On Mon, Mar 03, 2008 at 09:14:43AM +0100, Nagy Gabor wrote:

Bye PS: iso8859-2 is good for both non-utf8 users (like me) and utf8 users.

How is that possible? And why the obstination for not using utf8? If you still have problems with it, try to do something about it, at least by making sure the issues are known and reported.

On Mon, Mar 03, 2008 at 11:30:41AM +0100, Nagy Gabor wrote:
> > How is that possible?
> [*] I dunno, I asked utf8 users on #archlinux.hu channel, and I got 'no problem'
> feedback. (And I've never heart problems about Hungarian translation.)
> 
> > And why the obstination for not using utf8?
> * I don't really need that now, and most of my stuffs is in iso8859-2.
> * See this poll: http://hup.hu/node/50144
> * so I interpret [*] as for "everyone"
> 

Does that poll say that 82% of the users who voted use UTF-8? Well, if the
majority wins, that means the translation should stay in utf8.
If the reason wins, utf8 should be used as well. Not using a fine standard
when there is one is stupid.

Besides, the translation-help file has the following :
""In addition, for all new translations we would strongly recommend using UTF-8
encoding.""

You can try using convmv if you have filenames to convert, and recode if you
have text files to convert.

2008/3/3 Nagy Gabor :
> > On Mon, Mar 03, 2008 at 11:30:41AM +0100, Nagy Gabor wrote:
>  > > > How is that possible?
>  > > [*] I dunno, I asked utf8 users on #archlinux.hu channel, and I got 'no
>  > problem'
>  > > feedback. (And I've never heart problems about Hungarian translation.)
>  > >
>  > > > And why the obstination for not using utf8?
>  > > * I don't really need that now, and most of my stuffs is in iso8859-2.
>  > > * See this poll: http://hup.hu/node/50144
>  > > * so I interpret [*] as for "everyone"
>  > >
>  >
>  > Does that poll say that 82% of the users who voted use UTF-8? Well, if the
>  > majority wins, that means the translation should stay in utf8.
>  > If the reason wins, utf8 should be used as well. Not using a fine standard
>  > when there is one is stupid.
>  >
>  > Besides, the translation-help file has the following :
>  > ""In addition, for all new translations we would strongly recommend using
>  > UTF-8
>  > encoding.""
>  >
>  > You can try using convmv if you have filenames to convert, and recode if you
>  > have text files to convert.
>  >
>
>  Well, whatever you decide, please keep libalpm's and pacman's po in sync (now
>  one of them is utf8, one of them is latin2).
>
>  I leave the decision to you, but personally I (the translator;-) still vote to
>  iso8859-2.
>  Note: I didn't _changed_ the encoding, it was in iso8859-2, when I started to
>  work on vmiklos's translation.

Here is what I am going to say on this, and unless anyone has some
real good objections, we're going with this. Start by reading this:
http://www.gnu.org/software/gettext/manual/gettext.html#Charset-conversion

All of our translations are currently in UTF-8 except for hu. The
encoding of the translation has zero bearing on what you use on your
actual system, and for ease of use by others doing work on pacman and
the translations, it is best to use a universal encoding (rather than
assume we have this specific charset enabled on our machine, which my
VIM seems to balk on).

With this said, would it be a huge problem if we just went to the
UTF-8 standard for all translations? As you can see below, I
unfortunately did not catch that the conversion didn't change the
Content-Type heading which it should have. UTF-8 is universal and
works for everyone as a format for message files. We don't even need
to discuss your particular choice of locale as that is irrelevant at
this stage of the game. That is all handled internally by gettext.

