+	if rs.ByIgnore() {
+		r.NumIgnored++
+	} else {
+		if rs.Equal() {
+			r.NumSame++
+		} else {
+			r.NumDiff++
+		}
+	}
+	assert(r.NumSame+r.NumDiff+r.NumIgnored == 1)
+
+	if rs.ByMethod() {
+		r.NumCompared++
+	}
+	if rs.ByFunc() {
+		r.NumCompared++
+	}
+	assert(r.NumCompared <= 1)
+}
+
+func (child *valueNode) PopStep() (parent *valueNode) {
+	if child.parent == nil {
+		return nil
+	}
+	parent = child.parent
+	parent.NumSame += child.NumSame
+	parent.NumDiff += child.NumDiff
+	parent.NumIgnored += child.NumIgnored
+	parent.NumCompared += child.NumCompared
+	parent.NumTransformed += child.NumTransformed
+	parent.NumChildren += child.NumChildren + 1
+	if parent.MaxDepth < child.MaxDepth+1 {
+		parent.MaxDepth = child.MaxDepth + 1
+	}
+	return parent
+}
@@ -0,0 +1,3 @@
+.idea/
+.test/
+examples/_*
@@ -0,0 +1,56 @@
+# Version v1.4.7-v1.4.8
+* Documentation updates.
+* Small linter cleanups.
+* Added example in test. 
+
+# Version v1.4.6
+
+* Document the usage of Cleanup when re-reading a file (thanks to @lesovsky) for issue #18.
+* Add example directories with example and tests for issues.
+
+# Version v1.4.4-v1.4.5
+
+* Fix of checksum problem because of forced tag. No changes to the code.
+
+# Version v1.4.1
+
+* Incorporated PR 162 by by Mohammed902: "Simplify non-Windows build tag".
+
+# Version v1.4.0
+
+* Incorporated PR 9 by mschneider82: "Added seekinfo to Tail".
+
+# Version v1.3.1
+
+* Incorporated PR 7: "Fix deadlock when stopping on non-empty file/buffer",
+fixes upstream issue 93.
+
+
+# Version v1.3.0
+
+* Incorporated changes of unmerged upstream PR 149 by mezzi: "added line num
+to Line struct".
+
+# Version v1.2.1
+
+* Incorporated changes of unmerged upstream PR 128 by jadekler: "Compile-able
+code in readme".
+* Incorporated changes of unmerged upstream PR 130 by fgeller: "small change
+to comment wording".
+* Incorporated changes of unmerged upstream PR 133 by sm3142: "removed
+spurious newlines from log messages".
+
+# Version v1.2.0
+
+* Incorporated changes of unmerged upstream PR 126 by Code-Hex: "Solved the
+ problem for never return the last line if it's not followed by a newline".
+* Incorporated changes of unmerged upstream PR 131 by StoicPerlman: "Remove
+deprecated os.SEEK consts". The changes bumped the minimal supported Go
+release to 1.9.
+
+# Version v1.1.0
+
+* migration to go modules.
+* release of master branch of the dormant upstream, because it contains
+fixes and improvement no present in the tagged release.
+
@@ -0,0 +1,19 @@
+FROM golang
+
+RUN mkdir -p $GOPATH/src/github.com/nxadm/tail/
+ADD . $GOPATH/src/github.com/nxadm/tail/
+
+# expecting to fetch dependencies successfully.
+RUN go get -v github.com/nxadm/tail
+
+# expecting to run the test successfully.
+RUN go test -v github.com/nxadm/tail
+
+# expecting to install successfully
+RUN go install -v github.com/nxadm/tail
+RUN go install -v github.com/nxadm/tail/cmd/gotail
+
+RUN $GOPATH/bin/gotail -h || true
+
+ENV PATH $GOPATH/bin:$PATH
+CMD ["gotail"]
@@ -0,0 +1,21 @@
+# The MIT License (MIT)
+
+# © Copyright 2015 Hewlett Packard Enterprise Development LP
+Copyright (c) 2014 ActiveState
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
@@ -0,0 +1,44 @@
+![ci](https://github.com/nxadm/tail/workflows/ci/badge.svg)[![Go Reference](https://pkg.go.dev/badge/github.com/nxadm/tail.svg)](https://pkg.go.dev/github.com/nxadm/tail)
+
+# tail functionality in Go
+
+nxadm/tail provides a Go library that emulates the features of the BSD `tail`
+program. The library comes with full support for truncation/move detection as
+it is designed to work with log rotation tools. The library works on all
+operating systems supported by Go, including POSIX systems like Linux and
+*BSD, and MS Windows. Go 1.9 is the oldest compiler release supported.
+
+A simple example:
+
+```Go
+// Create a tail
+t, err := tail.TailFile(
+	"/var/log/nginx.log", tail.Config{Follow: true, ReOpen: true})
+if err != nil {
+    panic(err)
+}
+
+// Print the text of each received line
+for line := range t.Lines {
+    fmt.Println(line.Text)
+}
+```
+
+See [API documentation](https://pkg.go.dev/github.com/nxadm/tail).
+
+## Installing
+
+    go get github.com/nxadm/tail/...
+
+## History
+
+This project is an active, drop-in replacement for the
+[abandoned](https://en.wikipedia.org/wiki/HPE_Helion) Go tail library at
+[hpcloud](https://github.com/hpcloud/tail). Next to
+[addressing open issues/PRs of the original project](https://github.com/nxadm/tail/issues/6),
+nxadm/tail continues the development by keeping up to date with the Go toolchain
+(e.g. go modules) and dependencies, completing the documentation, adding features
+and fixing bugs.
+
+## Examples
+Examples, e.g. used to debug an issue, are kept in the [examples directory](/examples).
@@ -0,0 +1,7 @@
+Copyright (C) 2013 99designs
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
@@ -0,0 +1,97 @@
+// Package ratelimiter implements the Leaky Bucket ratelimiting algorithm with memcached and in-memory backends.
+package ratelimiter
+
+import (
+	"time"
+)
+
+type LeakyBucket struct {
+	Size         uint16
+	Fill         float64
+	LeakInterval time.Duration // time.Duration for 1 unit of size to leak
+	Lastupdate   time.Time
+	Now          func() time.Time
+}
+
+func NewLeakyBucket(size uint16, leakInterval time.Duration) *LeakyBucket {
+	bucket := LeakyBucket{
+		Size:         size,
+		Fill:         0,
+		LeakInterval: leakInterval,
+		Now:          time.Now,
+		Lastupdate:   time.Now(),
+	}
+
+	return &bucket
+}
+
+func (b *LeakyBucket) updateFill() {
+	now := b.Now()
+	if b.Fill > 0 {
+		elapsed := now.Sub(b.Lastupdate)
+
+		b.Fill -= float64(elapsed) / float64(b.LeakInterval)
+		if b.Fill < 0 {
+			b.Fill = 0
+		}
+	}
+	b.Lastupdate = now
+}
+
+func (b *LeakyBucket) Pour(amount uint16) bool {
+	b.updateFill()
+
+	var newfill float64 = b.Fill + float64(amount)
+
+	if newfill > float64(b.Size) {
+		return false
+	}
+
+	b.Fill = newfill
+
+	return true
+}
+
+// The time at which this bucket will be completely drained
+func (b *LeakyBucket) DrainedAt() time.Time {
+	return b.Lastupdate.Add(time.Duration(b.Fill * float64(b.LeakInterval)))
+}
+
+// The duration until this bucket is completely drained
+func (b *LeakyBucket) TimeToDrain() time.Duration {
+	return b.DrainedAt().Sub(b.Now())
+}
+
+func (b *LeakyBucket) TimeSinceLastUpdate() time.Duration {
+	return b.Now().Sub(b.Lastupdate)
+}
+
+type LeakyBucketSer struct {
+	Size         uint16
+	Fill         float64
+	LeakInterval time.Duration // time.Duration for 1 unit of size to leak
+	Lastupdate   time.Time
+}
+
+func (b *LeakyBucket) Serialise() *LeakyBucketSer {
+	bucket := LeakyBucketSer{
+		Size:         b.Size,
+		Fill:         b.Fill,
+		LeakInterval: b.LeakInterval,
+		Lastupdate:   b.Lastupdate,
+	}
+
+	return &bucket
+}
+
+func (b *LeakyBucketSer) DeSerialise() *LeakyBucket {
+	bucket := LeakyBucket{
+		Size:         b.Size,
+		Fill:         b.Fill,
+		LeakInterval: b.LeakInterval,
+		Lastupdate:   b.Lastupdate,
+		Now:          time.Now,
+	}
+
+	return &bucket
+}
@@ -0,0 +1,60 @@
+package ratelimiter
+
+import (
+	"errors"
+	"time"
+)
+
+const (
+	GC_SIZE   int           = 100
+	GC_PERIOD time.Duration = 60 * time.Second
+)
+
+type Memory struct {
+	store           map[string]LeakyBucket
+	lastGCCollected time.Time
+}
+
+func NewMemory() *Memory {
+	m := new(Memory)
+	m.store = make(map[string]LeakyBucket)
+	m.lastGCCollected = time.Now()
+	return m
+}
+
+func (m *Memory) GetBucketFor(key string) (*LeakyBucket, error) {
+
+	bucket, ok := m.store[key]
+	if !ok {
+		return nil, errors.New("miss")
+	}
+
+	return &bucket, nil
+}
+
+func (m *Memory) SetBucketFor(key string, bucket LeakyBucket) error {
+
+	if len(m.store) > GC_SIZE {
+		m.GarbageCollect()
+	}
+
+	m.store[key] = bucket
+
+	return nil
+}
+
+func (m *Memory) GarbageCollect() {
+	now := time.Now()
+
+	// rate limit GC to once per minute
+	if now.Unix() >= m.lastGCCollected.Add(GC_PERIOD).Unix() {
+		for key, bucket := range m.store {
+			// if the bucket is drained, then GC
+			if bucket.DrainedAt().Unix() < now.Unix() {
+				delete(m.store, key)
+			}
+		}
+
+		m.lastGCCollected = now
+	}
+}
@@ -0,0 +1,6 @@
+package ratelimiter
+
+type Storage interface {
+	GetBucketFor(string) (*LeakyBucket, error)
+	SetBucketFor(string, LeakyBucket) error
+}
@@ -0,0 +1,455 @@
+// Copyright (c) 2019 FOSS contributors of https://github.com/nxadm/tail
+// Copyright (c) 2015 HPE Software Inc. All rights reserved.
+// Copyright (c) 2013 ActiveState Software Inc. All rights reserved.
+
+//nxadm/tail provides a Go library that emulates the features of the BSD `tail`
+//program. The library comes with full support for truncation/move detection as
+//it is designed to work with log rotation tools. The library works on all
+//operating systems supported by Go, including POSIX systems like Linux and
+//*BSD, and MS Windows. Go 1.9 is the oldest compiler release supported.
+package tail
+
+import (
+	"bufio"
+	"errors"
+	"fmt"
+	"io"
+	"io/ioutil"
+	"log"
+	"os"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/nxadm/tail/ratelimiter"
+	"github.com/nxadm/tail/util"
+	"github.com/nxadm/tail/watch"
+	"gopkg.in/tomb.v1"
+)
+
+var (
+	// ErrStop is returned when the tail of a file has been marked to be stopped.
+	ErrStop = errors.New("tail should now stop")
+)
+
+type Line struct {
+	Text     string    // The contents of the file
+	Num      int       // The line number
+	SeekInfo SeekInfo  // SeekInfo
+	Time     time.Time // Present time
+	Err      error     // Error from tail
+}
+
+// Deprecated: this function is no longer used internally and it has little of no
+// use in the API. As such, it will be removed from the API in a future major
+// release.
+//
+// NewLine returns a * pointer to a Line struct.
+func NewLine(text string, lineNum int) *Line {
+	return &Line{text, lineNum, SeekInfo{}, time.Now(), nil}
+}
+
+// SeekInfo represents arguments to io.Seek. See: https://golang.org/pkg/io/#SectionReader.Seek
+type SeekInfo struct {
+	Offset int64
+	Whence int
+}
+
+type logger interface {
+	Fatal(v ...interface{})
+	Fatalf(format string, v ...interface{})
+	Fatalln(v ...interface{})
+	Panic(v ...interface{})
+	Panicf(format string, v ...interface{})
+	Panicln(v ...interface{})
+	Print(v ...interface{})
+	Printf(format string, v ...interface{})
+	Println(v ...interface{})
+}
+
+// Config is used to specify how a file must be tailed.
+type Config struct {
+	// File-specifc
+	Location  *SeekInfo // Tail from this location. If nil, start at the beginning of the file
+	ReOpen    bool      // Reopen recreated files (tail -F)
+	MustExist bool      // Fail early if the file does not exist
+	Poll      bool      // Poll for file changes instead of using the default inotify
+	Pipe      bool      // The file is a named pipe (mkfifo)
+
+	// Generic IO
+	Follow      bool // Continue looking for new lines (tail -f)
+	MaxLineSize int  // If non-zero, split longer lines into multiple lines
+
+	// Optionally, use a ratelimiter (e.g. created by the ratelimiter/NewLeakyBucket function)
+	RateLimiter *ratelimiter.LeakyBucket
+
+	// Optionally use a Logger. When nil, the Logger is set to tail.DefaultLogger.
+	// To disable logging, set it to tail.DiscardingLogger
+	Logger logger
+}
+
+type Tail struct {
+	Filename string     // The filename
+	Lines    chan *Line // A consumable channel of *Line
+	Config              // Tail.Configuration
+
+	file    *os.File
+	reader  *bufio.Reader
+	lineNum int
+
+	watcher watch.FileWatcher
+	changes *watch.FileChanges
+
+	tomb.Tomb // provides: Done, Kill, Dying
+
+	lk sync.Mutex
+}
+
+var (
+	// DefaultLogger logs to os.Stderr and it is used when Config.Logger == nil
+	DefaultLogger = log.New(os.Stderr, "", log.LstdFlags)
+	// DiscardingLogger can be used to disable logging output
+	DiscardingLogger = log.New(ioutil.Discard, "", 0)
+)
+
+// TailFile begins tailing the file. And returns a pointer to a Tail struct
+// and an error. An output stream is made available via the Tail.Lines
+// channel (e.g. to be looped and printed). To handle errors during tailing,
+// after finishing reading from the Lines channel, invoke the `Wait` or `Err`
+// method on the returned *Tail.
+func TailFile(filename string, config Config) (*Tail, error) {
+	if config.ReOpen && !config.Follow {
+		util.Fatal("cannot set ReOpen without Follow.")
+	}
+
+	t := &Tail{
+		Filename: filename,
+		Lines:    make(chan *Line),
+		Config:   config,
+	}
+
+	// when Logger was not specified in config, use default logger
+	if t.Logger == nil {
+		t.Logger = DefaultLogger
+	}
+
+	if t.Poll {
+		t.watcher = watch.NewPollingFileWatcher(filename)
+	} else {
+		t.watcher = watch.NewInotifyFileWatcher(filename)
+	}
+
+	if t.MustExist {
+		var err error
+		t.file, err = OpenFile(t.Filename)
+		if err != nil {
+			return nil, err
+		}
+	}
+
+	go t.tailFileSync()
+
+	return t, nil
+}
+
+// Tell returns the file's current position, like stdio's ftell() and an error.
+// Beware that this value may not be completely accurate because one line from
+// the chan(tail.Lines) may have been read already.
+func (tail *Tail) Tell() (offset int64, err error) {
+	if tail.file == nil {
+		return
+	}
+	offset, err = tail.file.Seek(0, io.SeekCurrent)
+	if err != nil {
+		return
+	}
+
+	tail.lk.Lock()
+	defer tail.lk.Unlock()
+	if tail.reader == nil {
+		return
+	}
+
+	offset -= int64(tail.reader.Buffered())
+	return
+}
+
+// Stop stops the tailing activity.
+func (tail *Tail) Stop() error {
+	tail.Kill(nil)
+	return tail.Wait()
+}
+
+// StopAtEOF stops tailing as soon as the end of the file is reached. The function
+// returns an error,
+func (tail *Tail) StopAtEOF() error {
+	tail.Kill(errStopAtEOF)
+	return tail.Wait()
+}
+
+var errStopAtEOF = errors.New("tail: stop at eof")
+
+func (tail *Tail) close() {
+	close(tail.Lines)
+	tail.closeFile()
+}
+
+func (tail *Tail) closeFile() {
+	if tail.file != nil {
+		tail.file.Close()
+		tail.file = nil
+	}
+}
+
+func (tail *Tail) reopen() error {
+	tail.closeFile()
+	tail.lineNum = 0
+	for {
+		var err error
+		tail.file, err = OpenFile(tail.Filename)
+		if err != nil {
+			if os.IsNotExist(err) {
+				tail.Logger.Printf("Waiting for %s to appear...", tail.Filename)
+				if err := tail.watcher.BlockUntilExists(&tail.Tomb); err != nil {
+					if err == tomb.ErrDying {
+						return err
+					}
+					return fmt.Errorf("Failed to detect creation of %s: %s", tail.Filename, err)
+				}
+				continue
+			}
+			return fmt.Errorf("Unable to open file %s: %s", tail.Filename, err)
+		}
+		break
+	}
+	return nil
+}
+
+func (tail *Tail) readLine() (string, error) {
+	tail.lk.Lock()
+	line, err := tail.reader.ReadString('\n')
+	tail.lk.Unlock()
+	if err != nil {
+		// Note ReadString "returns the data read before the error" in
+		// case of an error, including EOF, so we return it as is. The
+		// caller is expected to process it if err is EOF.
+		return line, err
+	}
+
+	line = strings.TrimRight(line, "\n")
+
+	return line, err
+}
+
+func (tail *Tail) tailFileSync() {
+	defer tail.Done()
+	defer tail.close()
+
+	if !tail.MustExist {
+		// deferred first open.
+		err := tail.reopen()
+		if err != nil {
+			if err != tomb.ErrDying {
+				tail.Kill(err)
+			}
+			return
+		}
+	}
+
+	// Seek to requested location on first open of the file.
+	if tail.Location != nil {
+		_, err := tail.file.Seek(tail.Location.Offset, tail.Location.Whence)
+		if err != nil {
+			tail.Killf("Seek error on %s: %s", tail.Filename, err)
+			return
+		}
+	}
+
+	tail.openReader()
+
+	// Read line by line.
+	for {
+		// do not seek in named pipes
+		if !tail.Pipe {
+			// grab the position in case we need to back up in the event of a half-line
+			if _, err := tail.Tell(); err != nil {
+				tail.Kill(err)
+				return
+			}
+		}
+
+		line, err := tail.readLine()
+
+		// Process `line` even if err is EOF.
+		if err == nil {
+			cooloff := !tail.sendLine(line)
+			if cooloff {
+				// Wait a second before seeking till the end of
+				// file when rate limit is reached.
+				msg := ("Too much log activity; waiting a second before resuming tailing")
+				offset, _ := tail.Tell()
+				tail.Lines <- &Line{msg, tail.lineNum, SeekInfo{Offset: offset}, time.Now(), errors.New(msg)}
+				select {
+				case <-time.After(time.Second):
+				case <-tail.Dying():
+					return
+				}
+				if err := tail.seekEnd(); err != nil {
+					tail.Kill(err)
+					return
+				}
+			}
+		} else if err == io.EOF {
+			if !tail.Follow {
+				if line != "" {
+					tail.sendLine(line)
+				}
+				return
+			}
+
+			if tail.Follow && line != "" {
+				tail.sendLine(line)
+				if err := tail.seekEnd(); err != nil {
+					tail.Kill(err)
+					return
+				}
+			}
+
+			// When EOF is reached, wait for more data to become
+			// available. Wait strategy is based on the `tail.watcher`
+			// implementation (inotify or polling).
+			err := tail.waitForChanges()
+			if err != nil {
+				if err != ErrStop {
+					tail.Kill(err)
+				}
+				return
+			}
+		} else {
+			// non-EOF error
+			tail.Killf("Error reading %s: %s", tail.Filename, err)
+			return
+		}
+
+		select {
+		case <-tail.Dying():
+			if tail.Err() == errStopAtEOF {
+				continue
+			}
+			return
+		default:
+		}
+	}
+}
+
+// waitForChanges waits until the file has been appended, deleted,
+// moved or truncated. When moved or deleted - the file will be
+// reopened if ReOpen is true. Truncated files are always reopened.
+func (tail *Tail) waitForChanges() error {
+	if tail.changes == nil {
+		pos, err := tail.file.Seek(0, io.SeekCurrent)
+		if err != nil {
+			return err
+		}
+		tail.changes, err = tail.watcher.ChangeEvents(&tail.Tomb, pos)
+		if err != nil {
+			return err
+		}
+	}
+
+	select {
+	case <-tail.changes.Modified:
+		return nil
+	case <-tail.changes.Deleted:
+		tail.changes = nil
+		if tail.ReOpen {
+			// XXX: we must not log from a library.
+			tail.Logger.Printf("Re-opening moved/deleted file %s ...", tail.Filename)
+			if err := tail.reopen(); err != nil {
+				return err
+			}
+			tail.Logger.Printf("Successfully reopened %s", tail.Filename)
+			tail.openReader()
+			return nil
+		}
+		tail.Logger.Printf("Stopping tail as file no longer exists: %s", tail.Filename)
+		return ErrStop
+	case <-tail.changes.Truncated:
+		// Always reopen truncated files (Follow is true)
+		tail.Logger.Printf("Re-opening truncated file %s ...", tail.Filename)
+		if err := tail.reopen(); err != nil {
+			return err
+		}
+		tail.Logger.Printf("Successfully reopened truncated %s", tail.Filename)
+		tail.openReader()
+		return nil
+	case <-tail.Dying():
+		return ErrStop
+	}
+}
+
+func (tail *Tail) openReader() {
+	tail.lk.Lock()
+	if tail.MaxLineSize > 0 {
+		// add 2 to account for newline characters
+		tail.reader = bufio.NewReaderSize(tail.file, tail.MaxLineSize+2)
+	} else {
+		tail.reader = bufio.NewReader(tail.file)
+	}
+	tail.lk.Unlock()
+}
+
+func (tail *Tail) seekEnd() error {
+	return tail.seekTo(SeekInfo{Offset: 0, Whence: io.SeekEnd})
+}
+
+func (tail *Tail) seekTo(pos SeekInfo) error {
+	_, err := tail.file.Seek(pos.Offset, pos.Whence)
+	if err != nil {
+		return fmt.Errorf("Seek error on %s: %s", tail.Filename, err)
+	}
+	// Reset the read buffer whenever the file is re-seek'ed
+	tail.reader.Reset(tail.file)
+	return nil
+}
+
+// sendLine sends the line(s) to Lines channel, splitting longer lines
+// if necessary. Return false if rate limit is reached.
+func (tail *Tail) sendLine(line string) bool {
+	now := time.Now()
+	lines := []string{line}
+
+	// Split longer lines
+	if tail.MaxLineSize > 0 && len(line) > tail.MaxLineSize {
+		lines = util.PartitionString(line, tail.MaxLineSize)
+	}
+
+	for _, line := range lines {
+		tail.lineNum++
+		offset, _ := tail.Tell()
+		select {
+		case tail.Lines <- &Line{line, tail.lineNum, SeekInfo{Offset: offset}, now, nil}:
+		case <-tail.Dying():
+			return true
+		}
+	}
+
+	if tail.Config.RateLimiter != nil {
+		ok := tail.Config.RateLimiter.Pour(uint16(len(lines)))
+		if !ok {
+			tail.Logger.Printf("Leaky bucket full (%v); entering 1s cooloff period.",
+				tail.Filename)
+			return false
+		}
+	}
+
+	return true
+}
+
+// Cleanup removes inotify watches added by the tail package. This function is
+// meant to be invoked from a process's exit handler. Linux kernel may not
+// automatically remove inotify watches after the process exits.
+// If you plan to re-read a file, don't call Cleanup in between.
+func (tail *Tail) Cleanup() {
+	watch.Cleanup(tail.Filename)
+}
@@ -0,0 +1,17 @@
+// Copyright (c) 2019 FOSS contributors of https://github.com/nxadm/tail
+// +build !windows
+
+package tail
+
+import (
+	"os"
+)
+
+// Deprecated: this function is only useful internally and, as such,
+// it will be removed from the API in a future major release.
+//
+// OpenFile proxies a os.Open call for a file so it can be correctly tailed
+// on POSIX and non-POSIX OSes like MS Windows.
+func OpenFile(name string) (file *os.File, err error) {
+	return os.Open(name)
+}
@@ -0,0 +1,19 @@
+// Copyright (c) 2019 FOSS contributors of https://github.com/nxadm/tail
+// +build windows
+
+package tail
+
+import (
+	"os"
+
+	"github.com/nxadm/tail/winfile"
+)
+
+// Deprecated: this function is only useful internally and, as such,
+// it will be removed from the API in a future major release.
+//
+// OpenFile proxies a os.Open call for a file so it can be correctly tailed
+// on POSIX and non-POSIX OSes like MS Windows.
+func OpenFile(name string) (file *os.File, err error) {
+	return winfile.OpenFile(name, os.O_RDONLY, 0)
+}
@@ -0,0 +1,49 @@
+// Copyright (c) 2019 FOSS contributors of https://github.com/nxadm/tail
+// Copyright (c) 2015 HPE Software Inc. All rights reserved.
