Closed
Bug 7715
Opened 27 years ago
Closed 27 years ago
Messenger to set force charset to webshell
Categories
(MailNews Core :: Internationalization, defect, P3)
MailNews Core
Internationalization
Tracking
(Not tracked)
VERIFIED
FIXED
M7
People
(Reporter: nhottanscp, Assigned: nhottanscp)
Details
Since messenger always generates UTF-8 HTML (and XML), it is preferred to set
force charset as
UTF-8 instead of putting META (which causes a redraw).
Currently, nsParser checks hard coded messenger's temp file name to set the
force charset. Instead,
messenger should set force charset to UTF-8 for its webshell.
Once messenger's change completes, please reassign to me (nhotta@netscape.com).
So I can remove
the hack in the parser.
Updated•27 years ago
|
Assignee: mscott → nhotta
Target Milestone: M7
Comment 1•27 years ago
|
||
Naoki, I just checked in a change to nsMessenger.cpp which sets the character
set of the messenger webshell to UTF-8 (on the webshell).
Just so we are clear: this isn't the webshell used when you go to display a
message. But it is a parent of the webshell(s) we use for displaying messsages.
So if the character set is passed down to the child webshells we should be okay.
Re-assigning to nhotta as my part here is done.
(Also setting TFV to M7)
| Assignee | ||
Updated•27 years ago
|
Status: NEW → ASSIGNED
| Assignee | ||
Comment 2•27 years ago
|
||
Scott, the webshell you set a charset is a parent of two (header and body),
is this correct?
Or webshell of whole messenger (including thread pane and folder list)?
Comment 3•27 years ago
|
||
I set the character set of the webshell for the message window in Messenger.
The two web shells (header & body) are created by layout but I believe are
children of the this webshell.
Basically we load some frame HTML code into the message pane webshell which
causes layout to indirectly create a webshell for the header frame and a
webshell for the body frame. They are created in the context of the message pane
webshell though.
| Assignee | ||
Comment 4•27 years ago
|
||
Thanks, that's good. I was worried about that might had some effects to .XUL
(it's UTF-8 but can specify something else). So far, I tested on Windows and see
no problem. I probably check in to remove the hack tomorrow.
Comment 5•27 years ago
|
||
I think we really need to test this against Linux or Mac before checking this
in. Windows wasn't the big trouble maker with the re-entrancy problem that the
code you are backing out was trying to fix. Linux & Mac were the ones that had
the problem the worse. So we should check against one of those platforms.
We can try a fix out for you if you need us to.
| Assignee | ||
Comment 6•27 years ago
|
||
Thanks, I need your help. Here is the diff.
Index: nsParser.cpp
===================================================================
RCS file: /cvsroot/mozilla/htmlparser/src/nsParser.cpp,v
retrieving revision 3.103
diff -c -r3.103 nsParser.cpp
*** nsParser.cpp 1999/06/03 22:38:33 3.103
--- nsParser.cpp 1999/06/08 00:47:47
***************
*** 624,636 ****
if (rv != NS_OK) return rv;
nsAutoString theName(spec);
- // Temp hack to prevent reload on mail message
- if(-1 != theName.Find("tempMessage.eml"))
- {
- nsAutoString utf8("UTF-8");
- SetDocumentCharset(utf8, kCharsetFromPreviousLoading);
- }
-
CParserContext* pc=new CParserContext(new nsScanner(theName,PR_FALSE,
mCharset, mCharsetSource),aKey,aListener);
if(pc) {
pc->mMultipart=PR_TRUE;
--- 624,629 ----
| Assignee | ||
Comment 7•27 years ago
|
||
I tested Macintosh using the i18n smoke test. I also tested a html mail with
ISO-8859-1 META tag (which had the infinite reload problem).
My linux build crashes at start up. It would be great if someone in mail/news
could try this change for me. Thanks.
Comment 8•27 years ago
|
||
I am trying it out on Linux. I just haven't gotten to it yet. Sorry if I wasn't
clear about that earlier in the bug report.
I haven't heard about any start up problems with todays build on Linux. I'm
looking into that now.
| Assignee | ||
Comment 9•27 years ago
|
||
Ok, I was finally able to test it on Unix. I tested the same cases as Macintosh.
I didn't see any problem caused by this change (at least didn't see infinite
redraw). I will check this in tomorrow.
Comment 10•27 years ago
|
||
My linux build with naoki's patch in layout looked good too. Ship it =).....
| Assignee | ||
Updated•27 years ago
|
Status: ASSIGNED → RESOLVED
Closed: 27 years ago
Resolution: --- → FIXED
| Assignee | ||
Comment 11•27 years ago
|
||
Checked in mozilla/htmlparser/src/nsParser.cpp (rev. 3.104).
Verification note, i18n smoke test. HTML mails (or mail with attachment) with
ISO-8859-1 (or whatever) META tag.
Notte: Cross platform verification needed but IQA is only available for Windows
verification (correct me if I am wrong).
Updated•27 years ago
|
Status: RESOLVED → VERIFIED
Comment 12•27 years ago
|
||
** Checked with 6/14/99 Win32, Mac and Linux builds **
I've been using the above builds and other recent builds for the
past several days.
I'm now reasonably satisfied that the original problems which
necessitated all the fixes including this one do not occur any more.
Specifically,
1. Messages which contain a met charset tag in the main body of the message
or in attachments do not cause inifinite loop in Win or Mac.
2. On Linux, in addition to the data in 1, pratically any message was
causing a crash. This does not happen any more.
Based on these results, I'll mark this ix verified.
Updated•21 years ago
|
Product: MailNews → Core
Updated•18 years ago
|
Product: Core → MailNews Core
You need to log in
before you can comment on or make changes to this bug.
Description
•