Re: Stable update to newer GCC versions soon?

Hamish McIntyre-Bhatty via Cygwin-apps <[email protected]> Wed, 1 Apr 2026 17:58:49 +0100
Newsgroups gmane.os.cygwin.applications
Message-ID <[email protected]>
On 21/03/2026 20:36, ASSI via Cygwin-apps wrote:
> Hamish McIntyre-Bhatty via Cygwin-apps writes:
>> I see that there are test packages for newer versions of GCC, but the
>> stable packages remains on the older GCC 13.
> I generally switch the system compiler only when GCC reaches the final
> stage, that is x.4 (sometimes there will be an x.5 if there are still
> important fixes coming late).
>
>> I've encountered internal compiler errors when compiling unit tests
>> for python-pybind11, and I was wondering why the test versions haven't
>> been marked stable yet, as they might fix the issue.
> But you don't say what those are, if there's any switches they are
> associated with and apparently you haven't reported them upstream.
>
>> If the answer is that more testing needs to be done then I may
>> volunteer myself for that as it would probably benefit me to have a
>> newer stable compiler.
> Well, if you see it's fixed in 14.3.1 and you can identify the patch
> responsible for it, then that fix could probably be backported.
> Otherwise, 14.4 is going to be released around Juli this year if the
> recent schedule is kept.
>
>> Also, can I report the internal compiler errors
>> upstream or can I not due to downstream Cygwin patches?
> If it's our patches that's going to be fun, but an internal compiler
> error is more likely to be caused upstream.

I have reported this upstream at: 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124753

I also tried the newer testing snapshots of GCC available in the Cygwin 
repo, but they didn't make any difference. I was also able to confirm I 
don't run into the issue on Linux.

Given their bug reporting guidelines say they're not interested in 
unofficial releases or snapshots of GCC, I suspect they'll just tell me 
to report it back to you, but we'll see.

Best,

Hamish