Re: ui-extract.pl

["Kenneth Christiansen" <kenneth gnu org>]

>         The problem is how best to extract translation information from   
> xml files, after much discussion with the translation team, Kenneth
> implemented a nice perl script to deal with the UI xml files ( it is used   
> for this in several CVS modules ). At the same time he also added some
> capability to deal with glade files; unfortunately the glade file support 
> seems to be badly broken.

It was some time ago, but it isn't any more! This has been comfirmed by a
lot of translators. THIS SHOULD BE WORKING!

>         The easy fix, is to return the .glade.h files to CVS, and knobble
> update.pl so it doesn't touch them; that is until there is a good fix for
> the problem.

This is working, so it's no problem! Please check it yourselves. The letter
miguel replied to, is pretty old, and the problem was fixed like the day after

>         Adding them to the repository is the _correct_ thing to do; many
> translators do their work on mimimal systems, and they do not wish to   
> build all the sourcecode first; consider the evils of having to build    
> eg. all of evolution, just to translate 1 new string.

THE DON'T have to build the source AT ALL! the update.pl script generated
the .h files in like few seconds, and then that is all what has to be done!

>     Either we fix ui-extract.pl, or we put them back in CVS; compile
> time generation using libglade-whatever also involves adding custom rules
> to automake files which is nasty too.

It's working already, so nothing has to be done!

Cheers, Kenneth