+// Copyright (c) 2013 ActiveState Software Inc. All rights reserved.
+
+package util
+
+import (
+	"fmt"
+	"log"
+	"os"
+	"runtime/debug"
+)
+
+type Logger struct {
+	*log.Logger
+}
+
+var LOGGER = &Logger{log.New(os.Stderr, "", log.LstdFlags)}
+
+// fatal is like panic except it displays only the current goroutine's stack.
+func Fatal(format string, v ...interface{}) {
+	// https://github.com/nxadm/log/blob/master/log.go#L45
+	LOGGER.Output(2, fmt.Sprintf("FATAL -- "+format, v...)+"\n"+string(debug.Stack()))
+	os.Exit(1)
+}
+
+// partitionString partitions the string into chunks of given size,
+// with the last chunk of variable size.
+func PartitionString(s string, chunkSize int) []string {
+	if chunkSize <= 0 {
+		panic("invalid chunkSize")
+	}
+	length := len(s)
+	chunks := 1 + length/chunkSize
+	start := 0
+	end := chunkSize
+	parts := make([]string, 0, chunks)
+	for {
+		if end > length {
+			end = length
+		}
+		parts = append(parts, s[start:end])
+		if end == length {
+			break
+		}
+		start, end = end, end+chunkSize
+	}
+	return parts
+}
@@ -0,0 +1,37 @@
+// Copyright (c) 2019 FOSS contributors of https://github.com/nxadm/tail
+package watch
+
+type FileChanges struct {
+	Modified  chan bool // Channel to get notified of modifications
+	Truncated chan bool // Channel to get notified of truncations
+	Deleted   chan bool // Channel to get notified of deletions/renames
+}
+
+func NewFileChanges() *FileChanges {
+	return &FileChanges{
+		make(chan bool, 1), make(chan bool, 1), make(chan bool, 1)}
+}
+
+func (fc *FileChanges) NotifyModified() {
+	sendOnlyIfEmpty(fc.Modified)
+}
+
+func (fc *FileChanges) NotifyTruncated() {
+	sendOnlyIfEmpty(fc.Truncated)
+}
+
+func (fc *FileChanges) NotifyDeleted() {
+	sendOnlyIfEmpty(fc.Deleted)
+}
+
+// sendOnlyIfEmpty sends on a bool channel only if the channel has no
+// backlog to be read by other goroutines. This concurrency pattern
+// can be used to notify other goroutines if and only if they are
+// looking for it (i.e., subsequent notifications can be compressed
+// into one).
+func sendOnlyIfEmpty(ch chan bool) {
+	select {
+	case ch <- true:
+	default:
+	}
+}
@@ -0,0 +1,136 @@
+// Copyright (c) 2019 FOSS contributors of https://github.com/nxadm/tail
+// Copyright (c) 2015 HPE Software Inc. All rights reserved.
+// Copyright (c) 2013 ActiveState Software Inc. All rights reserved.
+
+package watch
+
+import (
+	"fmt"
+	"os"
+	"path/filepath"
+
+	"github.com/nxadm/tail/util"
+
+    "github.com/fsnotify/fsnotify"
+	"gopkg.in/tomb.v1"
+)
+
+// InotifyFileWatcher uses inotify to monitor file changes.
+type InotifyFileWatcher struct {
+	Filename string
+	Size     int64
+}
+
+func NewInotifyFileWatcher(filename string) *InotifyFileWatcher {
+	fw := &InotifyFileWatcher{filepath.Clean(filename), 0}
+	return fw
+}
+
+func (fw *InotifyFileWatcher) BlockUntilExists(t *tomb.Tomb) error {
+	err := WatchCreate(fw.Filename)
+	if err != nil {
+		return err
+	}
+	defer RemoveWatchCreate(fw.Filename)
+
+	// Do a real check now as the file might have been created before
+	// calling `WatchFlags` above.
+	if _, err = os.Stat(fw.Filename); !os.IsNotExist(err) {
+		// file exists, or stat returned an error.
+		return err
+	}
+
+	events := Events(fw.Filename)
+
+	for {
+		select {
+		case evt, ok := <-events:
+			if !ok {
+				return fmt.Errorf("inotify watcher has been closed")
+			}
+			evtName, err := filepath.Abs(evt.Name)
+			if err != nil {
+				return err
+			}
+			fwFilename, err := filepath.Abs(fw.Filename)
+			if err != nil {
+				return err
+			}
+			if evtName == fwFilename {
+				return nil
+			}
+		case <-t.Dying():
+			return tomb.ErrDying
+		}
+	}
+	panic("unreachable")
+}
+
+func (fw *InotifyFileWatcher) ChangeEvents(t *tomb.Tomb, pos int64) (*FileChanges, error) {
+	err := Watch(fw.Filename)
+	if err != nil {
+		return nil, err
+	}
+
+	changes := NewFileChanges()
+	fw.Size = pos
+
+	go func() {
+
+		events := Events(fw.Filename)
+
+		for {
+			prevSize := fw.Size
+
+			var evt fsnotify.Event
+			var ok bool
+
+			select {
+			case evt, ok = <-events:
+				if !ok {
+					RemoveWatch(fw.Filename)
+					return
+				}
+			case <-t.Dying():
+				RemoveWatch(fw.Filename)
+				return
+			}
+
+			switch {
+			case evt.Op&fsnotify.Remove == fsnotify.Remove:
+				fallthrough
+
+			case evt.Op&fsnotify.Rename == fsnotify.Rename:
+				RemoveWatch(fw.Filename)
+				changes.NotifyDeleted()
+				return
+
+			//With an open fd, unlink(fd) - inotify returns IN_ATTRIB (==fsnotify.Chmod)
+			case evt.Op&fsnotify.Chmod == fsnotify.Chmod:
+				fallthrough
+
+			case evt.Op&fsnotify.Write == fsnotify.Write:
+				fi, err := os.Stat(fw.Filename)
+				if err != nil {
+					if os.IsNotExist(err) {
+						RemoveWatch(fw.Filename)
+						changes.NotifyDeleted()
+						return
+					}
+					// XXX: report this error back to the user
+					util.Fatal("Failed to stat file %v: %v", fw.Filename, err)
+				}
+				fw.Size = fi.Size()
+
+				if prevSize > 0 && prevSize > fw.Size {
+					changes.NotifyTruncated()
+				} else {
+					changes.NotifyModified()
+				}
+				prevSize = fw.Size
+			}
+		}
+	}()
+
+	return changes, nil
+}
@@ -0,0 +1,249 @@
+// Copyright (c) 2019 FOSS contributors of https://github.com/nxadm/tail
+// Copyright (c) 2015 HPE Software Inc. All rights reserved.
+// Copyright (c) 2013 ActiveState Software Inc. All rights reserved.
+
+package watch
+
+import (
+	"log"
+	"os"
+	"path/filepath"
+	"sync"
+	"syscall"
+
+	"github.com/nxadm/tail/util"
+
+    "github.com/fsnotify/fsnotify"
+)
+
+type InotifyTracker struct {
+	mux       sync.Mutex
+	watcher   *fsnotify.Watcher
+	chans     map[string]chan fsnotify.Event
+	done      map[string]chan bool
+	watchNums map[string]int
+	watch     chan *watchInfo
+	remove    chan *watchInfo
+	error     chan error
+}
+
+type watchInfo struct {
+	op    fsnotify.Op
+	fname string
+}
+
+func (this *watchInfo) isCreate() bool {
+	return this.op == fsnotify.Create
+}
+
+var (
+	// globally shared InotifyTracker; ensures only one fsnotify.Watcher is used
+	shared *InotifyTracker
+
+	// these are used to ensure the shared InotifyTracker is run exactly once
+	once  = sync.Once{}
+	goRun = func() {
+		shared = &InotifyTracker{
+			mux:       sync.Mutex{},
+			chans:     make(map[string]chan fsnotify.Event),
+			done:      make(map[string]chan bool),
+			watchNums: make(map[string]int),
+			watch:     make(chan *watchInfo),
+			remove:    make(chan *watchInfo),
+			error:     make(chan error),
+		}
+		go shared.run()
+	}
+
+	logger = log.New(os.Stderr, "", log.LstdFlags)
+)
+
+// Watch signals the run goroutine to begin watching the input filename
+func Watch(fname string) error {
+	return watch(&watchInfo{
+		fname: fname,
+	})
+}
+
+// Watch create signals the run goroutine to begin watching the input filename
+// if call the WatchCreate function, don't call the Cleanup, call the RemoveWatchCreate
+func WatchCreate(fname string) error {
+	return watch(&watchInfo{
+		op:    fsnotify.Create,
+		fname: fname,
+	})
+}
+
+func watch(winfo *watchInfo) error {
+	// start running the shared InotifyTracker if not already running
+	once.Do(goRun)
+
+	winfo.fname = filepath.Clean(winfo.fname)
+	shared.watch <- winfo
+	return <-shared.error
+}
+
+// RemoveWatch signals the run goroutine to remove the watch for the input filename
+func RemoveWatch(fname string) error {
+	return remove(&watchInfo{
+		fname: fname,
+	})
+}
+
+// RemoveWatch create signals the run goroutine to remove the watch for the input filename
+func RemoveWatchCreate(fname string) error {
+	return remove(&watchInfo{
+		op:    fsnotify.Create,
+		fname: fname,
+	})
+}
+
+func remove(winfo *watchInfo) error {
+	// start running the shared InotifyTracker if not already running
+	once.Do(goRun)
+
+	winfo.fname = filepath.Clean(winfo.fname)
+	shared.mux.Lock()
+	done := shared.done[winfo.fname]
+	if done != nil {
+		delete(shared.done, winfo.fname)
+		close(done)
+	}
+	shared.mux.Unlock()
+
+	shared.remove <- winfo
+	return <-shared.error
+}
+
+// Events returns a channel to which FileEvents corresponding to the input filename
+// will be sent. This channel will be closed when removeWatch is called on this
+// filename.
+func Events(fname string) <-chan fsnotify.Event {
+	shared.mux.Lock()
+	defer shared.mux.Unlock()
+
+	return shared.chans[fname]
+}
+
+// Cleanup removes the watch for the input filename if necessary.
+func Cleanup(fname string) error {
+	return RemoveWatch(fname)
+}
+
+// watchFlags calls fsnotify.WatchFlags for the input filename and flags, creating
+// a new Watcher if the previous Watcher was closed.
+func (shared *InotifyTracker) addWatch(winfo *watchInfo) error {
+	shared.mux.Lock()
+	defer shared.mux.Unlock()
+
+	if shared.chans[winfo.fname] == nil {
+		shared.chans[winfo.fname] = make(chan fsnotify.Event)
+	}
+	if shared.done[winfo.fname] == nil {
+		shared.done[winfo.fname] = make(chan bool)
+	}
+
+	fname := winfo.fname
+	if winfo.isCreate() {
+		// Watch for new files to be created in the parent directory.
+		fname = filepath.Dir(fname)
+	}
+
+	var err error
+	// already in inotify watch
+	if shared.watchNums[fname] == 0 {
+		err = shared.watcher.Add(fname)
+	}
+	if err == nil {
+		shared.watchNums[fname]++
+	}
+	return err
+}
+
+// removeWatch calls fsnotify.RemoveWatch for the input filename and closes the
+// corresponding events channel.
+func (shared *InotifyTracker) removeWatch(winfo *watchInfo) error {
+	shared.mux.Lock()
+
+	ch := shared.chans[winfo.fname]
+	if ch != nil {
+		delete(shared.chans, winfo.fname)
+		close(ch)
+	}
+
+	fname := winfo.fname
+	if winfo.isCreate() {
+		// Watch for new files to be created in the parent directory.
+		fname = filepath.Dir(fname)
+	}
+	shared.watchNums[fname]--
+	watchNum := shared.watchNums[fname]
+	if watchNum == 0 {
+		delete(shared.watchNums, fname)
+	}
+	shared.mux.Unlock()
+
+	var err error
+	// If we were the last ones to watch this file, unsubscribe from inotify.
+	// This needs to happen after releasing the lock because fsnotify waits
+	// synchronously for the kernel to acknowledge the removal of the watch
+	// for this file, which causes us to deadlock if we still held the lock.
+	if watchNum == 0 {
+		err = shared.watcher.Remove(fname)
+	}
+
+	return err
+}
+
+// sendEvent sends the input event to the appropriate Tail.
+func (shared *InotifyTracker) sendEvent(event fsnotify.Event) {
+	name := filepath.Clean(event.Name)
+
+	shared.mux.Lock()
+	ch := shared.chans[name]
+	done := shared.done[name]
+	shared.mux.Unlock()
+
+	if ch != nil && done != nil {
+		select {
+		case ch <- event:
+		case <-done:
+		}
+	}
+}
+
+// run starts the goroutine in which the shared struct reads events from its
+// Watcher's Event channel and sends the events to the appropriate Tail.
+func (shared *InotifyTracker) run() {
+	watcher, err := fsnotify.NewWatcher()
+	if err != nil {
+		util.Fatal("failed to create Watcher")
+	}
+	shared.watcher = watcher
+
+	for {
+		select {
+		case winfo := <-shared.watch:
+			shared.error <- shared.addWatch(winfo)
+
+		case winfo := <-shared.remove:
+			shared.error <- shared.removeWatch(winfo)
+
+		case event, open := <-shared.watcher.Events:
+			if !open {
+				return
+			}
+			shared.sendEvent(event)
+
+		case err, open := <-shared.watcher.Errors:
+			if !open {
+				return
+			} else if err != nil {
+				sysErr, ok := err.(*os.SyscallError)
+				if !ok || sysErr.Err != syscall.EINTR {
+					logger.Printf("Error in Watcher Error channel: %s", err)
+				}
+			}
+		}
+	}
+}
@@ -0,0 +1,119 @@
+// Copyright (c) 2019 FOSS contributors of https://github.com/nxadm/tail
+// Copyright (c) 2015 HPE Software Inc. All rights reserved.
+// Copyright (c) 2013 ActiveState Software Inc. All rights reserved.
+
+package watch
+
+import (
+	"os"
+	"runtime"
+	"time"
+
+	"github.com/nxadm/tail/util"
+	"gopkg.in/tomb.v1"
+)
+
+// PollingFileWatcher polls the file for changes.
+type PollingFileWatcher struct {
+	Filename string
+	Size     int64
+}
+
+func NewPollingFileWatcher(filename string) *PollingFileWatcher {
+	fw := &PollingFileWatcher{filename, 0}
+	return fw
+}
+
+var POLL_DURATION time.Duration
+
+func (fw *PollingFileWatcher) BlockUntilExists(t *tomb.Tomb) error {
+	for {
+		if _, err := os.Stat(fw.Filename); err == nil {
+			return nil
+		} else if !os.IsNotExist(err) {
+			return err
+		}
+		select {
+		case <-time.After(POLL_DURATION):
+			continue
+		case <-t.Dying():
+			return tomb.ErrDying
+		}
+	}
+	panic("unreachable")
+}
+
+func (fw *PollingFileWatcher) ChangeEvents(t *tomb.Tomb, pos int64) (*FileChanges, error) {
+	origFi, err := os.Stat(fw.Filename)
+	if err != nil {
+		return nil, err
+	}
+
+	changes := NewFileChanges()
+	var prevModTime time.Time
+
+	// XXX: use tomb.Tomb to cleanly manage these goroutines. replace
+	// the fatal (below) with tomb's Kill.
+
+	fw.Size = pos
+
+	go func() {
+		prevSize := fw.Size
+		for {
+			select {
+			case <-t.Dying():
+				return
+			default:
+			}
+
+			time.Sleep(POLL_DURATION)
+			fi, err := os.Stat(fw.Filename)
+			if err != nil {
+				// Windows cannot delete a file if a handle is still open (tail keeps one open)
+				// so it gives access denied to anything trying to read it until all handles are released.
+				if os.IsNotExist(err) || (runtime.GOOS == "windows" && os.IsPermission(err)) {
+					// File does not exist (has been deleted).
+					changes.NotifyDeleted()
+					return
+				}
+
+				// XXX: report this error back to the user
+				util.Fatal("Failed to stat file %v: %v", fw.Filename, err)
+			}
+
+			// File got moved/renamed?
+			if !os.SameFile(origFi, fi) {
+				changes.NotifyDeleted()
+				return
+			}
+
+			// File got truncated?
+			fw.Size = fi.Size()
+			if prevSize > 0 && prevSize > fw.Size {
+				changes.NotifyTruncated()
+				prevSize = fw.Size
+				continue
+			}
+			// File got bigger?
+			if prevSize > 0 && prevSize < fw.Size {
+				changes.NotifyModified()
+				prevSize = fw.Size
+				continue
+			}
+			prevSize = fw.Size
+
+			// File was appended to (changed)?
+			modTime := fi.ModTime()
+			if modTime != prevModTime {
+				prevModTime = modTime
+				changes.NotifyModified()
+			}
+		}
+	}()
+
+	return changes, nil
+}
+
+func init() {
+	POLL_DURATION = 250 * time.Millisecond
+}
@@ -0,0 +1,21 @@
+// Copyright (c) 2019 FOSS contributors of https://github.com/nxadm/tail
+// Copyright (c) 2015 HPE Software Inc. All rights reserved.
+// Copyright (c) 2013 ActiveState Software Inc. All rights reserved.
+
+package watch
+
+import "gopkg.in/tomb.v1"
+
+// FileWatcher monitors file-level events.
+type FileWatcher interface {
+	// BlockUntilExists blocks until the file comes into existence.
+	BlockUntilExists(*tomb.Tomb) error
+
+	// ChangeEvents reports on changes to a file, be it modification,
+	// deletion, renames or truncations. Returned FileChanges group of
+	// channels will be closed, thus become unusable, after a deletion
+	// or truncation event.
+	// In order to properly report truncations, ChangeEvents requires
+	// the caller to pass their current offset in the file.
+	ChangeEvents(*tomb.Tomb, int64) (*FileChanges, error)
+}
@@ -0,0 +1,93 @@
+// Copyright (c) 2019 FOSS contributors of https://github.com/nxadm/tail
+// +build windows
+
+package winfile
+
+import (
+	"os"
+	"syscall"
+	"unsafe"
+)
+
+// issue also described here
+//https://codereview.appspot.com/8203043/
+
+// https://github.com/jnwhiteh/golang/blob/master/src/pkg/syscall/syscall_windows.go#L218
+func Open(path string, mode int, perm uint32) (fd syscall.Handle, err error) {
+	if len(path) == 0 {
+		return syscall.InvalidHandle, syscall.ERROR_FILE_NOT_FOUND
+	}
+	pathp, err := syscall.UTF16PtrFromString(path)
+	if err != nil {
+		return syscall.InvalidHandle, err
+	}
+	var access uint32
+	switch mode & (syscall.O_RDONLY | syscall.O_WRONLY | syscall.O_RDWR) {
+	case syscall.O_RDONLY:
+		access = syscall.GENERIC_READ
+	case syscall.O_WRONLY:
+		access = syscall.GENERIC_WRITE
+	case syscall.O_RDWR:
+		access = syscall.GENERIC_READ | syscall.GENERIC_WRITE
+	}
+	if mode&syscall.O_CREAT != 0 {
+		access |= syscall.GENERIC_WRITE
+	}
+	if mode&syscall.O_APPEND != 0 {
+		access &^= syscall.GENERIC_WRITE
+		access |= syscall.FILE_APPEND_DATA
+	}
+	sharemode := uint32(syscall.FILE_SHARE_READ | syscall.FILE_SHARE_WRITE | syscall.FILE_SHARE_DELETE)
+	var sa *syscall.SecurityAttributes
+	if mode&syscall.O_CLOEXEC == 0 {
+		sa = makeInheritSa()
+	}
+	var createmode uint32
+	switch {
+	case mode&(syscall.O_CREAT|syscall.O_EXCL) == (syscall.O_CREAT | syscall.O_EXCL):
+		createmode = syscall.CREATE_NEW
+	case mode&(syscall.O_CREAT|syscall.O_TRUNC) == (syscall.O_CREAT | syscall.O_TRUNC):
+		createmode = syscall.CREATE_ALWAYS
+	case mode&syscall.O_CREAT == syscall.O_CREAT:
+		createmode = syscall.OPEN_ALWAYS
+	case mode&syscall.O_TRUNC == syscall.O_TRUNC:
+		createmode = syscall.TRUNCATE_EXISTING
+	default:
+		createmode = syscall.OPEN_EXISTING
+	}
+	h, e := syscall.CreateFile(pathp, access, sharemode, sa, createmode, syscall.FILE_ATTRIBUTE_NORMAL, 0)
+	return h, e
+}
+
+// https://github.com/jnwhiteh/golang/blob/master/src/pkg/syscall/syscall_windows.go#L211
+func makeInheritSa() *syscall.SecurityAttributes {
+	var sa syscall.SecurityAttributes
+	sa.Length = uint32(unsafe.Sizeof(sa))
+	sa.InheritHandle = 1
+	return &sa
+}
+
+// https://github.com/jnwhiteh/golang/blob/master/src/pkg/os/file_windows.go#L133
+func OpenFile(name string, flag int, perm os.FileMode) (file *os.File, err error) {
+	r, e := Open(name, flag|syscall.O_CLOEXEC, syscallMode(perm))
+	if e != nil {
+		return nil, e
+	}
+	return os.NewFile(uintptr(r), name), nil
+}
+
+// https://github.com/jnwhiteh/golang/blob/master/src/pkg/os/file_posix.go#L61
+func syscallMode(i os.FileMode) (o uint32) {
+	o |= uint32(i.Perm())
+	if i&os.ModeSetuid != 0 {
+		o |= syscall.S_ISUID
+	}
+	if i&os.ModeSetgid != 0 {
+		o |= syscall.S_ISGID
+	}
+	if i&os.ModeSticky != 0 {
+		o |= syscall.S_ISVTX
+	}
+	// No mapping for Go's ModeTemporary (plan9 only).
+	return
+}
@@ -0,0 +1,7 @@
+.DS_Store
+TODO
+tmp/**/*
+*.coverprofile
+.vscode
+.idea/
+*.log
@@ -0,0 +1,24 @@
+language: go
+go:
+  - tip
+  - 1.16.x
+  - 1.15.x
+
+cache:
+  directories:
+    - $GOPATH/pkg/mod
+
+# allow internal package imports, necessary for forked repositories
+go_import_path: github.com/onsi/ginkgo
+
+install:
+  - GO111MODULE="off" go get -v -t ./...
+  - GO111MODULE="off" go get golang.org/x/tools/cmd/cover
+  - GO111MODULE="off" go get github.com/onsi/gomega
+  - GO111MODULE="off" go install github.com/onsi/ginkgo/ginkgo
+  - export PATH=$GOPATH/bin:$PATH
+
+script:
+  - GO111MODULE="on" go mod tidy && git diff --exit-code go.mod go.sum
+  - go vet
+  - ginkgo -r --randomizeAllSpecs --randomizeSuites --race --trace
@@ -0,0 +1,393 @@
+## 1.16.5
+
+Ginkgo 2.0 now has a Release Candidate.  1.16.5 advertises the existence of the RC.
+1.16.5 deprecates GinkgoParallelNode in favor of GinkgoParallelProcess
+
+You can silence the RC advertisement by setting an `ACK_GINKG_RC=true` environment variable or creating a file in your home directory called `.ack-ginkgo-rc`
+
+## 1.16.4
+
+### Fixes
+1.16.4 retracts 1.16.3.  There are no code changes.  The 1.16.3 tag was associated with the wrong commit and an attempt to change it after-the-fact has proven problematic.  1.16.4 retracts 1.16.3 in Ginkgo's go.mod and creates a new, correctly tagged, release.
+
+## 1.16.3
+
+### Features
+- Measure is now deprecated and emits a deprecation warning.
+
+## 1.16.2
+
+### Fixes
+- Deprecations can be suppressed by setting an `ACK_GINKGO_DEPRECATIONS=<semver>` environment variable.
+
+## 1.16.1
+
+### Fixes
+- Supress --stream deprecation warning on windows (#793)
+
+## 1.16.0
+
+### Features
+- Advertise Ginkgo 2.0.  Introduce deprecations. [9ef1913]
+    - Update README.md to advertise that Ginkgo 2.0 is coming.
+    - Backport the 2.0 DeprecationTracker and start alerting users
+    about upcoming deprecations.
