nvetrini wrote: The motivation for this was not the diagnostic, but rather the use of the API to get an accurate location: the diagnostic is used just to have an easy way to test it.
@ojhunt It seems to me that you are proposing to get the end loc of the previous expression, then starting from that token re-parse the content of the remaining expr to find the first comma, to avoid passing it down in the ParenListExpr directly, right? In that way the change would be limited to `ParseSimpleExpressionList` and `MaybeConvertParenListExprToParenExpr`. I can see how that would work for comma exprs (there is only one comma), though I'm a bit wary about its precision on other kinds of expressions. Am I understanding your suggestion correctly? https://github.com/llvm/llvm-project/pull/199411 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
