[llvm-dev] Tracking miscompilations in LLVM using its own unit tests

[Eli Friedman via llvm-dev]

Can you give an example of a bug related to pointer comparisons?  As far as I know, pointer icmp is the same as integer icmp, and all transforms are consistent with that.  And it would be a serious issue if std::less was broken.

(There are a couple bugs where we can rewrite a pointer with one origin to a pointer with a different origin, but that's not really the same issue.)

-Eli

> -----Original Message-----
> From: llvm-dev <llvm-dev-bounces at lists.llvm.org> On Behalf Of Nuno Lopes
> via llvm-dev
> Sent: Wednesday, June 3, 2020 8:06 AM
> To: 'llvm-dev' <llvm-dev at lists.llvm.org>
> Cc: John Regehr <regehr at cs.utah.edu>
> Subject: [EXT] [llvm-dev] Tracking miscompilations in LLVM using its own unit
> tests
>
> Hi,
>
> John Regehr & I wrote a blog post on tracking miscompilations in LLVM using
> its own unit tests & Alive2.
> We have been tracking existing bugs & regressions in LLVM and we wrote a
> summary of the remaining issues we are aware of. Some issues are simpler,
> other require discussion around IR semantics.
>
> https://blog.regehr.org/archives/1737
>
> Thanks to everybody that has contributed so far in fixing bugs & LangRef.
>
> Nuno
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev