When Replying or Forwarding message with non-ascii in the body it is garbled

VERIFIED FIXED in M12

Status

defect
P3
normal
VERIFIED FIXED
21 years ago
11 years ago

People

(Reporter: marina, Assigned: rhp)

Tracking

Trunk
x86
Windows NT
Dependency tree / graph

Firefox Tracking Flags

(Not tracked)

Details

(Whiteboard: DEPEND - Intl - need to retest with latest builds?)

Steps to reproduce:
open an e-mail with non-ascii chars in the mail body|click Reply or Forward
//note: there is no proper conversion done for the string
question: do you see the text broken before send in the editor or broken in the
received message?
I see it broken before sending.
Status: NEW → ASSIGNED
Target Milestone: M6
I will try to fix the display problem. But looks like reply/forward is not
officially implemeneted for M5 (see 5493). We also have Ender related bug for
the body
sending.
Marking as M6.
Assignee: nhotta → ducarroz
Status: ASSIGNED → NEW
This is what I saw in the view where =C9 supposed to be É.
> abc=C9xyz>>
In above case, the text should be QP decoded. The same kind of problem could
happen if we support HTML (e.g. named entity like &Eacute).
I think we need more generic solution for these kind of problems.
Reassigning to ducarroz@netscape.com.
Status: NEW → ASSIGNED
Reply and Forward are totally broken since few days. I will take a look at it.
Reply/Forward are back now in plain text and HTML. marina, momoi could you test
again to see if the problem still exists.
Whiteboard: need to retest with latest builds?
If we think this is fixed lets mark it so, get it on the test radar,
and then re-open if its not..
Whiteboard: need to retest with latest builds? → DEPEND - Intl - need to retest with latest builds?
Forward and Reply of:

1. Latin 1 message body is now working for both Plain Text & HTML Mail.
2. Japanese message body is broken in display and also as a result of
   "send":
   a. The body is quoted in raw JIS when composer displays it.
   b. It seems that the JIS code point is mistakenly interpreted as
      "<" of HTML. Thus the body text after this point is not
      readable as Japanese. See this example,
      ("Kore ha Nihongo no meeru desu') below:

     $B$3$l$OF|K\8l$N%a!&lt;%k$G$9!#(B   (Note "&lt;"�j

     The correct sequence is:

     $B$3$l$OF|K\8l$N%a!<%k$G$9!#(B

     This problem of "<" is already visible in the Composer window.

Naoki, does the lack of Frank's Meta-tag code contribute to these
2 problems or are they independent?

The results for Japanese were obtained for both Plain text and HTML
mail settings.

Since both Latin 1 and Japanese Reply/Forward are M6 International
requirements, I can't mark this "resolved".
** The above results were obtained with the 5/22/99 Win32 build **
Regarding the question by momoi, I think sending is not affected by the META
problem.
Last time I mentioned this quoting problem, I was told this was not in feature
but now it's actually in (see http://www.mozilla.org/mailnews/milestones.html).

I checked the code nsComposeAppCore::HackToGetBody (nsComposeAppCore.cpp),
it has mMsgCompFields->SetBody(nsAutoCString(msgBody), NULL);
which only works for Latin1.
I don't know this is a temporary soultion or not (it was when I asked last
time).
Also, if this part is going to changed for the Ender related bug (6672 - M7)
then we need to change the milestone.
Target Milestone: M6 → M7
moved to M7
Now i'm getting Parsing error in the message header with 8-bit chars after
receiving forwarded or replyied messages
****observed with 5/24 Win build****
Blocks: 7228
Target Milestone: M7 → M8
Assignee: ducarroz → rhp
Status: ASSIGNED → NEW
Reassign to rhp as it should be a backend issue.
Status: NEW → ASSIGNED
Target Milestone: M8 → M9
It looks like the Latin-1 chars conversion bug is back.Forwarding a message that
was sent  as "Inline" or with the option "As attachement" turn all non-ascii
chars into garbage before sending (so it is not a send but i guess a display
problem). Obsereved with 1999072608 windows build.
Status: ASSIGNED → RESOLVED
Closed: 20 years ago
Resolution: --- → WORKSFORME
maybe I'm reading this wrong, but this seems to work for me. Can you verify if
this is still broken.

- rhp
I don't know how to verify this bug sisnce Reply and Forwarding is broken (bug
#11094). It might be that the reopening of this bug (5498) was a regression and
build problem but i guess Naoki fixed so after i can really verify it i would
say resolution RESOLVED as FIXED would be more appropriate.
sounds reasonable to me :-)

- rhp
QA Contact: momoi → marina
Assigning to marina for verification.
Status: RESOLVED → REOPENED
I verified this bug with M9candidate 1999-08-24 on Windows and here are the
results:
1.send a message with a string containing non-ascii in the body using Latin1
encoding.
 -get message back,select it and click Reply (or Forward);
//note: all high ascii are garbled or the string is truncated before non-ascii
char.
(same is true for Newsgroup Reply or Forward)
2. repeating all of the above steps with UTF8 encoding returns the correft
results (it is true for Newsgroups). So i think we have to clear the resolution
WORKSFORME, i'm reopening this bug
Resolution: WORKSFORME → ---
Additional comments: those results are incorrect for Latin1 only on IMAP and
News server. (Latin-1 string are either truncated or garbled in the message
body) The POP protocol doesn't have this problem.
This was observed with 1999-08-24-12 M9 build for Windows. I would say " a
stopper". The only good thing is it is working correctly for POP3 protocol and
with UTF8 encoding on IMAP and News server.
Target Milestone: M9 → M10
per both msanz and momoi...this will not be a stopper for M9.  Puting on M10
radar.  marina, you could add text for a Release Note to:
http://bugzilla.mozilla.org/show_bug.cgi?id=11352
This will be added to M9 Intl Release notes -- which will be done later today. Thanks!
Target Milestone: M10 → M11
Bulk move rhp bugs to M11 for now
Target Milestone: M11 → M15
Since Rich is out, and this bug is not mentioned in 12907, I'm moving to M15
Status: REOPENED → ASSIGNED
Bulk change to assigned.

- rhp
Target Milestone: M15 → M12
This is the same technical issue we are attacking in Bug #17034. When I fix
one, I should fix the other.

- rhp
Status: ASSIGNED → RESOLVED
Closed: 20 years ago20 years ago
Resolution: --- → FIXED
I believe that most quoting/forwarding issues should be fixed with my latest
checkin.

- rhp
Status: RESOLVED → VERIFIED
Verifiyng with 1999-12-14 win build. I don't see it happening any more.
Product: MailNews → Core
Product: Core → MailNews Core
You need to log in before you can comment on or make changes to this bug.