Re: Translating plural forms with different c-formats

[Clytie Siddall <clytie riverland net au>]

On 28/08/2006, at 3:34 PM, Theppitak Karoonboonyanan wrote:

> Hi Abel,
>
> On 8/28/06, Abel Cheung <abelcheung gmail com> wrote:
> > On 8/27/06, Theppitak Karoonboonyanan <thep linux thai net> wrote:
> > > My language has no plural form. And I recently got some
> > > complaint that having more than one msgstr[] while
> > > declaring nplurals=1 blocks the build due to  gettext error.
> > > So, I try to examine if there are other such errors left in
> > > any other packages.
> >
> > I'm afraid you are using gettext >= 0.15, as it is more strict in
> > checking plural form and mark inconsistent Plural-Forms and
> > actually translation as error. gettext < 0.15 just ignore it and
> > wouldn't complain.
>
> It's not me who find the complaints. It's just from others.

I have encountered the same problems with those files. These errors  
may be caught by gettext-0.15's useful capacity to point out spurious  
plural forms, but they are about string format.
> > For the real fix, you can remove any entry containing
> > msgstr[1], leaving only msgstr[0].
>
> Yes, I know that, as I said I had tried to do so to all PO's
> under my maintenance. However, my question is for certain
> cases where c-format string with plural form is shared
> with some non-c-format. And there is actually _no_way_
> to cover both with a single msgstr[0] entry without messing
> one of the two. Please read my first post for the details.
Exactly. If we kludge it and create a msgstr[0] which includes the  
placeholder, we still have the conflict between the two different  
types of strings.

from Clytie (vi-VN, Vietnamese free-software translation team / nhóm  
Việt hóa phần mềm tự do)
http://groups-beta.google.com/group/vi-VN