We no longer store the docs under a project name (since all the docs are now
built using the root conf.py). This mean the name and version are only used for
packaging and uploading, which typically is only used in CI.
This allows us to lazy load the package name and version, so we only read the
conf.py when we need to.
MozReview-Commit-ID: DV5Jxrbskoh
--HG--
extra : source : d03f75986f626938142bd5bde293773506a2fc14
extra : amend_source : d6d0ff5673515cac3a6c7cf03bfce4ffcd09991f
Previously, running |mach doc <subtree>| would use whatever conf.py file
happened to live in the subtree. For example, running:
./mach doc tools/lint
Would build with tools/lint/docs/conf.py. This is bad because it means the
generated docs will look different from the docs that eventually will be
published to firefox-source-docs.mozilla.com.
This patch makes sure we always use tools/docs/conf.py for building, even when
only generating a subtree.
Furthermore, this sets things up such that when you modify a file, only the
subtree containing the modified file will be re-generated. This cuts down
rebuild times from ~2 minutes to ~20 seconds.
There is one caveat. When rebuilding a subtree, the index of other trees will
be overwritten in that particular subtree. I couldn't figure out anyway around
this. This tradeoff for *much* faster rebuild times seems worth it.
MozReview-Commit-ID: Ly88mvHKpo7
--HG--
extra : source : 47fc3e2238676a40a3adc84239baed1ce873e95e
The current mechanism for reading SPHINX variables assumes we always want to
read metadata for the entire tree. Now that we have the ability to rebuild
specific subtrees, this assumption is false.
This patch allows us to specify a path that find_sphinx_variables can use to
filter down the set of moz.build variables it will traverse, yielding only
moz.builds that could potentially impact the specified path.
MozReview-Commit-ID: ALrCFLFgMLH
--HG--
extra : source : 22f2dc60e6d859d3ca411826c77002d87c1a49bd
Now that we can rebuild docs with the liveserver, there are some optimizations
we should make. One of those is processing the sphinx moz.build variables. This
patch makes sure we don't re-process moz.build if we've already done so in a
previous rebuild.
MozReview-Commit-ID: 2AIr1KeAPQV
--HG--
extra : source : 30b4083534b51213a1b9fe0d86f996cfa0e7fa54
In the mozbuild.sphinx extension, we create a new SphinxManager instance each
time. However this isn't ideal now that we can rebuild the docs within the same
interpreter using the livereload server.
This makes use of a singleton so that we can share state not only between
multiple invocations of sphinx-build, but also with the mach command. This will
be taken advantage of more heavily in future commits in this series.
MozReview-Commit-ID: 7ERYeN5BPeI
--HG--
extra : source : 8309212d820bcca29aa95b7892d39940437f2aa8
These two functions are typically only used by CI for packaging/uploading the
documentation. This is a minor re-organiztion for clarity.
MozReview-Commit-ID: 62UhQhSSkOs
--HG--
extra : source : 9942d2df371989c5dd67a75fd7a695533141dd89
We no longer store the docs under a project name (since all the docs are now
built using the root conf.py). This mean the name and version are only used for
packaging and uploading, which typically is only used in CI.
This allows us to lazy load the package name and version, so we only read the
conf.py when we need to.
MozReview-Commit-ID: DV5Jxrbskoh
--HG--
extra : rebase_source : d227b8a800715c677602cbbc9b868df5dbd31131
Previously, running |mach doc <subtree>| would use whatever conf.py file
happened to live in the subtree. For example, running:
./mach doc tools/lint
Would build with tools/lint/docs/conf.py. This is bad because it means the
generated docs will look different from the docs that eventually will be
published to firefox-source-docs.mozilla.com.
This patch makes sure we always use tools/docs/conf.py for building, even when
only generating a subtree.
Furthermore, this sets things up such that when you modify a file, only the
subtree containing the modified file will be re-generated. This cuts down
rebuild times from ~2 minutes to ~20 seconds.
There is one caveat. When rebuilding a subtree, the index of other trees will
be overwritten in that particular subtree. I couldn't figure out anyway around
this. This tradeoff for *much* faster rebuild times seems worth it.
MozReview-Commit-ID: Ly88mvHKpo7
--HG--
extra : rebase_source : 379ffb482a4d15f5f6b394f93adab2b536c7ce00
The current mechanism for reading SPHINX variables assumes we always want to
read metadata for the entire tree. Now that we have the ability to rebuild
specific subtrees, this assumption is false.
This patch allows us to specify a path that find_sphinx_variables can use to
filter down the set of moz.build variables it will traverse, yielding only
moz.builds that could potentially impact the specified path.
MozReview-Commit-ID: ALrCFLFgMLH
--HG--
extra : rebase_source : d0c26a006bb4dbc429be5eedad7825d4412dc2a4
Now that we can rebuild docs with the liveserver, there are some optimizations
we should make. One of those is processing the sphinx moz.build variables. This
patch makes sure we don't re-process moz.build if we've already done so in a
previous rebuild.
MozReview-Commit-ID: 2AIr1KeAPQV
--HG--
extra : rebase_source : d8034ef5be416975c19a473d5a13f073e95aba80
In the mozbuild.sphinx extension, we create a new SphinxManager instance each
time. However this isn't ideal now that we can rebuild the docs within the same
interpreter using the livereload server.
This makes use of a singleton so that we can share state not only between
multiple invocations of sphinx-build, but also with the mach command. This will
be taken advantage of more heavily in future commits in this series.
MozReview-Commit-ID: 7ERYeN5BPeI
--HG--
extra : rebase_source : 44aee637ea9b828b43b82e8639ddc3cc7f68c797
These two functions are typically only used by CI for packaging/uploading the
documentation. This is a minor re-organiztion for clarity.
MozReview-Commit-ID: 62UhQhSSkOs
--HG--
extra : rebase_source : c5b800734041689453f965e50ce356c2e7a623d2
Also fixes existing code which fails the rule.
MozReview-Commit-ID: CkLFgsspGMU
--HG--
extra : rebase_source : 86a43837659aa2ad83a87eab53b7aa8d39ccf55b
Also use package-lock.json rather than npm-shrinkwrap.json as we no longer need to support the latter.
MozReview-Commit-ID: 4yaFcElvhF7
--HG--
rename : npm-shrinkwrap.json => package-lock.json
extra : rebase_source : ae6fcadd8b4a5213fb94cd6426909c2efebfb6a1
For some reason, GNU as is not happy with the assembly generated after
bug 1238661 anymore on Debian armel.
OTOH, as mentioned in bug 1238661 comment 4, we actually don't need this
workaround anymore, so let's just kill it.
--HG--
extra : rebase_source : 6fd06832136d4f840c65f74b63f1c1bec48d525d
LUL::mAdminThreadId is used only to assert that certain calls into the LUL
object are made on the correct thread. Because those assertions are done
using MOZ_ASSERT, some compilers spot that LUL::mAdminThreadId is unused in
non-debug builds, emit a warning to that effect, and then cause the build to
fail due to the presence of -Werror.
Given that (1) it's unlikely that people will use the profiler in debug
builds, (2) failure of these assertions is likely to lead to deadlocking or
crashing in the profiler, and (3) they don't occur on high-frequency paths,
a good solution seems to be to convert them to MOZ_RELEASE_ASSERTs, hence
causing LUL::mAdminThreadId to be used in even in non-debug builds.
In some cases, related MOZ_ASSERTs relating to LUL::mAdminMode have also
been upgraded to MOZ_RELEASE_ASSERTs, for consistency with the
LUL::mAdminThreadId changes.
--HG--
extra : rebase_source : 554a31060a828db01246ece6d1e3afbcc0b42cd2
Since aarch64's DWARF doesn't have pc register, I use x29 (link register) if
not first frame.
I test by gtest on Linux/aarch64, and profiler works on Android/aarch64.
EM_AARCH64 might not be defined on our builders since headers are old, so
this define is needed.
MozReview-Commit-ID: 8VDb5i0vwBT
--HG--
extra : rebase_source : abfe58624dabc2551deb03527db4be3b93490206
This has the effect of exposing explicit imports (e.g., defineModuleGetter),
but not implicit imports (e.g., Cu.import), to rules like no-unused-vars and
no-shadow.
MozReview-Commit-ID: C8oXoSKMU1s
--HG--
extra : rebase_source : 21e87b48514c2075d09eca7e20a1671b206d871f
extra : histedit_source : 5080aef443f0a21f3a1e3e2bb398aadcc5a05402%2Cc472717700446079b1f7ac552a4c3339d207ad61
This file is not part of the build. (it's used to generate
generated-mochitest.ini)
What's tripping up this lint is this line:
# [gl:0D6DE000] mozilla::gl::GLContext::raw_fDrawArrays: Generated unexpected GL_INVALID_OPERATION error. (0x0502)
This line is picked up as a false-positive, since the too-naive regex
(effectively /#.*\[/) interprets this as a skipped test.
MozReview-Commit-ID: 2bMIkCH8WK9
Going through XPConnect for JS-to-JS access in the blocklist service adds no
benefit, but does add a lot of overhead and maintenance burden.
MozReview-Commit-ID: Lf1mDK0b0B0
--HG--
extra : rebase_source : 410ed3fcf999d7c7775ef4926c89f67d9e342da8
This changes the default to opening a livereload webserver after doc generation
(as opposed to opening the index file). Any changes to the specified path will
result in a rebuild and refresh of the browser.
For example, if you run:
./mach doc tools/lint
The linting docs will be built, served and opened in a browser. Modifying any
file under 'tools/lint/docs' will refresh the browser with your changes.
To disable this behaviour and simply open the index file, you can pass in
'--no-serve'. The '--no-open' flag will continue to work (both with http and
the file system).
One caveat to this patch is that when generating the root docs (by running
|mach doc|), we don't watch all possible doc paths (just the root one under
'tools/docs/'). This will probably be fixed in the follow-up bug 1454640.
MozReview-Commit-ID: FQecuePM0zZ
--HG--
extra : rebase_source : 3240402d7505e99a4f64dada309b1baec78306e1
This removes the ability to specify multiple doc paths at the same time with
|mach doc|. We will be changing the default from opening index files to serving
the documentation with a webserver. Supporting multiple doc roots would mean
spinning up multiple servers in different threads.
This would add a lot of complexity for a feature which I don't think is very
useful. It's very rare that one would need to edit more than one doc location
at the same time. And if this is ever needed, the developer can just build the
entire doctree (by running |mach doc|) or run |mach doc <path>| in multiple
different terminals.
MozReview-Commit-ID: GXEZJSgLpgF
--HG--
extra : rebase_source : 2eda23274eb6c2be82f7e77ca577072386bada34
Some requirements.txt are very large and result in a lot of package already
installed messages. Would be nice to hide this.
MozReview-Commit-ID: FQecuePM0zZ
--HG--
extra : rebase_source : 58eaa7324775cfaa39077871be0be0ef39ad7c11
ProfilerScreenshots encodes these surfaces to JPG data URLs, and submits them
as profiler markers. The encoding is done on a separate thread.
MozReview-Commit-ID: 7CKDBqUsLny
--HG--
extra : rebase_source : 45b608544b4ddf8502302cf974ca4e8b306de64e
These lines of code are within the #else section of the #ifndef MOZ_GECKO_PROFILER
at the top of the file, so MOZ_GECKO_PROFILER is always defined for them.
MozReview-Commit-ID: IxRYexzZH0G
--HG--
extra : rebase_source : 302e5515323a63f145eed75a2b66a04fbde052e5
This defers SIGINT handling to the underlying mozlint subprocess. Mozlint *should*
return quickly and gracefully. This way the hook process itself will not abort, just
the linting. After hitting Ctrl-C, the vcs operation should continue as if the lint
were successful.
If mozlint had already found some errors before hitting Ctrl-C, those should get
printed and could potentially abort the hook.
MozReview-Commit-ID: BKXq1dXuMlB
--HG--
extra : rebase_source : a5f2a80f13ead6b31fc58de815fcf5b749e35d73
Previously, using --workdir or --outgoing could miss errors when modifying a
support file since those could affect unmodified files.
This patch allows linters to define support-files in their .yml configuration.
If using --outgoing or --workdir and a file matching one of those patterns was
modified, we'll lint the entire tree.
MozReview-Commit-ID: CuGLYwQwiWr
--HG--
extra : rebase_source : 00d4107c41404f5e6ab05e0106d5cd377e25652f
Since I left the next two patches to bitrot, I realized that a path I had added
didn't exist anymore. We should definitely error out if non-existant paths are
specified, otherwise the lists will become outdated and it will be possible to
accidentally disable linting on some files.
I discovered a few instances of this already in our existing definitions.
MozReview-Commit-ID: 8jsTKLI0nFE
--HG--
extra : rebase_source : acceb0b129fc472fb456ff527e4c8c52228edd59
Enables |./mach try fuzzy --talos-profile|. This template only applies to talos
tasks. It also provides --geckoProfile for consistency with |mach try syntax|,
but I don't like this name so it's hidden from the help.
The 'talos-profile.yml' template is also very specific (only applies to Talos
tasks). Ideally I'd like a general 'command.yml' template that just appends
arguments to the command for any arbitrary tasks. But then we'd need to invent
an expression syntax in try_task_config.json so we could make sure it only
applies to Talos. Then I thought rather than implement it for a specific
template, we should have a general way of doing this which could apply to any
and all of the templates.
Needless to say, it's a rabbit hole and something that's best left to a
follow-up so we don't delay this bug.
MozReview-Commit-ID: GhllZ7sr0ar
--HG--
extra : rebase_source : 1de4deecc2f73130904d7c95d4ff12f85883cd91
This is gated by the `--chemspill-prio` flag, which should at least make anyone
abusing it to get faster results feel sorry for what they've done.
MozReview-Commit-ID: J4EwH45IkMX
--HG--
extra : rebase_source : 1bfbfafd7de914aaab52f48f0e37c09c0df05dd7
FxAccountsManager has been removed from the tree and now one else listened to
this event.
MozReview-Commit-ID: By7kkWbwBvW
--HG--
extra : rebase_source : 35b1ce32859e21b92d060835362400d1a356e81d
To add test for startProfiler's parameters on devtools, nsIProfilerStartParams
changes to scriptable.
MozReview-Commit-ID: 3Rf39NzsND1
--HG--
extra : rebase_source : 2b4fc908e7c611a36ac8766b9ce7ccd55f2c5620
In order to remove support for non-bootstrapped extensions, the remaining test
automation extensions need to be migrated to bootstrapped extensions. These
extensions all work by loading a single component, either with a
profile-after-change or command line handler. This is a straightforward
conversion of those components to bootstrap.js scripts.
MozReview-Commit-ID: 5uyNSqRPIVR
--HG--
rename : services/sync/tps/extensions/tps/components/tps-cmdline.js => services/sync/tps/extensions/tps/bootstrap.js
rename : testing/talos/talos/pageloader/components/tp-cmdline.js => testing/talos/talos/pageloader/bootstrap.js
rename : testing/talos/talos/startup_test/sessionrestore/addon/SessionRestoreTalosTest.js => testing/talos/talos/startup_test/sessionrestore/addon/bootstrap.js
rename : testing/talos/talos/talos-powers/components/TalosPowersService.js => testing/talos/talos/talos-powers/bootstrap.js
rename : tools/quitter/QuitterObserver.js => tools/quitter/bootstrap.js
extra : rebase_source : 738264d0ec8701aa5cf93c6dac958d88756c1e55
extra : source : 71d65d6b64997bc608a1cb3c924be35bbd460b39
Move and rename the server's process script (and accompanying JSM) that starts
DevTools for an entire content process from `content-process-debugger-server.js`
to `startup/content-process.js`. `connectToContent` also becomes the more
specific `connectToContentProcess`.
These code paths will likely change more as Site Isolation work continues, but
for now, we have this light cleanup to gather startup-related paths together.
MozReview-Commit-ID: 1evbZMB8T7r
--HG--
rename : devtools/server/content-process-debugger-server.js => devtools/server/startup/content-process.js
rename : devtools/server/content-server.jsm => devtools/server/startup/content-process.jsm
extra : rebase_source : e077dd3dc915ec274f866d53d3539909f8440de1
In order to remove support for non-bootstrapped extensions, the remaining test
automation extensions need to be migrated to bootstrapped extensions. These
extensions all work by loading a single component, either with a
profile-after-change or command line handler. This is a straightforward
conversion of those components to bootstrap.js scripts.
MozReview-Commit-ID: 5uyNSqRPIVR
--HG--
rename : services/sync/tps/extensions/tps/components/tps-cmdline.js => services/sync/tps/extensions/tps/bootstrap.js
rename : testing/talos/talos/pageloader/components/tp-cmdline.js => testing/talos/talos/pageloader/bootstrap.js
rename : testing/talos/talos/startup_test/sessionrestore/addon/SessionRestoreTalosTest.js => testing/talos/talos/startup_test/sessionrestore/addon/bootstrap.js
rename : testing/talos/talos/talos-powers/components/TalosPowersService.js => testing/talos/talos/talos-powers/bootstrap.js
rename : tools/quitter/QuitterObserver.js => tools/quitter/bootstrap.js
extra : rebase_source : 2df7dce1afe4e9e4054d12adfc1553e5f5d417e2
You can still run them on a --disable-stylo build, as long as that works
(presumably not for long).
