Email goes to la la land!
Change small sset api causes my mailbox messages listing to be 0 long.
Rolling back to finish initial sorted-set implementation is working. Tested
and had the problem with Tbird 3.1.2 and my Win Mobile 6.1 Phone - IMAP.
-Jon
On 8/10/2010 7:42 AM, Paul J Stevens
Hold it - I was mistaken. It was not the sset change, it was the
forward port fix for #851 change that caused the problems. I'm out of
brain power tonight to debug, but just so you know.
-Jon
On 8/11/2010 12:34 AM, Jonathan Feally wrote:
Email goes to la la land!
Change small sset api
New fixes seem to have fixed the issue. Chewing on new dog food now, yum!
Thanks,
-Jon
On 8/11/2010 1:12 AM, Paul J Stevens wrote:
Thanks for the heads-up. I'll fix this now.
On 08/11/2010 09:43 AM, Jonathan Feally wrote:
Hold it - I was mistaken. It was not the sset change
Nov 20 10:50:10 mail.consult-scs dbmail-imapd[99549]: [0x80373a080]
Debug:[server] server_daemonize(+302): sid: [99548]
Nov 20 10:50:10 mail.consult-scs dbmail-imapd[99549]: [0x80373a080]
Notice:[server] server_run(+594): starting main service loop for [IMAP]
Nov 20 10:50:10 mail.consult-scs
Did you also backport the adjustments I made for the IPv6 stuff or was
that just your initial commit? It appears not.
see
http://git.dbmail.eu/cgit/cgit.cgi/paul/dbmail/commit/?id=2b3944f4e9617388910a851b9d964b3e859c4099
and tweak
Did you autoreconf -i ??? That will usually create your configure script
and some other stuff.
-Jon
Michael Monnerie wrote:
On Mittwoch 07 Oktober 2009 Paul J Stevens wrote:
Download:
http://git.dbmail.eu/cgit/cgit.cgi/paul/dbmail/snapshot/dbmail-2.2.12
.tar.bz2
Thanks. Just
The following is a view for MySQL that can be used with postfix to allow
a user to authenticate as one user, but be able to send mail from on of
the attached aliases. Aka, I authenticate as vult...@netvulture.com, but
can send email as jfea...@netvulture.com. With out this sender login
map,
/dbmail'
make: *** [all] Error 2
lira:/usr/local/src/postfix/git/jon/dbmail#
-Original Message-
From: dbmail-dev-boun...@dbmail.org [mailto:dbmail-dev-
boun...@dbmail.org] On Behalf Of Jonathan Feally
Sent: quarta-feira, 16 de Setembro de 2009 2:38
To: DBMAIL Developers Mailinglist
Jorge Bastos wrote:
On Dienstag 15 September 2009 Jorge Bastos wrote:
IMAPD crashed, and when I go to start him this happen.
Then the diff seems that the TCP connections are not closed on crash in
the one version, while the other does. Or the crash happens in a
different path,
I'm working on the more correct way to handle getaddrinfo as it can
return more than 1 address that should have a socket on it. This should
have the sockets with the REUSE flag on it and work correctly. Since the
IPv6 code, my box is completely unhappy with those and SEGFAULTS upon
start up. I
Your multiple IP's could be handled better with my testing branch. It
all depends on what you are putting in bindip.
My patch make no attempt to detect that we are already listening on an
IP:port pair, so some adjustment may be needed on bindip in some cases.
I was able to test a simple
bindip
I have not added anything additional to track this issue. Did you try
your current binaries against a freshly created database? You can start
a 2nd instance on a different port using a separate dbmail.conf to
quickly test this.
