Re: non-utf8 po files breaking the build

[Christian Rose <menthos menthos com>]

mån 2002-11-25 klockan 00.02 skrev Keld Jørn Simonsen:
> > > I don't think you should do it, it would create some burdens for 
> > > translators that are unnecesary.
> > 
> > Can you explain this in more detail? What specific burdens, and why do
> > you think they are unnecessary? Vague comments without rationales don't
> > help.
> 
> I have discussed this earlier.

Ok, I thought you were to bring up some new input on this issue.


> One ting is people doing corrections and not observing that the encoding
> is UTF-8, and just correcting it in iso-8859-1. I have seen a number
> of examples of this in kde, these errors are cumbersome to solve,
> and these problems are unnecessary, and a waiste of my time.

These errors will probably always exist, irregardless of what encodings
we use. There will always be mistakes. But the tools can check for those
errors and they already do (msgfmt certainly does).


> Another thing is that my tools for translation, sed scripts,
> spellcheckers, are based on the encoding I use, and they dont work with
> utf-8.

I don't think we should base the future of GNOME internationalization on
the bugs that exist in some individual's private scripts. We certainly
need to look at a bigger picture than that.


> To me Linux stands for freedom - in the choice of tools to use, and 
> I do not see the need to impose use of specific tools when there is no
> need to. The gnome cvs system for translated files have worked fine
> with different encodings the two years I have used it, so I see no need
> to change it.

Then you are most likely not looking closely enough (which might be part
of the reason why we have this discussion over and over). The people
that do the daily development and work with the GNOME i18n
infrastructure encounter problems with non-UTF-8 encodings all the time.
The mail that started this thread was just one in the countless examples
of that.

Certainly, a global solution that would seemlessly convert between,
detect and handle all possible encodings in all possible situations
would certainly be nice and the "optimal" solution, but we don't have
it, and we need a solution *now* and already have needed it for several
years. Thus, standardizing on only one encoding is the only sane way to
go, especially since all GNOME translators already have to do it anyway,
most people don't report any problems at all with using it, and those
who do only report minor personal inconveniences. And minor personal
inconveniences we have to accept -- otherwise we can't build a desktop
with new code and new features. There's no way anyone can have developed
and enhanced everything already so that everything in the development
process is seamless and perfect. To some extent we build new roads here,
so you're bound to get your hands dirty and have inconveniences somehow.

Also, I think personal freedom is an exceptionally bad argument. Noone
is stopping you from encoding your files on your hard drive in whatever
encoding or format you prefer. You still can enjoy that freedom in every
way. The only thing is that we ask you to convert it to UTF-8 before you
commit it to GNOME cvs. There never ever was any "freedom" to commit
anything you'd like in any format you'd like to GNOME CVS to begin with,
so we're not taking that freedom away.

The decision to use UTF-8 as the mandatory encoding for all GNOME 2.x po
files has been made. Please accept that.


Christian