23635 - Mails with wrapped subject header shows wrong subject in both thread pane and message pane

Status: VERIFIED FIXED

(SeaMonkey :: MailNews: Message Display, defect, P3)

Description

[Henrik Gemal]

If you send a mail with this subject:
1234 5678 1234 5678 1234 5678 1234 5678 1234 5678 1234 5678 1234 5678 1234 5678
1234 5678 1234 5678 1234 5678 1234 5678 1234 5678 1234 5678 1234 5678 1234 5678
1234 5678 1234 5678 1234 5678 1234 5678 1234 5678 1234 5678 1234 5678 1234 5678
1234 5678 1234 5678 1234 5678 1234 5678 1234 5678 1234 5678 1234 5678

And read the mail in Mozilla there are some squares inserted into the subject
after the 14'th block of numbers.

Using build 2000011008

Attached is a image of the subject problem

Comment 1

[Henrik Gemal]

Attachment: Very long subject with squares!

Comment 2

[Henrik Gemal]

This also affects the subject when replying. Then there seems to be inserted
"spaces".
A new bug or an effect of this?

Comment 3

[lchiang]

henrik - probably the same cause.

Comment 4

[Phil Peterson]

Reassign to ducarroz, cc rhp. Marking M16

Comment 5

[Henrik Gemal]

The "[]" is inserted if the Subject is wrapped into two lines in the mail 
headers. So if you have a mail with the following headers:
---
To: [e-mail]
Subject: [Fwd: [OFFTOPIC] Re: Where is store the address book in netscape 
 messanger?]
MIME-Version: 1.0
---

A square is inserted after the word "netscape". This must be a easy fix!

Comment 6

[Henrik Gemal]

Still not fix on this bug...?

Comment 7

[Jean-Francois Ducarroz]

nope.

Comment 8

[lchiang]

this is marked M16 so should be fixed by then.  Right now, focus is on very 
"bad" bugs for M14

Comment 9

[Jean-Francois Ducarroz]

maybe for you rich?

Comment 10

[rhp (gone)]

I guess it could be but I'm kind of swamped right now...

- rhp

Comment 11

[selmer (gone)]

Not beta2 stopper.  Marking M18.

Comment 12

[selmer (gone)]

Oops.  Really marking M18 now...

Comment 13

[Henrik Gemal]

Just giving a better bug-subject since this doesnt have anythning to do with 
very long subjects. It has to do with wrapped subjects that doesn't get 
unwrapped properly somewhere deepdown in some read header stuff.
A wrap To field works fine...

According to 3.1.1. 
http://info.internet.isi.edu:80/in-notes/rfc/files/rfc822.txt mozilla doesn't 
unfold correct since:
Subject: test of wrap
should be the same as:
Subject: test
                    of wrap
Subject: test
 of
     wrap
Subject: test
           of
           wrap

Comment 14

[Henrik Gemal]

Attachment: Mail with wrapped subject

Comment 15

[sol]

Moving to future milestone. If this is a high frequency problem, please indicate 
this in the bug so we can reconsider.

Comment 16

[Henrik Gemal]

This is very high frequency. How do I indicate this...?

Comment 17

[Jean-Francois Ducarroz]

Reassign to rhp, this looks like a mime parser problem.

Comment 18

[Henrik Gemal]

BTW: this also affects sorting of messages. See attached screenshot of messages 
sorted by Subject.

Comment 19

[Henrik Gemal]

Attachment: Screenshot of sorting by Subject...

Comment 20

[pmock]

Changing qa assigned to myself.

Comment 21

[John G. Myers]

Currently header unfolding is only done in the presence of MIME encoded-words.  
This is incorrect, headers should always be unfolded.

Blocks a nsbeta1 bug, so nominating.

Comment 22

[(not reading, please use [e-mail] instead)]

taking from rhp.

Comment 23

[John G. Myers]

The header unfolding algorithm is described in RFC 822 section 3.1.1:

	   Unfolding  is  accomplished  by
        regarding   CRLF   immediately  followed  by  a  LWSP-char  as
        equivalent to the LWSP-char.

And in section 3.3:

     LWSP-char   =  SPACE / HTAB                 ; semantics = SPACE

Comment 24

[scottputterman]

marking nsbeta1+ and reassigning to ducarroz.

Comment 25

[John G. Myers]

Attachment: Proposed fix

Comment 26

[John G. Myers]

My proposed fix causes the MIME encoded-word decoder to unfold header 
continuations even if there are no encoded-words present in the input.  It also 
changes MIME_StripContinuations() to conform to RFC 822.

Note that this will cause a temporary regression in the handling of MIME encoded 
words.  Whitespace will now appear between successive encoded-words where it did 
not before.  This temporary regression will be fixed by the patch to bug 58114.

Comment 27

[Jean-Francois Ducarroz]

accepting the bug.

Comment 28

[scottputterman]

moving to mozilla0.9

Comment 29

[scottputterman]

though, if we can get this reviewed and checked in, it can of course go in
before mozilla0.9

Comment 30

[Jean-Francois Ducarroz]

looks good. R=ducarroz.

Comment 31

[David :Bienvenu]

sr=bienvenu

Comment 32

[Jean-Francois Ducarroz]

one more thing, DON'T REMOVE YET MOZ_SECURITY CODE. Thanks

Comment 33

[John G. Myers]

reassign to self

Comment 34

[John G. Myers]

Fix checked into tip.

Comment 35

[Henrik Gemal]

looks fine now!!!!