Bug 2188479
Summary: | Arabic translation disappeared from 39.11 | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Adam Williamson <awilliam> |
Component: | anaconda | Assignee: | Anaconda Maintenance Team <anaconda-maint-list> |
Status: | CLOSED RAWHIDE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
Severity: | medium | Docs Contact: | |
Priority: | unspecified | ||
Version: | rawhide | CC: | anaconda-maint-list, vponcova, vslavik, w |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | Unspecified | ||
OS: | Linux | ||
Whiteboard: | openqa | ||
Fixed In Version: | anaconda-39.16-1 | Doc Type: | If docs needed, set a value |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2023-05-30 14:00:28 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
Adam Williamson
2023-04-20 21:39:09 UTC
The translation is malformed somehow. I will look at it, but there's nothing apparent in Weblate :-/ Logs from today's scheduled release indicating the same - https://github.com/rhinstaller/anaconda/actions/runs/4922471265/jobs/8793287126 (...) for p in anaconda-39.14/po/*.po ; do \ if [ -e "$p" ]; then \ PYTHONPATH=./translation-canary /usr/bin/python3 -m translation_canary.translated \ --release anaconda-39.14/po ; \ fi ; \ break ; \ done test_percentage warned on my.po: amount translated of 4% below threshold of 10% test_percentage warned on vi.po: amount translated of 6% below threshold of 10% test_percentage warned on nn.po: amount translated of 4% below threshold of 10% test_percentage warned on am.po: amount translated of 3% below threshold of 10% test_percentage warned on cy.po: amount translated of 4% below threshold of 10% test_percentage warned on en_GB.po: amount translated of 2% below threshold of 10% test_percentage warned on fil.po: amount translated of 5% below threshold of 10% test_percentage warned on ur.po: amount translated of 5% below threshold of 10% test_percentage warned on si.po: amount translated of 2% below threshold of 10% test_percentage warned on lv.po: amount translated of 3% below threshold of 10% test_percentage warned on is.po: amount translated of 7% below threshold of 10% test_msgfmt failed on ar.po: Unable to compile anaconda-39.14/po/ar.po: anaconda-39.14/po/ar.po:346: number of format specifications in 'msgid_plural' and 'msgstr[0]' does not match msgfmt: found 3 fatal errors 868 translated messages, 12 fuzzy translations, 571 untranslated messages. Removing anaconda-39.14/po/ar.po make[4]: Leaving directory '/anaconda' (...) Ah, got it now via poedit. The message is "%d _storage device selected". The check for this is enforced in code but somehow is not showing up in Weblate. I will see if it's a Weblate misconfiguration on our side or something more. I will also comment in Weblate. Findings: - Message is 10 years old - Message is marked as python-format - Weblate has the corresponding check turned on as mandatory - Arabic translation changed 3 weeks ago - Arabic translation is technically wrong according to this rule - Weblate shows nothing in the UI Side note: The error is rather terse so it's not immediately clear: This is about matching amount of format placeholders (%d), not matching amount of plurals. So, there was some discussion on May 15 that this was "fixed", but we still don't see Arabic in current Rawhide composes with anaconda-39.15-1.fc39 from May 16. Was the "fix" supposed to be included in that, or not? Sorry - turns out that the change in Weblate did not propagate fast enough to get into the build. I assumed it would sync +- daily, as it usually does. Once I found this happened, I had Weblate sync the repos manually, but it was after the release of course. Next week's release 39.16 on May 23 should be finally fine - the change . (Of course, if somebody comes in the meantime, does not read the comment and puts the broken string back, it's broken again...) |