Closed Bug 167634 Opened 23 years ago Closed 22 years ago

text isn't wrapped when sending email

Categories

(MailNews Core :: Composition, defect)

x86
Windows 2000
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED INVALID

People

(Reporter: mno2go, Assigned: bugzilla)

References

Details

Attachments

(1 file)

When I compose email, the auto-wrap feature works perfectly. However, when I send the message, all wrapping is lost and the text auto-wraps to screen width. Executing Rewrap before sending fixes this issue, however, text should be sent as it is when composing.
*** Bug 167635 has been marked as a duplicate of this bug. ***
BUILD ID ?
Sorry :) Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.1) Gecko/20020826
Reporter, is this still happening? Can you test on a newer build?
Yup. I'm using BUILD Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.2b) Gecko/20021016 at this moment.
I can confirm exactly this behaviour on Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.2) Gecko/20021126. Yes, I need to upgrade to 1.2.1. But as this bug is around for a long time, and not even confirmed here, I guess it's still unfixed. I consider this the second most annoying issue in Mail. The most annoying being the incredible slowness of the search facility, of course. So please mod this bug up.
Confirming per reporters.
Status: UNCONFIRMED → NEW
Ever confirmed: true
I have 1.2.1 installed now. It's still there...
The Problem also exists in 1.3a and 1.3b. Currently I am using Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.3b) Gecko/20030210 and it still is in there. It is deterministic so it should not be hard to reproduce it. There are two workarounds: - Perform "rewrap" action before sending - "Send later" -> "Edit As New" -> "Send" (and delete duplicate) To me (and some others) this problem is the most annyoing one in Mozilla. My settings for mail: Composition: Wrap text messages at 72 characters (use quoted printable MIME ecncoding setting does not matter) Send format: 'Convert the message to plain text'
Yes, I have 1.3b installed now and the problem still exists. The workaround is too long for me to really bother :) Do you still want a before-and-after screenshot?
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030529 No wrapping in outgoing Mail :(
Windows XP Mozilla 1.4b, still there.
I can no longer confirm this bug. Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030529 I'm not sure if it really is fixed, so I won't close. Anyone else? Thanks to all volunteers for their great work! Mozilla really is a fine product.
What an insidious bug. After I noticed that Markus uses the same build as I do (only on X instead of Windows) and still experiences the problem, I decided to investigate further. I'm doing so for two long hours now and I've been misguided a couple of times, but I think I finally know the real problem now. Surprise: it's Mozilla's Mail Viewer, not the Compositor! For those in a hurry: This bug can be closed, wrapping of outgoing mail works fine. (Please confirm) The fact that Mozilla concatenates text lines in plaintext mails is probably responsible for people (perhaps including me) believing that there is something wrong with wrapping. Not sure if this applies to the whole lifetime of this bug, but this seems to be the case in 1.4. Read on for details. The following tested on Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030529. Please confirm that you can reproduce this behaviour. 1. Outgoing Mail - Use Edit->Preferences->Mail->Composition and set the wrap width to your liking. Note that this setting doesn't affect any Compose windows already open. - Compose and send a mail to yourself. Write a couple of paragraphs without inserting hard breaks. - Use "View->Message Source" on either the "Sent" message or the one you'll find in your Inbox, and you should see the body properly formatted using the wrap width you specified. So, there is no bug in outgoing mail wrapping, at least no longer! 2. Incoming Mail - Make sure that "Preferences->Mail->Message Display->Wrap text to fit window width" is disabled (!!) - For testing purposes, send a mail to yourself with a couple of paragraphs wrapped at only 20 or 30 characters and open it using Mozilla's Mail viewer. - Play around with the right window border, and watch how Mozilla wraps the short lines in your test mail according to the width of the message window. - You should notice the following "feature": if the Window is wide enough, Mozilla tries to concatenate consecutive lines of text when they are not separated by an empty line!
Confirming as per Comment 14. The weird thing is, though, that other people's messages ARE NOT "unwrapped" - only the ones I send. Very interesting. Anyone else confirming this?
Err... not "confirming" but "seeing"... :)
