From aebf2dfaed4f0e7e92b21312d87ae1cbcb3fafff Mon Sep 17 00:00:00 2001
From: Fangrui Song <i@maskray.me>
Date: Mon, 3 Sep 2018 12:21:32 -0700
Subject: [PATCH] Add index.multiVersion{Black,White}list

They allow files matching specified patterns to be indexed only once
even if index.multiVersion > 0. An example is to exclude system headers
with index.multiVersionBlacklist: ["^/usr/include"]
---
 src/config.h               |  8 +++++++-
 src/indexer.cc             | 18 +++++++++++++++++-
 src/indexer.h              |  1 +
 src/messages/initialize.cc |  1 +
 4 files changed, 26 insertions(+), 2 deletions(-)

diff --git a/src/config.h b/src/config.h
index 67cf0972..c22a99f8 100644
--- a/src/config.h
+++ b/src/config.h
@@ -191,6 +191,11 @@ struct Config {
     // If not 0, a file will be indexed in each tranlation unit that includes it.
     int multiVersion = 0;
 
+    // If multiVersion != 0, files that match blacklist but not whitelist will
+    // still only be indexed for one version.
+    std::vector<std::string> multiVersionBlacklist;
+    std::vector<std::string> multiVersionWhitelist;
+
     // Allow indexing on textDocument/didChange.
     // May be too slow for big projects, so it is off by default.
     bool onDidChange = false;
@@ -237,7 +242,8 @@ MAKE_REFLECT_STRUCT(Config::Diagnostics, blacklist, frequencyMs, onOpen,
                     onType, whitelist)
 MAKE_REFLECT_STRUCT(Config::Highlight, lsRanges, blacklist, whitelist)
 MAKE_REFLECT_STRUCT(Config::Index, blacklist, comments, enabled, multiVersion,
-                    onDidChange, reparseForDependency, threads, whitelist);
+                    multiVersionBlacklist, multiVersionWhitelist, onDidChange,
+                    reparseForDependency, threads, whitelist);
 MAKE_REFLECT_STRUCT(Config::WorkspaceSymbol, caseSensitivity, maxNum, sort);
 MAKE_REFLECT_STRUCT(Config::Xref, container, maxNum);
 MAKE_REFLECT_STRUCT(Config, compilationDatabaseCommand,
diff --git a/src/indexer.cc b/src/indexer.cc
index 4f3849e2..ff711a0c 100644
--- a/src/indexer.cc
+++ b/src/indexer.cc
@@ -17,6 +17,7 @@ limitations under the License.
 
 #include "clang_tu.h"
 #include "log.hh"
+#include "match.h"
 #include "platform.h"
 #include "serializer.h"
 using ccls::Intern;
@@ -44,8 +45,11 @@ namespace {
 
 constexpr int kInitializerMaxLines = 3;
 
+GroupMatch *multiVersionMatcher;
+
 struct IndexParam {
   std::unordered_map<llvm::sys::fs::UniqueID, std::string> SeenUniqueID;
+  std::unordered_map<llvm::sys::fs::UniqueID, bool> UID2multi;
   std::unordered_map<std::string, FileContents> file_contents;
   std::unordered_map<std::string, int64_t> file2write_time;
   struct DeclInfo {
@@ -81,6 +85,13 @@ struct IndexParam {
     SeenFile(FE);
     return file_consumer->TryConsumeFile(FE, &file_contents);
   }
+
+  bool UseMultiVersion(const FileEntry &FE) {
+    auto it = UID2multi.try_emplace(FE.getUniqueID());
+    if (it.second)
+      it.first->second = multiVersionMatcher->IsMatch(FileName(FE));
+    return it.first->second;
+  }
 };
 
 StringRef GetSourceInRange(const SourceManager &SM, const LangOptions &LangOpts,
@@ -639,7 +650,7 @@ public:
       return true;
     int lid = -1;
     IndexFile *db;
-    if (g_config->index.multiVersion) {
+    if (g_config->index.multiVersion && param.UseMultiVersion(*FE)) {
       db = param.ConsumeFile(*SM.getFileEntryForID(SM.getMainFileID()));
       if (!db)
         return true;
@@ -1180,6 +1191,11 @@ template <typename T> void Uniquify(std::vector<T> &a) {
 }
 
 namespace ccls::idx {
+void Init() {
+  multiVersionMatcher = new GroupMatch(g_config->index.multiVersionWhitelist,
+                                       g_config->index.multiVersionBlacklist);
+}
+
 std::vector<std::unique_ptr<IndexFile>>
 Index(VFS *vfs, const std::string &opt_wdir, const std::string &file,
       const std::vector<std::string> &args,
diff --git a/src/indexer.h b/src/indexer.h
index d1b1e712..355b5a98 100644
--- a/src/indexer.h
+++ b/src/indexer.h
@@ -277,6 +277,7 @@ struct IndexFile {
 };
 
 namespace ccls::idx {
+void Init();
 std::vector<std::unique_ptr<IndexFile>>
 Index(VFS *vfs, const std::string &opt_wdir, const std::string &file,
       const std::vector<std::string> &args,
diff --git a/src/messages/initialize.cc b/src/messages/initialize.cc
index b687b6a9..cfbca19e 100644
--- a/src/messages/initialize.cc
+++ b/src/messages/initialize.cc
@@ -478,6 +478,7 @@ struct Handler_Initialize : BaseMessageHandler<In_InitializeRequest> {
                                 EscapeFileName(g_config->projectRoot));
 
     diag_pub->Init();
+    idx::Init();
     semantic_cache->Init();
 
     // Open up / load the project.