[Xfce-i18n] Possible to translate devel branch of xfce4-panel

[?????]

From: "Nick Schermer" <nickschermer at gmail.com>
Sent: Friday, February 19, 2010 1:37 PM
To: "Xfce translations mailing list" <xfce-i18n at xfce.org>
Subject: Re: [Xfce-i18n] Possible to translate devel branch of xfce4-panel

> On Fri, Feb 19, 2010 at 12:18 AM, <eskaer_spamsink at ngs.ru> wrote:
>>
>> So we need new %2 element which is a string equal to ordinal number for hour+1.
> 
> Although not documented, %2 should already work (anything between 0
> and 9 will do), but i doubt you need that, because in 5:55, %0 = 5, %1
> = 6.

No, it's slightly different: because 6:00 "%0 o'clock" with %0=6 or "six" should be translated like "*six* hours", but 5:15 "a quarter past %1", if %1=6 or "six" should be translated like "a quarter of *sixth*", so if both %0 and %1 comes from same set of 1..12 or "one"..."twelve", there're problems arising. And we cannot write it in Russian as "6th" or "6.", we need a special set of "first","second",..."twelfth" strings corresponding to "one".."twelve", and indexing by (hour+1).
> 
>> Also "%1 o'clock" strings should be realized via gettextn to get correct multiple plural forms for languages with >1 plural.
> 
> That is also there, but not using gettextn. The strings with the "one"
> context are used for the first hour (1 0'clock) and the others for >1.

A bunch of European languages, Russian included, have more than one plural form, f. ex. "1 o'clock", "2 o'clock" and "7 o'clock" are THREE different templates, not just TWO for %1==1 and %1>1, so current set of strings won't suffice.
See http://www.gnu.org/software/hello/manual/gettext/Plural-forms.html

P.S. Maybe it's online editor limitation, but "%1 o'clock" stuff really should be stored as msgid/msgid_plural pair, to be translated by  msgstr[0..2] triplet.