I'm seeing this too. I tried wrapping the text at 30 characters and then send a message to myself. It did wrap the text while composing, but when it when I recieved it, it didn't show as wrapped. I then tried to send a similar message to an address using MS Outlook and it arrives wrapped there. Using "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.4) Gecko/20030529"
Julian is right, the mails are correctly wrapped but expanded in the mail viewer. The mail source viewer shows them ok.
I guess then that this bug has to be reassigned to those working on the viewer, not the composer. Regarding Comment 14, I don't think that "Wrap text to window width" has to be unchecked. From what I can tell from that description, it should only wrap text to fit width when needed, ie when there are no line breaks in the text. It shouldn't "unwrap" text that is wrapped. Is that a bug?
Max is right, this option doesn't affect the behaviour. I was emphasizing this only because one would expect that Mozilla leaves the text layout alone when not told to wrap lines. I think the Mail Viewer *removing* linefeeds is definitely a bug. No text viewer does something similar on purpose, correct me if I'm wrong. Now that Max, Joergen and Martin have confirmed my findings and we seem to agree that this behaviour is at least "strange", we should either mark this bug as invalid and open a new one, or mark it as a duplicate of the real bug, if such a bug is already filed. I'll browse Bugzilla later to see if this is the case.
Found it - we have been fooled by bug #101877 which seems to be known since late 2001. I wanted to mark this bug as a duplicate, but only the owner can do that. Jean-Francois? I'm going to add a pointer to this bug to #101877 now.
I think this Feature/Bug has to do with Content-Type: format=flowed More info on this can be found here: http://www.ietf.org/rfc/rfc2646.txt http://www.eudora.com/techsupport/kb/1626hq.html If "format=flowed" is removed from the content-type, the viewer shows the message with "correct" wrappings. I actually think that the composer and the viewer work correctly, but maybe it should be possible to disable "format=flowed" in the composer?
It seems that this is an invalid bug, but I will wait another few days for anyone to make a point as to why this bug isn't invalid. If no comments are added, I'll mark this bug as invalid. Regarding Comment 22, I think that this comment should also be posted under bug 101877 as well as it seems to refer more to that bug rather than this one.
Comment 22 explains everything. I was misguided again. Thanks Mathias, this should finally end the guessing! However, let me add that Mozilla's current behaviour is NOT correct. Let me try to put this together, I think I get it now. This text too long? Brief summary follows in next comment. It's in fact a bug in outgoing mail, which doesn't implement RFC2646 correctly. It claims to be compatible by declaring the "format=flowed" header, but it wraps to the user specified width and at the same time adds a blank to the end of each line - therefore rendering all lines "flowed" in terms of RFC2646, though they have been pre-wrapped. And pre-wrapping doesn't seem to be the intention of the flowed format. An easy way to verify the fact that Mozilla appends a blank to each wrapped (non-wrapped too?) line is by looking at the source of a sent message and selecting a couple of lines with the mouse - the inverted text will show a blank at the end of each line. That's not the only proof I have though, see below. The viewer component is correct in interpreting these lines as flowed lines (because the header value is present and the lines end in a blank) and combining them to paragraphs. The relevant quotes from the RFC are: "If the line ends in one or more spaces, the line is flowed. Otherwise it is fixed." (4.2) and "A Format=Flowed message consists of zero or more paragraphs, each containing one or more flowed lines followed by one fixed line. The usual case is a series of flowed text lines with blank (empty) fixed lines between them." (4.2) and "Consecutive [flowed] lines with the same quoting depth are considered one paragraph and are reformatted together." (4.5) Therefore, incoming is OK and the solution for outgoing is obvious: choose fixed format by removing the trailing blank and/or by omitting the header, or be damned and choose flowed format by not wrapping altogether and leaving everything else as-is (make sure to append that blank though). The best solution would be to let the user choose of course. I can imagine one radio button "Flowed format (no wrap)" and one radio button "Fixed format" which enables the text field for entering the desired wrap width. I prefer fixed format. Besides, this bug would be pointless if we were talking flowed. But I guess there are reasons for flowed format as well, why else would I use this fscking small HTML textbox. BTW that also explains why the odd behaviour only occurs with mails generated by Mozilla, as Max noticed. Seems like our foes implement RFC2646 better in this respect... until now! <evil laugh> I have checked my facts. Qmail users can inject the demo attached by doing 'qmail-inject youraccount < rfc2646demo.txt'. Don't know how it works with Sendmail et al. Note that when you remove the "format=flowed" header, you'll see that Mozilla Mail Viewer regards this change and doesn't interpret "flowed" lines any longer. Again, it seems the viewer is absolutely correct. I'll notify bug 101877 as well - thanks to Mathias' hint, it looks like "our" bug is now more relevant than "theirs". Ok, just kidding. Say, isn't this another demonstration of what huge a difference a tiny, even invisible blank can make?
This bug, as well as bug 101877 can be solved simply by stripping all whitespace at the end of each line of outgoing mail. More information in comment 24.
Qmail users can send this test mail using the following command: qmail-inject youraccount < rfc2646demo.txt
Very interesting. I hate these small bugs that don't seem to be bugs but really are. They're usually very difficult to find but have amazingly simple solutions :)
Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030612 Outgoing mail still mixes flowed and fixed format.
Oh my... I'm only adding more and more confusion to this discussion instead of helping to clarify the situation. After re-reading RFC 2646, it turns out I've been mistaken once again! Fourth time or so. Go ahead and laugh at me. Actually, now it looks to me like Mathias was right in the first place: both the Composer and the Viewer are correct, though the Viewer doesn't expose the ideal behaviour. In particular, I failed to read the following quote from section 4.2: "A series of one or more flowed lines followed by one fixed line is considered a paragraph, and MAY be flowed (wrapped *AND UNWRAPPED*) as appropriate on display and in the construction of new messages (see section 4.5)." (emphasis added) and, more importantly, the following from section 4.1: "A generating agent SHOULD: 1. Ensure ALL lines (fixed *AND FLOWED*) are 79 characters or fewer in length [...]" (emphasis added) So I was wrong when saying that in flowed format, lines shouldn't be wrapped at all. Composer seems to be correct in regarding the user-specified wrap column even for format=flowed messages (which makes a lot of sense for downward compatibility.) Furthermore, I was assuming that the viewer is correct in concatenating two consecutive flowed lines to one line when there is enough room, and leaving the lines as-is when the window width isn't enough. However, while this behaviour doesn't directly violate the RFC, it's not the ideal behaviour. The ideal (and also compliant) behaviour would be to reorganize "flowed" paragraphs regardless of where "soft breaks" (i.e. SP CRLF sequences) are found in the paragraph. In short, it looks like nothing is really broken, but Viewer should ideally reorganize flowed paragraphs freely. And as Mathias already said, you should ideally be able to switch between format=fixed and format=flowed in Composer. I justed wanted to point out this my mistake. It's probably better if I shut up now before I have to backup once again, so I'll leave the discussion for now. I *think* I should now understand the whole situation well enough to write a detailed "problem/solution" comment, but I thought so three times before so I'll refrain unless someone explicitly asks me to do so. Sorry for this mess.
See bug 168420. Seems like there's a LOT to say about format=flowed. I still think something is wrong with the system, probably the viewer. It's kind of unnatural that the viewer concatenates only full text lines, not on the word level. I think that's not how RFC2646 is meant.
I'm resolving this bug as Invalid, per Max Nokhrin's comment 23. People voting for this bug are invited to check out bug 168420, which includes a list of the bugs that *do* exist with f=f; if one of those strikes your fancy, move your vote to that. (If not, I've got a couple of pet bugs that I'd like to see gain a few more votes... :) Julian: I really do not know what you're talking about in comment 30, when you say "the viewer concatenates only full text lines, not on the word level." The viewer concatenates lines (for a f=f message, and with f=f display enabled) when the first line ends in a space, no matter how long either line is. As far as I've been able to tell with my testing, the viewing of f=f messages works as it should; the problems I'm aware of are glitches in mail composition.
Status: NEW → RESOLVED
Closed: 22 years ago
Resolution: --- → INVALID
Product: MailNews → Core
Product: Core → MailNews Core
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: