[LLVMdev] Incorrect code generated for arm64

[Simon Taylor]

> On 4 May 2015, at 21:56, Tim Northover <t.p.northover at gmail.com> wrote:
> 
> On 4 May 2015 at 10:37, Simon Taylor <simontaylor1 at ntlworld.com> wrote:
>> I’ve gone ahead and created a bug for this, as it seems to be a genuine
>> issue rather than me just overlooking something obvious.
>> https://llvm.org/bugs/show_bug.cgi?id=23408
> 
> Thanks for taking the trouble to report this Simon. I think I've found
> and fixed the issue now (r236457). It should make its way into Xcode's
> clang without too much delay.

Happy to help, and thanks very much for the rapid investigation and fix!

Do you think it’s worth getting into 3.6.1? Is the bug limited to tail call stuff (so any functions that make use of the extended vec3 will be fine - that’s my usual case), or does it potentially have wider effects?

Cheers

Simon