From 3b95ab569345028a1a8fe521d5ecd81fa97f2653 Mon Sep 17 00:00:00 2001
From: Aaron Orenstein <aorenste@fb.com>
Date: Thu, 17 Aug 2017 18:27:03 -0700
Subject: [PATCH] Performance: Improve efficiency of source file lookup in
 cmMakefile

Add an unordered map to cmMakefile to speed up GetSource() lookups.
---
 Source/cmMakefile.cxx | 50 ++++++++++++++++++++++++++++++++++++++++---
 Source/cmMakefile.h   | 11 ++++++++++
 Source/cmTarget.cxx   |  1 +
 3 files changed, 59 insertions(+), 3 deletions(-)

diff --git a/Source/cmMakefile.cxx b/Source/cmMakefile.cxx
index e51cfccaa6..c96b89207f 100644
--- a/Source/cmMakefile.cxx
+++ b/Source/cmMakefile.cxx
@@ -3082,9 +3082,18 @@ void cmMakefile::SetArgcArgv(const std::vector<std::string>& args)
 cmSourceFile* cmMakefile::GetSource(const std::string& sourceName) const
 {
   cmSourceFileLocation sfl(this, sourceName);
-  for (cmSourceFile* sf : this->SourceFiles) {
-    if (sf->Matches(sfl)) {
-      return sf;
+
+#if defined(_WIN32) || defined(__APPLE__)
+  const auto& name = cmSystemTools::LowerCase(sfl.GetName());
+#else
+  const auto& name = sfl.GetName();
+#endif
+  auto sfsi = this->SourceFileSearchIndex.find(name);
+  if (sfsi != this->SourceFileSearchIndex.end()) {
+    for (auto sf : sfsi->second) {
+      if (sf->Matches(sfl)) {
+        return sf;
+      }
     }
   }
   return nullptr;
@@ -3098,6 +3107,41 @@ cmSourceFile* cmMakefile::CreateSource(const std::string& sourceName,
     sf->SetProperty("GENERATED", "1");
   }
   this->SourceFiles.push_back(sf);
+
+  auto name = sf->GetLocation().GetName();
+#if defined(_WIN32) || defined(__APPLE__)
+  name = cmSystemTools::LowerCase(name);
+#endif
+
+  // For a file in the form "a.b.c" add the cmSourceFile to the index
+  // at "a.b.c", "a.b" and "a".
+  auto partial = name;
+  while (true) {
+    this->SourceFileSearchIndex[partial].insert(sf);
+    auto i = partial.rfind('.');
+    if (i == std::string::npos) {
+      break;
+    }
+    partial = partial.substr(0, i);
+  }
+
+  if (sf->GetLocation().ExtensionIsAmbiguous()) {
+    // For an ambiguous extension also add the various "known"
+    // extensions to the original filename.
+
+    const auto& srcExts = this->GetCMakeInstance()->GetSourceExtensions();
+    for (const auto& ext : srcExts) {
+      auto name_ext = name + "." + cmSystemTools::LowerCase(ext);
+      this->SourceFileSearchIndex[name_ext].insert(sf);
+    }
+
+    const auto& hdrExts = this->GetCMakeInstance()->GetHeaderExtensions();
+    for (const auto& ext : hdrExts) {
+      auto name_ext = name + "." + cmSystemTools::LowerCase(ext);
+      this->SourceFileSearchIndex[name_ext].insert(sf);
+    }
+  }
+
   return sf;
 }
 
diff --git a/Source/cmMakefile.h b/Source/cmMakefile.h
index 272522c9ec..2cae659b65 100644
--- a/Source/cmMakefile.h
+++ b/Source/cmMakefile.h
@@ -13,6 +13,7 @@
 #include <stddef.h>
 #include <string>
 #include <unordered_map>
+#include <unordered_set>
 #include <vector>
 
 #include "cmAlgorithms.h"
@@ -808,7 +809,17 @@ protected:
   // libraries, classes, and executables
   mutable cmTargets Targets;
   std::map<std::string, std::string> AliasTargets;
+
   std::vector<cmSourceFile*> SourceFiles;
+  // Because cmSourceFile names are compared in a fuzzy way (see
+  // cmSourceFileLocation::Match()) we can't have a straight mapping from
+  // filename to cmSourceFile.  To make lookups more efficient we store the
+  // Name portion of the cmSourceFileLocation and then compare on the list of
+  // cmSourceFiles that might match that name.  Note that on platforms which
+  // have a case-insensitive filesystem we store the key in all lowercase.
+  typedef std::unordered_set<cmSourceFile*> SourceFileSet;
+  typedef std::unordered_map<std::string, SourceFileSet> SourceFileMap;
+  SourceFileMap SourceFileSearchIndex;
 
   // Tests
   std::map<std::string, cmTest*> Tests;
diff --git a/Source/cmTarget.cxx b/Source/cmTarget.cxx
index cb09156b96..49b0664640 100644
--- a/Source/cmTarget.cxx
+++ b/Source/cmTarget.cxx
@@ -5,6 +5,7 @@
 #include "cmsys/RegularExpression.hxx"
 #include <algorithm>
 #include <assert.h>
+#include <iterator>
 #include <map>
 #include <set>
 #include <sstream>