On Fri, Aug 5, 2016 at 6:33 PM, Richard Smith <rich...@metafoo.co.uk> wrote:
> These all look OK for the branch. Hans, the second one fixes a crasher, so > I'd like to take at least that one. The others fix warning false positives > (which are not regressions) and look fairly safe. > Correction: looks like the final one is in fact a fix for a diagnostic regression. > On Fri, Aug 5, 2016 at 4:39 PM, Richard Trieu <rtr...@google.com> wrote: > >> Hans, Richard, >> >> These are some last minute diagnostic improvements for 3.9. >> >> http://llvm.org/viewvc/llvm-project?rev=277796&view=rev >> r277796 >> Don't warn when negative values, like -1, are used to initialize unsigned >> bit fields. >> >> http://llvm.org/viewvc/llvm-project?rev=277797&view=rev >> r277797 >> Fix crash with type alias in Template Type Diffing >> > > (You have a typo Internalal in this.) > > >> http://llvm.org/viewvc/llvm-project?rev=277866&view=rev >> r277866 >> Fix false positive in -Wunsequenced and templates >> >> http://llvm.org/viewvc/llvm-project?rev=277889&view=rev >> r277889 >> Fixes for two false positives in -Wreturn-stack-address dealing with >> parameter variables and reference to pointer types. >> > > (Test has some x3 variables that are unused.) >
_______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits