Re: bug report

[Eli Zaretskii]

> From: istry <address@hidden>
> Date: Sat, 26 Jan 2002 00:17:20 +0300
> 
> There are 2 obvious workarounds:
> 1) change declaration from KOI8-R to WINDOWS-1251
> 2) recode the file itself (mv ru.po ru.po.win && iconv -f WINDOWS-1251 
> -t KOI8-R ru.po.win > ru.po)
> 
> Which one is better/more correct? None.

I don't agree: as long as the charset declaration is consistent with
the file's actual encoding, the translations will work, since current
versions of gettext support on-the-fly recoding.  That is, if the
encoding requested by the target machine (via LANG and other related
environment variables) is different from the one used in ru.po, every
message will be converted to the requested encoding before it is
displayed.

Personally, I think recoding into KOI8-R is a better alternative,
since WINDOWS-1251 is less likely to be supported outside Windows
(this is important, e.g., if someone wants to edit ru.po with their
editor and rebuild the package).  But it's up to the translator to
decide, so I CC'ed Oleg.