[clangd] Fix a crash when reading an empty index file.

[hokein]

Summary

Unfortunately, yaml::Input::setCurrentDocument() and yaml::Input::nextDocument() are
internal APIs, the way we use them may cause a nullptr accessing when
processing an empty YAML file.

[ilya-biryukov]

LGTM. Thanks for fixing this.

A few questions for the long-term:

• Should we consider removing the YAML support altogether?
• If we want to keep, are there non-private APIs that we could use? (I assume we need them to avoid keeping the whole parsed YAML tree in memory, right?)

[hokein]

In D56442#1349666, @ilya-biryukov wrote:

LGTM. Thanks for fixing this.

A few questions for the long-term:

• Should we consider removing the YAML support altogether?

The motivation of the YAML is for debugging purpose (compared with the binary format), we might want to switch to another human-readable format like json.

• If we want to keep, are there non-private APIs that we could use? (I assume we need them to avoid keeping the whole parsed YAML tree in memory, right?)

This is the best way we've found so far. The public YAML APIs don't seem provide enough flexibility to read an Variant element (but I may miss something).

clang-tools-extra/trunk/clangd/index/YAMLSerialization.cpp

 
 llvm::Expected<IndexFileIn> readYAML(llvm::StringRef Data) {
   SymbolSlab::Builder Symbols;
   RefSlab::Builder Refs;
   llvm::BumpPtrAllocator
       Arena; // store the underlying data of Position::FileURI.
   llvm::UniqueStringSaver Strings(Arena);
   llvm::yaml::Input Yin(Data, &Strings);
-  do {
+  while (Yin.setCurrentDocument()) {
+    llvm::yaml::EmptyContext Ctx;
     VariantEntry Variant;
-    Yin >> Variant;
+    yamlize(Yin, Variant, true, Ctx);
     if (Yin.error())
       return llvm::errorCodeToError(Yin.error());
+
     if (Variant.Symbol)
       Symbols.insert(*Variant.Symbol);
     if (Variant.Refs)
       for (const auto &Ref : Variant.Refs->second)
         Refs.insert(Variant.Refs->first, Ref);
-  } while (Yin.nextDocument());
+    Yin.nextDocument();
+  }
 
   IndexFileIn Result;
   Result.Symbols.emplace(std::move(Symbols).build());
   Result.Refs.emplace(std::move(Refs).build());
   return std::move(Result);
 }
 
 std::string toYAML(const Symbol &S) {

clang-tools-extra/trunk/unittests/clangd/SerializationTests.cpp

 )";
 
 MATCHER_P(ID, I, "") { return arg.ID == cantFail(SymbolID::fromStr(I)); }
 MATCHER_P(QName, Name, "") { return (arg.Scope + arg.Name).str() == Name; }
 MATCHER_P2(IncludeHeaderWithRef, IncludeHeader, References, "") {
   return (arg.IncludeHeader == IncludeHeader) && (arg.References == References);
 }
 
+TEST(SerializationTest, NoCrashOnEmptyYAML) {
+  EXPECT_TRUE(bool(readIndexFile("")));
+}
+
 TEST(SerializationTest, YAMLConversions) {
   auto In = readIndexFile(YAML);
   EXPECT_TRUE(bool(In)) << In.takeError();
 
   auto ParsedYAML = readIndexFile(YAML);
   ASSERT_TRUE(bool(ParsedYAML)) << ParsedYAML.takeError();
   ASSERT_TRUE(bool(ParsedYAML->Symbols));
   EXPECT_THAT(