ojhunt wrote:

> @ojhunt Thank you for your feedback. This fix addresses the specific case of 
> using attributes within a `#pragma`. The change prevents crashes and 
> maintains the fix-it experience. I think retaining the fix-it, when possible, 
> is preferable to omitting it.
> 
> If you know of any other situations where a similar crash could happen, I’d 
> really appreciate it if you could point them out — I’m happy to address them 
> as part of this PR. Based on my review of the current 
> `ParseCXX11AttributeSpecifier` usage and the flow of `EndLoc` throughout the 
> codebase, I haven’t encountered any other problematic scenarios so far, so I 
> would appreciate any suggestions for additional test cases I should add. 
> Thanks.

Yes, but if the optional location argument is not passed somewhere else this 
will still crash, so it needs the null check *as well*.

https://github.com/llvm/llvm-project/pull/182362
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to