As Joerg mentioned, he believes that the error behavior is the way to go. I think that it is not useful behavior to error out on these warnings. If we wanted this behavior, we would not use linker warnings, we'd outright remove the deprecated symbols.
christos > On Jun 11, 2020, at 6:58 PM, Kamil Rytarowski <ka...@netbsd.org> wrote: > > Signed PGP part > On 12.06.2020 00:55, Christos Zoulas wrote: >> In article <20200611222544.6d3a6f...@cvs.netbsd.org>, >> Joerg Sonnenberger <source-changes-d@NetBSD.org> wrote: >>> -=-=-=-=-=- >>> >>> Module Name: src >>> Committed By: joerg >>> Date: Thu Jun 11 22:25:44 UTC 2020 >>> >>> Modified Files: >>> src/common/lib/libprop: prop_object_impl.h >>> >>> Log Message: >>> Unbreak clang builds by removing questionable linker warning sections >>> trggered all over the place. >> >> Why don't you do this just for clang, so that we can use gcc to track the >> remaining ones down and finish the job? Now we can't easily find them. >> >> christos >> > > Can we please revert this and fix clang? > > I'm strongly for linker warnings as they catch real issues. > > > <sanitizer.log>
Description: Message signed with OpenPGP