[PATCH] D65928: [clangd] Added an early return from VisitMemberExpr in SemanticHighlighting if the MemberLoc is invalid.

[Haojian Wu via Phabricator via cfe-commits]

hokein accepted this revision.
hokein added inline comments.
This revision is now accepted and ready to land.


================
Comment at: clang-tools-extra/clangd/SemanticHighlighting.cpp:56
+    if (isa<CXXConversionDecl>(MD))
+      // If the member is a conversion operator the loc will be invalid. This
+      // causes the addToken function to emit a lot of error logs about trying
----------------
jvikstrom wrote:
> hokein wrote:
> > Maybe just `The MemberLoc is invalid for C++ conversion operato , we don't attempt to add a token for an invalid location`?
> > 
> > 
> > Does the location is always invalid? or just for builtin types? e.g.
> > ```
> > class Foo {};
> > struct Bar {
> >   explicit operator Foo*() const; // 1
> >   explicit operator int() const; // 2
> > };
> > ```
> Builtin types has nothing to do with it. It's invalid for every conversion operator. Will actually add a test just to make sure that everything else is still highlighted correctly.
ah, I thought it is the operator declaration, but here we mean expression.


================
Comment at: clang-tools-extra/clangd/unittests/SemanticHighlightingTests.cpp:268
+        $Class[[Bar]] $Variable[[B]];
+        $Class[[Foo]] $Variable[[F]] = $Variable[[B]];
+        $Class[[Foo]] *$Variable[[FP]] = ($Class[[Foo]]*)$Variable[[B]];
----------------
nit: could you add a comment describing the purpose of this test?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D65928/new/

https://reviews.llvm.org/D65928