+
+- Add slim-sprig template functions to bootstrap/generate (#775) [9162b86]
+
+### Fixes
+- Fix accidental reference to 1488 (#784) [9fb7fe4]
+
+## 1.15.2
+
+### Fixes
+- ignore blank `-focus` and `-skip` flags (#780) [e90a4a0]
+
+## 1.15.1
+
+### Fixes
+- reporters/junit: Use `system-out` element instead of `passed` (#769) [9eda305]
+
+## 1.15.0
+
+### Features
+- Adds 'outline' command to print the outline of specs/containers in a file (#754) [071c369] [6803cc3] [935b538] [06744e8] [0c40583]
+- Add support for using template to generate tests (#752) [efb9e69]
+- Add a Chinese Doc #755 (#756) [5207632]
+- cli: allow multiple -focus and -skip flags (#736) [9a782fb]
+
+### Fixes
+- Add _internal to filename of tests created with internal flag (#751) [43c12da]
+
+## 1.14.2
+
+### Fixes
+- correct handling windows backslash in import path (#721) [97f3d51]
+- Add additional methods to GinkgoT() to improve compatibility with the testing.TB interface [b5fe44d]
+
+## 1.14.1
+
+### Fixes
+- Discard exported method declaration when running ginkgo bootstrap (#558) [f4b0240]
+
+## 1.14.0
+
+### Features
+- Defer running top-level container nodes until RunSpecs is called [d44dedf]
+- [Document Ginkgo lifecycle](http://onsi.github.io/ginkgo/#understanding-ginkgos-lifecycle)
+- Add `extensions/globals` package (#692) [3295c8f] - this can be helpful in contexts where you are test-driving your test-generation code (see [#692](https://github.com/onsi/ginkgo/pull/692))
+- Print Skip reason in JUnit reporter if one was provided [820dfab]
+
+## 1.13.0
+
+### Features
+- Add a version of table.Entry that allows dumping the entry parameters. (#689) [21eaef2]
+
+### Fixes
+- Ensure integration tests pass in an environment sans GOPATH [606fba2]
+- Add books package (#568) [fc0e44e]
+- doc(readme): installation via "tools package" (#677) [83bb20e]
+- Solve the undefined: unix.Dup2 compile error on mips64le (#680) [0624f75]
+- Import package without dot (#687) [6321024]
+- Fix integration tests to stop require GOPATH (#686) [a912ec5]
+
+## 1.12.3
+
+### Fixes
+- Print correct code location of failing table test (#666) [c6d7afb]
+
+## 1.12.2
+
+### Fixes
+- Update dependencies [ea4a036]
+
+## 1.12.1
+
+### Fixes
+- Make unfocus ("blur") much faster (#674) [8b18061]
+- Fix typo (#673) [7fdcbe8]
+- Test against 1.14 and remove 1.12 [d5c2ad6]
+- Test if a coverprofile content is empty before checking its latest character (#670) [14d9fa2]
+- replace tail package with maintained one. this fixes go get errors (#667) [4ba33d4]
+- improve ginkgo performance - makes progress on #644 [a14f98e]
+- fix convert integration tests [1f8ba69]
+- fix typo succesful -> successful (#663) [1ea49cf]
+- Fix invalid link (#658) [b886136]
+- convert utility : Include comments from source (#657) [1077c6d]
+- Explain what BDD means [d79e7fb]
+- skip race detector test on unsupported platform (#642) [f8ab89d]
+- Use Dup2 from golang.org/x/sys/unix instead of syscallDup (#638) [5d53c55]
+- Fix missing newline in combined coverage file (#641) [6a07ea2]
+- check if a spec is run before returning SpecSummary (#645) [8850000]
+
+## 1.12.0
+
+### Features
+- Add module definition (#630) [78916ab]
+
+## 1.11.0
+
+### Features
+- Add syscall for riscv64 architecture [f66e896]
+- teamcity reporter: output location of test failure as well as test definition (#626) [9869142]
+- teamcity reporter: output newline after every service message (#625) [3cfa02d]
+- Add support for go module when running `generate` command (#578) [9c89e3f]
+
+## 1.10.3
+
+### Fixes
+- Set go_import_path in travis.yml to allow internal packages in forks (#607) [3b721db]
+- Add integration test [d90e0dc]
+- Fix coverage files combining [e5dde8c]
+- A new CLI option: -ginkgo.reportFile <file path> (#601) [034fd25]
+
+## 1.10.2
+
+### Fixes
+- speed up table entry generateIt() (#609) [5049dc5]
+- Fix. Write errors to stderr instead of stdout (#610) [7bb3091]
+
+## 1.10.1
+
+### Fixes
+- stack backtrace: fix skipping (#600) [2a4c0bd]
+
+## 1.10.0
+
+### Fixes
+- stack backtrace: fix alignment and skipping [66915d6]
+- fix typo in documentation [8f97b93]
+
+## 1.9.0
+
+### Features
+- Option to print output into report, when tests have passed [0545415]
+
+### Fixes
+- Fixed typos in comments [0ecbc58]
+- gofmt code [a7f8bfb]
+- Simplify code [7454d00]
+- Simplify concatenation, incrementation and function assignment [4825557]
+- Avoid unnecessary conversions [9d9403c]
+- JUnit: include more detailed information about panic [19cca4b]
+- Print help to stdout when the user asks for help [4cb7441]
+
+
+## 1.8.0
+
+### New Features
+- allow config of the vet flag for `go test` (#562) [3cd45fa]
+- Support projects using go modules [d56ee76]
+
+### Fixes and Minor Improvements
+- chore(godoc): fixes typos in Measurement funcs [dbaca8e]
+- Optimize focus to avoid allocations [f493786]
+- Ensure generated test file names are underscored [505cc35]
+
+## 1.7.0
+
+### New Features
+- Add JustAfterEach (#484) [0d4f080]
+
+### Fixes
+- Correctly round suite time in junit reporter [2445fc1]
+- Avoid using -i argument to go test for Golang 1.10+ [46bbc26]
+
+## 1.6.0
+
+### New Features
+- add --debug flag to emit node output to files (#499) [39febac]
+
+### Fixes
+- fix: for `go vet` to pass [69338ec]
+- docs: fix for contributing instructions [7004cb1]
+- consolidate and streamline contribution docs (#494) [d848015]
+- Make generated Junit file compatable with "Maven Surefire" (#488) [e51bee6]
+- all: gofmt [000d317]
+- Increase eventually timeout to 30s [c73579c]
+- Clarify asynchronous test behaviour [294d8f4]
+- Travis badge should only show master [26d2143]
+
+## 1.5.0 5/10/2018
+
+### New Features
+- Supports go v1.10 (#443, #446, #451) [e873237, 468e89e, e37dbfe, a37f4c0, c0b857d, bca5260, 4177ca8]
+- Add a When() synonym for Context() (#386) [747514b, 7484dad, 7354a07, dd826c8]
+- Re-add noisySkippings flag [652e15c]
+- Allow coverage to be displayed for focused specs (#367) [11459a8]
+- Handle -outputdir flag (#364) [228e3a8]
+- Handle -coverprofile flag (#355) [43392d5]
+
+### Fixes
+- When using custom reporters register the custom reporters *before* the default reporter.  This allows users to see the output of any print statements in their customer reporters. (#365) [8382b23]
+- When running a test and calculating the coverage using the `-coverprofile` and `-outputdir` flags, Ginkgo fails with an error if the directory does not exist. This is due to an [issue in go 1.10](https://github.com/golang/go/issues/24588) (#446) [b36a6e0]
+- `unfocus` command ignores vendor folder (#459) [e5e551c, c556e43, a3b6351, 9a820dd]
+- Ignore packages whose tests are all ignored by go (#456) [7430ca7, 6d8be98]
+- Increase the threshold when checking time measuments (#455) [2f714bf, 68f622c]
+- Fix race condition in coverage tests (#423) [a5a8ff7, ab9c08b]
+- Add an extra new line after reporting spec run completion for test2json [874520d]
+- added name name field to junit reported testsuite [ae61c63]
+- Do not set the run time of a spec when the dryRun flag is used (#438) [457e2d9, ba8e856]
+- Process FWhen and FSpecify when unfocusing (#434) [9008c7b, ee65bd, df87dfe]
+- Synchronise the access to the state of specs to avoid race conditions (#430) [7d481bc, ae6829d]
+- Added Duration on GinkgoTestDescription (#383) [5f49dad, 528417e, 0747408, 329d7ed]
+- Fix Ginkgo stack trace on failure for Specify (#415) [b977ede, 65ca40e, 6c46eb8]
+- Update README with Go 1.6+, Golang -> Go (#409) [17f6b97, bc14b66, 20d1598]
+- Use fmt.Errorf instead of errors.New(fmt.Sprintf (#401) [a299f56, 44e2eaa]
+- Imports in generated code should follow conventions (#398) [0bec0b0, e8536d8]
+- Prevent data race error when Recording a benchmark value from multiple go routines (#390) [c0c4881, 7a241e9]
+- Replace GOPATH in Environment [4b883f0]
+
+
+## 1.4.0 7/16/2017
+
+- `ginkgo` now provides a hint if you accidentally forget to run `ginkgo bootstrap` to generate a `*_suite_test.go` file that actually invokes the Ginkgo test runner. [#345](https://github.com/onsi/ginkgo/pull/345)
+- thanks to improvements in `go test -c` `ginkgo` no longer needs to fix Go's compilation output to ensure compilation errors are expressed relative to the CWD. [#357]
+- `ginkgo watch -watchRegExp=...` allows you to specify a custom regular expression to watch.  Only files matching the regular expression are watched for changes (the default is `\.go$`) [#356]
+- `ginkgo` now always emits compilation output.  Previously, only failed compilation output was printed out. [#277]
+- `ginkgo -requireSuite` now fails the test run if there are `*_test.go` files but `go test` fails to detect any tests.  Typically this means you forgot to run `ginkgo bootstrap` to generate a suite file. [#344]
+- `ginkgo -timeout=DURATION` allows you to adjust the timeout for the entire test suite (default is 24 hours) [#248]
+
+## 1.3.0 3/28/2017
+
+Improvements:
+
+- Significantly improved parallel test distribution.  Now instead of pre-sharding test cases across workers (which can result in idle workers and poor test performance) Ginkgo uses a shared queue to keep all workers busy until all tests are complete.  This improves test-time performance and consistency.
+- `Skip(message)` can be used to skip the current test.
+- Added `extensions/table` - a Ginkgo DSL for [Table Driven Tests](http://onsi.github.io/ginkgo/#table-driven-tests)
+- Add `GinkgoRandomSeed()` - shorthand for `config.GinkgoConfig.RandomSeed`
+- Support for retrying flaky tests with `--flakeAttempts`
+- `ginkgo ./...` now recurses as you'd expect
+- Added `Specify` a synonym for `It`
+- Support colorise on Windows
+- Broader support for various go compilation flags in the `ginkgo` CLI
+
+Bug Fixes:
+
+- Ginkgo tests now fail when you `panic(nil)` (#167)
+
+## 1.2.0 5/31/2015
+
+Improvements
+
+- `ginkgo -coverpkg` calls down to `go test -coverpkg` (#160)
+- `ginkgo -afterSuiteHook COMMAND` invokes the passed-in `COMMAND` after a test suite completes (#152)
+- Relaxed requirement for Go 1.4+.  `ginkgo` now works with Go v1.3+ (#166)
+
+## 1.2.0-beta
+
+Ginkgo now requires Go 1.4+
+
+Improvements:
+
+- Call reporters in reverse order when announcing spec completion -- allows custom reporters to emit output before the default reporter does.
+- Improved focus behavior.  Now, this:
+
+    ```golang
+    FDescribe("Some describe", func() {
+        It("A", func() {})
+
+        FIt("B", func() {})
+    })
+    ```
+
+  will run `B` but *not* `A`.  This tends to be a common usage pattern when in the thick of writing and debugging tests.
+- When `SIGINT` is received, Ginkgo will emit the contents of the `GinkgoWriter` before running the `AfterSuite`.  Useful for debugging stuck tests.
+- When `--progress` is set, Ginkgo will write test progress (in particular, Ginkgo will say when it is about to run a BeforeEach, AfterEach, It, etc...) to the `GinkgoWriter`.  This is useful for debugging stuck tests and tests that generate many logs.
+- Improved output when an error occurs in a setup or teardown block.
+- When `--dryRun` is set, Ginkgo will walk the spec tree and emit to its reporter *without* actually running anything.  Best paired with `-v` to understand which specs will run in which order.
+- Add `By` to help document long `It`s.  `By` simply writes to the `GinkgoWriter`.
+- Add support for precompiled tests:
+    - `ginkgo build <path-to-package>` will now compile the package, producing a file named `package.test`
+    - The compiled `package.test` file can be run directly.  This runs the tests in series.
+    - To run precompiled tests in parallel, you can run: `ginkgo -p package.test`
+- Support `bootstrap`ping and `generate`ing [Agouti](http://agouti.org) specs.
+- `ginkgo generate` and `ginkgo bootstrap` now honor the package name already defined in a given directory
+- The `ginkgo` CLI ignores `SIGQUIT`.  Prevents its stack dump from interlacing with the underlying test suite's stack dump.
+- The `ginkgo` CLI now compiles tests into a temporary directory instead of the package directory.  This necessitates upgrading to Go v1.4+.
+- `ginkgo -notify` now works on Linux
+
+Bug Fixes:
+
+- If --skipPackages is used and all packages are skipped, Ginkgo should exit 0.
+- Fix tempfile leak when running in parallel
+- Fix incorrect failure message when a panic occurs during a parallel test run
+- Fixed an issue where a pending test within a focused context (or a focused test within a pending context) would skip all other tests.
+- Be more consistent about handling SIGTERM as well as SIGINT
+- When interupted while concurrently compiling test suites in the background, Ginkgo now cleans up the compiled artifacts.
+- Fixed a long standing bug where `ginkgo -p` would hang if a process spawned by one of the Ginkgo parallel nodes does not exit. (Hooray!)
+
+## 1.1.0 (8/2/2014)
+
+No changes, just dropping the beta.
+
+## 1.1.0-beta (7/22/2014)
+New Features:
+
+- `ginkgo watch` now monitors packages *and their dependencies* for changes.  The depth of the dependency tree can be modified with the `-depth` flag.
+- Test suites with a programmatic focus (`FIt`, `FDescribe`, etc...) exit with non-zero status code, even when they pass.  This allows CI systems to detect accidental commits of focused test suites.
+- `ginkgo -p` runs the testsuite in parallel with an auto-detected number of nodes.
+- `ginkgo -tags=TAG_LIST` passes a list of tags down to the `go build` command.
+- `ginkgo --failFast` aborts the test suite after the first failure.
+- `ginkgo generate file_1 file_2` can take multiple file arguments.
+- Ginkgo now summarizes any spec failures that occurred at the end of the test run. 
+- `ginkgo --randomizeSuites` will run tests *suites* in random order using the generated/passed-in seed.
+
+Improvements:
+
+- `ginkgo -skipPackage` now takes a comma-separated list of strings.  If the *relative path* to a package matches one of the entries in the comma-separated list, that package is skipped.
+- `ginkgo --untilItFails` no longer recompiles between attempts.
+- Ginkgo now panics when a runnable node (`It`, `BeforeEach`, `JustBeforeEach`, `AfterEach`, `Measure`) is nested within another runnable node.  This is always a mistake.  Any test suites that panic because of this change should be fixed.
+
+Bug Fixes:
+
+- `ginkgo boostrap` and `ginkgo generate` no longer fail when dealing with `hyphen-separated-packages`.
+- parallel specs are now better distributed across nodes - fixed a crashing bug where (for example) distributing 11 tests across 7 nodes would panic
+
+## 1.0.0 (5/24/2014)
+New Features:
+
+- Add `GinkgoParallelNode()` - shorthand for `config.GinkgoConfig.ParallelNode`
+
+Improvements:
+
+- When compilation fails, the compilation output is rewritten to present a correct *relative* path.  Allows ⌘-clicking in iTerm open the file in your text editor.
+- `--untilItFails` and `ginkgo watch` now generate new random seeds between test runs, unless a particular random seed is specified.
+
+Bug Fixes:
+
+- `-cover` now generates a correctly combined coverprofile when running with in parallel with multiple `-node`s.
+- Print out the contents of the `GinkgoWriter` when `BeforeSuite` or `AfterSuite` fail.
+- Fix all remaining race conditions in Ginkgo's test suite.
+
+## 1.0.0-beta (4/14/2014)
+Breaking changes:
+
+- `thirdparty/gomocktestreporter` is gone.  Use `GinkgoT()` instead
+- Modified the Reporter interface 
+- `watch` is now a subcommand, not a flag.
+
+DSL changes:
+
+- `BeforeSuite` and `AfterSuite` for setting up and tearing down test suites.
+- `AfterSuite` is triggered on interrupt (`^C`) as well as exit.
+- `SynchronizedBeforeSuite` and `SynchronizedAfterSuite` for setting up and tearing down singleton resources across parallel nodes.
+
+CLI changes:
+
+- `watch` is now a subcommand, not a flag
+- `--nodot` flag can be passed to `ginkgo generate` and `ginkgo bootstrap` to avoid dot imports.  This explicitly imports all exported identifiers in Ginkgo and Gomega.  Refreshing this list can be done by running `ginkgo nodot`
+- Additional arguments can be passed to specs.  Pass them after the `--` separator
+- `--skipPackage` flag takes a regexp and ignores any packages with package names passing said regexp.
+- `--trace` flag prints out full stack traces when errors occur, not just the line at which the error occurs.
+
+Misc:
+
+- Start using semantic versioning
+- Start maintaining changelog
+
+Major refactor:
+
+- Pull out Ginkgo's internal to `internal`
+- Rename `example` everywhere to `spec`
+- Much more!
@@ -0,0 +1,33 @@
+# Contributing to Ginkgo
+
+Your contributions to Ginkgo are essential for its long-term maintenance and improvement.
+
+- Please **open an issue first** - describe what problem you are trying to solve and give the community a forum for input and feedback ahead of investing time in writing code!
+- Ensure adequate test coverage:
+    - When adding to the Ginkgo library, add unit and/or integration tests (under the `integration` folder).
+    - When adding to the Ginkgo CLI, note that there are very few unit tests.  Please add an integration test.
+- Update the documentation. Ginko uses `godoc` comments and documentation on the `gh-pages` branch.
+  If relevant, please submit a docs PR to that branch alongside your code PR.
+
+Thanks for supporting Ginkgo!
+
+## Setup
+
+Fork the repo, then:
+
+```
+go get github.com/onsi/ginkgo
+go get github.com/onsi/gomega/...
+cd $GOPATH/src/github.com/onsi/ginkgo
+git remote add fork [email protected]:<NAME>/ginkgo.git
+
+ginkgo -r -p   # ensure tests are green
+go vet ./...   # ensure linter is happy
+```
+
+## Making the PR
+ - go to a new branch `git checkout -b my-feature`
+ - make your changes
+ - run tests and linter again (see above)
+ - `git push fork`
+ - open PR 🎉
@@ -0,0 +1,20 @@
+Copyright (c) 2013-2014 Onsi Fakhouri
+
+Permission is hereby granted, free of charge, to any person obtaining
+a copy of this software and associated documentation files (the
+"Software"), to deal in the Software without restriction, including
+without limitation the rights to use, copy, modify, merge, publish,
+distribute, sublicense, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so, subject to
+the following conditions:
+
+The above copyright notice and this permission notice shall be
+included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
+LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
@@ -0,0 +1,169 @@
+![Ginkgo: A Go BDD Testing Framework](https://onsi.github.io/ginkgo/images/ginkgo.png)
+
+[![test](https://github.com/onsi/ginkgo/workflows/test/badge.svg?branch=master)](https://github.com/onsi/ginkgo/actions?query=workflow%3Atest+branch%3Amaster)
+
+Jump to the [docs](https://onsi.github.io/ginkgo/) | [中文文档](https://ke-chain.github.io/ginkgodoc) to learn more.  To start rolling your Ginkgo tests *now* [keep reading](#set-me-up)!
+
+If you have a question, comment, bug report, feature request, etc. please open a GitHub issue, or visit the [Ginkgo Slack channel](https://app.slack.com/client/T029RQSE6/CQQ50BBNW).
+
+# Ginkgo 2.0 Release Candidate is available!
+
+An effort is underway to develop and deliver Ginkgo 2.0.  The work is happening in the [ver2](https://github.com/onsi/ginkgo/tree/ver2) branch and a changelog and migration guide is being maintained on that branch [here](https://github.com/onsi/ginkgo/blob/ver2/docs/MIGRATING_TO_V2.md).  Issue [#711](https://github.com/onsi/ginkgo/issues/711) is the central place for discussion.
+
+As described in the [changelog](https://github.com/onsi/ginkgo/blob/ver2/docs/MIGRATING_TO_V2.md) and [proposal](https://docs.google.com/document/d/1h28ZknXRsTLPNNiOjdHIO-F2toCzq4xoZDXbfYaBdoQ/edit#), Ginkgo 2.0 will clean up the Ginkgo codebase, deprecate and remove some v1 functionality, and add several new much-requested features.  To help users get ready for the migration, Ginkgo v1 has started emitting deprecation warnings for features that will no longer be supported with links to documentation for how to migrate away from these features.  If you have concerns or comments please chime in on [#711](https://github.com/onsi/ginkgo/issues/711).
+
+Please start exploring and using the V2 release!  To get started follow the [Using the Release Candidate](https://github.com/onsi/ginkgo/blob/ver2/docs/MIGRATING_TO_V2.md#using-the-beta) directions in the migration guide.
+
+## TLDR
+Ginkgo builds on Go's `testing` package, allowing expressive [Behavior-Driven Development](https://en.wikipedia.org/wiki/Behavior-driven_development) ("BDD") style tests.
+It is typically (and optionally) paired with the [Gomega](https://github.com/onsi/gomega) matcher library.
+
+```go
+Describe("the strings package", func() {
+  Context("strings.Contains()", func() {
+    When("the string contains the substring in the middle", func() {
+      It("returns `true`", func() {
+        Expect(strings.Contains("Ginkgo is awesome", "is")).To(BeTrue())
+      })
+    })
+  })
+})
+```
+
+## Feature List
+
+- Ginkgo uses Go's `testing` package and can live alongside your existing `testing` tests.  It's easy to [bootstrap](https://onsi.github.io/ginkgo/#bootstrapping-a-suite) and start writing your [first tests](https://onsi.github.io/ginkgo/#adding-specs-to-a-suite)
+
+- Ginkgo allows you to write tests in Go using expressive [Behavior-Driven Development](https://en.wikipedia.org/wiki/Behavior-driven_development) ("BDD") style:
+    - Nestable [`Describe`, `Context` and `When` container blocks](https://onsi.github.io/ginkgo/#organizing-specs-with-containers-describe-and-context)
+    - [`BeforeEach` and `AfterEach` blocks](https://onsi.github.io/ginkgo/#extracting-common-setup-beforeeach) for setup and teardown
+    - [`It` and `Specify` blocks](https://onsi.github.io/ginkgo/#individual-specs-it) that hold your assertions
+    - [`JustBeforeEach` blocks](https://onsi.github.io/ginkgo/#separating-creation-and-configuration-justbeforeeach) that separate creation from configuration (also known as the subject action pattern).
+    - [`BeforeSuite` and `AfterSuite` blocks](https://onsi.github.io/ginkgo/#global-setup-and-teardown-beforesuite-and-aftersuite) to prep for and cleanup after a suite.
+
+- A comprehensive test runner that lets you:
+    - Mark specs as [pending](https://onsi.github.io/ginkgo/#pending-specs)
+    - [Focus](https://onsi.github.io/ginkgo/#focused-specs) individual specs, and groups of specs, either programmatically or on the command line
+    - Run your tests in [random order](https://onsi.github.io/ginkgo/#spec-permutation), and then reuse random seeds to replicate the same order.
+    - Break up your test suite into parallel processes for straightforward [test parallelization](https://onsi.github.io/ginkgo/#parallel-specs)
+
+- `ginkgo`: a command line interface with plenty of handy command line arguments for [running your tests](https://onsi.github.io/ginkgo/#running-tests) and [generating](https://onsi.github.io/ginkgo/#generators) test files.  Here are a few choice examples:
+    - `ginkgo -nodes=N` runs your tests in `N` parallel processes and print out coherent output in realtime
+    - `ginkgo -cover` runs your tests using Go's code coverage tool
+    - `ginkgo convert` converts an XUnit-style `testing` package to a Ginkgo-style package
+    - `ginkgo -focus="REGEXP"` and `ginkgo -skip="REGEXP"` allow you to specify a subset of tests to run via regular expression
+    - `ginkgo -r` runs all tests suites under the current directory
+    - `ginkgo -v` prints out identifying information for each tests just before it runs
+
+    And much more: run `ginkgo help` for details!
+
+    The `ginkgo` CLI is convenient, but purely optional -- Ginkgo works just fine with `go test`
+
+- `ginkgo watch` [watches](https://onsi.github.io/ginkgo/#watching-for-changes) packages *and their dependencies* for changes, then reruns tests.  Run tests immediately as you develop!
+
+- Built-in support for testing [asynchronicity](https://onsi.github.io/ginkgo/#asynchronous-tests)
+
+- Built-in support for [benchmarking](https://onsi.github.io/ginkgo/#benchmark-tests) your code.  Control the number of benchmark samples as you gather runtimes and other, arbitrary, bits of numerical information about your code. 
+
+- [Completions for Sublime Text](https://github.com/onsi/ginkgo-sublime-completions): just use [Package Control](https://sublime.wbond.net/) to install `Ginkgo Completions`.
+
+- [Completions for VSCode](https://github.com/onsi/vscode-ginkgo): just use VSCode's extension installer to install `vscode-ginkgo`.
+
+- [Ginkgo tools for VSCode](https://marketplace.visualstudio.com/items?itemName=joselitofilho.ginkgotestexplorer): just use VSCode's extension installer to install `ginkgoTestExplorer`.
+
+- Straightforward support for third-party testing libraries such as [Gomock](https://code.google.com/p/gomock/) and [Testify](https://github.com/stretchr/testify).  Check out the [docs](https://onsi.github.io/ginkgo/#third-party-integrations) for details.
