Follow-up Comment #45, bug #1981 (project savane):
I can't create new bug w/out registering so I'll post comment here
"The on-the-fly conversion is done by the gettext routines, which will check
the encoding of the .mo file (e.g. KOI8-R) and the encoding of the output
locale (e.g. ru_RU.UTF-8). In the example case, the msgstr will be converted
from KOI8-R to UTF-8. If the output locale is ru_RU, the implied encoding is
KOI8-R, so no conversion would be done."
Last statement is completely wrong. Then no codepage is specified in current
locale default LOCALE'S codepage is used, which in case of ru_RU is
ISO-8859-5.
Thus currently if I set Russian as prefered language in my browser
(Opera/7.52.3834) all page is unreadable and seems to be in ISO-8859-5. In
Opera's "info page" it shows that server tell is to use ISO-8859-15. However I
can read content after forcing ISO-8859-5.
Also on some pages there is text in KOI8-R (seems related with bug #1986). For
example "%d matching items" on search result page.
Server must send to browser real codepage which in case is ISO-8859-5 and not
to mix different encodings in sigle page. As I suppose this occurs due to
invalid assumption in quoted text.
WBR, Profic
_______________________________________________________
Reply to this item at:
<http://gna.org/bugs/?func=detailitem&item_id=1981>
_______________________________________________
Message sent via/by Gna!
http://gna.org/
_______________________________________________
Savane-dev mailing list
[email protected]
https://mail.gna.org/listinfo/savane-dev
