My RT3 I18N nightmare (UTF-8)

Are you using perl 5.6 or perl 5.8? There are known issues with 5.6 and
I18N.On Mon, Feb 17, 2003 at 03:29:58PM +0100, Jan Okrouhly wrote:

Hi,

I wanted to make corrections of my translations, but found there
is still problem, that with the web ui I18N coding IS something VERY
BAD/WRONG.
The core of problem lies somewhere in RT which sometimes doubles
conversion into UTF-8 - its strange output you could see in the
attached file.

before_login - all works Ok (my locale is CS - Czech)
after_login - everything is wrong.

I assume that nobody has noticed it during translations, may be due
e.i. DE - German has not so much accuted characters? How about ZH,
RU, JA?

An example:
á (a accute)
should be hex C3 A1
RT outputs C3 83 C2 A1

Could somebody take closer look at this show stopper?
Maybe the problem is only somewhere in HTTP headers?

I proved that the problem is not in the apache nor browser,
but six hours spent playing with the RT code weren’t enough… :frowning:

Jan Okrouhly
-----------------------------------------±----okrouhly@civ.zcu.cz—
Laboratory for Computer Science | phone: +420 377 632 837
University of West Bohemia | location: Univerzitni 22
Americka 42, 306 14 Pilsen, Czech Republic | room: UI402
------------------------------------------73!-de-OK1INC@OK0PPL.#BOH.CZE.EU-

�|� http://www.bestpractical.com/rt – Trouble Ticketing. Free.

Are you using perl 5.6 or perl 5.8? There are known issues with 5.6 and
I18N.

Yes, one of things to do for me is to try 5.8 -
my mod_perl returns version 5.006001.

I dunno perl version on rt3.fsck.com (supposed you are already on 5.8).

Hi,

I wanted to make corrections of my translations, but found there
is still problem, that with the web ui I18N coding IS something VERY
BAD/WRONG.
The core of problem lies somewhere in RT which sometimes doubles
conversion into UTF-8 - its strange output you could see in the
attached file.

before_login - all works Ok (my locale is CS - Czech)
after_login - everything is wrong.

I assume that nobody has noticed it during translations, may be due
e.i. DE - German has not so much accuted characters? How about ZH,
RU, JA?

An example:
ĂĄ (a accute)
should be hex C3 A1
RT outputs C3 83 C2 A1

Could somebody take closer look at this show stopper?
Maybe the problem is only somewhere in HTTP headers?

I proved that the problem is not in the apache nor browser,
but six hours spent playing with the RT code weren’t enough… :frowning:

Jan Okrouhly
-----------------------------------------±----okrouhly@civ.zcu.cz—
Laboratory for Computer Science | phone: +420 377 632 837
University of West Bohemia | location: Univerzitni 22
Americka 42, 306 14 Pilsen, Czech Republic | room: UI402
------------------------------------------73!-de-OK1INC@OK0PPL.#BOH.CZE.EU-


ďż˝|ďż˝ http://www.bestpractical.com/rt – Trouble Ticketing. Free.

Jan Okrouhly
-----------------------------------------±----okrouhly@civ.zcu.cz—
Laboratory for Computer Science | phone: +420 377 632 837
University of West Bohemia | location: Univerzitni 22
Americka 42, 306 14 Pilsen, Czech Republic | room: UI402
------------------------------------------73!-de-OK1INC@OK0PPL.#BOH.CZE.EU-

Are you using perl 5.6 or perl 5.8? There are known issues with 5.6 and
I18N.

Yes, one of things to do for me is to try 5.8 -
my mod_perl returns version 5.006001.

I dunno perl version on rt3.fsck.com (supposed you are already on 5.8).

fsck.com is actually running perl 5.6.1, as it’s a debian stable (woody)
box. Autrijus has said that he’ll have a look at the issues with I18N
and 5.6.1 at some point, but it’s definitely a low priority. At this
point, perl 5.8 has proven itself to be so much better at handling
utf-8 that I’d strongly encourage anyone who needs non-us-ascii support
in RT to run perl 5.8, rather than 5.6.1.

-j

�|� http://www.bestpractical.com/rt – Trouble Ticketing. Free.