+
+- A modular architecture that lets you easily:
+    - Write [custom reporters](https://onsi.github.io/ginkgo/#writing-custom-reporters) (for example, Ginkgo comes with a [JUnit XML reporter](https://onsi.github.io/ginkgo/#generating-junit-xml-output) and a TeamCity reporter).
+    - [Adapt an existing matcher library (or write your own!)](https://onsi.github.io/ginkgo/#using-other-matcher-libraries) to work with Ginkgo
+
+## [Gomega](https://github.com/onsi/gomega): Ginkgo's Preferred Matcher Library
+
+Ginkgo is best paired with Gomega.  Learn more about Gomega [here](https://onsi.github.io/gomega/)
+
+## [Agouti](https://github.com/sclevine/agouti): A Go Acceptance Testing Framework
+
+Agouti allows you run WebDriver integration tests.  Learn more about Agouti [here](https://agouti.org)
+
+## Getting Started
+
+You'll need the Go command-line tools. Follow the [installation instructions](https://golang.org/doc/install) if you don't have it installed.
+
+### Global installation
+To install the Ginkgo command line interface:
+```bash
+go get -u github.com/onsi/ginkgo/ginkgo
+```
+Note that this will install it to `$GOBIN`, which will need to be in the `$PATH` (or equivalent). Run `go help install` for more information.
+
+### Go module ["tools package"](https://github.com/golang/go/issues/25922):
+Create (or update) a file called `tools/tools.go` with the following contents:
+```go
+// +build tools
+
+package tools
+
+import (
+	_ "github.com/onsi/ginkgo/ginkgo"
+)
+
+// This file imports packages that are used when running go generate, or used
+// during the development process but not otherwise depended on by built code.
+```
+The Ginkgo command can then be run via `go run github.com/onsi/ginkgo/ginkgo`.
+This approach allows the version of Ginkgo to be maintained under source control for reproducible results,
+and is well suited to automated test pipelines.
+
+### Bootstrapping
+```bash
+cd path/to/package/you/want/to/test
+
+ginkgo bootstrap # set up a new ginkgo suite
+ginkgo generate  # will create a sample test file.  edit this file and add your tests then...
+
+go test # to run your tests
+
+ginkgo  # also runs your tests
+
+```
+
+## I'm new to Go: What are my testing options?
+
+Of course, I heartily recommend [Ginkgo](https://github.com/onsi/ginkgo) and [Gomega](https://github.com/onsi/gomega).  Both packages are seeing heavy, daily, production use on a number of projects and boast a mature and comprehensive feature-set.
+
+With that said, it's great to know what your options are :)
+
+### What Go gives you out of the box
+
+Testing is a first class citizen in Go, however Go's built-in testing primitives are somewhat limited: The [testing](https://golang.org/pkg/testing) package provides basic XUnit style tests and no assertion library.
+
+### Matcher libraries for Go's XUnit style tests
+
+A number of matcher libraries have been written to augment Go's built-in XUnit style tests.  Here are two that have gained traction:
+
+- [testify](https://github.com/stretchr/testify)
+- [gocheck](https://labix.org/gocheck)
+
+You can also use Ginkgo's matcher library [Gomega](https://github.com/onsi/gomega) in [XUnit style tests](https://onsi.github.io/gomega/#using-gomega-with-golangs-xunitstyle-tests)
+
+### BDD style testing frameworks
+
+There are a handful of BDD-style testing frameworks written for Go.  Here are a few:
+
+- [Ginkgo](https://github.com/onsi/ginkgo) ;)
+- [GoConvey](https://github.com/smartystreets/goconvey) 
+- [Goblin](https://github.com/franela/goblin)
+- [Mao](https://github.com/azer/mao)
+- [Zen](https://github.com/pranavraja/zen)
+
+Finally, @shageman has [put together](https://github.com/shageman/gotestit) a comprehensive comparison of Go testing libraries.
+
+Go explore!
+
+## License
+
+Ginkgo is MIT-Licensed
+
+## Contributing
+
+See [CONTRIBUTING.md](CONTRIBUTING.md)
@@ -0,0 +1,17 @@
+A Ginkgo release is a tagged git sha and a GitHub release.  To cut a release:
+
+1. Ensure CHANGELOG.md is up to date.
+  - Use `git log --pretty=format:'- %s [%h]' HEAD...vX.X.X` to list all the commits since the last release
+  - Categorize the changes into
+    - Breaking Changes (requires a major version)
+    - New Features (minor version)
+    - Fixes (fix version)
+    - Maintenance (which in general should not be mentioned in `CHANGELOG.md` as they have no user impact)
+1. Update `VERSION` in `config/config.go`
+1. Commit, push, and release:
+  ```
+  git commit -m "vM.m.p"
+  git push
+  gh release create "vM.m.p"
+  git fetch --tags origin master
+  ```
@@ -0,0 +1,232 @@
+/*
+Ginkgo accepts a number of configuration options.
+
+These are documented [here](http://onsi.github.io/ginkgo/#the-ginkgo-cli)
+
+You can also learn more via
+
+	ginkgo help
+
+or (I kid you not):
+
+	go test -asdf
+*/
+package config
+
+import (
+	"flag"
+	"time"
+
+	"fmt"
+)
+
+const VERSION = "1.16.5"
+
+type GinkgoConfigType struct {
+	RandomSeed         int64
+	RandomizeAllSpecs  bool
+	RegexScansFilePath bool
+	FocusStrings       []string
+	SkipStrings        []string
+	SkipMeasurements   bool
+	FailOnPending      bool
+	FailFast           bool
+	FlakeAttempts      int
+	EmitSpecProgress   bool
+	DryRun             bool
+	DebugParallel      bool
+
+	ParallelNode  int
+	ParallelTotal int
+	SyncHost      string
+	StreamHost    string
+}
+
+var GinkgoConfig = GinkgoConfigType{}
+
+type DefaultReporterConfigType struct {
+	NoColor           bool
+	SlowSpecThreshold float64
+	NoisyPendings     bool
+	NoisySkippings    bool
+	Succinct          bool
+	Verbose           bool
+	FullTrace         bool
+	ReportPassed      bool
+	ReportFile        string
+}
+
+var DefaultReporterConfig = DefaultReporterConfigType{}
+
+func processPrefix(prefix string) string {
+	if prefix != "" {
+		prefix += "."
+	}
+	return prefix
+}
+
+type flagFunc func(string)
+
+func (f flagFunc) String() string     { return "" }
+func (f flagFunc) Set(s string) error { f(s); return nil }
+
+func Flags(flagSet *flag.FlagSet, prefix string, includeParallelFlags bool) {
+	prefix = processPrefix(prefix)
+	flagSet.Int64Var(&(GinkgoConfig.RandomSeed), prefix+"seed", time.Now().Unix(), "The seed used to randomize the spec suite.")
+	flagSet.BoolVar(&(GinkgoConfig.RandomizeAllSpecs), prefix+"randomizeAllSpecs", false, "If set, ginkgo will randomize all specs together.  By default, ginkgo only randomizes the top level Describe, Context and When groups.")
+	flagSet.BoolVar(&(GinkgoConfig.SkipMeasurements), prefix+"skipMeasurements", false, "If set, ginkgo will skip any measurement specs.")
+	flagSet.BoolVar(&(GinkgoConfig.FailOnPending), prefix+"failOnPending", false, "If set, ginkgo will mark the test suite as failed if any specs are pending.")
+	flagSet.BoolVar(&(GinkgoConfig.FailFast), prefix+"failFast", false, "If set, ginkgo will stop running a test suite after a failure occurs.")
+
+	flagSet.BoolVar(&(GinkgoConfig.DryRun), prefix+"dryRun", false, "If set, ginkgo will walk the test hierarchy without actually running anything.  Best paired with -v.")
+
+	flagSet.Var(flagFunc(flagFocus), prefix+"focus", "If set, ginkgo will only run specs that match this regular expression. Can be specified multiple times, values are ORed.")
+	flagSet.Var(flagFunc(flagSkip), prefix+"skip", "If set, ginkgo will only run specs that do not match this regular expression. Can be specified multiple times, values are ORed.")
+
+	flagSet.BoolVar(&(GinkgoConfig.RegexScansFilePath), prefix+"regexScansFilePath", false, "If set, ginkgo regex matching also will look at the file path (code location).")
+
+	flagSet.IntVar(&(GinkgoConfig.FlakeAttempts), prefix+"flakeAttempts", 1, "Make up to this many attempts to run each spec. Please note that if any of the attempts succeed, the suite will not be failed. But any failures will still be recorded.")
+
+	flagSet.BoolVar(&(GinkgoConfig.EmitSpecProgress), prefix+"progress", false, "If set, ginkgo will emit progress information as each spec runs to the GinkgoWriter.")
+
+	flagSet.BoolVar(&(GinkgoConfig.DebugParallel), prefix+"debug", false, "If set, ginkgo will emit node output to files when running in parallel.")
+
+	if includeParallelFlags {
+		flagSet.IntVar(&(GinkgoConfig.ParallelNode), prefix+"parallel.node", 1, "This worker node's (one-indexed) node number.  For running specs in parallel.")
+		flagSet.IntVar(&(GinkgoConfig.ParallelTotal), prefix+"parallel.total", 1, "The total number of worker nodes.  For running specs in parallel.")
+		flagSet.StringVar(&(GinkgoConfig.SyncHost), prefix+"parallel.synchost", "", "The address for the server that will synchronize the running nodes.")
+		flagSet.StringVar(&(GinkgoConfig.StreamHost), prefix+"parallel.streamhost", "", "The address for the server that the running nodes should stream data to.")
+	}
+
+	flagSet.BoolVar(&(DefaultReporterConfig.NoColor), prefix+"noColor", false, "If set, suppress color output in default reporter.")
+	flagSet.Float64Var(&(DefaultReporterConfig.SlowSpecThreshold), prefix+"slowSpecThreshold", 5.0, "(in seconds) Specs that take longer to run than this threshold are flagged as slow by the default reporter.")
+	flagSet.BoolVar(&(DefaultReporterConfig.NoisyPendings), prefix+"noisyPendings", true, "If set, default reporter will shout about pending tests.")
+	flagSet.BoolVar(&(DefaultReporterConfig.NoisySkippings), prefix+"noisySkippings", true, "If set, default reporter will shout about skipping tests.")
+	flagSet.BoolVar(&(DefaultReporterConfig.Verbose), prefix+"v", false, "If set, default reporter print out all specs as they begin.")
+	flagSet.BoolVar(&(DefaultReporterConfig.Succinct), prefix+"succinct", false, "If set, default reporter prints out a very succinct report")
+	flagSet.BoolVar(&(DefaultReporterConfig.FullTrace), prefix+"trace", false, "If set, default reporter prints out the full stack trace when a failure occurs")
+	flagSet.BoolVar(&(DefaultReporterConfig.ReportPassed), prefix+"reportPassed", false, "If set, default reporter prints out captured output of passed tests.")
+	flagSet.StringVar(&(DefaultReporterConfig.ReportFile), prefix+"reportFile", "", "Override the default reporter output file path.")
+
+}
+
+func BuildFlagArgs(prefix string, ginkgo GinkgoConfigType, reporter DefaultReporterConfigType) []string {
+	prefix = processPrefix(prefix)
+	result := make([]string, 0)
+
+	if ginkgo.RandomSeed > 0 {
+		result = append(result, fmt.Sprintf("--%sseed=%d", prefix, ginkgo.RandomSeed))
+	}
+
+	if ginkgo.RandomizeAllSpecs {
+		result = append(result, fmt.Sprintf("--%srandomizeAllSpecs", prefix))
+	}
+
+	if ginkgo.SkipMeasurements {
+		result = append(result, fmt.Sprintf("--%sskipMeasurements", prefix))
+	}
+
+	if ginkgo.FailOnPending {
+		result = append(result, fmt.Sprintf("--%sfailOnPending", prefix))
+	}
+
+	if ginkgo.FailFast {
+		result = append(result, fmt.Sprintf("--%sfailFast", prefix))
+	}
+
+	if ginkgo.DryRun {
+		result = append(result, fmt.Sprintf("--%sdryRun", prefix))
+	}
+
+	for _, s := range ginkgo.FocusStrings {
+		result = append(result, fmt.Sprintf("--%sfocus=%s", prefix, s))
+	}
+
+	for _, s := range ginkgo.SkipStrings {
+		result = append(result, fmt.Sprintf("--%sskip=%s", prefix, s))
+	}
+
+	if ginkgo.FlakeAttempts > 1 {
+		result = append(result, fmt.Sprintf("--%sflakeAttempts=%d", prefix, ginkgo.FlakeAttempts))
+	}
+
+	if ginkgo.EmitSpecProgress {
+		result = append(result, fmt.Sprintf("--%sprogress", prefix))
+	}
+
+	if ginkgo.DebugParallel {
+		result = append(result, fmt.Sprintf("--%sdebug", prefix))
+	}
+
+	if ginkgo.ParallelNode != 0 {
+		result = append(result, fmt.Sprintf("--%sparallel.node=%d", prefix, ginkgo.ParallelNode))
+	}
+
+	if ginkgo.ParallelTotal != 0 {
+		result = append(result, fmt.Sprintf("--%sparallel.total=%d", prefix, ginkgo.ParallelTotal))
+	}
+
+	if ginkgo.StreamHost != "" {
+		result = append(result, fmt.Sprintf("--%sparallel.streamhost=%s", prefix, ginkgo.StreamHost))
+	}
+
+	if ginkgo.SyncHost != "" {
+		result = append(result, fmt.Sprintf("--%sparallel.synchost=%s", prefix, ginkgo.SyncHost))
+	}
+
+	if ginkgo.RegexScansFilePath {
+		result = append(result, fmt.Sprintf("--%sregexScansFilePath", prefix))
+	}
+
+	if reporter.NoColor {
+		result = append(result, fmt.Sprintf("--%snoColor", prefix))
+	}
+
+	if reporter.SlowSpecThreshold > 0 {
+		result = append(result, fmt.Sprintf("--%sslowSpecThreshold=%.5f", prefix, reporter.SlowSpecThreshold))
+	}
+
+	if !reporter.NoisyPendings {
+		result = append(result, fmt.Sprintf("--%snoisyPendings=false", prefix))
+	}
+
+	if !reporter.NoisySkippings {
+		result = append(result, fmt.Sprintf("--%snoisySkippings=false", prefix))
+	}
+
+	if reporter.Verbose {
+		result = append(result, fmt.Sprintf("--%sv", prefix))
+	}
+
+	if reporter.Succinct {
+		result = append(result, fmt.Sprintf("--%ssuccinct", prefix))
+	}
+
+	if reporter.FullTrace {
+		result = append(result, fmt.Sprintf("--%strace", prefix))
+	}
+
+	if reporter.ReportPassed {
+		result = append(result, fmt.Sprintf("--%sreportPassed", prefix))
+	}
+
+	if reporter.ReportFile != "" {
+		result = append(result, fmt.Sprintf("--%sreportFile=%s", prefix, reporter.ReportFile))
+	}
+
+	return result
+}
+
+// flagFocus implements the -focus flag.
+func flagFocus(arg string) {
+	if arg != "" {
+		GinkgoConfig.FocusStrings = append(GinkgoConfig.FocusStrings, arg)
+	}
+}
+
+// flagSkip implements the -skip flag.
+func flagSkip(arg string) {
+	if arg != "" {
+		GinkgoConfig.SkipStrings = append(GinkgoConfig.SkipStrings, arg)
+	}
+}
@@ -0,0 +1,190 @@
+package formatter
+
+import (
+	"fmt"
+	"regexp"
+	"strings"
+)
+
+const COLS = 80
+
+type ColorMode uint8
+
+const (
+	ColorModeNone ColorMode = iota
+	ColorModeTerminal
+	ColorModePassthrough
+)
+
+var SingletonFormatter = New(ColorModeTerminal)
+
+func F(format string, args ...interface{}) string {
+	return SingletonFormatter.F(format, args...)
+}
+
+func Fi(indentation uint, format string, args ...interface{}) string {
+	return SingletonFormatter.Fi(indentation, format, args...)
+}
+
+func Fiw(indentation uint, maxWidth uint, format string, args ...interface{}) string {
+	return SingletonFormatter.Fiw(indentation, maxWidth, format, args...)
+}
+
+type Formatter struct {
+	ColorMode                ColorMode
+	colors                   map[string]string
+	styleRe                  *regexp.Regexp
+	preserveColorStylingTags bool
+}
+
+func NewWithNoColorBool(noColor bool) Formatter {
+	if noColor {
+		return New(ColorModeNone)
+	}
+	return New(ColorModeTerminal)
+}
+
+func New(colorMode ColorMode) Formatter {
+	f := Formatter{
+		ColorMode: colorMode,
+		colors: map[string]string{
+			"/":         "\x1b[0m",
+			"bold":      "\x1b[1m",
+			"underline": "\x1b[4m",
+
+			"red":          "\x1b[38;5;9m",
+			"orange":       "\x1b[38;5;214m",
+			"coral":        "\x1b[38;5;204m",
+			"magenta":      "\x1b[38;5;13m",
+			"green":        "\x1b[38;5;10m",
+			"dark-green":   "\x1b[38;5;28m",
+			"yellow":       "\x1b[38;5;11m",
+			"light-yellow": "\x1b[38;5;228m",
+			"cyan":         "\x1b[38;5;14m",
+			"gray":         "\x1b[38;5;243m",
+			"light-gray":   "\x1b[38;5;246m",
+			"blue":         "\x1b[38;5;12m",
+		},
+	}
+	colors := []string{}
+	for color := range f.colors {
+		colors = append(colors, color)
+	}
+	f.styleRe = regexp.MustCompile("{{(" + strings.Join(colors, "|") + ")}}")
+	return f
+}
+
+func (f Formatter) F(format string, args ...interface{}) string {
+	return f.Fi(0, format, args...)
+}
+
+func (f Formatter) Fi(indentation uint, format string, args ...interface{}) string {
+	return f.Fiw(indentation, 0, format, args...)
+}
+
+func (f Formatter) Fiw(indentation uint, maxWidth uint, format string, args ...interface{}) string {
+	out := fmt.Sprintf(f.style(format), args...)
+
+	if indentation == 0 && maxWidth == 0 {
+		return out
+	}
+
+	lines := strings.Split(out, "\n")
+
+	if maxWidth != 0 {
+		outLines := []string{}
+
+		maxWidth = maxWidth - indentation*2
+		for _, line := range lines {
+			if f.length(line) <= maxWidth {
+				outLines = append(outLines, line)
+				continue
+			}
+			outWords := []string{}
+			length := uint(0)
+			words := strings.Split(line, " ")
+			for _, word := range words {
+				wordLength := f.length(word)
+				if length+wordLength <= maxWidth {
+					length += wordLength
+					outWords = append(outWords, word)
+					continue
+				}
+				outLines = append(outLines, strings.Join(outWords, " "))
+				outWords = []string{word}
+				length = wordLength
+			}
+			if len(outWords) > 0 {
+				outLines = append(outLines, strings.Join(outWords, " "))
+			}
+		}
+
+		lines = outLines
+	}
+
+	if indentation == 0 {
+		return strings.Join(lines, "\n")
+	}
+
+	padding := strings.Repeat("  ", int(indentation))
+	for i := range lines {
+		if lines[i] != "" {
+			lines[i] = padding + lines[i]
+		}
+	}
+
+	return strings.Join(lines, "\n")
+}
+
+func (f Formatter) length(styled string) uint {
+	n := uint(0)
+	inStyle := false
+	for _, b := range styled {
+		if inStyle {
+			if b == 'm' {
+				inStyle = false
+			}
+			continue
+		}
+		if b == '\x1b' {
+			inStyle = true
+			continue
+		}
+		n += 1
+	}
+	return n
+}
+
+func (f Formatter) CycleJoin(elements []string, joiner string, cycle []string) string {
+	if len(elements) == 0 {
+		return ""
+	}
+	n := len(cycle)
+	out := ""
+	for i, text := range elements {
+		out += cycle[i%n] + text
+		if i < len(elements)-1 {
+			out += joiner
+		}
+	}
+	out += "{{/}}"
+	return f.style(out)
+}
+
+func (f Formatter) style(s string) string {
+	switch f.ColorMode {
+	case ColorModeNone:
+		return f.styleRe.ReplaceAllString(s, "")
+	case ColorModePassthrough:
+		return s
+	case ColorModeTerminal:
+		return f.styleRe.ReplaceAllStringFunc(s, func(match string) string {
+			if out, ok := f.colors[strings.Trim(match, "{}")]; ok {
+				return out
+			}
+			return match
+		})
+	}
+
+	return ""
+}
@@ -0,0 +1,681 @@
+/*
+Ginkgo is a BDD-style testing framework for Golang
+
+The godoc documentation describes Ginkgo's API.  More comprehensive documentation (with examples!) is available at http://onsi.github.io/ginkgo/
+
+Ginkgo's preferred matcher library is [Gomega](http://github.com/onsi/gomega)
+
+Ginkgo on Github: http://github.com/onsi/ginkgo
+
+Ginkgo is MIT-Licensed
+*/
+package ginkgo
+
+import (
+	"flag"
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+	"reflect"
+	"strings"
+	"time"
+
+	"github.com/onsi/ginkgo/config"
+	"github.com/onsi/ginkgo/internal/codelocation"
+	"github.com/onsi/ginkgo/internal/global"
+	"github.com/onsi/ginkgo/internal/remote"
+	"github.com/onsi/ginkgo/internal/testingtproxy"
+	"github.com/onsi/ginkgo/internal/writer"
+	"github.com/onsi/ginkgo/reporters"
+	"github.com/onsi/ginkgo/reporters/stenographer"
+	colorable "github.com/onsi/ginkgo/reporters/stenographer/support/go-colorable"
+	"github.com/onsi/ginkgo/types"
+)
+
+var deprecationTracker = types.NewDeprecationTracker()
+
+const GINKGO_VERSION = config.VERSION
+const GINKGO_PANIC = `
+Your test failed.
+Ginkgo panics to prevent subsequent assertions from running.
+Normally Ginkgo rescues this panic so you shouldn't see it.
+
+But, if you make an assertion in a goroutine, Ginkgo can't capture the panic.
+To circumvent this, you should call
+
+	defer GinkgoRecover()
+
+at the top of the goroutine that caused this panic.
+`
+
+func init() {
+	config.Flags(flag.CommandLine, "ginkgo", true)
+	GinkgoWriter = writer.New(os.Stdout)
+}
+
+//GinkgoWriter implements an io.Writer
+//When running in verbose mode any writes to GinkgoWriter will be immediately printed
+//to stdout.  Otherwise, GinkgoWriter will buffer any writes produced during the current test and flush them to screen
+//only if the current test fails.
+var GinkgoWriter io.Writer
+
+//The interface by which Ginkgo receives *testing.T
+type GinkgoTestingT interface {
+	Fail()
+}
+
+//GinkgoRandomSeed returns the seed used to randomize spec execution order.  It is
+//useful for seeding your own pseudorandom number generators (PRNGs) to ensure
+//consistent executions from run to run, where your tests contain variability (for
+//example, when selecting random test data).
+func GinkgoRandomSeed() int64 {
+	return config.GinkgoConfig.RandomSeed
+}
+
+//GinkgoParallelNode is deprecated, use GinkgoParallelProcess instead
+func GinkgoParallelNode() int {
+	deprecationTracker.TrackDeprecation(types.Deprecations.ParallelNode(), codelocation.New(1))
+	return GinkgoParallelProcess()
+}
+
+//GinkgoParallelProcess returns the parallel process number for the current ginkgo process
+//The process number is 1-indexed
+func GinkgoParallelProcess() int {
+	return config.GinkgoConfig.ParallelNode
+}
+
+//Some matcher libraries or legacy codebases require a *testing.T
+//GinkgoT implements an interface analogous to *testing.T and can be used if
+//the library in question accepts *testing.T through an interface
+//
+// For example, with testify:
+// assert.Equal(GinkgoT(), 123, 123, "they should be equal")
+//
+// Or with gomock:
+// gomock.NewController(GinkgoT())
+//
+// GinkgoT() takes an optional offset argument that can be used to get the
+// correct line number associated with the failure.
+func GinkgoT(optionalOffset ...int) GinkgoTInterface {
+	offset := 3
+	if len(optionalOffset) > 0 {
+		offset = optionalOffset[0]
+	}
+	failedFunc := func() bool {
+		return CurrentGinkgoTestDescription().Failed
+	}
+	nameFunc := func() string {
+		return CurrentGinkgoTestDescription().FullTestText
+	}
+	return testingtproxy.New(GinkgoWriter, Fail, Skip, failedFunc, nameFunc, offset)
+}
+
+//The interface returned by GinkgoT().  This covers most of the methods
+//in the testing package's T.
+type GinkgoTInterface interface {
+	Cleanup(func())
+	Setenv(key, value string)
+	Error(args ...interface{})
+	Errorf(format string, args ...interface{})
+	Fail()
+	FailNow()
+	Failed() bool
+	Fatal(args ...interface{})
+	Fatalf(format string, args ...interface{})
+	Helper()
+	Log(args ...interface{})
+	Logf(format string, args ...interface{})
+	Name() string
+	Parallel()
+	Skip(args ...interface{})
+	SkipNow()
+	Skipf(format string, args ...interface{})
+	Skipped() bool
+	TempDir() string
+}
+
+//Custom Ginkgo test reporters must implement the Reporter interface.
