rsmith added a comment.
In D59885#1496734 <https://reviews.llvm.org/D59885#1496734>, @ilya-biryukov
wrote:
> Overall, the performance cost in empty-callback case seems to be lower than
> `5%`. This is significant, but hopefully acceptable.
5% is a lot to pay for something that we won't be using at all during normal
preprocessing / compilation. But I think we can simplify this a bit, and that
might help.
================
Comment at: clang/lib/Lex/PPCaching.cpp:64
ExitCachingLexMode();
- Lex(Result);
+ Lex(Result, Report);
----------------
This change seems redundant: `LexLevel` is always `1` here, so this token would
never be reported anyway. And with that gone I think you can remove the
`Report` parameter entirely.
================
Comment at: clang/lib/Lex/Preprocessor.cpp:892
do {
switch (CurLexerKind) {
case CLK_Lexer:
----------------
Doesn't this always set `Report` to the same value as `IsNewToken`? (The only
case we set `Report` to `false` is when we call `CachingLex` and it sets
`IsNewToken` to `false`, and `CachingLex` can't be recursively called twice, so
its recursive call to `Lex` can't set `Report` to `false`..)
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D59885/new/
https://reviews.llvm.org/D59885
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
