71541 - Should use default charset for UNKNOWN-8BIT

Status: RESOLVED FIXED

(MailNews Core :: Internationalization, defect)

Description

[John G. Myers]

RFC 1428 defines the charset "UNKNOWN-8BIT" to refer to data for which the 
encoder does not know the charset.  The MIME encoded-word decoder should use the 
default charset for decoding this data.  Currently the decoder passes 
"unknown-8bit" down to the charset converter, which in turn treats the 
unrecognized charset as iso-8859-1.

Comment 1

[John G. Myers]

Attachment: Proposed fix

Comment 2

[John G. Myers]

Attachment: Corrected proposed fix

Comment 3

[Katsuhiko Momoi]

The onyl exception to this is that when an auto-detection
module is tunred ON, that should take precedence over
the view default charset.
So please extend the applicability of auto-detection
to cases of charset=unknown-8bit.

Should we file a separate bug for this?

Comment 4

[John G. Myers]

Autodetection doesn't apply to headers.  But we should file a separate bug or
extend this bug to address message bodies.  Changing the summary to extend this
bug to message bodies.  I will address message bodies in a followup patch.

Comment 5

[Jaime Rodriguez, Jr.]

Adding intl keyword, and jenm to cc: list.

Comment 6

[Alexey Chernyak]

I think at the moment Messageview uses specified Folder encoding to display
headers with no MIME set.

John, please have a look at bug 77903.

Comment 7

[John G. Myers]

Bug 77943 is orthogonal to this bug.  This bug refers to headers which are
encoded in MIME encoded-words, but have a charset label of "UNKNOWN-8BIT".

Could I please have a review?  This bug has been blocked waiting for review for
a year and a half.

Comment 8

[ji]

QA contact to marina. Thanks.

Comment 9

[nhottanscp]

Comment on attachment 27343 [details] [diff] [review]
Corrected proposed fix

r=nhotta

Comment 10

[Christopher Blizzard (:blizzard)]

Comment on attachment 27343 [details] [diff] [review]
Corrected proposed fix

sr=blizzard

Comment 11

[John G. Myers]

Fixed on trunk.