+//
+//The custom reporter is passed in a SuiteSummary when the suite begins and ends,
+//and a SpecSummary just before a spec begins and just after a spec ends
+type Reporter reporters.Reporter
+
+//Asynchronous specs are given a channel of the Done type.  You must close or write to the channel
+//to tell Ginkgo that your async test is done.
+type Done chan<- interface{}
+
+//GinkgoTestDescription represents the information about the current running test returned by CurrentGinkgoTestDescription
+//	FullTestText: a concatenation of ComponentTexts and the TestText
+//	ComponentTexts: a list of all texts for the Describes & Contexts leading up to the current test
+//	TestText: the text in the actual It or Measure node
+//	IsMeasurement: true if the current test is a measurement
+//	FileName: the name of the file containing the current test
+//	LineNumber: the line number for the current test
+//	Failed: if the current test has failed, this will be true (useful in an AfterEach)
+type GinkgoTestDescription struct {
+	FullTestText   string
+	ComponentTexts []string
+	TestText       string
+
+	IsMeasurement bool
+
+	FileName   string
+	LineNumber int
+
+	Failed   bool
+	Duration time.Duration
+}
+
+//CurrentGinkgoTestDescripton returns information about the current running test.
+func CurrentGinkgoTestDescription() GinkgoTestDescription {
+	summary, ok := global.Suite.CurrentRunningSpecSummary()
+	if !ok {
+		return GinkgoTestDescription{}
+	}
+
+	subjectCodeLocation := summary.ComponentCodeLocations[len(summary.ComponentCodeLocations)-1]
+
+	return GinkgoTestDescription{
+		ComponentTexts: summary.ComponentTexts[1:],
+		FullTestText:   strings.Join(summary.ComponentTexts[1:], " "),
+		TestText:       summary.ComponentTexts[len(summary.ComponentTexts)-1],
+		IsMeasurement:  summary.IsMeasurement,
+		FileName:       subjectCodeLocation.FileName,
+		LineNumber:     subjectCodeLocation.LineNumber,
+		Failed:         summary.HasFailureState(),
+		Duration:       summary.RunTime,
+	}
+}
+
+//Measurement tests receive a Benchmarker.
+//
+//You use the Time() function to time how long the passed in body function takes to run
+//You use the RecordValue() function to track arbitrary numerical measurements.
+//The RecordValueWithPrecision() function can be used alternatively to provide the unit
+//and resolution of the numeric measurement.
+//The optional info argument is passed to the test reporter and can be used to
+// provide the measurement data to a custom reporter with context.
+//
+//See http://onsi.github.io/ginkgo/#benchmark_tests for more details
+type Benchmarker interface {
+	Time(name string, body func(), info ...interface{}) (elapsedTime time.Duration)
+	RecordValue(name string, value float64, info ...interface{})
+	RecordValueWithPrecision(name string, value float64, units string, precision int, info ...interface{})
+}
+
+//RunSpecs is the entry point for the Ginkgo test runner.
+//You must call this within a Golang testing TestX(t *testing.T) function.
+//
+//To bootstrap a test suite you can use the Ginkgo CLI:
+//
+//	ginkgo bootstrap
+func RunSpecs(t GinkgoTestingT, description string) bool {
+	specReporters := []Reporter{buildDefaultReporter()}
+	if config.DefaultReporterConfig.ReportFile != "" {
+		reportFile := config.DefaultReporterConfig.ReportFile
+		specReporters[0] = reporters.NewJUnitReporter(reportFile)
+		specReporters = append(specReporters, buildDefaultReporter())
+	}
+	return runSpecsWithCustomReporters(t, description, specReporters)
+}
+
+//To run your tests with Ginkgo's default reporter and your custom reporter(s), replace
+//RunSpecs() with this method.
+func RunSpecsWithDefaultAndCustomReporters(t GinkgoTestingT, description string, specReporters []Reporter) bool {
+	deprecationTracker.TrackDeprecation(types.Deprecations.CustomReporter())
+	specReporters = append(specReporters, buildDefaultReporter())
+	return runSpecsWithCustomReporters(t, description, specReporters)
+}
+
+//To run your tests with your custom reporter(s) (and *not* Ginkgo's default reporter), replace
+//RunSpecs() with this method.  Note that parallel tests will not work correctly without the default reporter
+func RunSpecsWithCustomReporters(t GinkgoTestingT, description string, specReporters []Reporter) bool {
+	deprecationTracker.TrackDeprecation(types.Deprecations.CustomReporter())
+	return runSpecsWithCustomReporters(t, description, specReporters)
+}
+
+func runSpecsWithCustomReporters(t GinkgoTestingT, description string, specReporters []Reporter) bool {
+	writer := GinkgoWriter.(*writer.Writer)
+	writer.SetStream(config.DefaultReporterConfig.Verbose)
+	reporters := make([]reporters.Reporter, len(specReporters))
+	for i, reporter := range specReporters {
+		reporters[i] = reporter
+	}
+	passed, hasFocusedTests := global.Suite.Run(t, description, reporters, writer, config.GinkgoConfig)
+
+	if deprecationTracker.DidTrackDeprecations() {
+		fmt.Fprintln(colorable.NewColorableStderr(), deprecationTracker.DeprecationsReport())
+	}
+
+	if passed && hasFocusedTests && strings.TrimSpace(os.Getenv("GINKGO_EDITOR_INTEGRATION")) == "" {
+		fmt.Println("PASS | FOCUSED")
+		os.Exit(types.GINKGO_FOCUS_EXIT_CODE)
+	}
+	return passed
+}
+
+func buildDefaultReporter() Reporter {
+	remoteReportingServer := config.GinkgoConfig.StreamHost
+	if remoteReportingServer == "" {
+		stenographer := stenographer.New(!config.DefaultReporterConfig.NoColor, config.GinkgoConfig.FlakeAttempts > 1, colorable.NewColorableStdout())
+		return reporters.NewDefaultReporter(config.DefaultReporterConfig, stenographer)
+	} else {
+		debugFile := ""
+		if config.GinkgoConfig.DebugParallel {
+			debugFile = fmt.Sprintf("ginkgo-node-%d.log", config.GinkgoConfig.ParallelNode)
+		}
+		return remote.NewForwardingReporter(config.DefaultReporterConfig, remoteReportingServer, &http.Client{}, remote.NewOutputInterceptor(), GinkgoWriter.(*writer.Writer), debugFile)
+	}
+}
+
+//Skip notifies Ginkgo that the current spec was skipped.
+func Skip(message string, callerSkip ...int) {
+	skip := 0
+	if len(callerSkip) > 0 {
+		skip = callerSkip[0]
+	}
+
+	global.Failer.Skip(message, codelocation.New(skip+1))
+	panic(GINKGO_PANIC)
+}
+
+//Fail notifies Ginkgo that the current spec has failed. (Gomega will call Fail for you automatically when an assertion fails.)
+func Fail(message string, callerSkip ...int) {
+	skip := 0
+	if len(callerSkip) > 0 {
+		skip = callerSkip[0]
+	}
+
+	global.Failer.Fail(message, codelocation.New(skip+1))
+	panic(GINKGO_PANIC)
+}
+
+//GinkgoRecover should be deferred at the top of any spawned goroutine that (may) call `Fail`
+//Since Gomega assertions call fail, you should throw a `defer GinkgoRecover()` at the top of any goroutine that
+//calls out to Gomega
+//
+//Here's why: Ginkgo's `Fail` method records the failure and then panics to prevent
+//further assertions from running.  This panic must be recovered.  Ginkgo does this for you
+//if the panic originates in a Ginkgo node (an It, BeforeEach, etc...)
+//
+//Unfortunately, if a panic originates on a goroutine *launched* from one of these nodes there's no
+//way for Ginkgo to rescue the panic.  To do this, you must remember to `defer GinkgoRecover()` at the top of such a goroutine.
+func GinkgoRecover() {
+	e := recover()
+	if e != nil {
+		global.Failer.Panic(codelocation.New(1), e)
+	}
+}
+
+//Describe blocks allow you to organize your specs.  A Describe block can contain any number of
+//BeforeEach, AfterEach, JustBeforeEach, It, and Measurement blocks.
+//
+//In addition you can nest Describe, Context and When blocks.  Describe, Context and When blocks are functionally
+//equivalent.  The difference is purely semantic -- you typically Describe the behavior of an object
+//or method and, within that Describe, outline a number of Contexts and Whens.
+func Describe(text string, body func()) bool {
+	global.Suite.PushContainerNode(text, body, types.FlagTypeNone, codelocation.New(1))
+	return true
+}
+
+//You can focus the tests within a describe block using FDescribe
+func FDescribe(text string, body func()) bool {
+	global.Suite.PushContainerNode(text, body, types.FlagTypeFocused, codelocation.New(1))
+	return true
+}
+
+//You can mark the tests within a describe block as pending using PDescribe
+func PDescribe(text string, body func()) bool {
+	global.Suite.PushContainerNode(text, body, types.FlagTypePending, codelocation.New(1))
+	return true
+}
+
+//You can mark the tests within a describe block as pending using XDescribe
+func XDescribe(text string, body func()) bool {
+	global.Suite.PushContainerNode(text, body, types.FlagTypePending, codelocation.New(1))
+	return true
+}
+
+//Context blocks allow you to organize your specs.  A Context block can contain any number of
+//BeforeEach, AfterEach, JustBeforeEach, It, and Measurement blocks.
+//
+//In addition you can nest Describe, Context and When blocks.  Describe, Context and When blocks are functionally
+//equivalent.  The difference is purely semantic -- you typical Describe the behavior of an object
+//or method and, within that Describe, outline a number of Contexts and Whens.
+func Context(text string, body func()) bool {
+	global.Suite.PushContainerNode(text, body, types.FlagTypeNone, codelocation.New(1))
+	return true
+}
+
+//You can focus the tests within a describe block using FContext
+func FContext(text string, body func()) bool {
+	global.Suite.PushContainerNode(text, body, types.FlagTypeFocused, codelocation.New(1))
+	return true
+}
+
+//You can mark the tests within a describe block as pending using PContext
+func PContext(text string, body func()) bool {
+	global.Suite.PushContainerNode(text, body, types.FlagTypePending, codelocation.New(1))
+	return true
+}
+
+//You can mark the tests within a describe block as pending using XContext
+func XContext(text string, body func()) bool {
+	global.Suite.PushContainerNode(text, body, types.FlagTypePending, codelocation.New(1))
+	return true
+}
+
+//When blocks allow you to organize your specs.  A When block can contain any number of
+//BeforeEach, AfterEach, JustBeforeEach, It, and Measurement blocks.
+//
+//In addition you can nest Describe, Context and When blocks.  Describe, Context and When blocks are functionally
+//equivalent.  The difference is purely semantic -- you typical Describe the behavior of an object
+//or method and, within that Describe, outline a number of Contexts and Whens.
+func When(text string, body func()) bool {
+	global.Suite.PushContainerNode("when "+text, body, types.FlagTypeNone, codelocation.New(1))
+	return true
+}
+
+//You can focus the tests within a describe block using FWhen
+func FWhen(text string, body func()) bool {
+	global.Suite.PushContainerNode("when "+text, body, types.FlagTypeFocused, codelocation.New(1))
+	return true
+}
+
+//You can mark the tests within a describe block as pending using PWhen
+func PWhen(text string, body func()) bool {
+	global.Suite.PushContainerNode("when "+text, body, types.FlagTypePending, codelocation.New(1))
+	return true
+}
+
+//You can mark the tests within a describe block as pending using XWhen
+func XWhen(text string, body func()) bool {
+	global.Suite.PushContainerNode("when "+text, body, types.FlagTypePending, codelocation.New(1))
+	return true
+}
+
+//It blocks contain your test code and assertions.  You cannot nest any other Ginkgo blocks
+//within an It block.
+//
+//Ginkgo will normally run It blocks synchronously.  To perform asynchronous tests, pass a
+//function that accepts a Done channel.  When you do this, you can also provide an optional timeout.
+func It(text string, body interface{}, timeout ...float64) bool {
+	validateBodyFunc(body, codelocation.New(1))
+	global.Suite.PushItNode(text, body, types.FlagTypeNone, codelocation.New(1), parseTimeout(timeout...))
+	return true
+}
+
+//You can focus individual Its using FIt
+func FIt(text string, body interface{}, timeout ...float64) bool {
+	validateBodyFunc(body, codelocation.New(1))
+	global.Suite.PushItNode(text, body, types.FlagTypeFocused, codelocation.New(1), parseTimeout(timeout...))
+	return true
+}
+
+//You can mark Its as pending using PIt
+func PIt(text string, _ ...interface{}) bool {
+	global.Suite.PushItNode(text, func() {}, types.FlagTypePending, codelocation.New(1), 0)
+	return true
+}
+
+//You can mark Its as pending using XIt
+func XIt(text string, _ ...interface{}) bool {
+	global.Suite.PushItNode(text, func() {}, types.FlagTypePending, codelocation.New(1), 0)
+	return true
+}
+
+//Specify blocks are aliases for It blocks and allow for more natural wording in situations
+//which "It" does not fit into a natural sentence flow. All the same protocols apply for Specify blocks
+//which apply to It blocks.
+func Specify(text string, body interface{}, timeout ...float64) bool {
+	validateBodyFunc(body, codelocation.New(1))
+	global.Suite.PushItNode(text, body, types.FlagTypeNone, codelocation.New(1), parseTimeout(timeout...))
+	return true
+}
+
+//You can focus individual Specifys using FSpecify
+func FSpecify(text string, body interface{}, timeout ...float64) bool {
+	validateBodyFunc(body, codelocation.New(1))
+	global.Suite.PushItNode(text, body, types.FlagTypeFocused, codelocation.New(1), parseTimeout(timeout...))
+	return true
+}
+
+//You can mark Specifys as pending using PSpecify
+func PSpecify(text string, is ...interface{}) bool {
+	global.Suite.PushItNode(text, func() {}, types.FlagTypePending, codelocation.New(1), 0)
+	return true
+}
+
+//You can mark Specifys as pending using XSpecify
+func XSpecify(text string, is ...interface{}) bool {
+	global.Suite.PushItNode(text, func() {}, types.FlagTypePending, codelocation.New(1), 0)
+	return true
+}
+
+//By allows you to better document large Its.
+//
+//Generally you should try to keep your Its short and to the point.  This is not always possible, however,
+//especially in the context of integration tests that capture a particular workflow.
+//
+//By allows you to document such flows.  By must be called within a runnable node (It, BeforeEach, Measure, etc...)
+//By will simply log the passed in text to the GinkgoWriter.  If By is handed a function it will immediately run the function.
+func By(text string, callbacks ...func()) {
+	preamble := "\x1b[1mSTEP\x1b[0m"
+	if config.DefaultReporterConfig.NoColor {
+		preamble = "STEP"
+	}
+	fmt.Fprintln(GinkgoWriter, preamble+": "+text)
+	if len(callbacks) == 1 {
+		callbacks[0]()
+	}
+	if len(callbacks) > 1 {
+		panic("just one callback per By, please")
+	}
+}
+
+//Measure blocks run the passed in body function repeatedly (determined by the samples argument)
+//and accumulate metrics provided to the Benchmarker by the body function.
+//
+//The body function must have the signature:
+//	func(b Benchmarker)
+func Measure(text string, body interface{}, samples int) bool {
+	deprecationTracker.TrackDeprecation(types.Deprecations.Measure(), codelocation.New(1))
+	global.Suite.PushMeasureNode(text, body, types.FlagTypeNone, codelocation.New(1), samples)
+	return true
+}
+
+//You can focus individual Measures using FMeasure
+func FMeasure(text string, body interface{}, samples int) bool {
+	deprecationTracker.TrackDeprecation(types.Deprecations.Measure(), codelocation.New(1))
+	global.Suite.PushMeasureNode(text, body, types.FlagTypeFocused, codelocation.New(1), samples)
+	return true
+}
+
+//You can mark Measurements as pending using PMeasure
+func PMeasure(text string, _ ...interface{}) bool {
+	deprecationTracker.TrackDeprecation(types.Deprecations.Measure(), codelocation.New(1))
+	global.Suite.PushMeasureNode(text, func(b Benchmarker) {}, types.FlagTypePending, codelocation.New(1), 0)
+	return true
+}
+
+//You can mark Measurements as pending using XMeasure
+func XMeasure(text string, _ ...interface{}) bool {
+	deprecationTracker.TrackDeprecation(types.Deprecations.Measure(), codelocation.New(1))
+	global.Suite.PushMeasureNode(text, func(b Benchmarker) {}, types.FlagTypePending, codelocation.New(1), 0)
+	return true
+}
+
+//BeforeSuite blocks are run just once before any specs are run.  When running in parallel, each
+//parallel node process will call BeforeSuite.
+//
+//BeforeSuite blocks can be made asynchronous by providing a body function that accepts a Done channel
+//
+//You may only register *one* BeforeSuite handler per test suite.  You typically do so in your bootstrap file at the top level.
+func BeforeSuite(body interface{}, timeout ...float64) bool {
+	validateBodyFunc(body, codelocation.New(1))
+	global.Suite.SetBeforeSuiteNode(body, codelocation.New(1), parseTimeout(timeout...))
+	return true
+}
+
+//AfterSuite blocks are *always* run after all the specs regardless of whether specs have passed or failed.
+//Moreover, if Ginkgo receives an interrupt signal (^C) it will attempt to run the AfterSuite before exiting.
+//
+//When running in parallel, each parallel node process will call AfterSuite.
+//
+//AfterSuite blocks can be made asynchronous by providing a body function that accepts a Done channel
+//
+//You may only register *one* AfterSuite handler per test suite.  You typically do so in your bootstrap file at the top level.
+func AfterSuite(body interface{}, timeout ...float64) bool {
+	validateBodyFunc(body, codelocation.New(1))
+	global.Suite.SetAfterSuiteNode(body, codelocation.New(1), parseTimeout(timeout...))
+	return true
+}
+
+//SynchronizedBeforeSuite blocks are primarily meant to solve the problem of setting up singleton external resources shared across
+//nodes when running tests in parallel.  For example, say you have a shared database that you can only start one instance of that
+//must be used in your tests.  When running in parallel, only one node should set up the database and all other nodes should wait
+//until that node is done before running.
+//
+//SynchronizedBeforeSuite accomplishes this by taking *two* function arguments.  The first is only run on parallel node #1.  The second is
+//run on all nodes, but *only* after the first function completes successfully.  Ginkgo also makes it possible to send data from the first function (on Node 1)
+//to the second function (on all the other nodes).
+//
+//The functions have the following signatures.  The first function (which only runs on node 1) has the signature:
+//
+//	func() []byte
+//
+//or, to run asynchronously:
+//
+//	func(done Done) []byte
+//
+//The byte array returned by the first function is then passed to the second function, which has the signature:
+//
+//	func(data []byte)
+//
+//or, to run asynchronously:
+//
+//	func(data []byte, done Done)
+//
+//Here's a simple pseudo-code example that starts a shared database on Node 1 and shares the database's address with the other nodes:
+//
+//	var dbClient db.Client
+//	var dbRunner db.Runner
+//
+//	var _ = SynchronizedBeforeSuite(func() []byte {
+//		dbRunner = db.NewRunner()
+//		err := dbRunner.Start()
+//		Ω(err).ShouldNot(HaveOccurred())
+//		return []byte(dbRunner.URL)
+//	}, func(data []byte) {
+//		dbClient = db.NewClient()
+//		err := dbClient.Connect(string(data))
+//		Ω(err).ShouldNot(HaveOccurred())
+//	})
+func SynchronizedBeforeSuite(node1Body interface{}, allNodesBody interface{}, timeout ...float64) bool {
+	global.Suite.SetSynchronizedBeforeSuiteNode(
+		node1Body,
+		allNodesBody,
+		codelocation.New(1),
+		parseTimeout(timeout...),
+	)
+	return true
+}
+
+//SynchronizedAfterSuite blocks complement the SynchronizedBeforeSuite blocks in solving the problem of setting up
+//external singleton resources shared across nodes when running tests in parallel.
+//
+//SynchronizedAfterSuite accomplishes this by taking *two* function arguments.  The first runs on all nodes.  The second runs only on parallel node #1
+//and *only* after all other nodes have finished and exited.  This ensures that node 1, and any resources it is running, remain alive until
+//all other nodes are finished.
+//
+//Both functions have the same signature: either func() or func(done Done) to run asynchronously.
+//
+//Here's a pseudo-code example that complements that given in SynchronizedBeforeSuite.  Here, SynchronizedAfterSuite is used to tear down the shared database
+//only after all nodes have finished:
+//
+//	var _ = SynchronizedAfterSuite(func() {
+//		dbClient.Cleanup()
+//	}, func() {
+//		dbRunner.Stop()
+//	})
+func SynchronizedAfterSuite(allNodesBody interface{}, node1Body interface{}, timeout ...float64) bool {
+	global.Suite.SetSynchronizedAfterSuiteNode(
+		allNodesBody,
+		node1Body,
+		codelocation.New(1),
+		parseTimeout(timeout...),
+	)
+	return true
+}
+
+//BeforeEach blocks are run before It blocks.  When multiple BeforeEach blocks are defined in nested
+//Describe and Context blocks the outermost BeforeEach blocks are run first.
+//
+//Like It blocks, BeforeEach blocks can be made asynchronous by providing a body function that accepts
+//a Done channel
+func BeforeEach(body interface{}, timeout ...float64) bool {
+	validateBodyFunc(body, codelocation.New(1))
+	global.Suite.PushBeforeEachNode(body, codelocation.New(1), parseTimeout(timeout...))
+	return true
+}
+
+//JustBeforeEach blocks are run before It blocks but *after* all BeforeEach blocks.  For more details,
+//read the [documentation](http://onsi.github.io/ginkgo/#separating_creation_and_configuration_)
+//
+//Like It blocks, BeforeEach blocks can be made asynchronous by providing a body function that accepts
+//a Done channel
+func JustBeforeEach(body interface{}, timeout ...float64) bool {
+	validateBodyFunc(body, codelocation.New(1))
+	global.Suite.PushJustBeforeEachNode(body, codelocation.New(1), parseTimeout(timeout...))
+	return true
+}
+
+//JustAfterEach blocks are run after It blocks but *before* all AfterEach blocks.  For more details,
+//read the [documentation](http://onsi.github.io/ginkgo/#separating_creation_and_configuration_)
+//
+//Like It blocks, JustAfterEach blocks can be made asynchronous by providing a body function that accepts
+//a Done channel
+func JustAfterEach(body interface{}, timeout ...float64) bool {
+	validateBodyFunc(body, codelocation.New(1))
+	global.Suite.PushJustAfterEachNode(body, codelocation.New(1), parseTimeout(timeout...))
+	return true
+}
+
+//AfterEach blocks are run after It blocks.   When multiple AfterEach blocks are defined in nested
+//Describe and Context blocks the innermost AfterEach blocks are run first.
+//
+//Like It blocks, AfterEach blocks can be made asynchronous by providing a body function that accepts
+//a Done channel
+func AfterEach(body interface{}, timeout ...float64) bool {
+	validateBodyFunc(body, codelocation.New(1))
+	global.Suite.PushAfterEachNode(body, codelocation.New(1), parseTimeout(timeout...))
+	return true
+}
+
+func validateBodyFunc(body interface{}, cl types.CodeLocation) {
+	t := reflect.TypeOf(body)
+	if t.Kind() != reflect.Func {
+		return
+	}
+
+	if t.NumOut() > 0 {
+		return
+	}
+
+	if t.NumIn() == 0 {
+		return
+	}
+
+	if t.In(0) == reflect.TypeOf(make(Done)) {
+		deprecationTracker.TrackDeprecation(types.Deprecations.Async(), cl)
+	}
+}
+
+func parseTimeout(timeout ...float64) time.Duration {
+	if len(timeout) == 0 {
+		return global.DefaultTimeout
+	} else {
+		return time.Duration(timeout[0] * float64(time.Second))
+	}
+}
@@ -0,0 +1,48 @@
+package codelocation
+
+import (
+	"regexp"
+	"runtime"
+	"runtime/debug"
+	"strings"
+
+	"github.com/onsi/ginkgo/types"
+)
+
+func New(skip int) types.CodeLocation {
+	_, file, line, _ := runtime.Caller(skip + 1)
+	stackTrace := PruneStack(string(debug.Stack()), skip+1)
+	return types.CodeLocation{FileName: file, LineNumber: line, FullStackTrace: stackTrace}
+}
+
+// PruneStack removes references to functions that are internal to Ginkgo
+// and the Go runtime from a stack string and a certain number of stack entries
+// at the beginning of the stack. The stack string has the format
+// as returned by runtime/debug.Stack. The leading goroutine information is
+// optional and always removed if present. Beware that runtime/debug.Stack
+// adds itself as first entry, so typically skip must be >= 1 to remove that
+// entry.
+func PruneStack(fullStackTrace string, skip int) string {
+	stack := strings.Split(fullStackTrace, "\n")
+	// Ensure that the even entries are the method names and the
+	// the odd entries the source code information.
+	if len(stack) > 0 && strings.HasPrefix(stack[0], "goroutine ") {
+		// Ignore "goroutine 29 [running]:" line.
