Paul Hinze <paul.t.hinze@gmail.com> on 2008-12-22 at 19:16:
> I did a bit of debugging and asking around in #python and it turns out
> that because Maildir.py uses a conditional import of md5 from either
> hashlib or the md5 module, the md5.new() call is not always correct.
> The attached patch should fix this problem.
Thanks to Daniel Rall for pointing out that the attachment got stripped
off my last message. Patch follows.
Paul
----
From e01fdfbf62c61c31b514e826e6cc00fb98d88808 Mon Sep 17 00:00:00 2001
From: Paul Hinze <paul.t.hinze@gmail.com>
Date: Mon, 22 Dec 2008 18:42:44 -0600
Subject: [PATCH] Fixing md5 import when hashlib is available
Thanks to habnabit and Rhamphoryncus in #python for help debugging
Here's the way I'd like to use offlineimap on my laptop:
1. Have a regular cron job running infrequently. The cron job
checks to see
if I'm online, plugged in, and that no other copy of offlineimap is
running. If
all of these conditions are satisfied, it runs offlineimap just once:
"offlineimap -o -u Noninteractive.Quiet"
2. When I start up mutt, I do it by calling a wrapper script that
delays
until cron-started copies of offlineimap have finished, then starts
offlineimap
on its regular, stay-alive and keep checking schedule. When I quit
mutt, the
wrapper script tells offlineimap to stop.
This way I get frequent regular checks while I have mutt running, but
I don't
waste my battery/cpu checking frequently for mail when I'm not
interested in
it.
To make this work, though, it'd be nicer if it were easier to tell
offlineimap,
from the outside, things like "terminate cleanly now" and "when you've
finished
synching, then terminate instead of sleeping and synching again."
OK, to put my money where my mouth is, I attach two patches against
offlineimap
6.0.3.
The first, "cleanup.patch", cleans up a few spots that tend to throw
exceptions
for me as offlineimap is exiting from a KeyboardInterrupt.
The second adds signaling capabilities to offlineimap.
* sending a SIGTERM tells offlineimap to terminate immediately but
cleanly,
just as if "q" had been pressed in the GUI interface
* sending a SIGUSR1 tells every account to do a full sync asap: if
it's
sleeping, then wake up and do the sync now. If it's mid-sync, then
re-synch
any folders whose syncing has already been started or completed, and
continue
to synch the other, queued but not-yet-synched folders.
* sending a SIGHUP tells every account to die as soon as it can (but
not
immediately: only after finishing any synch it's now engaged in)
* sending a SIGUSR2 tells every account to do a full sync asap (as
with
SIGUSR1), then die
It's tricky to mix signals with threads, but I think I've done this
correctly.
I've been using it now for a few weeks without any obvious
problems. But I'm passing it
on so that others can review the code and test it out on their
systems. I developed the
patch when I was running Python 2.5.2, but to my knowledge I don't use
any Python 2.5-specific
code. Now I'm using the patch with Python 2.6.
Although I said "without any obvious problems," let me confess that
I'm
seeing offlineimap regularly choke when I do things like this: start
up
my offlineimap-wrapped copy of mutt, wait a while, put the machine to
sleep (not sure if offlineimap is active in the background or idling),
move to a different spot, wake the machine up again and it acquires a
new network, sometimes a wired network instead of wifi. Offlineimap
doesn't like that so much. I don't yet have any reason to think the
problems here come from my patches. But I'm just acknowledging them,
so
that if others are able to use offlineimap without any difficulty in
situations like I described, then maybe the fault is with my patches.
From Jim Pryor
The first, "cleanup.patch", cleans up a few spots that tend to throw
exceptions for me as offlineimap is exiting from a KeyboardInterrupt.
Frankly, the original code doesn't even make much sense, and
moreover it's not forward compatible with python 2.6, and
furthermore:
DeprecationWarning: socket.ssl() is deprecated.
So, this commit is just a temporary fix.
Signed-off-by: Michael Witten <mfwitten@mit.edu>
module threading should be taken out back and shot.
