ymandel added a comment. Sorry for the delay. Read through the whole thread today. I agree with Aaron -- we shouldn't add this until
1. we have a general interface 2. we're sure this serves a general audience As I mentioned in my previous reply, I have a working general interface that I shared and whose implementation I'd be happy to upstream. But, that brings us to point 2, for which I'm on the fence. I think it serves a general audience but I also suspect its really just a frustrating "not quite good enough" solution. To the point of opening up matchers to a wider audience and the cost of `ASTMatchers.h` -- this sounds like a broader issue that we should find a way to address, well beyond this patch. Happy to be in on that discussion, but I don't think I have the cycles now to drive it. :) Regarding CFG matchers -- none that I know of. We've had it as project on our "interesting clang projects" list (in our heads) for a while, but have yet to invest in it. What we *have* started is a dataflow analysis framework: clang/Analysis/FlowSensitive (see the unittests for some example uses). My main use of `hasSubstatementSequence` was to find pointer variables which could be safely converted to `unique_ptr`. I converted that into a dataflow analysis built on the new framework and it's a lot cleaner. CHANGES SINCE LAST ACTION https://reviews.llvm.org/D116518/new/ https://reviews.llvm.org/D116518 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
