Testing current pootle translation

l10n
#1

Hello!

How can i download and test the latest
translation from the pootle server in libreoffice?

Best regards!

#2

Hello Jon,

Hello!

How can i download and test the latest
translation from the pootle server in libreoffice?

You need to build LibreOffice from the source code. It is not (yet)
possible to snap-in translations from Pootle to an existing
LibreOffice install.

For howtos start reading here:
http://wiki.documentfoundation.org/Development/Native_Build

Best regards,
Andras

#3

Hi andras,

Would it be possible to implement an easier procedure to test new translations?
Maybe integrating them into nightly builds or similar?

It is very hard to know exactly where in the program a specific phrase will be shown,
and what other phrases will be along it to create a good total result.

On a side note: Why are we using an email list instead of a forums for this? Email
lists are great for small communities that need to gain attention from its members,
now this list is just becoming flooded since we have so many contributors.

A forum would be easier to keep clean and reuse information that is already available.
A mailing list keeps users that are not used to it away from the project too (it sure
took me a while to join my first list, when i on the other hand didnt hesitate one second
when joining a forum)

Best regards Jon

Hello Jon,

Hello!

How can i
download and test the latest
translation from the pootle server in
libreoffice?

You need to build LibreOffice from the source code. It is

not (yet)

possible to snap-in translations from Pootle to an

existing

LibreOffice install.

For howtos start reading

here:

http://wiki.documentfoundation.org/Development/Native_Build

Best

regards,

Andras

#4

Pre-release along with language packs are good way to see the
translation result. However, it is kind of late to see the translation
error in that build. Problems are often caught after final release
because of the time to see if translations are well being short.

By the way, Key-ID language build will be awesome to help translators
find out which string is what we see. It will save considerable time
for us. I hope it can be downloaded more easily for translators.

#5

Pre-release along with language packs are good way to see thetranslation
result. However, it is kind of late to see the translationerror in that
build. Problems are often caught after final releasebecause of the time to
see if translations are well being short.

I think we should have nightly builds, EVERY night. That way we can test out
new features and new translations on a daily basis insted of just before a new
release.

By the way, Key-ID language build
will be awesome to help translators find out which string is what we see. It
will save considerable time for us. I hope it can be downloaded more easily
for translators.

Yes, the KeyID is a great addition, combined with nightly builds of your latest translations i think we might be onto something really good!

Best regards!

#6

Of course the best option would be if LO switched to pure gettext so
that we could simply put updated MO files to the language folder and
immediately see the changes. But converting the whole thing to gettext
would probably be a pretty big (and tedious and uninteresting)
undertaking.

#7

If it makes technical sense and adds value to the project, you will be
surprised to see the swarm of amateur developers eager to make the move...

But converting the whole thing to gettext
would probably be a pretty big (and tedious and uninteresting)
undertaking.

- --
Olivier Hallot
Founder, Board of Directors Member - The Document Foundation
LibreOffice translation leader for Brazilian Portuguese
+55-21-8822-8812

#8

Hi,

Pre-release along with language packs are good way to see thetranslation
result. However, it is kind of late to see the translationerror in that
build. Problems are often caught after final releasebecause of the time to
see if translations are well being short.

I think we should have nightly builds, EVERY night. That way we can test out
new features and new translations on a daily basis insted of just before a new
release.

By the way, Key-ID language build
will be awesome to help translators find out which string is what we see. It
will save considerable time for us. I hope it can be downloaded more easily
for translators.

Yes, the KeyID is a great addition, combined with nightly builds of your latest translations i think we might be onto something really good!

It is not realistic to release nightly builds with all languages.
* Translators do not update their language daily
* Tree is not buildable every night
* It takes hours to export files from Pootle
* It takes hours to fix build breaking bugs in translations
* It takes hours to build --with-language=ALL (what platforms do you
want? Linux, Windows, Mac? I guess you want them all. :slight_smile:
* It takes hours to upload binaries to a public ftp/www server

So, It hink it will not happen anytime soon. On the other hand *you*
can always try to build from the source (only your language), it
requires a few gigabytes of space on hard drive and some time.

Also, TDF releases translated builds quite often. Not nightly, but for
3.5.0 we released a beta, rc or a final version in every 1-2 weeks.
Strings do not change - in theory - in the libreoffice-3-5 brach, and
there will be many releases from that branch, so you can keep
improving translations.

KeyID builds were provided with 3.5 betas. I don't know, if they are
still available somewhere. They will be provided again for 3.6 betas
in a few months.

Best regards,
Andras

#9

Þann fös 24.feb 2012 19:56, skrifaði Olivier Hallot:

If it makes technical sense and adds value to the project, you will be
surprised to see the swarm of amateur developers eager to make the move...

But converting the whole thing to gettext
would probably be a pretty big (and tedious and uninteresting)
undertaking.

- --
Olivier Hallot

At the LibreOffice conference in Paris, Andras did mention conversion to gettext as a future possibility;
has there been any feedback from the devs about that ?

From what I've seen, there are other pending restructurations of the codebase; another toolkit (GTK?), simplification of icon locations, skinning, etc.

One at a time or one big surgery ?

Regards,
Sveinn í Felli

#10

2012.02.25. 10:02 keltezéssel, Sveinn í Felli írta:
At the LibreOffice conference in Paris, Andras did mention conversion to

gettext as a future possibility;
has there been any feedback from the devs about that ?

From what I've seen, there are other pending restructurations of the
codebase; another toolkit (GTK?), simplification of icon locations,
skinning, etc.

Caolán also mentioned gettext in his FOSDEM 2012 presentation.
See http://people.redhat.com/caolanm/FOSDEM2012.odp
All slides from FOSDEM 2012 LibreOffice devroom:
http://wiki.documentfoundation.org/Marketing/Events/Fosdem2012

A started to write a program that converted po files to LibreOffice resources. Unfortunately I did not finish it and I have not worked on it since last summer. I also had a gettext patch but it had strange character corruption issues and nodoby could tell why. So it was dropped. I'm sure that Caolán's solution will be more robust. I'm not a programmer after all. :slight_smile:

Best regards,
Andras