I think I haven't missed anything, but please double-check.
MozReview-Commit-ID: 3BIAEjgTLo5
Also change eslint-plugin-mozilla's globals.js to use eslint-scope rather than escope as
ESLint 4.x changed to eslint-scope. This avoids dependency issues that were masked by the
depedencies of eslint-plugin-no-unsanitized's previous version.
MozReview-Commit-ID: 6wLY7Oj0am3
--HG--
extra : rebase_source : 863313d7907e8246d1f5b0ee83164cc0f943f8f6
The shims that this rule tests for no longer exist.
MozReview-Commit-ID: DMgP7Hczavc
--HG--
extra : rebase_source : 765ddd5c62c9449c07ed050e44d86a3bd5c0ae64
extra : amend_source : 627a7694ac07182200f876901ded7a34721cd228
There a few pieces needed here to properly handle KeyboardInterrupts.
1. All in-progress work needs to abort. Ideally the underlying linters will be
able to catch KeyboardInterrupt, and return partial results (like the flake8
linter does). Linters may alternatively allow the KeyboardInterrupt to
propagate up. Mozlint will catch and handle this appropriately, though any
results found will be lost. The only change to this behaviour was fixing a bug
in the flake8 linter.
2. Any unstarted jobs need to be canceled. In concurrent.futures, there are two
different queues. First, jobs are placed on the work queue, which is just a list
maintained by the parent process. As workers become available, jobs are moved
off the work queue, and onto the call queue (which is a multiprocessing.Queue).
Jobs that live on the work queue can be canceled with 'future.cancel()', whereas
jobs that live on the call queue cannot. The number of extra jobs that are stored
on the call queue is determined by this variable:
https://hg.mozilla.org/mozilla-central/file/deb7714a7bcd/third_party/python/futures/concurrent/futures/process.py#l86
In this patch, the parent process' sigint handler (which will be called on Ctrl-C)
is responsible for canceling all the jobs on the work queue. For the jobs on the
call queue, the best we can do is set a global variable that tells workers to
abort early.
3. Idle workers should exit gracefully. When there are no more jobs left, workers
will block on the call queue (either waiting for more jobs, or waiting for the
executor to send the shutdown signal). If a KeyboardInterrupt is received while a
worker is blocking, it isn't possible to intercept that anywhere (due to quirks
of how concurrent.futures is implemented). The InterruptableQueue class was
created to solve this problem. It will return None instead of propagating
KeyboardInterrupt. A None value will wake the worker up and tell it to gracefully
shutdown. This way, we avoid cryptic tracebacks in the output.
With all of these various pieces solved, pressing Ctrl-C appears to always exit
gracefully, sometimes even printing partial results.
MozReview-Commit-ID: 36Pe3bbUKmk
--HG--
extra : rebase_source : d4c312ee5cc3679eeee1407c5521aed679f84ad4
extra : source : a93a00141bf62f6bc9e30934c0e56f6b2e434bf0
Thunderbird is currently still building on macOS, where `mktemp` requires an
argument.
MozReview-Commit-ID: DCoZAYA6tTI
--HG--
extra : rebase_source : 48775eba4f934fd244a74da615bc7754370b504e
The overlay was responsible for script loading and defining three elements
(bhTooltip, placesCommands, placesContext). In the majority of places where
the overlay was included only part of it was used. To remove the overlay, the
elements were each split into include files and moved into where they
were used. For the scripts, a JS file was added that defines all the lazy
modules and then this script, globalOverlay.js and utilityOverlay.js were
inlined to everywhere that would have included them from the overlay.
MozReview-Commit-ID: 8T5D46oYWLn
--HG--
rename : browser/components/places/content/placesOverlay.xul => browser/components/places/content/placesCommands.inc.xul
rename : browser/components/places/content/placesOverlay.xul => browser/components/places/content/placesContextMenu.inc.xul
extra : rebase_source : c1071af4ea264a95183cbc65caae98feb23d58e5
Move the main contents of editBookmarkOverlay.xul into an include file and
inline the DTD and CSS files where used. Convert several chrome tests to
browser tests since the preprocessor is hard to use within the testing
framework.
