[PATCH] D77233: [NFC] Refactoring PropertyAttributeKind for ObjCPropertyDecl and ObjCDeclSpec.

Puyan Lotfi via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Tue Apr 14 11:16:26 PDT 2020


plotfi marked an inline comment as done.
plotfi added inline comments.


================
Comment at: clang/include/clang-c/Index.h:4503
   CXObjCPropertyAttr_unsafe_unretained = 0x800,
-  CXObjCPropertyAttr_class = 0x1000
+  CXObjCPropertyAttr_classattr = 0x1000
 } CXObjCPropertyAttrKind;
----------------
erik.pilkington wrote:
> plotfi wrote:
> > erik.pilkington wrote:
> > > plotfi wrote:
> > > > @erik.pilkington  Do you think we should be adding the C-API analogs of nullability / null_resettable (and direct) to clang/include/clang-c/Index.h? 
> > > > 
> > > > I noticed those are missing.
> > > > 
> > > > 
> > > Oh, sorry, I missed this. The C API is supposed to be stable, so I don't think we should change the name of this enumerator (@arphaman can you confirm?). Adding the missing attributes seem fine though. 
> > @erik.pilkington  @arphaman Any chance I can land the C API changes in a follow up commit? That's what I was thinking. 
> I think this enumerator should stay named `CXObjCPropertyAttr_class`, changing the name to `CXObjCPropertyAttr_classattr` is API breaking, and the C API should be stable.
Ah yeah! Good catch. Makes total sense. What do you think about dropping the C macro usage in clang/tools/c-index-test/c-index-test.c (ie PRINT_PROP_ATTR)? I can't drop the prefix in the c++ enum completely without changing class to classattr since it is a keyword. 


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D77233/new/

https://reviews.llvm.org/D77233





More information about the cfe-commits mailing list