-Jon
Jorge Bastos wrote:
Hi Jon,
To keep tracking the source
Part of fixing this bug was to make sure that the session is properly
closed and deleted. I can't find what has been missed on session cleanup
when comparing a session that is in an idle loop vs. a session that did
a logout. If I left the session cleanup broken (doesn't do it) then I
never got
Jorge Bastos wrote:
Hi Jon,
Jorge Bastos wrote:
John Paul,
To recreate the complete cache, can i:
---
Delete from dbmail_header;
Delete from dbmail_headername;
Delete from dbmail_headervalue;
Delete from dbmail_envelope;
Dbmail-util -by
Yes - delete from all 4 tables. Do
Michael Monnerie wrote:
--sysconfdir=/etc/dbmail
I do not know if the man page compilation can use that setting to insert
the correct value. The man page will read:
-f configfile::
Specify an alternate config file. The utilities are currently
hardcoded to use /etc/dbmail.conf for their
Jorge Bastos wrote:
Hi Jon,
Did you missed my email? :P
I didn't miss it. Just haven't had time to reproduce yet. The bt will
show self=0x??? on imap_idle_loop()
The errorlog should also show that session.
-Jon
--
Scanned for viruses and dangerous content by MailScanner
Jorge Bastos wrote:
Howdy,
I’d like to confirm that i don’t have any lost mailbox’s on the DB,
can you give me an sql query example to check this?
Jorge,
I'm pretty sure that the dbmail-util -t already does that check. I have
a patch I'm testing to help with db locking during that type
Jorge Bastos wrote:
Jon,
No good,
Still happens, on g_free(D-data)
Well, I'm at a loss on this. For now just comment out the changes and be
aware of the memory leak so you can use it.
To reproduce the problem, you are just having a client login, select
mailbox, send idle command, then
Jorge Bastos wrote:
John Paul,
To recreate the complete cache, can i:
---
Delete from dbmail_header;
Delete from dbmail_headername;
Delete from dbmail_headervalue;
Delete from dbmail_envelope;
Dbmail-util -by
Yes - delete from all 4 tables. Do take your daemons down while doing
Jorge Bastos wrote:
PS: there are some other issues with current GIT but I'll create a bug form
them.
The most crazy one if this fix:
http://git.dbmail.eu/cgit/cgit.cgi/paul/dbmail/commit/?id=02b0fe06c4947fb50b
dcbe3ac39661006f78aa07
this causes an invalid pointer on dbmail-imapd (o posted
Jorge Bastos wrote:
Ops, still has the problem, check:
[New Thread 0xaecffb90 (LWP 19496)]
[New Thread 0xae4ffb90 (LWP 19497)]
*** glibc detected *** /usr/local/sbin/dbmail-imapd: free(): invalid
pointer: 0x0805cca0 ***
=== Backtrace: =
/lib/libc.so.6[0xb7c0cac5]
All,
I finally got around to playing with timsieved and found that the
avelsieve was slightly busted.
Patch at:
http://www.netvulture.com/avelsieve_fix_prototype.diff
svn co
https://email.uoa.gr/repos/squirrelmail/avelsieve/main_plugin/trunk
avelsieve
You will also need the patch just
Jorge,
Please fill in answers inline:
Prior to moving to post 2.3.6 git head, what version were you running?
What upgrade sql scripts did you run?
Did you run the latest 2.3.5-2.3.6 upgrade script? It is different than
previous 2.3.5-2.3.6 in pre-2.3.6 git head
Version of gmime?
Way to
Jorge Bastos wrote:
@Jonathan: Do you have the script/the mail kept in you sent folder?
I'm not sure what is needed. The script for Uwe was based upon his
current database schema and i'm not sure what version level that script
took him to.
Send me offlist a mysqldump -d of your
With INNODB you need to run optimize table dbmail_mimeparts to reclaim
space. Deleted rows do not get overwritten, just lost from the index.
Optimizing will copy all active rows to a new file and then swap them,
and delete the original table. So you need to have enough free space for
the data
Paul J Stevens wrote:
New features in this release:
There is also a new option -M on dbmail-util to migrate 2.2 messageblks
stored messages to the 2.3 single-instance storage. Messages are moved
1 per run of dbmail-util. You can change the 1 limit with a -m #
argument.
-Jon
--
Jake Anderson wrote:
Any chance of debs of this?
It makes upgrading much easier ;-
Oh and is there anything to migrate headers over to the new storage as
well?
Debs would be Paul's dept.