Condition.wait() is a busywait loop that has negative implications for
battery consumption on laptops.
Queue.get() in blocking mode blocks SIGINT from being delivered.
Argh, argh, argh.
Closes: #493801.
This should improve power-management abilities some more
The catch is that we can't wait any longer for the kathread to
terminate. We were waiting for this in some cases. This is probably
not a big deal.
fixes deb#434074
fixes#66
This patch adds sslclientcert and sslclientkey configuration
options which are passed through to imaplib in order to allow
the use of client certificates for authentication.
Tue Nov 13 14:44:17 CST 2007 Mark Hymers <mhy@debian.org>
Noted thie code:
statusfolder =
statusrepos.getfolder(remotefolder.getvisiblename().\
replace(remoterepos.getsep(),
statusrepos.getsep()))
in accounts.py. Should have been using the sep of the LocalStatus all
along.
refs deb#479798, #68
fixes#20
04/27/2008 09:34 AM - Vincent Beffara
* File darwin.patch added
OK, I finally installed an IMAPS server on my iMac, so I could test a
little further. For some reason, fixing the read() method in
imaplibutil.py did not seem to work (it hung on connecting to the
server) - however, modifying the file imapserver.py similarly to the
non-SSL case fixed the crash.
I also reduced the chunk size to 100k instead of 1M, as 1M seemed to
still trigger the memory error in some cases. Ah, and I added a
platform test, so that the patch does essentially nothing on
non-Darwin machines ...
So, still no guarantee, but the attached patch works for me. Any
comments ?
(If noone here screams in horror at my code, I will include the patch
into the Fink package and see what happens there.)
I created new IMAP folders on my server. In some moment I found offlineimap
crashed. The reason were empty LocalStatus files. From looking at the code I
have no explanation how those files came into existence.
* John Goerzen (jgoerzen@complete.org) wrote:
> tages 359213 help
> thanks
>
> Hi folks,
>
> Unfortunately, I lack both the expertise to add Kerberos
> authentication and a way to test this. A patch would be very welcome
> :-)
Well it was a bit unclear but there's a patch for offlineimap attached
to Guido's message:
http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=10;filename=support-GSSAPI-via-pykerberos.diff;att=1;bug=359213
I've tried it out and it works great! Except that if the Kerberos
layer throws an exception it wouldn't fail over to plain
authentication properly. Attached is a slightly reworked patch that I
tested it in the following situations:
1. python-kerberos uninstalled, with kerberos credentials.
Prompted for password authentication as expected.
2. python-kerberos installed, with kerberos credentials.
Keberos authentication was successful and my mail was fetched
without password prompting. Yeehaw!
3. python-kerberos installed, with no kerberos credentials.
Prompted for password authentication as expected.
I think those are all the interesting situations. So I think this is
ready to go in, with a Suggest on python-kerberos (>=
1.0+mk080218-1).
--
Eric Dorland <eric@kuroneko.ca>
ICQ: #61138586, Jabber: hooty@jabber.com
refs deb#469598
Bug report received in Debian. User is having occasional trouble with Gmail
driver crashing, indicates problem may have started after switching to Gmail.
Gmail driver merged to tree in 81b86fb74 on 2008-01-03 and recently released
in beta tarballs.
BT shows:
File "/var/lib/python-support/python2.4/offlineimap/folder/Gmail.py", line 102, in processmessagesflags
attributehash = imaputil.flags2hash(imaputil.imapsplit(result)[1])
File "/var/lib/python-support/python2.4/offlineimap/imaputil.py", line 88, in imapsplit
for i in range(len(imapstring)):
TypeError: len() of unsized object
imap debug log shows:
imap: 29:58.16 > BEAL75 UID STORE 4887 +FLAGS (\Deleted)
imap: 29:58.47 < BEAL75 OK Success
imap: 29:58.47 matched r'(?P<tag>BEAL\d+) (?P<type>[A-Z]+) (?P<data>.*)' => ('BEAL75', 'OK', 'Success')
imap: imapsplit() called with input: None
imap: imapsplit() got a non-string input; working around.
looking at code for Gmail.py processmessagesflags, comments from Ricardo
indicate it was copied from the same function in folder/IMAP.py.