+		stack = stack[1:]
+	}
+	// The "+1" is for skipping over the initial entry, which is
+	// runtime/debug.Stack() itself.
+	if len(stack) > 2*(skip+1) {
+		stack = stack[2*(skip+1):]
+	}
+	prunedStack := []string{}
+	re := regexp.MustCompile(`\/ginkgo\/|\/pkg\/testing\/|\/pkg\/runtime\/`)
+	for i := 0; i < len(stack)/2; i++ {
+		// We filter out based on the source code file name.
+		if !re.Match([]byte(stack[i*2+1])) {
+			prunedStack = append(prunedStack, stack[i*2])
+			prunedStack = append(prunedStack, stack[i*2+1])
+		}
+	}
+	return strings.Join(prunedStack, "\n")
+}
@@ -0,0 +1,151 @@
+package containernode
+
+import (
+	"math/rand"
+	"sort"
+
+	"github.com/onsi/ginkgo/internal/leafnodes"
+	"github.com/onsi/ginkgo/types"
+)
+
+type subjectOrContainerNode struct {
+	containerNode *ContainerNode
+	subjectNode   leafnodes.SubjectNode
+}
+
+func (n subjectOrContainerNode) text() string {
+	if n.containerNode != nil {
+		return n.containerNode.Text()
+	} else {
+		return n.subjectNode.Text()
+	}
+}
+
+type CollatedNodes struct {
+	Containers []*ContainerNode
+	Subject    leafnodes.SubjectNode
+}
+
+type ContainerNode struct {
+	text         string
+	flag         types.FlagType
+	codeLocation types.CodeLocation
+
+	setupNodes               []leafnodes.BasicNode
+	subjectAndContainerNodes []subjectOrContainerNode
+}
+
+func New(text string, flag types.FlagType, codeLocation types.CodeLocation) *ContainerNode {
+	return &ContainerNode{
+		text:         text,
+		flag:         flag,
+		codeLocation: codeLocation,
+	}
+}
+
+func (container *ContainerNode) Shuffle(r *rand.Rand) {
+	sort.Sort(container)
+	permutation := r.Perm(len(container.subjectAndContainerNodes))
+	shuffledNodes := make([]subjectOrContainerNode, len(container.subjectAndContainerNodes))
+	for i, j := range permutation {
+		shuffledNodes[i] = container.subjectAndContainerNodes[j]
+	}
+	container.subjectAndContainerNodes = shuffledNodes
+}
+
+func (node *ContainerNode) BackPropagateProgrammaticFocus() bool {
+	if node.flag == types.FlagTypePending {
+		return false
+	}
+
+	shouldUnfocus := false
+	for _, subjectOrContainerNode := range node.subjectAndContainerNodes {
+		if subjectOrContainerNode.containerNode != nil {
+			shouldUnfocus = subjectOrContainerNode.containerNode.BackPropagateProgrammaticFocus() || shouldUnfocus
+		} else {
+			shouldUnfocus = (subjectOrContainerNode.subjectNode.Flag() == types.FlagTypeFocused) || shouldUnfocus
+		}
+	}
+
+	if shouldUnfocus {
+		if node.flag == types.FlagTypeFocused {
+			node.flag = types.FlagTypeNone
+		}
+		return true
+	}
+
+	return node.flag == types.FlagTypeFocused
+}
+
+func (node *ContainerNode) Collate() []CollatedNodes {
+	return node.collate([]*ContainerNode{})
+}
+
+func (node *ContainerNode) collate(enclosingContainers []*ContainerNode) []CollatedNodes {
+	collated := make([]CollatedNodes, 0)
+
+	containers := make([]*ContainerNode, len(enclosingContainers))
+	copy(containers, enclosingContainers)
+	containers = append(containers, node)
+
+	for _, subjectOrContainer := range node.subjectAndContainerNodes {
+		if subjectOrContainer.containerNode != nil {
+			collated = append(collated, subjectOrContainer.containerNode.collate(containers)...)
+		} else {
+			collated = append(collated, CollatedNodes{
+				Containers: containers,
+				Subject:    subjectOrContainer.subjectNode,
+			})
+		}
+	}
+
+	return collated
+}
+
+func (node *ContainerNode) PushContainerNode(container *ContainerNode) {
+	node.subjectAndContainerNodes = append(node.subjectAndContainerNodes, subjectOrContainerNode{containerNode: container})
+}
+
+func (node *ContainerNode) PushSubjectNode(subject leafnodes.SubjectNode) {
+	node.subjectAndContainerNodes = append(node.subjectAndContainerNodes, subjectOrContainerNode{subjectNode: subject})
+}
+
+func (node *ContainerNode) PushSetupNode(setupNode leafnodes.BasicNode) {
+	node.setupNodes = append(node.setupNodes, setupNode)
+}
+
+func (node *ContainerNode) SetupNodesOfType(nodeType types.SpecComponentType) []leafnodes.BasicNode {
+	nodes := []leafnodes.BasicNode{}
+	for _, setupNode := range node.setupNodes {
+		if setupNode.Type() == nodeType {
+			nodes = append(nodes, setupNode)
+		}
+	}
+	return nodes
+}
+
+func (node *ContainerNode) Text() string {
+	return node.text
+}
+
+func (node *ContainerNode) CodeLocation() types.CodeLocation {
+	return node.codeLocation
+}
+
+func (node *ContainerNode) Flag() types.FlagType {
+	return node.flag
+}
+
+//sort.Interface
+
+func (node *ContainerNode) Len() int {
+	return len(node.subjectAndContainerNodes)
+}
+
+func (node *ContainerNode) Less(i, j int) bool {
+	return node.subjectAndContainerNodes[i].text() < node.subjectAndContainerNodes[j].text()
+}
+
+func (node *ContainerNode) Swap(i, j int) {
+	node.subjectAndContainerNodes[i], node.subjectAndContainerNodes[j] = node.subjectAndContainerNodes[j], node.subjectAndContainerNodes[i]
+}
@@ -0,0 +1,92 @@
+package failer
+
+import (
+	"fmt"
+	"sync"
+
+	"github.com/onsi/ginkgo/types"
+)
+
+type Failer struct {
+	lock    *sync.Mutex
+	failure types.SpecFailure
+	state   types.SpecState
+}
+
+func New() *Failer {
+	return &Failer{
+		lock:  &sync.Mutex{},
+		state: types.SpecStatePassed,
+	}
+}
+
+func (f *Failer) Panic(location types.CodeLocation, forwardedPanic interface{}) {
+	f.lock.Lock()
+	defer f.lock.Unlock()
+
+	if f.state == types.SpecStatePassed {
+		f.state = types.SpecStatePanicked
+		f.failure = types.SpecFailure{
+			Message:        "Test Panicked",
+			Location:       location,
+			ForwardedPanic: fmt.Sprintf("%v", forwardedPanic),
+		}
+	}
+}
+
+func (f *Failer) Timeout(location types.CodeLocation) {
+	f.lock.Lock()
+	defer f.lock.Unlock()
+
+	if f.state == types.SpecStatePassed {
+		f.state = types.SpecStateTimedOut
+		f.failure = types.SpecFailure{
+			Message:  "Timed out",
+			Location: location,
+		}
+	}
+}
+
+func (f *Failer) Fail(message string, location types.CodeLocation) {
+	f.lock.Lock()
+	defer f.lock.Unlock()
+
+	if f.state == types.SpecStatePassed {
+		f.state = types.SpecStateFailed
+		f.failure = types.SpecFailure{
+			Message:  message,
+			Location: location,
+		}
+	}
+}
+
+func (f *Failer) Drain(componentType types.SpecComponentType, componentIndex int, componentCodeLocation types.CodeLocation) (types.SpecFailure, types.SpecState) {
+	f.lock.Lock()
+	defer f.lock.Unlock()
+
+	failure := f.failure
+	outcome := f.state
+	if outcome != types.SpecStatePassed {
+		failure.ComponentType = componentType
+		failure.ComponentIndex = componentIndex
+		failure.ComponentCodeLocation = componentCodeLocation
+	}
+
+	f.state = types.SpecStatePassed
+	f.failure = types.SpecFailure{}
+
+	return failure, outcome
+}
+
+func (f *Failer) Skip(message string, location types.CodeLocation) {
+	f.lock.Lock()
+	defer f.lock.Unlock()
+
+	if f.state == types.SpecStatePassed {
+		f.state = types.SpecStateSkipped
+		f.failure = types.SpecFailure{
+			Message:  message,
+			Location: location,
+		}
+	}
+}
@@ -0,0 +1,22 @@
+package global
+
+import (
+	"time"
+
+	"github.com/onsi/ginkgo/internal/failer"
+	"github.com/onsi/ginkgo/internal/suite"
+)
+
+const DefaultTimeout = time.Duration(1 * time.Second)
+
+var Suite *suite.Suite
+var Failer *failer.Failer
+
+func init() {
+	InitializeGlobals()
+}
+
+func InitializeGlobals() {
+	Failer = failer.New()
+	Suite = suite.New(Failer)
+}
@@ -0,0 +1,103 @@
+package leafnodes
+
+import (
+	"math"
+	"time"
+
+	"sync"
+
+	"github.com/onsi/ginkgo/types"
+)
+
+type benchmarker struct {
+	mu           sync.Mutex
+	measurements map[string]*types.SpecMeasurement
+	orderCounter int
+}
+
+func newBenchmarker() *benchmarker {
+	return &benchmarker{
+		measurements: make(map[string]*types.SpecMeasurement),
+	}
+}
+
+func (b *benchmarker) Time(name string, body func(), info ...interface{}) (elapsedTime time.Duration) {
+	t := time.Now()
+	body()
+	elapsedTime = time.Since(t)
+
+	b.mu.Lock()
+	defer b.mu.Unlock()
+	measurement := b.getMeasurement(name, "Fastest Time", "Slowest Time", "Average Time", "s", 3, info...)
+	measurement.Results = append(measurement.Results, elapsedTime.Seconds())
+
+	return
+}
+
+func (b *benchmarker) RecordValue(name string, value float64, info ...interface{}) {
+	b.mu.Lock()
+	measurement := b.getMeasurement(name, "Smallest", " Largest", " Average", "", 3, info...)
+	defer b.mu.Unlock()
+	measurement.Results = append(measurement.Results, value)
+}
+
+func (b *benchmarker) RecordValueWithPrecision(name string, value float64, units string, precision int, info ...interface{}) {
+	b.mu.Lock()
+	measurement := b.getMeasurement(name, "Smallest", " Largest", " Average", units, precision, info...)
+	defer b.mu.Unlock()
+	measurement.Results = append(measurement.Results, value)
+}
+
+func (b *benchmarker) getMeasurement(name string, smallestLabel string, largestLabel string, averageLabel string, units string, precision int, info ...interface{}) *types.SpecMeasurement {
+	measurement, ok := b.measurements[name]
+	if !ok {
+		var computedInfo interface{}
+		computedInfo = nil
+		if len(info) > 0 {
+			computedInfo = info[0]
+		}
+		measurement = &types.SpecMeasurement{
+			Name:          name,
+			Info:          computedInfo,
+			Order:         b.orderCounter,
+			SmallestLabel: smallestLabel,
+			LargestLabel:  largestLabel,
+			AverageLabel:  averageLabel,
+			Units:         units,
+			Precision:     precision,
+			Results:       make([]float64, 0),
+		}
+		b.measurements[name] = measurement
+		b.orderCounter++
+	}
+
+	return measurement
+}
+
+func (b *benchmarker) measurementsReport() map[string]*types.SpecMeasurement {
+	b.mu.Lock()
+	defer b.mu.Unlock()
+	for _, measurement := range b.measurements {
+		measurement.Smallest = math.MaxFloat64
+		measurement.Largest = -math.MaxFloat64
+		sum := float64(0)
+		sumOfSquares := float64(0)
+
+		for _, result := range measurement.Results {
+			if result > measurement.Largest {
+				measurement.Largest = result
+			}
+			if result < measurement.Smallest {
+				measurement.Smallest = result
+			}
+			sum += result
+			sumOfSquares += result * result
+		}
+
+		n := float64(len(measurement.Results))
+		measurement.Average = sum / n
+		measurement.StdDeviation = math.Sqrt(sumOfSquares/n - (sum/n)*(sum/n))
+	}
+
+	return b.measurements
+}
@@ -0,0 +1,19 @@
+package leafnodes
+
+import (
+	"github.com/onsi/ginkgo/types"
+)
+
+type BasicNode interface {
+	Type() types.SpecComponentType
+	Run() (types.SpecState, types.SpecFailure)
+	CodeLocation() types.CodeLocation
+}
+
+type SubjectNode interface {
+	BasicNode
+
+	Text() string
+	Flag() types.FlagType
+	Samples() int
+}
@@ -0,0 +1,47 @@
+package leafnodes
+
+import (
+	"time"
+
+	"github.com/onsi/ginkgo/internal/failer"
+	"github.com/onsi/ginkgo/types"
+)
+
+type ItNode struct {
+	runner *runner
+
+	flag types.FlagType
+	text string
+}
+
+func NewItNode(text string, body interface{}, flag types.FlagType, codeLocation types.CodeLocation, timeout time.Duration, failer *failer.Failer, componentIndex int) *ItNode {
+	return &ItNode{
+		runner: newRunner(body, codeLocation, timeout, failer, types.SpecComponentTypeIt, componentIndex),
+		flag:   flag,
+		text:   text,
+	}
+}
+
+func (node *ItNode) Run() (outcome types.SpecState, failure types.SpecFailure) {
+	return node.runner.run()
+}
+
+func (node *ItNode) Type() types.SpecComponentType {
+	return types.SpecComponentTypeIt
+}
+
+func (node *ItNode) Text() string {
+	return node.text
+}
+
+func (node *ItNode) Flag() types.FlagType {
+	return node.flag
+}
+
+func (node *ItNode) CodeLocation() types.CodeLocation {
+	return node.runner.codeLocation
+}
+
+func (node *ItNode) Samples() int {
+	return 1
+}
@@ -0,0 +1,62 @@
+package leafnodes
+
+import (
+	"reflect"
+
+	"github.com/onsi/ginkgo/internal/failer"
+	"github.com/onsi/ginkgo/types"
+)
+
+type MeasureNode struct {
+	runner *runner
+
+	text        string
+	flag        types.FlagType
+	samples     int
+	benchmarker *benchmarker
+}
+
+func NewMeasureNode(text string, body interface{}, flag types.FlagType, codeLocation types.CodeLocation, samples int, failer *failer.Failer, componentIndex int) *MeasureNode {
+	benchmarker := newBenchmarker()
+
+	wrappedBody := func() {
+		reflect.ValueOf(body).Call([]reflect.Value{reflect.ValueOf(benchmarker)})
+	}
+
+	return &MeasureNode{
+		runner: newRunner(wrappedBody, codeLocation, 0, failer, types.SpecComponentTypeMeasure, componentIndex),
+
+		text:        text,
+		flag:        flag,
+		samples:     samples,
+		benchmarker: benchmarker,
+	}
+}
+
+func (node *MeasureNode) Run() (outcome types.SpecState, failure types.SpecFailure) {
+	return node.runner.run()
+}
+
+func (node *MeasureNode) MeasurementsReport() map[string]*types.SpecMeasurement {
+	return node.benchmarker.measurementsReport()
+}
+
+func (node *MeasureNode) Type() types.SpecComponentType {
+	return types.SpecComponentTypeMeasure
+}
+
+func (node *MeasureNode) Text() string {
+	return node.text
+}
+
+func (node *MeasureNode) Flag() types.FlagType {
+	return node.flag
+}
+
+func (node *MeasureNode) CodeLocation() types.CodeLocation {
+	return node.runner.codeLocation
+}
+
+func (node *MeasureNode) Samples() int {
+	return node.samples
+}
@@ -0,0 +1,117 @@
+package leafnodes
+
+import (
+	"fmt"
+	"reflect"
+	"time"
+
+	"github.com/onsi/ginkgo/internal/codelocation"
+	"github.com/onsi/ginkgo/internal/failer"
+	"github.com/onsi/ginkgo/types"
+)
+
+type runner struct {
+	isAsync          bool
+	asyncFunc        func(chan<- interface{})
+	syncFunc         func()
+	codeLocation     types.CodeLocation
+	timeoutThreshold time.Duration
+	nodeType         types.SpecComponentType
+	componentIndex   int
+	failer           *failer.Failer
+}
+
+func newRunner(body interface{}, codeLocation types.CodeLocation, timeout time.Duration, failer *failer.Failer, nodeType types.SpecComponentType, componentIndex int) *runner {
+	bodyType := reflect.TypeOf(body)
+	if bodyType.Kind() != reflect.Func {
+		panic(fmt.Sprintf("Expected a function but got something else at %v", codeLocation))
+	}
+
+	runner := &runner{
+		codeLocation:     codeLocation,
+		timeoutThreshold: timeout,
+		failer:           failer,
+		nodeType:         nodeType,
+		componentIndex:   componentIndex,
+	}
+
+	switch bodyType.NumIn() {
+	case 0:
+		runner.syncFunc = body.(func())
+		return runner
+	case 1:
+		if !(bodyType.In(0).Kind() == reflect.Chan && bodyType.In(0).Elem().Kind() == reflect.Interface) {
+			panic(fmt.Sprintf("Must pass a Done channel to function at %v", codeLocation))
+		}
+
+		wrappedBody := func(done chan<- interface{}) {
+			bodyValue := reflect.ValueOf(body)
+			bodyValue.Call([]reflect.Value{reflect.ValueOf(done)})
+		}
+
+		runner.isAsync = true
+		runner.asyncFunc = wrappedBody
+		return runner
+	}
+
+	panic(fmt.Sprintf("Too many arguments to function at %v", codeLocation))
+}
+
+func (r *runner) run() (outcome types.SpecState, failure types.SpecFailure) {
+	if r.isAsync {
+		return r.runAsync()
+	} else {
+		return r.runSync()
+	}
+}
+
+func (r *runner) runAsync() (outcome types.SpecState, failure types.SpecFailure) {
+	done := make(chan interface{}, 1)
+
+	go func() {
+		finished := false
+
+		defer func() {
+			if e := recover(); e != nil || !finished {
+				r.failer.Panic(codelocation.New(2), e)
+				select {
+				case <-done:
+					break
+				default:
+					close(done)
+				}
+			}
+		}()
+
+		r.asyncFunc(done)
+		finished = true
+	}()
+
+	// If this goroutine gets no CPU time before the select block,
+	// the <-done case may complete even if the test took longer than the timeoutThreshold.
+	// This can cause flaky behaviour, but we haven't seen it in the wild.
+	select {
+	case <-done:
+	case <-time.After(r.timeoutThreshold):
+		r.failer.Timeout(r.codeLocation)
+	}
+
+	failure, outcome = r.failer.Drain(r.nodeType, r.componentIndex, r.codeLocation)
+	return
+}
+func (r *runner) runSync() (outcome types.SpecState, failure types.SpecFailure) {
+	finished := false
+
+	defer func() {
+		if e := recover(); e != nil || !finished {
+			r.failer.Panic(codelocation.New(2), e)
+		}
+
+		failure, outcome = r.failer.Drain(r.nodeType, r.componentIndex, r.codeLocation)
+	}()
+
+	r.syncFunc()
+	finished = true
+
+	return
+}
@@ -0,0 +1,48 @@
+package leafnodes
+
+import (
+	"time"
+
+	"github.com/onsi/ginkgo/internal/failer"
+	"github.com/onsi/ginkgo/types"
+)
+
+type SetupNode struct {
+	runner *runner
+}
+
+func (node *SetupNode) Run() (outcome types.SpecState, failure types.SpecFailure) {
+	return node.runner.run()
+}
+
+func (node *SetupNode) Type() types.SpecComponentType {
+	return node.runner.nodeType
+}
+
+func (node *SetupNode) CodeLocation() types.CodeLocation {
+	return node.runner.codeLocation
+}
+
+func NewBeforeEachNode(body interface{}, codeLocation types.CodeLocation, timeout time.Duration, failer *failer.Failer, componentIndex int) *SetupNode {
+	return &SetupNode{
+		runner: newRunner(body, codeLocation, timeout, failer, types.SpecComponentTypeBeforeEach, componentIndex),
+	}
+}
+
+func NewAfterEachNode(body interface{}, codeLocation types.CodeLocation, timeout time.Duration, failer *failer.Failer, componentIndex int) *SetupNode {
+	return &SetupNode{
+		runner: newRunner(body, codeLocation, timeout, failer, types.SpecComponentTypeAfterEach, componentIndex),
+	}
+}
+
+func NewJustBeforeEachNode(body interface{}, codeLocation types.CodeLocation, timeout time.Duration, failer *failer.Failer, componentIndex int) *SetupNode {
+	return &SetupNode{
+		runner: newRunner(body, codeLocation, timeout, failer, types.SpecComponentTypeJustBeforeEach, componentIndex),
+	}
+}
+
+func NewJustAfterEachNode(body interface{}, codeLocation types.CodeLocation, timeout time.Duration, failer *failer.Failer, componentIndex int) *SetupNode {
+	return &SetupNode{
+		runner: newRunner(body, codeLocation, timeout, failer, types.SpecComponentTypeJustAfterEach, componentIndex),
+	}
+}
@@ -0,0 +1,55 @@
+package leafnodes
+
+import (
+	"time"
+
+	"github.com/onsi/ginkgo/internal/failer"
+	"github.com/onsi/ginkgo/types"
+)
+
+type SuiteNode interface {
+	Run(parallelNode int, parallelTotal int, syncHost string) bool
+	Passed() bool
+	Summary() *types.SetupSummary
+}
+
+type simpleSuiteNode struct {
+	runner  *runner
+	outcome types.SpecState
+	failure types.SpecFailure
+	runTime time.Duration
+}
+
+func (node *simpleSuiteNode) Run(parallelNode int, parallelTotal int, syncHost string) bool {
+	t := time.Now()
+	node.outcome, node.failure = node.runner.run()
+	node.runTime = time.Since(t)
+
+	return node.outcome == types.SpecStatePassed
+}
+
+func (node *simpleSuiteNode) Passed() bool {
+	return node.outcome == types.SpecStatePassed
+}
+
+func (node *simpleSuiteNode) Summary() *types.SetupSummary {
+	return &types.SetupSummary{
+		ComponentType: node.runner.nodeType,
+		CodeLocation:  node.runner.codeLocation,
+		State:         node.outcome,
+		RunTime:       node.runTime,
+		Failure:       node.failure,
+	}
+}
+
+func NewBeforeSuiteNode(body interface{}, codeLocation types.CodeLocation, timeout time.Duration, failer *failer.Failer) SuiteNode {
+	return &simpleSuiteNode{
+		runner: newRunner(body, codeLocation, timeout, failer, types.SpecComponentTypeBeforeSuite, 0),
+	}
+}
+
+func NewAfterSuiteNode(body interface{}, codeLocation types.CodeLocation, timeout time.Duration, failer *failer.Failer) SuiteNode {
+	return &simpleSuiteNode{
+		runner: newRunner(body, codeLocation, timeout, failer, types.SpecComponentTypeAfterSuite, 0),
+	}
+}
@@ -0,0 +1,90 @@
+package leafnodes
+
+import (
+	"encoding/json"
+	"io/ioutil"
+	"net/http"
+	"time"
+
+	"github.com/onsi/ginkgo/internal/failer"
+	"github.com/onsi/ginkgo/types"
+)
+
+type synchronizedAfterSuiteNode struct {
+	runnerA *runner
+	runnerB *runner
+
+	outcome types.SpecState
+	failure types.SpecFailure
+	runTime time.Duration
+}
+
+func NewSynchronizedAfterSuiteNode(bodyA interface{}, bodyB interface{}, codeLocation types.CodeLocation, timeout time.Duration, failer *failer.Failer) SuiteNode {
+	return &synchronizedAfterSuiteNode{
+		runnerA: newRunner(bodyA, codeLocation, timeout, failer, types.SpecComponentTypeAfterSuite, 0),
+		runnerB: newRunner(bodyB, codeLocation, timeout, failer, types.SpecComponentTypeAfterSuite, 0),
+	}
+}
+
+func (node *synchronizedAfterSuiteNode) Run(parallelNode int, parallelTotal int, syncHost string) bool {
+	node.outcome, node.failure = node.runnerA.run()
+
+	if parallelNode == 1 {
+		if parallelTotal > 1 {
+			node.waitUntilOtherNodesAreDone(syncHost)
+		}
+
+		outcome, failure := node.runnerB.run()
+
+		if node.outcome == types.SpecStatePassed {
+			node.outcome, node.failure = outcome, failure
+		}
+	}
+
+	return node.outcome == types.SpecStatePassed
+}
+
+func (node *synchronizedAfterSuiteNode) Passed() bool {
+	return node.outcome == types.SpecStatePassed
+}
+
+func (node *synchronizedAfterSuiteNode) Summary() *types.SetupSummary {
+	return &types.SetupSummary{
+		ComponentType: node.runnerA.nodeType,
+		CodeLocation:  node.runnerA.codeLocation,
+		State:         node.outcome,
+		RunTime:       node.runTime,
+		Failure:       node.failure,
+	}
+}
+
+func (node *synchronizedAfterSuiteNode) waitUntilOtherNodesAreDone(syncHost string) {
+	for {
+		if node.canRun(syncHost) {
+			return
+		}
+
+		time.Sleep(50 * time.Millisecond)
+	}
+}
+
+func (node *synchronizedAfterSuiteNode) canRun(syncHost string) bool {
+	resp, err := http.Get(syncHost + "/RemoteAfterSuiteData")
+	if err != nil || resp.StatusCode != http.StatusOK {
+		return false
+	}
+
+	body, err := ioutil.ReadAll(resp.Body)
+	if err != nil {
+		return false
+	}
+	resp.Body.Close()
+
+	afterSuiteData := types.RemoteAfterSuiteData{}
+	err = json.Unmarshal(body, &afterSuiteData)
+	if err != nil {
+		return false
+	}
+
+	return afterSuiteData.CanRun
+}
@@ -0,0 +1,181 @@
+package leafnodes
+
+import (
+	"bytes"
+	"encoding/json"
+	"io/ioutil"
+	"net/http"
+	"reflect"
+	"time"
+
+	"github.com/onsi/ginkgo/internal/failer"
+	"github.com/onsi/ginkgo/types"
+)
+
+type synchronizedBeforeSuiteNode struct {
+	runnerA *runner
+	runnerB *runner
+
+	data []byte
+
+	outcome types.SpecState
+	failure types.SpecFailure
+	runTime time.Duration
+}
+
+func NewSynchronizedBeforeSuiteNode(bodyA interface{}, bodyB interface{}, codeLocation types.CodeLocation, timeout time.Duration, failer *failer.Failer) SuiteNode {
+	node := &synchronizedBeforeSuiteNode{}
+
+	node.runnerA = newRunner(node.wrapA(bodyA), codeLocation, timeout, failer, types.SpecComponentTypeBeforeSuite, 0)
+	node.runnerB = newRunner(node.wrapB(bodyB), codeLocation, timeout, failer, types.SpecComponentTypeBeforeSuite, 0)
+
+	return node
+}
+
+func (node *synchronizedBeforeSuiteNode) Run(parallelNode int, parallelTotal int, syncHost string) bool {
+	t := time.Now()
+	defer func() {
+		node.runTime = time.Since(t)
+	}()
+
+	if parallelNode == 1 {
+		node.outcome, node.failure = node.runA(parallelTotal, syncHost)
+	} else {
+		node.outcome, node.failure = node.waitForA(syncHost)
+	}
+
+	if node.outcome != types.SpecStatePassed {
+		return false
+	}
+	node.outcome, node.failure = node.runnerB.run()
+
+	return node.outcome == types.SpecStatePassed
+}
+
+func (node *synchronizedBeforeSuiteNode) runA(parallelTotal int, syncHost string) (types.SpecState, types.SpecFailure) {
+	outcome, failure := node.runnerA.run()
+
+	if parallelTotal > 1 {
+		state := types.RemoteBeforeSuiteStatePassed
+		if outcome != types.SpecStatePassed {
+			state = types.RemoteBeforeSuiteStateFailed
+		}
+		json := (types.RemoteBeforeSuiteData{
+			Data:  node.data,
+			State: state,
+		}).ToJSON()
+		http.Post(syncHost+"/BeforeSuiteState", "application/json", bytes.NewBuffer(json))
+	}
+
+	return outcome, failure
+}
+
+func (node *synchronizedBeforeSuiteNode) waitForA(syncHost string) (types.SpecState, types.SpecFailure) {
+	failure := func(message string) types.SpecFailure {
+		return types.SpecFailure{
+			Message:               message,
+			Location:              node.runnerA.codeLocation,
+			ComponentType:         node.runnerA.nodeType,
+			ComponentIndex:        node.runnerA.componentIndex,
+			ComponentCodeLocation: node.runnerA.codeLocation,
+		}
+	}
+	for {
+		resp, err := http.Get(syncHost + "/BeforeSuiteState")
+		if err != nil || resp.StatusCode != http.StatusOK {
+			return types.SpecStateFailed, failure("Failed to fetch BeforeSuite state")
+		}
+
+		body, err := ioutil.ReadAll(resp.Body)
+		if err != nil {
+			return types.SpecStateFailed, failure("Failed to read BeforeSuite state")
+		}
+		resp.Body.Close()
+
+		beforeSuiteData := types.RemoteBeforeSuiteData{}
+		err = json.Unmarshal(body, &beforeSuiteData)
+		if err != nil {
+			return types.SpecStateFailed, failure("Failed to decode BeforeSuite state")
+		}
+
+		switch beforeSuiteData.State {
+		case types.RemoteBeforeSuiteStatePassed:
+			node.data = beforeSuiteData.Data
+			return types.SpecStatePassed, types.SpecFailure{}
+		case types.RemoteBeforeSuiteStateFailed:
+			return types.SpecStateFailed, failure("BeforeSuite on Node 1 failed")
+		case types.RemoteBeforeSuiteStateDisappeared:
+			return types.SpecStateFailed, failure("Node 1 disappeared before completing BeforeSuite")
+		}
+
+		time.Sleep(50 * time.Millisecond)
+	}
+}
+
+func (node *synchronizedBeforeSuiteNode) Passed() bool {
+	return node.outcome == types.SpecStatePassed
+}
+
+func (node *synchronizedBeforeSuiteNode) Summary() *types.SetupSummary {
+	return &types.SetupSummary{
+		ComponentType: node.runnerA.nodeType,
+		CodeLocation:  node.runnerA.codeLocation,
+		State:         node.outcome,
+		RunTime:       node.runTime,
+		Failure:       node.failure,
+	}
+}
+
+func (node *synchronizedBeforeSuiteNode) wrapA(bodyA interface{}) interface{} {
+	typeA := reflect.TypeOf(bodyA)
+	if typeA.Kind() != reflect.Func {
+		panic("SynchronizedBeforeSuite expects a function as its first argument")
+	}
+
+	takesNothing := typeA.NumIn() == 0
+	takesADoneChannel := typeA.NumIn() == 1 && typeA.In(0).Kind() == reflect.Chan && typeA.In(0).Elem().Kind() == reflect.Interface
+	returnsBytes := typeA.NumOut() == 1 && typeA.Out(0).Kind() == reflect.Slice && typeA.Out(0).Elem().Kind() == reflect.Uint8
+
+	if !((takesNothing || takesADoneChannel) && returnsBytes) {
+		panic("SynchronizedBeforeSuite's first argument should be a function that returns []byte and either takes no arguments or takes a Done channel.")
