dyung wrote:

> > > https://lab.llvm.org/buildbot/#/builders/164/builds/20521 the msan bot 
> > > seems good now, but darwin still seems unhappy 
> > > https://lab.llvm.org/buildbot/#/builders/23/builds/19172. I have one more 
> > > PR I think could solve this, and if not, I'll have to revert the stack.
> > 
> > 
> > Which PR is this? I can take the bot offline and give it a try. 
> > Alternatively, is there anything I can easily do that might help to debug 
> > the issue?
> 
> The pr is in #191644
> 
> I have a revert up also in #191668, which is sad since it was quite the 
> undertaking to get that part of the stack landed.
> 
> As far as debugging ... IDK. I can't repro on an M4 macbook, so this seems 
> either specific to that builder or to x86 darwin. Its frustrating since the 
> msan bot is green now, so I'm not too sure what I can do other than back it 
> all out.
> 
> Likely there is something subtle going on with the memory management and I'm 
> missing it someplace.

I'll give your change a try and see what happens. FWIW, the failure also 
doesn't seem to happen on my AArch64 MacOS builder, so it does seem x86 
specific.

https://github.com/llvm/llvm-project/pull/190054
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to