On 2/18/26 1:55 PM, Olivier Dion wrote:
On Wed, 18 Feb 2026, "Bernhard M. Wiedemann" <[email protected]> wrote:
Today I reverted that work-around for a test and parallelism still
caused variations.

Interesting, so cross-modules inlining is not the only culprit.  Do you
have an idea?  If you have a .tar.gz, with two binaries so I could use
diffoscope on them, that would be great.

I uploaded 2 rpms to
https://rb.zq1.de/other/guile/

With rpm2cpio you can unpack them.
And I uploaded them to diffoscope for you:

https://try.diffoscope.org/ukqxxyguyeca.html

Attachment: OpenPGP_signature.asc
Description: OpenPGP digital signature

Reply via email to