+	}
+
+	if takesADoneChannel {
+		return func(done chan<- interface{}) {
+			out := reflect.ValueOf(bodyA).Call([]reflect.Value{reflect.ValueOf(done)})
+			node.data = out[0].Interface().([]byte)
+		}
+	}
+
+	return func() {
+		out := reflect.ValueOf(bodyA).Call([]reflect.Value{})
+		node.data = out[0].Interface().([]byte)
+	}
+}
+
+func (node *synchronizedBeforeSuiteNode) wrapB(bodyB interface{}) interface{} {
+	typeB := reflect.TypeOf(bodyB)
+	if typeB.Kind() != reflect.Func {
+		panic("SynchronizedBeforeSuite expects a function as its second argument")
+	}
+
+	returnsNothing := typeB.NumOut() == 0
+	takesBytesOnly := typeB.NumIn() == 1 && typeB.In(0).Kind() == reflect.Slice && typeB.In(0).Elem().Kind() == reflect.Uint8
+	takesBytesAndDone := typeB.NumIn() == 2 &&
+		typeB.In(0).Kind() == reflect.Slice && typeB.In(0).Elem().Kind() == reflect.Uint8 &&
+		typeB.In(1).Kind() == reflect.Chan && typeB.In(1).Elem().Kind() == reflect.Interface
+
+	if !((takesBytesOnly || takesBytesAndDone) && returnsNothing) {
+		panic("SynchronizedBeforeSuite's second argument should be a function that returns nothing and either takes []byte or ([]byte, Done)")
+	}
+
+	if takesBytesAndDone {
+		return func(done chan<- interface{}) {
+			reflect.ValueOf(bodyB).Call([]reflect.Value{reflect.ValueOf(node.data), reflect.ValueOf(done)})
+		}
+	}
+
+	return func() {
+		reflect.ValueOf(bodyB).Call([]reflect.Value{reflect.ValueOf(node.data)})
+	}
+}
@@ -0,0 +1,249 @@
+/*
+
+Aggregator is a reporter used by the Ginkgo CLI to aggregate and present parallel test output
+coherently as tests complete.  You shouldn't need to use this in your code.  To run tests in parallel:
+
+	ginkgo -nodes=N
+
+where N is the number of nodes you desire.
+*/
+package remote
+
+import (
+	"time"
+
+	"github.com/onsi/ginkgo/config"
+	"github.com/onsi/ginkgo/reporters/stenographer"
+	"github.com/onsi/ginkgo/types"
+)
+
+type configAndSuite struct {
+	config  config.GinkgoConfigType
+	summary *types.SuiteSummary
+}
+
+type Aggregator struct {
+	nodeCount    int
+	config       config.DefaultReporterConfigType
+	stenographer stenographer.Stenographer
+	result       chan bool
+
+	suiteBeginnings           chan configAndSuite
+	aggregatedSuiteBeginnings []configAndSuite
+
+	beforeSuites           chan *types.SetupSummary
+	aggregatedBeforeSuites []*types.SetupSummary
+
+	afterSuites           chan *types.SetupSummary
+	aggregatedAfterSuites []*types.SetupSummary
+
+	specCompletions chan *types.SpecSummary
+	completedSpecs  []*types.SpecSummary
+
+	suiteEndings           chan *types.SuiteSummary
+	aggregatedSuiteEndings []*types.SuiteSummary
+	specs                  []*types.SpecSummary
+
+	startTime time.Time
+}
+
+func NewAggregator(nodeCount int, result chan bool, config config.DefaultReporterConfigType, stenographer stenographer.Stenographer) *Aggregator {
+	aggregator := &Aggregator{
+		nodeCount:    nodeCount,
+		result:       result,
+		config:       config,
+		stenographer: stenographer,
+
+		suiteBeginnings: make(chan configAndSuite),
+		beforeSuites:    make(chan *types.SetupSummary),
+		afterSuites:     make(chan *types.SetupSummary),
+		specCompletions: make(chan *types.SpecSummary),
+		suiteEndings:    make(chan *types.SuiteSummary),
+	}
+
+	go aggregator.mux()
+
+	return aggregator
+}
+
+func (aggregator *Aggregator) SpecSuiteWillBegin(config config.GinkgoConfigType, summary *types.SuiteSummary) {
+	aggregator.suiteBeginnings <- configAndSuite{config, summary}
+}
+
+func (aggregator *Aggregator) BeforeSuiteDidRun(setupSummary *types.SetupSummary) {
+	aggregator.beforeSuites <- setupSummary
+}
+
+func (aggregator *Aggregator) AfterSuiteDidRun(setupSummary *types.SetupSummary) {
+	aggregator.afterSuites <- setupSummary
+}
+
+func (aggregator *Aggregator) SpecWillRun(specSummary *types.SpecSummary) {
+	//noop
+}
+
+func (aggregator *Aggregator) SpecDidComplete(specSummary *types.SpecSummary) {
+	aggregator.specCompletions <- specSummary
+}
+
+func (aggregator *Aggregator) SpecSuiteDidEnd(summary *types.SuiteSummary) {
+	aggregator.suiteEndings <- summary
+}
+
+func (aggregator *Aggregator) mux() {
+loop:
+	for {
+		select {
+		case configAndSuite := <-aggregator.suiteBeginnings:
+			aggregator.registerSuiteBeginning(configAndSuite)
+		case setupSummary := <-aggregator.beforeSuites:
+			aggregator.registerBeforeSuite(setupSummary)
+		case setupSummary := <-aggregator.afterSuites:
+			aggregator.registerAfterSuite(setupSummary)
+		case specSummary := <-aggregator.specCompletions:
+			aggregator.registerSpecCompletion(specSummary)
+		case suite := <-aggregator.suiteEndings:
+			finished, passed := aggregator.registerSuiteEnding(suite)
+			if finished {
+				aggregator.result <- passed
+				break loop
+			}
+		}
+	}
+}
+
+func (aggregator *Aggregator) registerSuiteBeginning(configAndSuite configAndSuite) {
+	aggregator.aggregatedSuiteBeginnings = append(aggregator.aggregatedSuiteBeginnings, configAndSuite)
+
+	if len(aggregator.aggregatedSuiteBeginnings) == 1 {
+		aggregator.startTime = time.Now()
+	}
+
+	if len(aggregator.aggregatedSuiteBeginnings) != aggregator.nodeCount {
+		return
+	}
+
+	aggregator.stenographer.AnnounceSuite(configAndSuite.summary.SuiteDescription, configAndSuite.config.RandomSeed, configAndSuite.config.RandomizeAllSpecs, aggregator.config.Succinct)
+
+	totalNumberOfSpecs := 0
+	if len(aggregator.aggregatedSuiteBeginnings) > 0 {
+		totalNumberOfSpecs = configAndSuite.summary.NumberOfSpecsBeforeParallelization
+	}
+
+	aggregator.stenographer.AnnounceTotalNumberOfSpecs(totalNumberOfSpecs, aggregator.config.Succinct)
+	aggregator.stenographer.AnnounceAggregatedParallelRun(aggregator.nodeCount, aggregator.config.Succinct)
+	aggregator.flushCompletedSpecs()
+}
+
+func (aggregator *Aggregator) registerBeforeSuite(setupSummary *types.SetupSummary) {
+	aggregator.aggregatedBeforeSuites = append(aggregator.aggregatedBeforeSuites, setupSummary)
+	aggregator.flushCompletedSpecs()
+}
+
+func (aggregator *Aggregator) registerAfterSuite(setupSummary *types.SetupSummary) {
+	aggregator.aggregatedAfterSuites = append(aggregator.aggregatedAfterSuites, setupSummary)
+	aggregator.flushCompletedSpecs()
+}
+
+func (aggregator *Aggregator) registerSpecCompletion(specSummary *types.SpecSummary) {
+	aggregator.completedSpecs = append(aggregator.completedSpecs, specSummary)
+	aggregator.specs = append(aggregator.specs, specSummary)
+	aggregator.flushCompletedSpecs()
+}
+
+func (aggregator *Aggregator) flushCompletedSpecs() {
+	if len(aggregator.aggregatedSuiteBeginnings) != aggregator.nodeCount {
+		return
+	}
+
+	for _, setupSummary := range aggregator.aggregatedBeforeSuites {
+		aggregator.announceBeforeSuite(setupSummary)
+	}
+
+	for _, specSummary := range aggregator.completedSpecs {
+		aggregator.announceSpec(specSummary)
+	}
+
+	for _, setupSummary := range aggregator.aggregatedAfterSuites {
+		aggregator.announceAfterSuite(setupSummary)
+	}
+
+	aggregator.aggregatedBeforeSuites = []*types.SetupSummary{}
+	aggregator.completedSpecs = []*types.SpecSummary{}
+	aggregator.aggregatedAfterSuites = []*types.SetupSummary{}
+}
+
+func (aggregator *Aggregator) announceBeforeSuite(setupSummary *types.SetupSummary) {
+	aggregator.stenographer.AnnounceCapturedOutput(setupSummary.CapturedOutput)
+	if setupSummary.State != types.SpecStatePassed {
+		aggregator.stenographer.AnnounceBeforeSuiteFailure(setupSummary, aggregator.config.Succinct, aggregator.config.FullTrace)
+	}
+}
+
+func (aggregator *Aggregator) announceAfterSuite(setupSummary *types.SetupSummary) {
+	aggregator.stenographer.AnnounceCapturedOutput(setupSummary.CapturedOutput)
+	if setupSummary.State != types.SpecStatePassed {
+		aggregator.stenographer.AnnounceAfterSuiteFailure(setupSummary, aggregator.config.Succinct, aggregator.config.FullTrace)
+	}
+}
+
+func (aggregator *Aggregator) announceSpec(specSummary *types.SpecSummary) {
+	if aggregator.config.Verbose && specSummary.State != types.SpecStatePending && specSummary.State != types.SpecStateSkipped {
+		aggregator.stenographer.AnnounceSpecWillRun(specSummary)
+	}
+
+	aggregator.stenographer.AnnounceCapturedOutput(specSummary.CapturedOutput)
+
+	switch specSummary.State {
+	case types.SpecStatePassed:
+		if specSummary.IsMeasurement {
+			aggregator.stenographer.AnnounceSuccessfulMeasurement(specSummary, aggregator.config.Succinct)
+		} else if specSummary.RunTime.Seconds() >= aggregator.config.SlowSpecThreshold {
+			aggregator.stenographer.AnnounceSuccessfulSlowSpec(specSummary, aggregator.config.Succinct)
+		} else {
+			aggregator.stenographer.AnnounceSuccessfulSpec(specSummary)
+		}
+
+	case types.SpecStatePending:
+		aggregator.stenographer.AnnouncePendingSpec(specSummary, aggregator.config.NoisyPendings && !aggregator.config.Succinct)
+	case types.SpecStateSkipped:
+		aggregator.stenographer.AnnounceSkippedSpec(specSummary, aggregator.config.Succinct || !aggregator.config.NoisySkippings, aggregator.config.FullTrace)
+	case types.SpecStateTimedOut:
+		aggregator.stenographer.AnnounceSpecTimedOut(specSummary, aggregator.config.Succinct, aggregator.config.FullTrace)
+	case types.SpecStatePanicked:
+		aggregator.stenographer.AnnounceSpecPanicked(specSummary, aggregator.config.Succinct, aggregator.config.FullTrace)
+	case types.SpecStateFailed:
+		aggregator.stenographer.AnnounceSpecFailed(specSummary, aggregator.config.Succinct, aggregator.config.FullTrace)
+	}
+}
+
+func (aggregator *Aggregator) registerSuiteEnding(suite *types.SuiteSummary) (finished bool, passed bool) {
+	aggregator.aggregatedSuiteEndings = append(aggregator.aggregatedSuiteEndings, suite)
+	if len(aggregator.aggregatedSuiteEndings) < aggregator.nodeCount {
+		return false, false
+	}
+
+	aggregatedSuiteSummary := &types.SuiteSummary{}
+	aggregatedSuiteSummary.SuiteSucceeded = true
+
+	for _, suiteSummary := range aggregator.aggregatedSuiteEndings {
+		if !suiteSummary.SuiteSucceeded {
+			aggregatedSuiteSummary.SuiteSucceeded = false
+		}
+
+		aggregatedSuiteSummary.NumberOfSpecsThatWillBeRun += suiteSummary.NumberOfSpecsThatWillBeRun
+		aggregatedSuiteSummary.NumberOfTotalSpecs += suiteSummary.NumberOfTotalSpecs
+		aggregatedSuiteSummary.NumberOfPassedSpecs += suiteSummary.NumberOfPassedSpecs
+		aggregatedSuiteSummary.NumberOfFailedSpecs += suiteSummary.NumberOfFailedSpecs
+		aggregatedSuiteSummary.NumberOfPendingSpecs += suiteSummary.NumberOfPendingSpecs
+		aggregatedSuiteSummary.NumberOfSkippedSpecs += suiteSummary.NumberOfSkippedSpecs
+		aggregatedSuiteSummary.NumberOfFlakedSpecs += suiteSummary.NumberOfFlakedSpecs
+	}
+
+	aggregatedSuiteSummary.RunTime = time.Since(aggregator.startTime)
+
+	aggregator.stenographer.SummarizeFailures(aggregator.specs)
+	aggregator.stenographer.AnnounceSpecRunCompletion(aggregatedSuiteSummary, aggregator.config.Succinct)
+
+	return true, aggregatedSuiteSummary.SuiteSucceeded
+}
@@ -0,0 +1,147 @@
+package remote
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+
+	"github.com/onsi/ginkgo/internal/writer"
+	"github.com/onsi/ginkgo/reporters"
+	"github.com/onsi/ginkgo/reporters/stenographer"
+
+	"github.com/onsi/ginkgo/config"
+	"github.com/onsi/ginkgo/types"
+)
+
+//An interface to net/http's client to allow the injection of fakes under test
+type Poster interface {
+	Post(url string, bodyType string, body io.Reader) (resp *http.Response, err error)
+}
+
+/*
+The ForwardingReporter is a Ginkgo reporter that forwards information to
+a Ginkgo remote server.
+
+When streaming parallel test output, this repoter is automatically installed by Ginkgo.
+
+This is accomplished by passing in the GINKGO_REMOTE_REPORTING_SERVER environment variable to `go test`, the Ginkgo test runner
+detects this environment variable (which should contain the host of the server) and automatically installs a ForwardingReporter
+in place of Ginkgo's DefaultReporter.
+*/
+
+type ForwardingReporter struct {
+	serverHost        string
+	poster            Poster
+	outputInterceptor OutputInterceptor
+	debugMode         bool
+	debugFile         *os.File
+	nestedReporter    *reporters.DefaultReporter
+}
+
+func NewForwardingReporter(config config.DefaultReporterConfigType, serverHost string, poster Poster, outputInterceptor OutputInterceptor, ginkgoWriter *writer.Writer, debugFile string) *ForwardingReporter {
+	reporter := &ForwardingReporter{
+		serverHost:        serverHost,
+		poster:            poster,
+		outputInterceptor: outputInterceptor,
+	}
+
+	if debugFile != "" {
+		var err error
+		reporter.debugMode = true
+		reporter.debugFile, err = os.Create(debugFile)
+		if err != nil {
+			fmt.Println(err.Error())
+			os.Exit(1)
+		}
+
+		if !config.Verbose {
+			//if verbose is true then the GinkgoWriter emits to stdout.  Don't _also_ redirect GinkgoWriter output as that will result in duplication.
+			ginkgoWriter.AndRedirectTo(reporter.debugFile)
+		}
+		outputInterceptor.StreamTo(reporter.debugFile) //This is not working
+
+		stenographer := stenographer.New(false, true, reporter.debugFile)
+		config.Succinct = false
+		config.Verbose = true
+		config.FullTrace = true
+		reporter.nestedReporter = reporters.NewDefaultReporter(config, stenographer)
+	}
+
+	return reporter
+}
+
+func (reporter *ForwardingReporter) post(path string, data interface{}) {
+	encoded, _ := json.Marshal(data)
+	buffer := bytes.NewBuffer(encoded)
+	reporter.poster.Post(reporter.serverHost+path, "application/json", buffer)
+}
+
+func (reporter *ForwardingReporter) SpecSuiteWillBegin(conf config.GinkgoConfigType, summary *types.SuiteSummary) {
+	data := struct {
+		Config  config.GinkgoConfigType `json:"config"`
+		Summary *types.SuiteSummary     `json:"suite-summary"`
+	}{
+		conf,
+		summary,
+	}
+
+	reporter.outputInterceptor.StartInterceptingOutput()
+	if reporter.debugMode {
+		reporter.nestedReporter.SpecSuiteWillBegin(conf, summary)
+		reporter.debugFile.Sync()
+	}
+	reporter.post("/SpecSuiteWillBegin", data)
+}
+
+func (reporter *ForwardingReporter) BeforeSuiteDidRun(setupSummary *types.SetupSummary) {
+	output, _ := reporter.outputInterceptor.StopInterceptingAndReturnOutput()
+	reporter.outputInterceptor.StartInterceptingOutput()
+	setupSummary.CapturedOutput = output
+	if reporter.debugMode {
+		reporter.nestedReporter.BeforeSuiteDidRun(setupSummary)
+		reporter.debugFile.Sync()
+	}
+	reporter.post("/BeforeSuiteDidRun", setupSummary)
+}
+
+func (reporter *ForwardingReporter) SpecWillRun(specSummary *types.SpecSummary) {
+	if reporter.debugMode {
+		reporter.nestedReporter.SpecWillRun(specSummary)
+		reporter.debugFile.Sync()
+	}
+	reporter.post("/SpecWillRun", specSummary)
+}
+
+func (reporter *ForwardingReporter) SpecDidComplete(specSummary *types.SpecSummary) {
+	output, _ := reporter.outputInterceptor.StopInterceptingAndReturnOutput()
+	reporter.outputInterceptor.StartInterceptingOutput()
+	specSummary.CapturedOutput = output
+	if reporter.debugMode {
+		reporter.nestedReporter.SpecDidComplete(specSummary)
+		reporter.debugFile.Sync()
+	}
+	reporter.post("/SpecDidComplete", specSummary)
+}
+
+func (reporter *ForwardingReporter) AfterSuiteDidRun(setupSummary *types.SetupSummary) {
+	output, _ := reporter.outputInterceptor.StopInterceptingAndReturnOutput()
+	reporter.outputInterceptor.StartInterceptingOutput()
+	setupSummary.CapturedOutput = output
+	if reporter.debugMode {
+		reporter.nestedReporter.AfterSuiteDidRun(setupSummary)
+		reporter.debugFile.Sync()
+	}
+	reporter.post("/AfterSuiteDidRun", setupSummary)
+}
+
+func (reporter *ForwardingReporter) SpecSuiteDidEnd(summary *types.SuiteSummary) {
+	reporter.outputInterceptor.StopInterceptingAndReturnOutput()
+	if reporter.debugMode {
+		reporter.nestedReporter.SpecSuiteDidEnd(summary)
+		reporter.debugFile.Sync()
+	}
+	reporter.post("/SpecSuiteDidEnd", summary)
+}
@@ -0,0 +1,13 @@
+package remote
+
+import "os"
+
+/*
+The OutputInterceptor is used by the ForwardingReporter to
+intercept and capture all stdin and stderr output during a test run.
