arphaman added inline comments.
================ Comment at: lib/Tooling/Refactoring/ASTSelection.cpp:164 + unsigned NumMatches = 0; + for (Decl *D : Context.getTranslationUnitDecl()->decls()) { + if (ObjCImplEndLoc.isValid() && ---------------- klimek wrote: > arphaman wrote: > > klimek wrote: > > > Why don't we do this as part of TraverseDecl() in the visitor? > > I think it's easier to handle the Objective-C `@implementation` logic here, > > unless there's some better way that I can't see ATM. > Ok, in that case, can you write a comment at the start of the loop explaining > that we basically only do that for the Objective-C @implementation? (I'd also > like to understand that better in general, as I have no clue about Obj-C :) Hmm, maybe it would be better to move this logic to another layer. Like a wrapper around `RecursiveASTVisitor` that ensures that iteration occurs in a lexical order. It can then be used by other things that might need this, this code will get simpler and I will be able to test it better. Repository: rL LLVM https://reviews.llvm.org/D35012 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits