On Fri, Feb 17, 2023 at 03:32:09PM -0500, Jason Merrill via Gcc-patches wrote: > Tested x86_64-pc-linux-gnu. This isn't really a regression fix, but it's very > safe, and fixes a longstanding annoyance, so I'm leaning toward putting it > into > GCC 13. Thoughts?
I think it is ok for GCC 13. I bet for the testsuite coverage of the whole DR2518 wording we have already tons of testcases which verify we actually error on #error directives and warn on #warning directives and error in failed static_assert unless in uninstantiated templates that it isn't worth repeating them in the dr2518*.C testcase(s). I'm a little bit surprised we didn't check for errors from static_asserts failed asserts in uninstantiated templates... Jakub