<<< Date Index >>>     <<< Thread Index >>>

mutt/2094: ESMTP patch: newline conversion and faulty error checking



>Number:         2094
>Notify-List:    
>Category:       mutt
>Synopsis:       ESMTP patch: newline conversion and faulty error checking
>Confidential:   no
>Severity:       normal
>Priority:       medium
>Responsible:    mutt-dev
>State:          open
>Keywords:       
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Tue Sep 27 11:33:40 +0200 2005
>Originator:     Ralf Wildenhues
>Release:        CVS HEAD + patch-1.5.11.bc.smtp.10
>Organization:
>Environment:
x86_64-unknown-linux-gnu
>Description:
The ESMTP patch proposed for mutt writes out the mail
header+body with plain \n instead of \r\n.  qmail does
not accept it.  Snippet from strace output:

write(4, "DATA\r\n", 6)                 = 6
read(4, "354 go ahead\r\n", 1024)       = 14
fstat(5, {st_dev=makedev(8, 5), st_ino=16476, st_mode=S_IFREG|0600, st_nlink=1, 
st_uid=1032, st_gid=1030, st_blksize=4096, st_blocks=16, st_size=387, 
st_atime=2005/09/27-10:50:54, st_mtime=2005/09/mmap(NULL, 4096, 
PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2aaaad4e5000
read(5, "Date: Tue, 27 Sep 2005 10:50:54 +0200\nFrom: Ralf Wildenhues 
<wildenhues@xxxxxxxxxxxxxxx>\nTo: Ralf Wildenhues 
<wildenhues@xxxxxxxxxxxxxxx>\nSubject: test8\nMessage-ID: <20050927085054.GA8
write(4, "Date: Tue, 27 Sep 2005 10:50:54 +0200\n", 38) = 38
lseek(5, 0, SEEK_CUR)                   = 387
write(4, "From: Ralf Wildenhues <wildenhues@xxxxxxxxxxxxxxx>\n", 51) = 51
write(4, "To: Ralf Wildenhues <wildenhues@xxxxxxxxxxxxxxx>\n", 49) = 49
write(4, "Subject: test8\n", 15)        = 15
write(4, "Message-ID: <20050927085054.GA832@xxxxxxxxxxxxxxx>\n", 51) = 51
write(4, "MIME-Version: 1.0\n", 18)     = 18
write(4, "Content-Type: text/plain; charset=us-ascii\n", 43) = 43
*snip*
write(4, ".\r\n", 3)                    = 3
read(4, "451 See http://pobox.com/~djb/docs/smtplf.html.\r\n";, 1024) = 49


After the send failure, I can edit/postpone the message,
but it is also wrongly stored in Fcc.


Is it ok, by the way, to use GNATS for not (yet?) accepted
patches?
>How-To-Repeat:
If desirable or necessary, I can produce a config file +
command sequence to reproduce this.
>Fix:
Unknown
>Add-To-Audit-Trail:

>Unformatted: