Steps to reproduce:
-sent from 4.6 the above page (URL is attached);
-get message in 5.0;
-click Forward (forward it to youself);
-click Send (manually dismiss the window);
-lcik Get new Mail:
//note there is no mail, message didn't get sent
(in the MS-DOS window read those comments:"JavaError uncaught
exception:component returned: failure###)
it doesn't occur when you forwarding html file from the local drive, it gets
sent fine, so my guess is MIME encoding (Multi-part?)
observed with 1999072208 M9 windows build
This is actually a problem in the XIF to DTD converter. We are getting lots of
asserts because of the content model for the message being composed. We can
display this data fine in the compose window, but it seems to have a real
problem trying to get it in HTML form.
Akkana...any ideas who I should assign something like this to?
I've inherited ownership of XIF. But I don't understand what the problem is
here -- Rich, can you explain why you say it's the XIF to HTML translation
causing a problem?
If I type an alt-X (Rich, please please could you consider adding a debug menu
like the one in the editor window to the mail compose window? Test Selection
and Output text/html/xif make debugging problems like this a LOT easier, esp.
for people on Windows where alt key events don't work) I do get lots of html
output, which has a valid-looking <head> and <body> and so forth.
Can you possibly add the menu that Akkana is talking about? This will help a
lot on message composition problems.
Are you just loading that URL in an editor window and looking at that? If so,
this is different than what the bug is. What happens is that we receive this as
an email attachment that we display inline. Then, we quote this into a new
message composition window. In the message compose window, it looks fine, but
when we hit Send and try to get the HTML, we get tons of asserts and it doesn't
It just looked like I was asserting in the DTD that did XIF to HTML..but I
could be off here.
No, I was testing this in the html mail window, forwarding the message as you
described, then using the hardwired editor alt-I accelerator (which may go away
once we get XUL keybindings, hence my request for the debug menu). OutputHTML
does produce html for this case.
Are there 2 bugs here? Here's what I see with the 8/16/99 Win32 build.
(Under HTML Mail send option)
1. Local mail folder: If the message with the abobe URL attached is in
POP3/local mail folder, then, the entire page seems to get
quoted -- though there is a nasty repaint problem (Bug 11573).
When this message is "sent", no message arrives at the recipient.
2. IMAP Mail folder:
When you reply/quote the message with the above URL attached (identical
test message as the 1 above), the original page is quoted only
partially -- actually it seems only the headers get quoted and then
the rest of the page is omitted.
When this message is sent, it arrives at the recipient but in
a partiallly quoted condition which it was in during the
(Also as a side note, the progress bar at the bottom left says "document
loading .." with this particular message either in local folder or
in IMAP folder. However, the attached page does seem to load entirely
and so this may not be relevant.)
Are both 1 and 2 due to the same cause or should we separate them out?
IMAP Quoting should be ignored. IMAP is a hack and is no indication of the
state of the editor...especially in cases with complex messages or non us-ascii
messages. IMAP will be covered when I get this working and I have a FEATURE bug
for that already.
I've changed the summary line to more accurately reflect the current
state of this bug.
Reassign to JF to add the menu Rich asked for on 7/26, and then hopefully we'll
be able to tell whether this is an editor bug or not.
Sounds like this needs to be fixed for PR1, so I added a note to the Status
I've already added the debug menu in M10. I will take a look at it soon...
(target milestone is M11 or M12 - add to mail beta tracking bug)
Rich, I assign this bug back to you. Message compose now have the debug menu you have requested for...
Bulk change to assigned.
This bug has been around for a while, but I believe it works now. I just went
through the steps to reproduce and I think everything behaved as it should.
Just in case, I am copying jefft since he worked on forwarding.
Verified as worksforme using the following build: