Ubuntu Docs Jaunty PO to XML errors, HTML files

[Alexey Balmashnov]

On Mon, Mar 16, 2009 at 8:58 PM, Adi Roiban <adi at roiban.ro> wrote:
> Hi,
> On Mon, 2009-03-16 at 20:31 +0100, Alexey Balmashnov wrote:
>> Nice, looks like exactly the same errors as ones that had been fixed a
>> while back for Intrepid.
>>
>> Are there any chances that tool validating somehow results of
>> translators work will be merged into Rosetta? Or may be used as an
>> intermediate step somewhere in the import process for translations of
>> the current+1 release?
>
> Basicaly I was just converting back from po to xml and the from xml to
> html.
>
> I don't think we will have those validations in Rosetta, as Rosetta is
> supposed to work with PO files... and there is a long feature list
> waiting.

Aha. Then one more point in the feature list won't make it longer
anymore, so why not to add this?

> Also right not the validation takes about 1 hour on my computer. The
> scripts could be optimized but I don't think you could get under 30
> minutes. I'm goind  to set up a cron job to check the documentation at
> the end of each week.
>
> What we could do, is to create these stats on a more central place.
> I have added a link on the Ubuntu Translation wiki page
> https://wiki.ubuntu.com/TranslatingUbuntu

I do not know enough about Rosetta, XML and other related internals,
but may be it is feasible to setup somehow generation of small peace
of documentation just for validity check upon submission of
documentation paragraph by translator, it is not even necessary for
every paragraph, but only for those using elements, which might
invalidate XML. Again, I am not an expert.

> Can you describe the use case as the "intermediate step" . I don't know
> how exactly you see the process.

No problem. AFAIK, import of documentation translations for a new
release cycle performed only once and its normally based on current
documentation for latest stable release. I think it is possible to
perform validity check as you do right now, just before the import of
translations to stable+1, so that errors introduced in stable do not
propagate further on.

Regards,
  Alexey