sfirke opened a new issue, #45377:
URL: https://github.com/apache/superset/issues/45377
### Bug description
A large part of the Slovak catalog
(`superset/translations/sk/LC_MESSAGES/messages.po`)
is Czech rather than Slovak. The two languages are closely related but
distinct, so
users who select Slovak see a mix of the two, sometimes within one phrase.
You can check this without reading either language: the letters **ě, ř and
ů** are
used in Czech but don't exist in Slovak. On `master`:
- **388 of 4,801** translated Slovak entries (8%) contain at least one of
those letters.
- **1,162** Slovak entries are identical to the corresponding Czech (`cs`)
entry. Some
words are spelled the same in both languages, so that number is an upper
bound.
The 388 are certain.
- Only 35 of the 388 are marked `#, fuzzy` and none came from the AI
backfill. The rest
are confirmed entries, so a review of fuzzy strings won't find them.
Examples:
| English source | Slovak catalog (`sk`) | Czech catalog (`cs`) | Note |
|---|---|---|---|
| `March` / `October` / `September` | `Březen` / `Říjen` / `Září` |
identical | Czech month names in Slovak date UI |
| `Sort` | `Řadit` | identical | Czech word |
| `Sort Descending` | `Řadit sestupne` | `Řadit sestupně` | Czech with one
letter changed; neither language |
| `Maximum value on the gauge axis` | `Maximální hodnota na ose meřidla` |
`…na ose měřidla` | Same pattern |
| `Duplicate column name(s): %(columns)s` | `Duplicitní názov(vy) stĺpce(ů):
…` | — | Slovak and Czech mixed in one phrase |
**Where it came from.** The Superset 6.1.0 Slovak catalog was mostly
untranslated
(4,290 of 4,624 entries empty) and has no Czech-only letters. Counting those
letters
across the catalog's history puts their arrival at #40219 ("feat(i18n): add
missing
Slovak translations"), after which 402 `msgstr` lines contained them. Later
catalog
maintenance brought that down to 276 lines (the 388 entries above), but most
remain.
The release branch for 7.0 and the 7.0.0rc1 tag include #40219.
**Reproduce the count:**
```bash
pip install polib
python - <<'PY'
import polib, re
cz = re.compile("[ěřůĚŘŮ]")
po = polib.pofile("superset/translations/sk/LC_MESSAGES/messages.po")
hits = [e for e in po if not e.obsolete and any(cz.search(s) for s in
(e.msgstr_plural.values() if e.msgstr_plural else [e.msgstr]))]
print(len(hits))
for e in hits[:20]:
print(repr(e.msgid), "->", repr(e.msgstr or e.msgstr_plural.get(0)))
PY
```
**Expected behavior:** Slovak UI text is Slovak. Strings with no Slovak
translation
show the English fallback.
**Possible remedies,** for maintainers to choose from:
1. Review by a native Slovak speaker, ideally covering the 1,162 entries
identical to `cs`.
2. As a stopgap, clear the 388 entries containing Czech-only letters (or
mark them
`#, fuzzy`) so they fall back to English or get regenerated by
`scripts/translations/backfill_po.py`, which is then reviewed. Clearing
is the more
conservative choice, because fuzzy entries are still served in the UI.
**Related, not Slovak-specific:** a few entries are mistranslated in both
catalogs,
e.g. `%s metric` → `Řadit metriku` ("Sort metric") and `Pin to top` → `Shora
dolů`
("Top to bottom"). They may deserve their own look in `cs`.
### Screenshots/recordings
_No response_
### Superset version
master / latest-dev (also 7.0.0rc1; not present in 6.1.0)
### Python version
3.11
### Node version
18 or greater
### Browser
Not applicable
### Additional context
Found while editing Slovak entries for #45376. I don't speak Slovak or
Czech; the
identification relies on the letters above and on direct comparison with the
`cs`
catalog.
### Checklist
- [x] I have searched Superset docs and Slack and didn't find a solution to
my problem.
- [x] I have searched the GitHub issue tracker and didn't find a similar bug
report.
- [x] I have checked Superset's logs for errors and if I found a relevant
Python stacktrace, I included it here as text in the "additional context"
section.
---
🤖 _Drafted by Claude Code, reviewed and approved by @sfirke._
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