However, folder/IMAP.py has checks for this, added 2002-07-12 in
5342dacc & 817a10ce. Suspect that Gmail author believed those checks
superfluous for Gmail.
Copied checks from folder/IMAP.py to folder/Gmail.py.
This reverts commit 71a76d9a61.
Zak Smith reported a problem at:
self.processmessagesflags('+', uidlist, flags)
File "/usr/lib/python2.5/site-packages/offlineimap/folder/IMAP.py",
line 372, in processmessagesflags
myrights = imapobj.myrights(self.getfullname())[1][0].split()[1]
IndexError: list index out of range
Conflicts:
offlineimap/folder/IMAP.py
Added WrappedIMAP4_SSL class to help fix up performance of SSL
Standard imaplib.py is really bad with this, since it reads one
character at a time.
Reported by Aaron Kaplan at
http://lists.complete.org/offlineimap@complete.org/2008/01/msg00012.html.gz
He wrote:
I just noticed that the version of offlineimap I've been using
(3.99.17) is well over four years old. How time flies. I haven't
had any problems with it, but out of curiosity I decided to pull in
5.99.2 from the fedora repository. It turns out to take
consistently over twice as long as the old version to sync the same
account. Is this expected?
He tracked it down at
http://lists.complete.org/offlineimap@complete.org/2008/02/msg00012.html.gz
The following changeset is the one responsible for the difference in
speed I was noticing between the imaplib.py that was packaged with
older versions of offlineimap and the one that comes with python:
* /offlineimap/head: changeset 169
More optimizations -- this time fix readline() to not work
character-by-character!
closes#22
from pistore in OfflineIMAP #22:
When an IMAP flag update is performed for multiple messages, some IMAP
servers (e.g. Exchange) return the UID attribute only for some of the
FETCH untagged responses, as shown in the following log:
21:19.04 > DCKF8 UID STORE 66050,50613,52164,40043,40055,25874 +FLAGS
(\Deleted)
21:19.36 < * 35 FETCH (FLAGS (\Seen \Deleted) UID 25874)
21:19.36 < * 321 FETCH (FLAGS (\Seen \Deleted))
21:19.57 < * 322 FETCH (FLAGS (\Seen \Deleted))
21:19.57 < * 560 FETCH (FLAGS (\Seen \Deleted))
21:19.57 < * 581 FETCH (FLAGS (\Seen \Deleted) UID 52164)
21:19.62 < * 1022 FETCH (FLAGS (\Seen \Deleted))
21:19.62 < DCKF8 OK STORE completed.
Function IMAPFolder.processmessagesflags is able to manage the servers
which return the UID and the servers which do not return it, but is
not able to deal with the mixed behavior shown above.
The problem is that the fragment of function
IMAPFolder.processmessagesflags that handles the responses with UID
attribute uses variable flags to store the list of flags of the
message in the IMAP format ("flags = attributehashFLAGS?"), while the
fragment that handles the responses without UID expects variable
"flags" to contain the list of modified flags passed to the function
in Maildir format ("self.messagelist[uid]flags?.append(flag)").
As a consequence, the wrong list of flags is used for the messages
without UID, leading to the addition of "strange" flags to the Maildir
messages:
Syncing messages IMAP[INBOX] -> Maildir[.]
Adding flags to 4 messages on Maildir[.]
Adding flags e to 4 messages on Maildir[.]
Adding flags d to 4 messages on Maildir[.]
Adding flags ) to 4 messages on Maildir[.]
Adding flags ( to 4 messages on Maildir[.]
Adding flags l to 4 messages on Maildir[.]
Adding flags n to 4 messages on Maildir[.]
Adding flags t to 4 messages on Maildir[.]
Adding flags \ to 4 messages on Maildir[.]
Adding flags D to 4 messages on Maildir[.]
Deleting flags T to 4 messages on Maildir[.]
Adding flags to 4 messages on LocalStatus[.]
Adding flags e to 4 messages on LocalStatus[.]
Adding flags d to 4 messages on LocalStatus[.]
Adding flags ) to 4 messages on LocalStatus[.]
Adding flags ( to 4 messages on LocalStatus[.]
Adding flags l to 4 messages on LocalStatus[.]
Adding flags n to 4 messages on LocalStatus[.]
Adding flags t to 4 messages on LocalStatus[.]
Adding flags \ to 4 messages on LocalStatus[.]
Adding flags D to 4 messages on LocalStatus[.]
Deleting flags T to 4 messages on LocalStatus[.]
Fix: use a different variable to store IMAP flags when managing
messages corresponding to responses with UID attribute, e.g.:
*** IMAP.py.orig Wed Aug 22 18:23:17 2007
--- IMAP.py Wed Aug 22 18:22:38 2007
*************** class IMAPFolder(BaseFolder):
*** 340,348 ****
if not ('UID' in attributehash and 'FLAGS' in
attributehash):
# Compensate for servers that don't return a UID
attribute.
continue
! flags = attributehash['FLAGS']
uid = long(attributehash['UID'])
! self.messagelist[uid]['flags'] =
imaputil.flagsimap2maildir(flags)
try:
needupdate.remove(uid)
except ValueError: # Let it slide if it's not
in the list
--- 340,348 ----
if not ('UID' in attributehash and 'FLAGS' in
attributehash):
# Compensate for servers that don't return a UID
attribute.
continue
! lflags = attributehash['FLAGS']
uid = long(attributehash['UID'])
! self.messagelist[uid]['flags'] =
imaputil.flagsimap2maildir(lflags)
try:
needupdate.remove(uid)
except ValueError: # Let it slide if it's not
in the list
02/03/08 14:04:35 changed by js
* attachment flags-fix.patch added.
Delete 02/03/08 14:05:24 changed by js
Unfortunately I have to fetch some of my mail from an Exchange server
(Microsoft Exchange Server 2003 IMAP4rev1 server version 6.5.7638.1)
and I can confirm that the analysis of the problem is correct, and the
patch given here fixes the problem.
Looking at the code of the processmessagesflags() method I think it
generally is a bug that the "flags" parameter is reused as a local
variable, since the final "for uid in needupdate:" loop needs the
original value of "flags". This only worked by accident.
I'm attaching a unidiff version of the patch which applies cleanly
against Debian unstable's offlineimap 5.99.4.
New repository/folder classes to support "real deletion" of messages
thorugh Gmail's IMAP interface: to really delete a message in Gmail,
one has to move it to the Trash folder, rather than EXPUNGE it.
This patch maneuvers around python imaplib's mysterious read-only detection
algorithm and correctly calls the UI's deletetoreadonly(), when trying to
delete/expunge in a mailbox without having the necessary rights.
This involves several changes at different places:
- syncfoldersto() takes statusfolder as an argument, and returns the
list of new folders and the list of folders that should be ingnored,
typically those that were deleted. Warns the user about folders that
are present only on one side and are not synced.
- syncfoldersto() is called both ways, and on folder creation
forgetfolders() is used to rebuild the list and take the new creation
into account. Probably not the most efficient, since it involves
talking to the IMAP server again, but it will rarely be used anyway.
- Locally created folders are treated separately in the synchronization,
namely the local messages are uploaded and then the normal sync still
occurs. If the same folder is created on both sides and contains
messages on both sides, a two-way sync occurs.
fixes deb#439384
From: martin f krafft
Subject: race condition in Maildir writing
The offlineimap Maildir code checks for file existence and then
opens a file. That's open to a race condition. It's better to open
the file and fail if it already exists. The following patch does
this. It catches OSError 17 (file exists) and re-raises all others.
I'll leave it up to you to decide whether this is appropriate.
This involves several changes at different places:
- syncfoldersto() takes statusfolder as an argument, and returns the
list of new folders and the list of folders that should be ingnored,
typically those that were deleted. Warns the user about folders that
are present only on one side and are not synced.
- syncfoldersto() is called both ways, and on folder creation
forgetfolders() is used to rebuild the list and take the new creation
into account. Probably not the most efficient, since it involves
talking to the IMAP server again, but it will rarely be used anyway.
- Locally created folders are treated separately in the synchronization,
namely the local messages are uploaded and then the normal sync still
occurs. If the same folder is created on both sides and contains
messages on both sides, a two-way sync occurs.
From: "Mark A. Hershberger"
https://bugs.launchpad.net/ubuntu/+source/offlineimap/+bug/96710
the locked() method isn't implemented for non-interactive UIs, so
exceptions are thrown on cron jobs. Ubuntu's new apport catches these
and ? well, you get the idea.
patch provided.
fixes deb#433732
Date: Sun, 30 Sep 2007 13:54:56 -0400
From: Daniel Jacobowitz <drow@false.org>
To: offlineimap@complete.org
Subject: Assorted patches
Here's the result of a lazy Sunday hacking on offlineimap. Sorry for
not breaking this into multiple patches. They're mostly logically
independent so just ask if that would make a difference.
First, a new -q (quick) option. The quick option means to only update
folders that seem to have had significant changes. For Maildir, any
change to any message UID or flags is significant, because checking
the flags doesn't add a significant cost. For IMAP, only a change to
the total number of messages or a change in the UID of the most recent
message is significant. This should catch everything except for
flags changes.
The difference in bandwidth is astonishing: a quick sync takes 80K
instead of 5.3MB, and 28 seconds instead of 90.
There's a configuration variable that lets you say every tenth sync
should update flags, but let all the intervening ones be lighter.
Second, a fix to the UID validity problems many people have been
reporting with Courier. As discussed in Debian bug #433732, I changed
the UID validity check to use SELECT unless the server complains that
the folder is read-only. This avoids the Courier bug (see the Debian
log for more details). This won't fix existing validity errors, you
need to remove the local status and validity files by hand and resync.
Third, some speedups in Maildir checking. It's still pretty slow
due to a combination of poor performance in os.listdir (never reads
more than 4K of directory entries at a time) and some semaphore that
leads to lots of futex wake operations, but at least this saves
20% or so of the CPU time running offlineimap on a single folder:
Time with quick refresh and md5 in loop: 4.75s user 0.46s system 12%
cpu 41.751 total
Time with quick refresh and md5 out of loop: 4.38s user 0.50s system
14% cpu 34.799 total
Time using string compare to check folder: 4.11s user 0.47s system 13%
cpu 34.788 total
And fourth, some display fixes for Curses.Blinkenlights. I made
warnings more visible, made the new quick sync message cyan, and
made all not explicitly colored messages grey. That last one was
really bugging me. Any time OfflineIMAP printed a warning in
this UI, it had even odds of coming out black on black!
Anyway, I hope these are useful. I'm happy to revise them if you see
a problem.
--
Daniel Jacobowitz
CodeSourcery
patch from Mike Gerber
Two times today I have found my offlineimap to have died with this same
situation. It appears as if certain types of messages (both spam in my
situation), cause offlineimap to choke. When it does it cannot proceed.
This means that when I run offlineimap, it pulls in messages from some
folders, then it hits the folder with the bad message and dies, leaving
undownloaded mail on the server. The only fix to this problem is to find
the problem message on the server and remove it by hand. This isn't such
a huge deal for me, since I run the server, but other people have to
come to me to ask me to delete these messages, and until I do they
cannot download their email.
I have captured the output by running script during one of these
incidents, this has been attached. Additionally, I have also attach the
problematic message.
The patch seems to work for me, might need some Python wizard and better
testing, though.
fixes deb#396443
r[1] is a list. In case it contains more than one 'str' they are concatenated
using '. '. In processmessagesflags the join is tested, for savemessagesflags I
assume that it is also needed.
