AaronBallman wrote: > @AaronBallman This is not the first time there is a need/desire of treating > linux differently (see #226775 for another recent example), So i really like > your idea of `-flinux-kernel`. > > I do think the behavior in that PR should not be exposed by any other mean, > there has been no evidence than this flag is useful outside of the linux > kernel, yet it has been exposed and documented as something folks might want > to use. > > Which I think is either confusing or dangerous, and certainly a bother to > maintain. > > So, I'd suggest reworking the pr to introduce `-flinux-kernel` instead, > which, afaict, would only requires renaming the flag and reworking the > documentation. > > It would further ease the discussions the next time something like that comes > around (of course discussions still will need to happen as there is still a > cost to adds behavior upstream even if there is a single flag for all of > them). > > Thanks!
I'm on the fence about requiring `-flinux-kernel` as part of this PR, but mostly because that feels like the sort of thing we'd have an RFC for. But thinking about that RFC, that's basically "do we want to support the Linux kernel?" and I believe the community has already answered that with "yes, we do." and so an RFC doesn't feel as necessary. If we all agree an RFC doesn't seem necessary, then I think that's a reasonable request for this PR (though I think the flag should be documented explicitly in the language extensions page, so there's a bit more work to be done here, just not much). My feeling is that we wouldn't want `-fint64-is-long-long` as a separate flag in that case, just `-flinux-kernel`. https://github.com/llvm/llvm-project/pull/234780 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
