435020 - A patch formatted by git-format-patch is sent incorrectly by Thunderbird; does not follow it's draft

Status: RESOLVED DUPLICATE of bug 204478

(Thunderbird :: General, defect)

Description

[Jeremy White]

User-Agent:       Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.8.1.12) Gecko/20080129 Iceweasel/2.0.0.12 (Debian-2.0.0.12-1)
Build Identifier: Multiple; on Ubuntu Hardy 2.0.0.14, and on Debian/Lenny 2.0.0.9

I use git-format-patch | git-imap-send to create an email in my drafts folder.

I edit the draft and review it; it is perfect; perfectly formatted, everything.

I then click send, and the actual email *sent* is *not* what was contained in the draft.  Specifically, spaces are injected prior to some lines, which essentially
makes the patch unusable via diff and it doesn't work.

Using git-format-patch --attach does work, but it seems it shouldn't be necessary.

I will attach an example of a draft I used and the resulting sent email

Reproducible: Sometimes

Steps to Reproduce:
git-format-patch .... --stdout | git-imap-send
click on drafts folder, review the draft email, confirm it is correct
click send
click on sent folder, review what actually was sent
Actual Results:  
The message actually sent will sometimes, not always, but often, have spaces
injected prior to lines, thereby mashing up the diff.

Expected Results:  
For the message actually sent to be identical to my draft.

I will attach a draft and a sent email that shows the issue.

Comment 1

[Jeremy White]

Attachment: A sample draft email

A sample draft email that fails.

Comment 2

[Jeremy White]

Attachment: A sample resulting sent file

Comment 3

[Jeremy White]

Arg!  I'm having a hard time getting an exact useful reproduction of this bug.

That is, if I make a draft message just as I have been all along, save that draft
(so as to attach it here), and then send the message back to myself, the
resulting message is correct.

The problem does not occur :-(.  I will make some effort to pin down exactly what varies to cause this (e.g. not saving the draft first or sending it outside my own domain).

Comment 4

[Jeremy White]

Ah.  This must be a side effect of flowing the text.  

Hrm.  Is there any chance the flow logic could be adapted to detect these
cases and not kick in?  (For that matter, is the flow logic documented anywhere; playing with sending email, I can't figure out just how it works...)

Comment 5

[Magnus Melin [:mkmelin]]

Bug 204478?

Comment 6

[Jeremy White]

Yup, looks like it.