118486 - Can't decode subject in GB2312

Status: VERIFIED INVALID

(MailNews Core :: Internationalization, defect)

Description

[Koike Kazuhiko]

Subject in GB2312(MIME-encoded) is displayed as garbage on Linux.
Body text has no problem.

Comment 1

[Koike Kazuhiko]

Attachment: mail file

Comment 2

[Koike Kazuhiko]

Screenshot:
http://bugzilla.mozilla.gr.jp/showattachment.cgi?attach_id=476

Original report in Bugzilla-jp:
http://bugzilla.mozilla.gr.jp/show_bug.cgi?id=1751

Comment 3

[ji]

With the attached testing mail, I did see the problem with the subject display.
However, the testing mail is a mix of Simplied Chinese and Japanese characters.
In this case, it is possible to have this problem: Japanese characters are
mapped to gb2312 codepoints. And please note that the testing mail is a reply mail. 

I don't have any problems when viewing a gb2312 mail sent from mozilla.
This testing mail is sent from Becky!, another possible cause of the problem.

Comment 4

[Katsuhiko Momoi]

The test case shows that the body is encoded in GB2312 but the
real encoding used for the subject header is Shift_JIS.
Try decoding the Base64 for the header and see which encoding makes sense.
See the attached image below which was obtained when the encoding
was swicthed to Shift_JIS from GB2312.

Comment 5

[Katsuhiko Momoi]

Attachment: The header B64 string decoded as Shift_JIS

The subeject seems to have been in Shift_JIS originally but
was encoded usign BG 2312. How this happened exactly is a mystery
but perhaps you can do this with Becky.

Comment 6

[Katsuhiko Momoi]

You can set the folder properties encoding to Shift_JIS and
put a checkmark to force that encoding. You will see the 
subject header correctly also this way.

Comment 7

[ji]

Yes, i can see the Japanese subject after forcing the folder charset to
Shift-JIS, another good reason to have folder charset feature for this kind of
mails. :-)

Comment 8

[nhottanscp]

So the message is broken and the bug is invalid?

Comment 9

[Katsuhiko Momoi]

My guess is that this is a Becky-side bug. 
Most probably the original message had a Shift_JIS subject 
header and as the user replied to this message using GB2312,
it encoded it without converting it to GB2312. 
If you use Mozilla from the beginning, the subject line
in the image I attached can be encoded OK with BG2312 and
you can read it under that encoding. 

Comment 10

[Katsuhiko Momoi]

Marking it invalid for now unless there is additinal data
to indicate otherwise.

Comment 11

[ji]

Verified as such.

Comment 12

[ji]

Marked it as verified.