Plural forms in the style KCM

[Ingo Klöcker]

On Mittwoch, 2. September 2026 10:22:50 Mitteleuropäische Sommerzeit Freek de 
Kruijf wrote:
> Op woensdag 2 september 2026 10:18:46 Midden-Europese zomertijd schreef
> 
> Kisaragi Hiu:
> > Good day,
> > 
> > 
> > AFAIK this should still work for any language.
> > 
> > https://invent.kde.org/plasma/plasma-workspace/-/blob/master/kcms/style/
> > 
> > kcmstyle.cpp?ref_type=heads#L225:
> >             message = i18ncp("@info:status",
> >             
> >                              "Successfully uninstalled “%2” but could not
> > 
> > uninstall others:\n",
> > 
> >                              "Successfully uninstalled %1 styles but could
> > 
> > not uninstall others:\n",
> > 
> >                              uninstalled.size(),
> >                              uninstalled.first());
> > 
> > i18ncp uses the first integer-valued argument as the number that is
> > plugged
> > into the plural forms expression. That expression is decided by the
> > particular PO file and does not assume Germatic plurals.
> > https://invent.kde.org/frameworks/ki18n/-/blob/master/docs/programmers-gui
> > d
> > e.md#gen_usage:~:text=The%20plural%20form%20is%20decided%20by%20the%20firs
> > t% 20integer%2Dvalued%20argument
> > 
> > 
> > (note: a guide should not be the only place some information is
> > documented,
> > but right now the ki18n documentation does not mention this fact in the
> > reference docs, only in this guide. I might try to open an MR for this...)
> > 
> > 
> > So, the size is passed into the plural-forms expression of the target
> > language to decide which msgstr to use, in addition to being the value to
> > fill into the %1 placeholder.
> > 
> > 
> > So it should still work fine with non-Germatic languages no matter how
> > complicated the plural-forms expression is.
> > 
> > 
> > (Gettext does assume 1 or 2 plurals for source text, which is one of the
> > only issues that Gettext has. This means while it supports translating
> > into
> > any language correctly, it only supports translating from languages with 1
> > or 2 plurals correctly. This is fine in KDE since we're translating from
> > English, but the issue does exist.)
> 
> Below is the error message about not accepting the translation:
> 
> Committing transaction...
> svn: E165001: Commit failed (details follow):
> svn: E165001: Commit blocked by pre-commit hook (exit code 1) with output:
> <stdin>:132: a format specification for argument 2, as in 'msgstr[0]',
> doesn't exist in 'msgid_plural'
> <stdin>:140: a format specification for argument 2, as in 'msgstr[0]',
> doesn't exist in 'msgid_plural'
> msgfmt: found 2 fatal errors
> 
> trunk/l10n-kf6/nl/messages/plasma-workspace/kcm_style.po:
> ERROR: There were syntax errors detected in a po file you committed.
> ERROR: Please correct them first.

Okay. As I suspected the problem is the missing %2 in the second string which 
our pre-commit hooks treat as a hard error (probably for a good reason). I 
still think that technically there's nothing wrong with this call and it would 
work if you wouldn't be blocked from committing the translation.

Looks like the code needs to be rewritten with a simple i18nc for the case
n == 1 and an i18ncp with a dummy n == 1 string.

Regards,
Ingo
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 265 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/kde-i18n-doc/attachments/20260902/90cbc92f/attachment.sig>