150530 - Character Coding is wrong when replying to multipart message

Status: RESOLVED FIXED

(MailNews Core :: MIME, defect)

Description

[Koike Kazuhiko]

Character Coding is wrong when replying to multipart message.

(1) Compose a message in ASCII with setting [View]-[Character Coding]
to ISO-8859-1, and save it as draft.

(2) Open the draft message with Forward button(set forward message 
setting to "As Attachment"), write Japanese text with setting 
[View]-[Character Coding] to ISO-2022-JP, and save it as draft.

(3) Open the draft message with Reply button.


Actual result: [View]-[Character Coding] is ISO-8859-1

Expected result: ISO-2022-JP


Original report in Bugzilla-jp:
http://bugzilla.mozilla.gr.jp/show_bug.cgi?id=2293

Comment 1

[nhottanscp]

Please specify mozilla build info.

Comment 2

[nhottanscp]

Reassign to nhotta.

Comment 3

[nhottanscp]

I think our current implementation for multipart takes a charset of the last
part which is correct for alternative type. For mixed type, we may want to do
differently.

Comment 4

[ji]

Attachment: A testing mail.

To reproduce the problem, save the mail to local folders, and then reply to the
mail.

Comment 5

[ji]

QA contact to myself.

Comment 6

[nhottanscp]

While debuging this, I found the charset is re-initialized by the mail font code.
I think that was not intended by bug 112904 (this also related to bug 107869).
cc to shanjian

Comment 7

[nhottanscp]

Attachment: Separated the menu charset code from the charset initialization function.

The font code should not change the menu check mark, so the menu code is
separated from the charset initialization function.

Comment 8

[Jean-Francois Ducarroz]

Comment on attachment 104592 [details] [diff] [review]
Separated the menu charset code from the charset initialization function.

R=ducarroz

Comment 9

[David :Bienvenu]

Comment on attachment 104592 [details] [diff] [review]
Separated the menu charset code from the charset initialization function.

sr=bienvenu

Comment 10

[nhottanscp]

Attachment: Make sure to set the menu charset for non multipart case.

Comment 11

[Jean-Francois Ducarroz]

Comment on attachment 104911 [details] [diff] [review]
Make sure to set the menu charset for non multipart case.

R=ducarroz

Comment 12

[ji]

Reopened, due to a regression

Comment 13

[David :Bienvenu]

Comment on attachment 104911 [details] [diff] [review]
Make sure to set the menu charset for non multipart case.

sr=bienvenu - I would use the ? operator here:
SetMailCharacterSetToMsgWindow(body, nsCRT::strcasecmp(text->charset,
"us-ascii") ?  NS_ConvertASCIItoUCS2(text->charset).get() ... since it might
generate more efficient code, but it would be a long line, so it's up to you.

Comment 14

[nhottanscp]

I filed a bug 178680 to move the charset name comparison to inside
SetMailCharacterSetToMsgWindow().

Comment 15

[nhottanscp]

checked in to the trunk