3 Mar
2008
3 Mar
'08
3:11 p.m.
2008/3/3 Nagy Gabor <ngaba@bibl.u-szeged.hu>: > > 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"