================
@@ -267,9 +267,11 @@ void Parser::ParseInnerNamespace(const
InnerNamespaceInfoList &InnerNSs,
while (!tryParseMisplacedModuleImport() && Tok.isNot(tok::r_brace) &&
Tok.isNot(tok::eof)) {
ParsedAttributes DeclAttrs(AttrFactory);
- MaybeParseCXX11Attributes(DeclAttrs);
- ParsedAttributes EmptyDeclSpecAttrs(AttrFactory);
- ParseExternalDeclaration(DeclAttrs, EmptyDeclSpecAttrs);
+ ParsedAttributes DeclSpecAttrs(AttrFactory);
+ while (MaybeParseCXX11Attributes(DeclAttrs) ||
+ MaybeParseGNUAttributes(DeclSpecAttrs))
+ ;
----------------
Sirraide wrote:
> If we factor out `ParseExternalDeclaration` in a function that also parses
> the attributes, factoring out attributes parsing is probably not as usefull
> right ?
Yeah, I think just making a new function that both parses attributes and then
calls `ParseExternalDeclaration` seems to be enough.
> Since it's quite a lot of changes, should I do it in two commits ?
The commit history of a pr isn’t really that important because we always squash
on merge anyway, so use however many commits you need—I’d recommend avoiding
force-pushing though because it can make reviews more complicated (I at least
find it harder to figure out what’s changed since my last review if there was a
force-push).
https://github.com/llvm/llvm-project/pull/117148
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
