137098 - character coding is not remembered

Status: NEW

(MailNews Core :: Internationalization, enhancement)

Description

[Helge Hielscher]

Steps to reproduce:
1. visit a newsgroup with a non-latin language, I used fido7.ru.student.talks
2. view a message and set the coding: KOI8, Autodetect Off
3. view next message
result: character coding is back to ISO-8859-1

expected result: character coding stays as set (KOI8) for the entire newsgroup
including headers.

I am using mozilla 0.9.9-2mdk/Linux. 

Auto-Detect is also unusable b/c only the body not the headers are encoded, this
is bug http://bugzilla.mozilla.org/show_bug.cgi?id=90584

Comment 1

[Helge Hielscher]

Attachment: russian newsgroup, character-coding=koi-8r

Comment 2

[nhottanscp]

>expected result: character coding stays as set (KOI8) for the entire newsgroup
>including headers.

The menu change is not supposed to change charsets of the folder. There is a
feature to set a charset per folder bases in folder property UI.

Comment 3

[Helge Hielscher]

Well I understood that there is a different UI for the folder. 

But if I change the encoding of one message I expect:
a) that the encoding affects the message as well as the appropriate headers in
the table above (maybe this is a seperate bug?)
b) that the encoding is durable, so that I dont have to change the encoding over
and over again everytime I view the (same) message

Comment 4

[nhottanscp]

reopen

Comment 5

[nhottanscp]

enhancement

Comment 6

[r-koifman]

Experiencing the exact same problem : 

Using Mozilla 1.3b
Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.3b) Gecko/20030210

Edit > Preferences > Languages > Default Character coding   is NOT respected 

it automatically reverts to Western (ISO-8859-1)

This might not sound like a big problem except for the fact that improper
character coding messes up the page layout in some cases (phpfrance.com/forums)

Comment 7

[Andreas Eibach]

Same problem here!

Browsing a RUSSIAN FTP server with both Latin and Russian characters, and
because coding jumps back to 'Western (ISO- 8859-1)' all the time, Cyrillic
(Windows-1251) coding has always to be reset MANUALLY if I move to another
folder! ARGL!

Please fix this. This *is* a bug and shouldn't behave that way.

[OS: Win2K SP 3, tried with latest nightly build 2003070312]

Comment 8

[Andreas Eibach]

Ehh .. for testing developers please contact me PRIVATELY so I can give the FTP
address. This shouldn't be public because ... well ...