On Sat, Sep 26, 2009 at 12:23 AM, Dimitris Glezos <dimitris(a)glezos.com> wrote:
On Fri, Sep 25, 2009 at 8:56 PM, Miloš Komarčević
<kmilos(a)gmail.com> wrote:
> Hi peeps,
>
> Hoping someone will be able to shed some light on the following. I
> have asked Dan to support ngettext in setroubleshoot [1] and found
> positive response and a fix already.
>
> However, the POT generated by Tx differs to the one checked in by the
> developer in Hg. So the questions are:
>
> 1. Does Tx POT generation at all support extraction of ngettext
> strings from an intltoolized project? If so, why doesn't it extract
> them in this case?
Transifex uses intltool-update under the hood to produce the new file.
Maybe a version mismatch?
> 2. Why does Tx even generate a POT file when there is one checked in?
> Difficult to decide which takes precedence, I know. If the project is
> intltoolized, then we can ask the developer not to check in the POT
> any more, but it would then be nice if ngettext extraction worked...
The developer can choose whether to use a static POT or intltool
(dynamically generated POT). In the case of setroubleshoot, the chosen
option is the latter. If the developer wants to produce the POT, we
(he) can just switch the i18n method to use a static POT file.
Hope this helps.
Thanks Dimitris, but to be honest, I was hoping someone could take a
closer look at the setroubleshoot setup and recent changes, as I
unfortunately just can't figure out the inner workings of Tx and
intltool.
Miloš