Re: gettext, plural form and translation
| От | Peter Eisentraut |
|---|---|
| Тема | Re: gettext, plural form and translation |
| Дата | |
| Msg-id | 49C25F76.3030604@gmx.net обсуждение исходный текст |
| Ответ на | Re: gettext, plural form and translation (Aidan Van Dyk <aidan@highrise.ca>) |
| Список | pgsql-hackers |
Aidan Van Dyk wrote: > * Peter Eisentraut <peter_e@gmx.net> [090319 04:21]: > >> It depends also on what we *want* to target. I originally omitted the >> plural support because it was a GNU extension, and I wanted to support >> "standard" gettext implementations as well. (There was also a licensing >> consideration.) Solaris is the "original" implementation of this API, >> so it can serve as a reference point. >> >> But it is open to debate whether that decision is more useful than the >> tradeoff it imposes. > > Of course, ngettext can easily be "ported", at it's most basic level > (equivalent to GNU's C locale implentation) it's: > > #ifdef NEED_NGETTEXT > #define ngettext(s,p,n) gettext((n)==1?(s):(p)) > #endif The more interesting question is whether such a msgfmt will correctly process message catalogs containing things like #: print.c:2351 #, c-format msgid "(1 row)" msgid_plural "(%lu rows)" msgstr[0] "(1 Zeile)" msgstr[1] "(%lu Zeilen)" My guess is not.
В списке pgsql-hackers по дате отправления: