https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71613
--- Comment #9 from joseph at codesourcery dot com <joseph at codesourcery dot com> --- On Mon, 29 Oct 2018, egallager at gcc dot gnu.org wrote: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71613 > > --- Comment #8 from Eric Gallager <egallager at gcc dot gnu.org> --- > (In reply to jos...@codesourcery.com from comment #6) > > Rather than piecemeal fixes with no evidence of completeness, I think we > > should disable the smarts around system header macro locations determining > > whether diagnostics appear (i.e., the system-header tests for disabling > > diagnostics should all use the expansion location), until we have a proper > > design for avoiding such issues and have systematically reviewed all > > diagnostics for conforming to such a design (if the design needs each > > diagnostic function call to be reviewed). > > Coming up with a proper design might never happen though. I'd rather just have > the piecemeal fixes in the meantime. Any such piecemeal fixes should I think be starting from a basis where the warnings appear even if sometimes the real issue is in the header, and then are selectively disabled, rather than from a basis of selectively trying to enable warnings that are wrongly disabled because a system header macro is involved.