================
@@ -1351,7 +1352,8 @@ def OSConsumesThis : InheritableAttr {
def Cleanup : InheritableAttr {
let Spellings = [GCC<"cleanup">];
- let Args = [DeclArgument<Function, "FunctionDecl">];
+ let Args = [DeclArgument<Function, "FunctionDecl">,
+ ExprArgument<"Expr", /*opt=*/0, /*fake=*/1>];
----------------
erichkeane wrote:
> Ah good call on those already being trailing objects on `ParsedAttr`, so it's
> an even bigger ask than I originally was thinking. Another good reason not to
> go that route yet, that's likely to be a complicated patch because of how
> much code changes (ClangAttrEmitter.cpp, SemaDeclAttr.cpp, probably the
> parser as well) and because of performance concerns (would definitely need to
> be tested for perf changes).
Yep. I'm somewhat talking myself into 2 above being acceptalbe, but I'll hate
it :D
> (I am very much a drive by contributor and this discussion is making me very
> scared of ever getting this landed :p)
*sigh* I don't think we can accept it as-is. But I think with some mild work
to how `DeclArgument` works in tablegen (in a way that causes us to need
exposes the `Expr` as well) would be completely acceptable to us as well.
I THINK I would be OK having us have this Attr store the `SourceLocation` in
`AdditionalMembers` for this one, as long as we had some sort of FIXME esque
comment on it that results in pointing out the above.
https://github.com/llvm/llvm-project/pull/140233
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
