On 5/1/26 1:37 PM, Oswald Buddenhagen via Development wrote:
Are there any other considerations that I should be aware of?
adding a special case for zero (or any other specific number) can always
be done at the application level. this inherently can't break anything.
it's just extra work, but overall much less work than if you impose that
extra case on all other instances that don't need it, or make the system
much more complex by supporting each edge case explicitly through some
overloads/configuration flags.
generally, the plural support found in the various i18n systems concerns
cardinal numerals, that is, strings that actually include the number of
items. cf. https://qt-project.atlassian.net/browse/QTBUG-5928
I agree with Oswald's arguments in the bug report. Compare the gettext
plurals documentation at
https://www.gnu.org/software/gettext/manual/html_node/Plural-forms.html
and look for "There is one case where using ngettext is not appropriate,
however: namely, when neither of the two strings contains a cardinal
number.". There's an example that illustrates why ngettext - or the tr
plurals form - is not appropriate.
Note that this is different from the "zero case" Nicolas asked about.
But like Masoud, I'm not so sure that this belongs to the plurals facility.
--
Jörg Bornemann | The Qt Company
--
Development mailing list
[email protected]
https://lists.qt-project.org/listinfo/development
