This commit https://reviews.llvm.org/rGe5158b52730d323bb8cd2cba6dc6c89b90cba452 introduced an accidental change, which renames -faccess-control and -fno-access-control to -fno-access-control and -fno-no-access-control.
Details
Details
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
Unit Tests
Unit Tests
Time | Test | |
---|---|---|
70 ms | x64 windows > LLVM.CodeGen/XCore::threads.ll |
Event Timeline
Comment Actions
LGTM. BTW, for partial reverts or potentially time-sensitive fixes like this, it's usually better to land first and ask questions later.
Comment Actions
Ah, I see the original commit was a while ago, so maybe not so time-sensitive.
(It'd be nice to have a test for this too if possible...)
(Or maybe OptOutFFLag could reject an option beginning with "no-"?)
Comment Actions
Thanks! Sorry for the breakage.
Actually, neither -fno-access-control nor -faccess-control has any test..
Comment Actions
(Or maybe OptOutFFLag could reject an option beginning with "no-"?)
I think this is reasonable. I'll reach out to the mailing list to see if there's a way to add assertions to TableGen that would catch this.
Thanks! Sorry for the breakage.
No problem.