Hi all,
I doubt I can add much useful info to what you all discovered, so maybe
I'm only adding questions and no new knowledge, but well, we all need to
start somewhere.
On 4/24/26 20:05, Christopher Baines wrote:
Andreas Enge <[email protected]> writes:
Well, it failed again, now in the compilation of Guix, with a message on
glib failing.
I am out of ideas now: Locally, I can "guix pull" to the commit and I can
build the glib package from this commit.
Chris, have you got an idea?
In comparison to the 2 previously failed jobs, this error is clearer,
but also a lot earlier.
For reference, I think this line is the first error in the 2026-04-24
14:46:15 run:
https://data.qa.guix.gnu.org/job/19775
▶ 201/389 /filenamecompleter/basic -
GLib-GIO:ERROR:../glib-2.86.0/gio/tests/filenamecompleter.c:83:run_test_cases:
assertion failed (all_expected_completions == all_results_completions):
first differing element at index 0:
"/tmp/guix-build-glib-2.86.0.drv-0/test_filenamecompleter_89EGO3/filenamecompleter/basic/.dirs/home/file_1"
does not equal
"/tmp/guix-build-glib-2.86.0.drv-0/test_filenamecompleter_89EGO3/filenamecompleter/basic/.dirs/home/folder_1/"
FAIL
It looks like a race condition maybe? Because it compares "file_1" with
"folder_1". Or maybe a file ordering difference? Is that directory
still around somewhere to look?
Could this non-determinism actually be caused by the new Python version?
That would surprise me, but I suppose we should check.
For completeness: I could build
/gnu/store/mn9ybb131n0kbhnhyjz7x42brvrm6v8s-glib-2.86.0.drv successfully
locally, several times, and could not get it to fail.
I've triggered the data service to retry processing the revision.
How long would 'processing the revision' take? That is not yet building
any packages yet right? Except guix, which depends on glib, which
depends on Python, and thus here we are?
