[libcxx-commits] [PATCH] D128146: [libc++] Use uninitialized algorithms for vector

Augusto Noronha via Phabricator via libcxx-commits libcxx-commits at lists.llvm.org
Thu Jul 21 16:09:51 PDT 2022


augusto2112 added a comment.

In D128146#3670190 <https://reviews.llvm.org/D128146#3670190>, @philnik wrote:

> @augusto2112 Thank you for the information. I've run lldb locally with my patch and it seems to be fine. I ran everything in the test and AFAICT everything ran and produced the correct results. My lldb is compiled from 36c9e9968affac543952e81637a0584a4b708597 <https://reviews.llvm.org/rG36c9e9968affac543952e81637a0584a4b708597> without assertions enabled. So the bug either doesn't show in the output (maybe the assertion is just wrong?), or the bug has been introduced within the last 3 weeks. Either way it's pretty clear to me that my code is fine. I'd like to re-land this patch with the LLDB test disabled, since D68365 <https://reviews.llvm.org/D68365> depends on it and I'd like to get that into LLVM 15. Do you have any objections to this?

Ok, I tested your change and removing the assert and it seems to work fine. It's not great that the assertion is failing though, the `ASTContext::getInjectedClassNameType` code hasn't been touched since 2010 so I doubt that it's broken (the assert checks the of the decl is an InjectedClassNameType. Before your change it's an `InjectedClassName` after your change it somehow becomes a `Record`, given the function is called `getInjectedClassNameType` that assert seems correct to me). If this is urgent you can xfail the test, and if you could include a bug report number with your xfail that'd be great.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D128146/new/

https://reviews.llvm.org/D128146



More information about the libcxx-commits mailing list