[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Easily grabbing all POs for one language



On Feb 17, 2008 8:06 PM, Dimitris Glezos <dimitris glezos com> wrote:
> One feature that is really useful for a translator is to be able to
> download all the PO files for their language easily.
>
> We've been receiving this request for *years* now, and rightly so!
> Here's why: On elvis, I estimate all language's POs totalled about
> 500MB. To download them, one needed 2 hours on an ADSL, or 20 hours on
> PSTN. A langpack of PO files for a single language weighted at about
> 10MB. That's 3 minutes on ADSL or just 25 minutes on PSTN. Now think
> of giving the chance to the translator to download the langpack
> zipped. It could go down to 2MB -- that's 30 seconds on ADSL, 4
> minutes on PSTN.
>
> Unless I'm mistaken in some calculations, does the above mean we could
> allow our translators on PSTN lines to get the files they need to
> contribute by waiting 4 minutes instead of 20 hours?
>
> Here's a handy single command that gets you all F9 language files
> (msgmerged) for a particular language from
> translate.fedoraproject.org. I've already added it to our FAQ page.
>
>   export MYLANG=el
>   wget -nv -r -N -l 1 --accept *$MYLANG.po -nd -nH -P pofiles
> http://translate.fedoraproject.org/languages/$MYLANG/fedora-9
>
> My dream is to have the tools to do all the dirty work for us, and
> save us time for enjoying life instead. So, I'd like to add support to
> Transifex to serve PO files, ready to be translated, from the command
> line. Something like this:
>
>   tx list-modules (see what modules we serve)
>   tx get <anaconda> (get all anaconda files)
>   tx get <anaconda> --lang=<el> (get iclandic anaconda PO)
>   tx get --lang=<is> --gz  (get all iclandic files, zipped)
>   tx get --tag=<fedora-websites> (get only files related to Fedora websites)
>   ...translate...
>   tx login
>   tx submit anaconda po/is.po
>
> Until we have such a dream tool to our disposal, we'll have to settle
> with the wget-style command.
>
> So after a lot of jabber, here's the rub. We could have the above wget
> command run daily on our servers, and produce a langpack for each
> language in .tar.gz with all its PO files, for easily download. Is
> this something we want to spend some time on, or are we OK with
> running the wget command for now?

Why isn't this something transifex can do already in the background
maybe?  Python has some great tools for writing such kind of
'shell-like' scripting.

That dream like command could interact with transfex via json-rpc
quite easily too.

-Yaakov


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]