MozReview-Commit-ID: DpPBOpZSuBN
--HG--
rename : browser/components/places/content/editBookmarkOverlay.js => browser/components/places/content/editBookmark.js
rename : browser/components/places/content/editBookmarkOverlay.xul => browser/components/places/content/editBookmarkPanel.inc.xul
rename : browser/components/places/tests/chrome/test_bug427633_no_newfolder_if_noip.xul => browser/components/places/tests/browser/browser_bug427633_no_newfolder_if_noip.js
rename : browser/components/places/tests/chrome/test_bug485100-change-case-loses-tag.xul => browser/components/places/tests/browser/browser_bug485100-change-case-loses-tag.js
rename : browser/components/places/tests/chrome/test_bug631374_tags_selector_scroll.xul => browser/components/places/tests/browser/browser_bug631374_tags_selector_scroll.js
rename : browser/components/places/tests/chrome/test_editBookmarkOverlay_keywords.xul => browser/components/places/tests/browser/browser_editBookmark_keywords.js
rename : browser/components/places/tests/chrome/test_editBookmarkOverlay_tags_liveUpdate.xul => browser/components/places/tests/browser/browser_editBookmark_tags_liveUpdate.js
rename : browser/themes/linux/places/editBookmarkOverlay.css => browser/themes/linux/places/editBookmark.css
rename : browser/themes/osx/places/editBookmarkOverlay.css => browser/themes/osx/places/editBookmark.css
rename : browser/themes/windows/places/editBookmarkOverlay.css => browser/themes/windows/places/editBookmark.css
extra : rebase_source : aca072691251c1a44e82ab8091796abd2b140e22
This essentially tests tools/lint/python/flake8.py. Though it also adds
a basic framework for testing all the other linters as well. Getting this
added now will allow others to collaborate on adding more tests without
needing to get to 100% coverage for all linters right off the bat.
All python tests under tools/lint/test will run as part of the 'ml' task
on Linux, and the build task on Windows (OSX coverage is currently missing
for python tests).
The flake8 linter currently has a bug where the 'exclude' argument is
ignored. This is why we need to also exclude 'tools/lint/test/files' in
topsrcdir/.flake8, even though it is already listed in the
'mach_commands.py'. Other linters shouldn't need to do this, the exclusion
in 'mach_commands.py' should be good enough. See bug 1277851 for more
details.
MozReview-Commit-ID: 9ho8C83eeuj
--HG--
extra : rebase_source : 0adb01aa81c06f26cb891ae9f92a17f8f9fdf111
MinGW headers are all lowercase, and because we build FF with MinGW
on Linux - capitalization matters. But if you're plugging along,
working on Windows, you might forget about that, and wind up
inadvertently breaking the MinGW build because you capitalized
an include file (as is common in Windows development.)
This lint job takes the list of header files from MinGW
(generated by cd mingw-w64/mingw-w64-headers &&
find . -name "*.h" | xargs -I bob -- basename bob | sort | uniq)
and will alert if they are in an #include statement but not
all-lowercase.
MozReview-Commit-ID: 4QFSdHx5Uak
--HG--
extra : rebase_source : d76c258191d4cbb7665775bd1b8d79028890727f
console.assert keeps the same semantics as NS_ASSERT in that it doesn't throw an exception,
but a lot of the places code was using it in a way that would be better served by throwing
an exception when the condition is false.
MozReview-Commit-ID: DEF5HSfYO36
This changeset changes behavior.
If the profile is streamed before any JSContext has gone away, we now iterate
over the entire buffer twice (per thread): First, to collect information about
JIT frames, and then again when we build the JSON for the samples. The first
traversal stores small pieces of JSON for JIT fromes in individual strings, and
the second iteration splices those strings into the thread JSON's frame table.
When the JSContext for a thread goes away, we no longer build JSON for samples,
and we don't reset the profiler buffer. We now only build the JSON for JIT
frames. Once the complete profile is requested and we build samples for it, we
iterate over the entire buffer, and look up the cached JIT frame information for
JitReturnAddr entries from the correct range. Different parts of the buffer may
correspond to the life time of different JSContexts: For each JSContext we will
have one range in the JITFrameInfo, and we can look up the correct range based
on the buffer position of the JitReturnAddr entry that we're processing.
This new way of doing things has multiple advantages:
- We no longer reset the buffer, so we no longer lose information about other
threads.
- All threads from a given process now always have sample data for the same
time range. Before this change, the "partial profile" from a thread that
lost its JSContext could extend further into the past than the other threads'
profiles.
- Requesting profiles multiple times now has more consistent results. Before
this change, the first requested profile would include the partial profile,
but then the partial profile was discarded. And the second requested profile
would not contain any data for the time before the JSContext went away.
- We now do less work when a thread's JSContext goes away. This should
decrease the interruption time.
MozReview-Commit-ID: 3KhnPtBijna
--HG--
extra : rebase_source : 2ef5ac933e4db1c98526a2b36147ff031893de9e
extra : intermediate-source : d63b04327077d1ef55f509b365cd2693905b0733
extra : source : f2d8c993aa2c16bc491c55179d545f5a2c727391
In an upcoming patch we'll want to initialize mUniqueStrings by stealing
another object's UniqueJSONStrings, but UniqueJSONStrings itself is not
move-constructible. But UniquePtr is.
Making UniqueJSONStrings itself move-constructible would be a bit tricky
because it has a SpliceableChunkedJSONWriter which is not move-constructible;
and making SpliceableChunkedJSONWriter move-constructible is hard because
there's no obvious "empty but valid" state that we could leave a moved-out-of
SpliceableChunkedJSONWriter in; for example, it expects to have a non-null
WriteFunc at all times.
MozReview-Commit-ID: Q6o61HFTiD
--HG--
extra : rebase_source : e1073be9892cf7cfd6ca4f2562ce939690b3b4d7
This also renames FlushSamplesAndMarkers to NotifyAboutToLoseJSContext.
MozReview-Commit-ID: FWinMi85yDZ
--HG--
extra : rebase_source : 6d8bfd6937ce757108f80f43e878a2dbead318bd
This makes it clear which combinations of fields are possible.
MozReview-Commit-ID: C3PriO7nWsJ
--HG--
extra : rebase_source : 68df01f11121b09b2f2762581dc28184262abfb8
I think this file was picking up such a declaration from a different file
already, through unified builds.
MozReview-Commit-ID: 9xTB2QA86U0
--HG--
extra : rebase_source : d7b614524ebdf25a185698b6632d6f925ce08da8
nsCString is the idiomatic string class we want to use anyway. There's no need
to think about std::string here.
MozReview-Commit-ID: 1rjZAyqp13d
--HG--
extra : rebase_source : 92c4eb22876c4fc89f00104c56f92a75053284e9
nsGenericHashKey never calls this operator, as far as I can tell.
MozReview-Commit-ID: NDybMI8DOA
--HG--
extra : rebase_source : 933c12e2ef4600ea33b6c7026f539e51447fc454
The file copies and renames make this patch look a bit confusing. Here's what
happens:
ThreadInfo.h:
Most of the code gets moved into RegisteredThread.h and ProfiledThreadData.h,
but a small piece remains in ThreadInfo.h.
ThreadInfo.cpp:
Gets split into RegisteredThread.cpp and ProfiledThreadData.cpp.
ThreadInfo.cpp itself goes away.
In the mercurial changeset, I've marked ThreadInfo.h as being copied to both
RegisteredThread.h and to ProfiledThreadData.h, and ThreadInfo.cpp as being
copied to RegisteredThread.cpp and as being renamed to ProfiledThreadData.cpp.
MozReview-Commit-ID: 1j1imAv9cTd
--HG--
rename : tools/profiler/core/ThreadInfo.cpp => tools/profiler/core/ProfiledThreadData.cpp
rename : tools/profiler/core/ThreadInfo.h => tools/profiler/core/ProfiledThreadData.h
rename : tools/profiler/core/ThreadInfo.cpp => tools/profiler/core/RegisteredThread.cpp
rename : tools/profiler/core/ThreadInfo.h => tools/profiler/core/RegisteredThread.h
extra : rebase_source : 91310691d5e46246ec03305511c1b86ab458b0b9
The overlay elements with children of editMenuOverlay.xul are moved into
include files (editMenuCommands.inc.xul and editMenuKeys.inc.xul). For
the other single elements in the overlay, the attributes are inlined
wherever they are used.
MozReview-Commit-ID: 792cuzUvQxT
--HG--
extra : rebase_source : 58e4c05bde16cee873d37c6198de102d048499c2
As of Bug 1425463 it's available in all contexts (including JSM). The following
patche is going to remove imports to Console.jsm, so prepare for this by allowing
it to be used as a global in eslint.
MozReview-Commit-ID: 3gCIspnlVgB
--HG--
extra : rebase_source : 6da75d42d7d937b628ef1d9c0d4e349fd76eda36