+*/
+type OutputInterceptor interface {
+	StartInterceptingOutput() error
+	StopInterceptingAndReturnOutput() (string, error)
+	StreamTo(*os.File)
+}
@@ -0,0 +1,82 @@
+// +build freebsd openbsd netbsd dragonfly darwin linux solaris
+
+package remote
+
+import (
+	"errors"
+	"io/ioutil"
+	"os"
+
+	"github.com/nxadm/tail"
+	"golang.org/x/sys/unix"
+)
+
+func NewOutputInterceptor() OutputInterceptor {
+	return &outputInterceptor{}
+}
+
+type outputInterceptor struct {
+	redirectFile *os.File
+	streamTarget *os.File
+	intercepting bool
+	tailer       *tail.Tail
+	doneTailing  chan bool
+}
+
+func (interceptor *outputInterceptor) StartInterceptingOutput() error {
+	if interceptor.intercepting {
+		return errors.New("Already intercepting output!")
+	}
+	interceptor.intercepting = true
+
+	var err error
+
+	interceptor.redirectFile, err = ioutil.TempFile("", "ginkgo-output")
+	if err != nil {
+		return err
+	}
+
+	// This might call Dup3 if the dup2 syscall is not available, e.g. on
+	// linux/arm64 or linux/riscv64
+	unix.Dup2(int(interceptor.redirectFile.Fd()), 1)
+	unix.Dup2(int(interceptor.redirectFile.Fd()), 2)
+
+	if interceptor.streamTarget != nil {
+		interceptor.tailer, _ = tail.TailFile(interceptor.redirectFile.Name(), tail.Config{Follow: true})
+		interceptor.doneTailing = make(chan bool)
+
+		go func() {
+			for line := range interceptor.tailer.Lines {
+				interceptor.streamTarget.Write([]byte(line.Text + "\n"))
+			}
+			close(interceptor.doneTailing)
+		}()
+	}
+
+	return nil
+}
+
+func (interceptor *outputInterceptor) StopInterceptingAndReturnOutput() (string, error) {
+	if !interceptor.intercepting {
+		return "", errors.New("Not intercepting output!")
+	}
+
+	interceptor.redirectFile.Close()
+	output, err := ioutil.ReadFile(interceptor.redirectFile.Name())
+	os.Remove(interceptor.redirectFile.Name())
+
+	interceptor.intercepting = false
+
+	if interceptor.streamTarget != nil {
+		interceptor.tailer.Stop()
+		interceptor.tailer.Cleanup()
+		<-interceptor.doneTailing
+		interceptor.streamTarget.Sync()
+	}
+
+	return string(output), err
+}
+
+func (interceptor *outputInterceptor) StreamTo(out *os.File) {
+	interceptor.streamTarget = out
+}
@@ -0,0 +1,36 @@
+// +build windows
+
+package remote
+
+import (
+	"errors"
+	"os"
+)
+
+func NewOutputInterceptor() OutputInterceptor {
+	return &outputInterceptor{}
+}
+
+type outputInterceptor struct {
+	intercepting bool
+}
+
+func (interceptor *outputInterceptor) StartInterceptingOutput() error {
+	if interceptor.intercepting {
+		return errors.New("Already intercepting output!")
+	}
+	interceptor.intercepting = true
+
+	// not working on windows...
+
+	return nil
+}
+
+func (interceptor *outputInterceptor) StopInterceptingAndReturnOutput() (string, error) {
+	// not working on windows...
+	interceptor.intercepting = false
+
+	return "", nil
+}
+
+func (interceptor *outputInterceptor) StreamTo(*os.File) {}
@@ -0,0 +1,224 @@
+/*
+
+The remote package provides the pieces to allow Ginkgo test suites to report to remote listeners.
+This is used, primarily, to enable streaming parallel test output but has, in principal, broader applications (e.g. streaming test output to a browser).
+
+*/
+
+package remote
+
+import (
+	"encoding/json"
+	"io/ioutil"
+	"net"
+	"net/http"
+	"sync"
+
+	"github.com/onsi/ginkgo/internal/spec_iterator"
+
+	"github.com/onsi/ginkgo/config"
+	"github.com/onsi/ginkgo/reporters"
+	"github.com/onsi/ginkgo/types"
+)
+
+/*
+Server spins up on an automatically selected port and listens for communication from the forwarding reporter.
+It then forwards that communication to attached reporters.
+*/
+type Server struct {
+	listener        net.Listener
+	reporters       []reporters.Reporter
+	alives          []func() bool
+	lock            *sync.Mutex
+	beforeSuiteData types.RemoteBeforeSuiteData
+	parallelTotal   int
+	counter         int
+}
+
+//Create a new server, automatically selecting a port
+func NewServer(parallelTotal int) (*Server, error) {
+	listener, err := net.Listen("tcp", "127.0.0.1:0")
+	if err != nil {
+		return nil, err
+	}
+	return &Server{
+		listener:        listener,
+		lock:            &sync.Mutex{},
+		alives:          make([]func() bool, parallelTotal),
+		beforeSuiteData: types.RemoteBeforeSuiteData{Data: nil, State: types.RemoteBeforeSuiteStatePending},
+		parallelTotal:   parallelTotal,
+	}, nil
+}
+
+//Start the server.  You don't need to `go s.Start()`, just `s.Start()`
+func (server *Server) Start() {
+	httpServer := &http.Server{}
+	mux := http.NewServeMux()
+	httpServer.Handler = mux
+
+	//streaming endpoints
+	mux.HandleFunc("/SpecSuiteWillBegin", server.specSuiteWillBegin)
+	mux.HandleFunc("/BeforeSuiteDidRun", server.beforeSuiteDidRun)
+	mux.HandleFunc("/AfterSuiteDidRun", server.afterSuiteDidRun)
+	mux.HandleFunc("/SpecWillRun", server.specWillRun)
+	mux.HandleFunc("/SpecDidComplete", server.specDidComplete)
+	mux.HandleFunc("/SpecSuiteDidEnd", server.specSuiteDidEnd)
+
+	//synchronization endpoints
+	mux.HandleFunc("/BeforeSuiteState", server.handleBeforeSuiteState)
+	mux.HandleFunc("/RemoteAfterSuiteData", server.handleRemoteAfterSuiteData)
+	mux.HandleFunc("/counter", server.handleCounter)
+	mux.HandleFunc("/has-counter", server.handleHasCounter) //for backward compatibility
+
+	go httpServer.Serve(server.listener)
+}
+
+//Stop the server
+func (server *Server) Close() {
+	server.listener.Close()
+}
+
+//The address the server can be reached it.  Pass this into the `ForwardingReporter`.
+func (server *Server) Address() string {
+	return "http://" + server.listener.Addr().String()
+}
+
+//
+// Streaming Endpoints
+//
+
+//The server will forward all received messages to Ginkgo reporters registered with `RegisterReporters`
+func (server *Server) readAll(request *http.Request) []byte {
+	defer request.Body.Close()
+	body, _ := ioutil.ReadAll(request.Body)
+	return body
+}
+
+func (server *Server) RegisterReporters(reporters ...reporters.Reporter) {
+	server.reporters = reporters
+}
+
+func (server *Server) specSuiteWillBegin(writer http.ResponseWriter, request *http.Request) {
+	body := server.readAll(request)
+
+	var data struct {
+		Config  config.GinkgoConfigType `json:"config"`
+		Summary *types.SuiteSummary     `json:"suite-summary"`
+	}
+
+	json.Unmarshal(body, &data)
+
+	for _, reporter := range server.reporters {
+		reporter.SpecSuiteWillBegin(data.Config, data.Summary)
+	}
+}
+
+func (server *Server) beforeSuiteDidRun(writer http.ResponseWriter, request *http.Request) {
+	body := server.readAll(request)
+	var setupSummary *types.SetupSummary
+	json.Unmarshal(body, &setupSummary)
+
+	for _, reporter := range server.reporters {
+		reporter.BeforeSuiteDidRun(setupSummary)
+	}
+}
+
+func (server *Server) afterSuiteDidRun(writer http.ResponseWriter, request *http.Request) {
+	body := server.readAll(request)
+	var setupSummary *types.SetupSummary
+	json.Unmarshal(body, &setupSummary)
+
+	for _, reporter := range server.reporters {
+		reporter.AfterSuiteDidRun(setupSummary)
+	}
+}
+
+func (server *Server) specWillRun(writer http.ResponseWriter, request *http.Request) {
+	body := server.readAll(request)
+	var specSummary *types.SpecSummary
+	json.Unmarshal(body, &specSummary)
+
+	for _, reporter := range server.reporters {
+		reporter.SpecWillRun(specSummary)
+	}
+}
+
+func (server *Server) specDidComplete(writer http.ResponseWriter, request *http.Request) {
+	body := server.readAll(request)
+	var specSummary *types.SpecSummary
+	json.Unmarshal(body, &specSummary)
+
+	for _, reporter := range server.reporters {
+		reporter.SpecDidComplete(specSummary)
+	}
+}
+
+func (server *Server) specSuiteDidEnd(writer http.ResponseWriter, request *http.Request) {
+	body := server.readAll(request)
+	var suiteSummary *types.SuiteSummary
+	json.Unmarshal(body, &suiteSummary)
+
+	for _, reporter := range server.reporters {
+		reporter.SpecSuiteDidEnd(suiteSummary)
+	}
+}
+
+//
+// Synchronization Endpoints
+//
+
+func (server *Server) RegisterAlive(node int, alive func() bool) {
+	server.lock.Lock()
+	defer server.lock.Unlock()
+	server.alives[node-1] = alive
+}
+
+func (server *Server) nodeIsAlive(node int) bool {
+	server.lock.Lock()
+	defer server.lock.Unlock()
+	alive := server.alives[node-1]
+	if alive == nil {
+		return true
+	}
+	return alive()
+}
+
+func (server *Server) handleBeforeSuiteState(writer http.ResponseWriter, request *http.Request) {
+	if request.Method == "POST" {
+		dec := json.NewDecoder(request.Body)
+		dec.Decode(&(server.beforeSuiteData))
+	} else {
+		beforeSuiteData := server.beforeSuiteData
+		if beforeSuiteData.State == types.RemoteBeforeSuiteStatePending && !server.nodeIsAlive(1) {
+			beforeSuiteData.State = types.RemoteBeforeSuiteStateDisappeared
+		}
+		enc := json.NewEncoder(writer)
+		enc.Encode(beforeSuiteData)
+	}
+}
+
+func (server *Server) handleRemoteAfterSuiteData(writer http.ResponseWriter, request *http.Request) {
+	afterSuiteData := types.RemoteAfterSuiteData{
+		CanRun: true,
+	}
+	for i := 2; i <= server.parallelTotal; i++ {
+		afterSuiteData.CanRun = afterSuiteData.CanRun && !server.nodeIsAlive(i)
+	}
+
+	enc := json.NewEncoder(writer)
+	enc.Encode(afterSuiteData)
+}
+
+func (server *Server) handleCounter(writer http.ResponseWriter, request *http.Request) {
+	c := spec_iterator.Counter{}
+	server.lock.Lock()
+	c.Index = server.counter
+	server.counter++
+	server.lock.Unlock()
+
+	json.NewEncoder(writer).Encode(c)
+}
+
+func (server *Server) handleHasCounter(writer http.ResponseWriter, request *http.Request) {
+	writer.Write([]byte(""))
+}
@@ -0,0 +1,247 @@
+package spec
+
+import (
+	"fmt"
+	"io"
+	"time"
+
+	"sync"
+
+	"github.com/onsi/ginkgo/internal/containernode"
+	"github.com/onsi/ginkgo/internal/leafnodes"
+	"github.com/onsi/ginkgo/types"
+)
+
+type Spec struct {
+	subject          leafnodes.SubjectNode
+	focused          bool
+	announceProgress bool
+
+	containers []*containernode.ContainerNode
+
+	state            types.SpecState
+	runTime          time.Duration
+	startTime        time.Time
+	failure          types.SpecFailure
+	previousFailures bool
+
+	stateMutex *sync.Mutex
+}
+
+func New(subject leafnodes.SubjectNode, containers []*containernode.ContainerNode, announceProgress bool) *Spec {
+	spec := &Spec{
+		subject:          subject,
+		containers:       containers,
+		focused:          subject.Flag() == types.FlagTypeFocused,
+		announceProgress: announceProgress,
+		stateMutex:       &sync.Mutex{},
+	}
+
+	spec.processFlag(subject.Flag())
+	for i := len(containers) - 1; i >= 0; i-- {
+		spec.processFlag(containers[i].Flag())
+	}
+
+	return spec
+}
+
+func (spec *Spec) processFlag(flag types.FlagType) {
+	if flag == types.FlagTypeFocused {
+		spec.focused = true
+	} else if flag == types.FlagTypePending {
+		spec.setState(types.SpecStatePending)
+	}
+}
+
+func (spec *Spec) Skip() {
+	spec.setState(types.SpecStateSkipped)
+}
+
+func (spec *Spec) Failed() bool {
+	return spec.getState() == types.SpecStateFailed || spec.getState() == types.SpecStatePanicked || spec.getState() == types.SpecStateTimedOut
+}
+
+func (spec *Spec) Passed() bool {
+	return spec.getState() == types.SpecStatePassed
+}
+
+func (spec *Spec) Flaked() bool {
+	return spec.getState() == types.SpecStatePassed && spec.previousFailures
+}
+
+func (spec *Spec) Pending() bool {
+	return spec.getState() == types.SpecStatePending
+}
+
+func (spec *Spec) Skipped() bool {
+	return spec.getState() == types.SpecStateSkipped
+}
+
+func (spec *Spec) Focused() bool {
+	return spec.focused
+}
+
+func (spec *Spec) IsMeasurement() bool {
+	return spec.subject.Type() == types.SpecComponentTypeMeasure
+}
+
+func (spec *Spec) Summary(suiteID string) *types.SpecSummary {
+	componentTexts := make([]string, len(spec.containers)+1)
+	componentCodeLocations := make([]types.CodeLocation, len(spec.containers)+1)
+
+	for i, container := range spec.containers {
+		componentTexts[i] = container.Text()
+		componentCodeLocations[i] = container.CodeLocation()
+	}
+
+	componentTexts[len(spec.containers)] = spec.subject.Text()
+	componentCodeLocations[len(spec.containers)] = spec.subject.CodeLocation()
+
+	runTime := spec.runTime
+	if runTime == 0 && !spec.startTime.IsZero() {
+		runTime = time.Since(spec.startTime)
+	}
+
+	return &types.SpecSummary{
+		IsMeasurement:          spec.IsMeasurement(),
+		NumberOfSamples:        spec.subject.Samples(),
+		ComponentTexts:         componentTexts,
+		ComponentCodeLocations: componentCodeLocations,
+		State:                  spec.getState(),
+		RunTime:                runTime,
+		Failure:                spec.failure,
+		Measurements:           spec.measurementsReport(),
+		SuiteID:                suiteID,
+	}
+}
+
+func (spec *Spec) ConcatenatedString() string {
+	s := ""
+	for _, container := range spec.containers {
+		s += container.Text() + " "
+	}
+
+	return s + spec.subject.Text()
+}
+
+func (spec *Spec) Run(writer io.Writer) {
+	if spec.getState() == types.SpecStateFailed {
+		spec.previousFailures = true
+	}
+
+	spec.startTime = time.Now()
+	defer func() {
+		spec.runTime = time.Since(spec.startTime)
+	}()
+
+	for sample := 0; sample < spec.subject.Samples(); sample++ {
+		spec.runSample(sample, writer)
+
+		if spec.getState() != types.SpecStatePassed {
+			return
+		}
+	}
+}
+
+func (spec *Spec) getState() types.SpecState {
+	spec.stateMutex.Lock()
+	defer spec.stateMutex.Unlock()
+	return spec.state
+}
+
+func (spec *Spec) setState(state types.SpecState) {
+	spec.stateMutex.Lock()
+	defer spec.stateMutex.Unlock()
+	spec.state = state
+}
+
+func (spec *Spec) runSample(sample int, writer io.Writer) {
+	spec.setState(types.SpecStatePassed)
+	spec.failure = types.SpecFailure{}
+	innerMostContainerIndexToUnwind := -1
+
+	defer func() {
+		for i := innerMostContainerIndexToUnwind; i >= 0; i-- {
+			container := spec.containers[i]
+			for _, justAfterEach := range container.SetupNodesOfType(types.SpecComponentTypeJustAfterEach) {
+				spec.announceSetupNode(writer, "JustAfterEach", container, justAfterEach)
+				justAfterEachState, justAfterEachFailure := justAfterEach.Run()
+				if justAfterEachState != types.SpecStatePassed && spec.state == types.SpecStatePassed {
+					spec.state = justAfterEachState
+					spec.failure = justAfterEachFailure
+				}
+			}
+		}
+
+		for i := innerMostContainerIndexToUnwind; i >= 0; i-- {
+			container := spec.containers[i]
+			for _, afterEach := range container.SetupNodesOfType(types.SpecComponentTypeAfterEach) {
+				spec.announceSetupNode(writer, "AfterEach", container, afterEach)
+				afterEachState, afterEachFailure := afterEach.Run()
+				if afterEachState != types.SpecStatePassed && spec.getState() == types.SpecStatePassed {
+					spec.setState(afterEachState)
+					spec.failure = afterEachFailure
+				}
+			}
+		}
+	}()
+
+	for i, container := range spec.containers {
+		innerMostContainerIndexToUnwind = i
+		for _, beforeEach := range container.SetupNodesOfType(types.SpecComponentTypeBeforeEach) {
+			spec.announceSetupNode(writer, "BeforeEach", container, beforeEach)
+			s, f := beforeEach.Run()
+			spec.failure = f
+			spec.setState(s)
+			if spec.getState() != types.SpecStatePassed {
+				return
+			}
+		}
+	}
+
+	for _, container := range spec.containers {
+		for _, justBeforeEach := range container.SetupNodesOfType(types.SpecComponentTypeJustBeforeEach) {
+			spec.announceSetupNode(writer, "JustBeforeEach", container, justBeforeEach)
+			s, f := justBeforeEach.Run()
+			spec.failure = f
+			spec.setState(s)
+			if spec.getState() != types.SpecStatePassed {
+				return
+			}
+		}
+	}
+
+	spec.announceSubject(writer, spec.subject)
+	s, f := spec.subject.Run()
+	spec.failure = f
+	spec.setState(s)
+}
+
+func (spec *Spec) announceSetupNode(writer io.Writer, nodeType string, container *containernode.ContainerNode, setupNode leafnodes.BasicNode) {
+	if spec.announceProgress {
+		s := fmt.Sprintf("[%s] %s\n  %s\n", nodeType, container.Text(), setupNode.CodeLocation().String())
+		writer.Write([]byte(s))
+	}
+}
+
+func (spec *Spec) announceSubject(writer io.Writer, subject leafnodes.SubjectNode) {
+	if spec.announceProgress {
+		nodeType := ""
+		switch subject.Type() {
+		case types.SpecComponentTypeIt:
+			nodeType = "It"
+		case types.SpecComponentTypeMeasure:
+			nodeType = "Measure"
+		}
+		s := fmt.Sprintf("[%s] %s\n  %s\n", nodeType, subject.Text(), subject.CodeLocation().String())
+		writer.Write([]byte(s))
+	}
+}
+
+func (spec *Spec) measurementsReport() map[string]*types.SpecMeasurement {
+	if !spec.IsMeasurement() || spec.Failed() {
+		return map[string]*types.SpecMeasurement{}
+	}
+
+	return spec.subject.(*leafnodes.MeasureNode).MeasurementsReport()
+}
@@ -0,0 +1,144 @@
+package spec
+
+import (
+	"math/rand"
+	"regexp"
+	"sort"
+	"strings"
+)
+
+type Specs struct {
+	specs []*Spec
+	names []string
+
+	hasProgrammaticFocus bool
+	RegexScansFilePath   bool
+}
+
+func NewSpecs(specs []*Spec) *Specs {
+	names := make([]string, len(specs))
+	for i, spec := range specs {
+		names[i] = spec.ConcatenatedString()
+	}
+	return &Specs{
+		specs: specs,
+		names: names,
+	}
+}
+
+func (e *Specs) Specs() []*Spec {
+	return e.specs
+}
+
+func (e *Specs) HasProgrammaticFocus() bool {
+	return e.hasProgrammaticFocus
+}
+
+func (e *Specs) Shuffle(r *rand.Rand) {
+	sort.Sort(e)
+	permutation := r.Perm(len(e.specs))
+	shuffledSpecs := make([]*Spec, len(e.specs))
+	names := make([]string, len(e.specs))
+	for i, j := range permutation {
+		shuffledSpecs[i] = e.specs[j]
+		names[i] = e.names[j]
+	}
+	e.specs = shuffledSpecs
+	e.names = names
+}
+
+func (e *Specs) ApplyFocus(description string, focus, skip []string) {
+	if len(focus)+len(skip) == 0 {
+		e.applyProgrammaticFocus()
+	} else {
+		e.applyRegExpFocusAndSkip(description, focus, skip)
+	}
+}
+
+func (e *Specs) applyProgrammaticFocus() {
+	e.hasProgrammaticFocus = false
+	for _, spec := range e.specs {
+		if spec.Focused() && !spec.Pending() {
+			e.hasProgrammaticFocus = true
+			break
+		}
+	}
+
+	if e.hasProgrammaticFocus {
+		for _, spec := range e.specs {
+			if !spec.Focused() {
+				spec.Skip()
+			}
+		}
+	}
+}
+
+// toMatch returns a byte[] to be used by regex matchers.  When adding new behaviours to the matching function,
+// this is the place which we append to.
+func (e *Specs) toMatch(description string, i int) []byte {
+	if i > len(e.names) {
+		return nil
+	}
+	if e.RegexScansFilePath {
+		return []byte(
+			description + " " +
+				e.names[i] + " " +
+				e.specs[i].subject.CodeLocation().FileName)
+	} else {
+		return []byte(
+			description + " " +
+				e.names[i])
+	}
+}
+
+func (e *Specs) applyRegExpFocusAndSkip(description string, focus, skip []string) {
+	var focusFilter, skipFilter *regexp.Regexp
+	if len(focus) > 0 {
+		focusFilter = regexp.MustCompile(strings.Join(focus, "|"))
+	}
+	if len(skip) > 0 {
+		skipFilter = regexp.MustCompile(strings.Join(skip, "|"))
+	}
+
+	for i, spec := range e.specs {
+		matchesFocus := true
+		matchesSkip := false
+
+		toMatch := e.toMatch(description, i)
+
+		if focusFilter != nil {
+			matchesFocus = focusFilter.Match(toMatch)
+		}
+
+		if skipFilter != nil {
+			matchesSkip = skipFilter.Match(toMatch)
+		}
+
+		if !matchesFocus || matchesSkip {
+			spec.Skip()
+		}
+	}
+}
+
+func (e *Specs) SkipMeasurements() {
+	for _, spec := range e.specs {
+		if spec.IsMeasurement() {
+			spec.Skip()
+		}
+	}
+}
+
+//sort.Interface
+
+func (e *Specs) Len() int {
+	return len(e.specs)
+}
+
+func (e *Specs) Less(i, j int) bool {
+	return e.names[i] < e.names[j]
+}
+
+func (e *Specs) Swap(i, j int) {
+	e.names[i], e.names[j] = e.names[j], e.names[i]
+	e.specs[i], e.specs[j] = e.specs[j], e.specs[i]
+}
Author	SHA1	Message	Date
rkervella	2fe9b56dc5	Add support for HTTPS stagers	2022-09-28 07:36:27 -07:00
Ronan Kervella	35907fa929	Merge pull request #916 from BishopFox/go/v1.19 Go/v1.19	2022-09-26 20:04:50 -07:00
moloch--	608de5213a	Update wg GOGARBLE mask	2022-09-26 18:40:29 -07:00
moloch--	b4b9d8d413	Update wg GOGARBLE mask	2022-09-26 18:31:46 -07:00
moloch--	9dfde9d5c9	Update implant vendor/	2022-09-26 18:00:10 -07:00
moloch--	f6d6e9e55e	Unit tests pass	2022-09-26 17:50:00 -07:00
moloch--	0fd10d0c59	go mod tidy/vendor	2022-09-26 16:25:27 -07:00
moloch--	35a7562980	Merge master	2022-09-26 16:24:34 -07:00
rkervella	8ceaff528b	Update garble	2022-09-07 11:04:36 -07:00
rkervella	a5b4e5310f	Update garble	2022-09-06 15:48:17 -07:00
rkervella	3b3987d19f	Update supported minor version	2022-09-06 12:23:26 -07:00
rkervella	6072affe38	Update to go1.19.1	2022-09-06 12:20:21 -07:00
rkervella	555e1d1123	Update Dockerfile and assets	2022-09-06 12:17:10 -07:00
rkervella	e7524d28d0	Implant vendor	2022-09-06 12:09:48 -07:00
rkervella	68515f0873	Update vendor	2022-09-06 12:09:28 -07:00
rkervella	8b55dca890	go1.19	2022-09-06 12:09:19 -07:00