-Dan

$ grep -RFI 'Content-Type' *
lib/libalpm/po/es.po:"Content-Type: text/plain; charset=UTF-8\n"
lib/libalpm/po/fr.po:"Content-Type: text/plain; charset=utf-8\n"
lib/libalpm/po/en_GB.po:"Content-Type: text/plain; charset=UTF-8\n"
lib/libalpm/po/pt_BR.po:"Content-Type: text/plain; charset=UTF-8\n"
lib/libalpm/po/pl.po:"Content-Type: text/plain; charset=UTF-8\n"
lib/libalpm/po/zh_CN.po:"Content-Type: text/plain; charset=UTF-8\n"
lib/libalpm/po/hu.po:"Content-Type: text/plain; charset=ISO-8859-2\n"
lib/libalpm/po/it.po:"Content-Type: text/plain; charset=utf-8\n"
lib/libalpm/po/cs.po:"Content-Type: text/plain; charset=UTF-8\n"
lib/libalpm/po/de.po:"Content-Type: text/plain; charset=UTF-8\n"
lib/libalpm/po/libalpm.pot:"Content-Type: text/plain; charset=CHARSET\n"
lib/libalpm/po/ru.po:"Content-Type: text/plain; charset=UTF-8\n"
po/es.po:"Content-Type: text/plain; charset=UTF-8\n"
po/fr.po:"Content-Type: text/plain; charset=utf-8\n"
po/en_GB.po:"Content-Type: text/plain; charset=UTF-8\n"
po/pt_BR.po:"Content-Type: text/plain; charset=UTF-8\n"
po/pacman.pot:"Content-Type: text/plain; charset=CHARSET\n"
po/pl.po:"Content-Type: text/plain; charset=UTF-8\n"
po/zh_CN.po:"Content-Type: text/plain; charset=utf-8\n"
po/hu.po:"Content-Type: text/plain; charset=ISO-8859-2\n"
po/it.po:"Content-Type: text/plain; charset=utf-8\n"
po/cs.po:"Content-Type: text/plain; charset=UTF-8\n"
po/de.po:"Content-Type: text/plain; charset=UTF-8\n"
po/ru.po:"Content-Type: text/plain; charset=UTF-8\n"

2008/3/3 Nagy Gabor :

...

Here is what I am going to say on this, and unless anyone has some real good objections, we're going with this. Start by reading this:

http://www.gnu.org/software/gettext/manual/gettext.html#Charset-conversion

If this means, that this will work for non-utf8 users, then there is nothing to discuss, go ahead, utf8 conversion acked.

I have thought about if it's right decision to use zh_CN.UTF-8 instead of zh_CN.GBK for Simplified Chinese translation, but in the end I found most users of our (chinese ArchLinux) forum use utf-8 instead of GBK, so I vote for utf-8.

Bye

---------------------------------------------------- SZTE Egyetemi Könyvtár - http://www.bibl.u-szeged.hu This mail sent through IMP: http://horde.org/imp/

_______________________________________________ pacman-dev mailing list pacman-dev@archlinux.org http://archlinux.org/mailman/listinfo/pacman-dev

On Mon, Mar 3, 2008 at 10:05 AM, Matthias Gorissen wrote:

Here is a minor cosmetic update which fixes an earlier problem with an annoying umlaut (it does not appear in libalpm/po, which seems pretty much up-to-date).

Best regards, Matt

Applied, thanks. -Dan

On Mon, Mar 03, 2008 at 09:11:26AM -0600, Dan McGee wrote:

Here is what I am going to say on this, and unless anyone has some real good objections, we're going with this. Start by reading this: http://www.gnu.org/software/gettext/manual/gettext.html#Charset-conversion

All of our translations are currently in UTF-8 except for hu. The encoding of the translation has zero bearing on what you use on your actual system, and for ease of use by others doing work on pacman and the translations, it is best to use a universal encoding (rather than assume we have this specific charset enabled on our machine, which my VIM seems to balk on).

With this said, would it be a huge problem if we just went to the UTF-8 standard for all translations? As you can see below, I unfortunately did not catch that the conversion didn't change the Content-Type heading which it should have. UTF-8 is universal and works for everyone as a format for message files. We don't even need to discuss your particular choice of locale as that is irrelevant at this stage of the game. That is all handled internally by gettext.

Oh, I see, it all doesn't matter that much then, it's all transparent for the user. So Nagy wouldn't be affected as an user by that rule. But he would still be affected as a translator, right? But well, files can be easily converted, so it shouldn't be a big problem. And any way, I still vote for having all translations in pacman repo in UTF-8, because that would be nicer, more consistent and universal.