[Diffusion] rL253518: Fix returned value for GVN: could return "false" even after modifying the IR

[Mehdi Amini via llvm-commits]

No, such test would be very fragile, it crashes only depending on what the inliner decides to do after GVN.
I hesitated about a c++ unit test that would verify the returned value on this IR fed to GVN, but it does not seems “usual”.

— 
Mehdi



> On Nov 18, 2015, at 5:08 PM, Daniel Berlin <dberlin at dberlin.org> wrote:
> 
> DId the test case for this get committed?
> (it was a crasher, so it should crash before the fix and not crash after).
> 
> 
> On Wed, Nov 18, 2015 at 5:00 PM, Mehdi AMINI via llvm-commits <llvm-commits at lists.llvm.org <mailto:llvm-commits at lists.llvm.org>> wrote:
> joker.eph added a subscriber: joker.eph.
> joker.eph added auditors: 3.7-release, llvm-commits, tstellarAMD.
> joker.eph added a comment.
> 
> This is very low risk and fix potential bug/crash.
> 
> 
> Users:
>   3.7-release (Auditor)
>   llvm-commits (Auditor)
>   tstellarAMD (Auditor)
> 
> http://reviews.llvm.org/rL253518 <http://reviews.llvm.org/rL253518>
> 
> 
> 
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org <mailto:llvm-commits at lists.llvm.org>
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits <http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits>
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20151118/0251b8d9/attachment.html>