I don't feel sufficiently strongly to insist - clang-tidy's mostly outside my wheelhouse anyway.
As for how to go about it - my rough approach would be to write a small test case that calls an implicitly-deleted-but-explicitly-defaulted move op, run it, check the diagnostic text, find that in DiagnosticSemaKinds, find where the diagnostic identifier is used in the code, set a breakpoint, break there - then trace back to when/where it got marked as deleted (find the member function of the AST node in question that is used to mark it as deleted - break in there, with the condition that 'this ==' the specific instance you're looking at, etc). Then see if, at the point at which the thing is marked as deleted, you can check that it's not dependent (not a template, etc) and emit the appropriate warning. - Dave On Mon, Apr 11, 2016 at 9:44 AM, Alex Pilkiewicz via cfe-commits < cfe-commits@lists.llvm.org> wrote: > pilki added a comment. > > In http://reviews.llvm.org/D18961#397163, @dblaikie wrote: > > > I'd consider just making this a compiler warning, perhaps? > > > I honestly don't feel competent right now to write compiler warnings: > clang-tidy has a nice, well defined interface. A compiler warning would > force me to dig in the internals of clang. But if you feel strongly about > it and give me a couple of pointers (like a diff that adds a similar > warning, and maybe a pointer to the general area in the code where I would > add it), I can maybe give it a try in the next few days/weeks. > > > http://reviews.llvm.org/D18961 > > > > _______________________________________________ > cfe-commits mailing list > cfe-commits@lists.llvm.org > http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits >
_______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
