boomanaiden154 wrote: > GCC has a related -fno-lifetime-dse option, possibly we should be integrating > with that flag instead of creating a new one? (Notably, LLVM sets that flag > because we have some hacks in the User implementation that rely on accessing > the object in overloaded operator delete, see > https://github.com/llvm/llvm-project/issues/24952).
Reusing the existing flag name sounds reasonable enough to me, but in #40040 it seems like people had opinions on the flag naming, with @zygoloid suggesting `-fno-strict-lifetimes` for the opt-out. It looks like a dummy implementation of the flag was proposed in https://reviews.llvm.org/D150930, but never landed. I know in the past we have tried to go for gcc compatibility, so I'm wondering what opinions here are. I think `-fstrict-lifetimes` makes a lot more sense to users, but it might be good to support the gcc naming. I think it should be pretty simple to support both and just alias them. https://github.com/llvm/llvm-project/pull/166276 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