Part of upgrading to 2.3.6 removes the old header caching tables and
creates new ones. dbmail-util then
: [Dbmail-dev] New Idea
Jonathan Feally wrote:
Jorge Bastos wrote:
166319344 bytes RX, well not possible!!
I assumed that the values would start at 0. I have set them to 0 at
the
clientbase_t creation so give that a shot.
Jon,
Not such a strange
Jesse Norell wrote:
That's exactly what dbmail_pbsp does already, just in a separate table.
Maybe you just need to enable pbsp in your config file? (I don't
remember that support being removed, but it's possible. It's still in
the sql schema though.)
The pbsp is only used with the pop3
Since I made the patch, I'll comment:
Messages at this point do not have to be moved to the new single storage
structure. Retrieving messages that are not found in the partlists table
will be attempted from the messageblks. This option is there to
convert/move the messages stored with the 2.2.x
content by MailScanner
From 00280656be2f31d3e620b30bf7960c0ad7463a22 Mon Sep 17 00:00:00 2001
From: Jonathan Feally vult...@netvulture.com
Date: Fri, 12 Jun 2009 17:16:32 -0700
Subject: Add dbmail_authlog table and code to record logins, logouts, r/w byte
counts. Fixes pop3 bailout.
---
sql/mysql
Paul J Stevens wrote:
But first, let's get 2.4 ready, starting with 2.3.6! Almost there, we are.
Someone been watching Star Wars has he?
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Paul J Stevens wrote:
Some kind of replication of the dbmail_users table across database
servers would still be required, like we do now between ldap and sql.
That would allow *any* of the database backends to function as the
primary backend - in fact, any of the active daemons/tools would be
Michael Monnerie wrote:
But how do you delete messageblks without their entire message? There
are lots of constraints all over the place. You'd need to redefine
those. OK, as there will be no more old inserts, it shouldn't really
hurt to loose those.
If per-user is not good - what would
Paul J Stevens wrote:
Jon,
there's no need to update the header cache.
Ok, I wasn't sure. I didn't think it needed to be done. Throttling the
movement would be difficult. While the processing could be done on
another box separate of the production daemons, the database load and
disk
Paul J Stevens wrote:
Other than the lack of any perceivable added benefit (for me at least),
my main objections are with regard to change management and security.
For example, I like being able to stick /etc/dbmail/ into a git
repository. And I *really* don't want my ldap bind_dn parameters
Michael Monnerie wrote:
Also, have a single upgrade script to upgrade any schema version to the
newest would be very appreciated. So people who have an early 2.2 would
also get the newset indices etc. A table version would be nice to have
where a single dbmail-schema value is included.
Jake Anderson wrote:
Aaron Stone wrote:
We dropped the dbmail_config table very early on, between DBMail 1.1 and
1.2. There were good reasons to have a configuration file, and once things
started going in there, having two places to configure the daemons seemed
like a bad idea. One
cff34685a26a080ef3f43e7c32c33dcc2839a595 Mon Sep 17 00:00:00 2001
From: Jonathan Feally vult...@netvulture.com
Date: Mon, 1 Jun 2009 09:42:57 -0700
Subject: [PATCH] remove auto_reply and auto_notification completely
---
dbmail.conf | 20 --
sql/mysql/2_3_5-2_3_6.mysql |4
Paul,
You missed the DATE sort option.
SORT has options for:
ARRIVAL (internal_date)
CC
DATE
FROM
SIZE
SUBJECT
TO
with the optional REVERSE tag in front to sort in opposite order.
I'm not sure it runs any faster with views than with the larger query.
Since you are left joining the view, it does
What is the command line you/your client is issuing? I do not think that
it is a sort, as message-id is not one of the fields for sort that can
be used.
-Jon
Maxim Podorov wrote:
I installed the latest code, fixed DB structure, rebuilt caches, and now
the IMAP search on text headers
Paul J Stevens wrote:
Jonathan Feally wrote:
What is the command line you/your client is issuing? I do not think that
it is a sort, as message-id is not one of the fields for sort that can
be used.
Still, the setup in dbmail-mailbox.c,_append_join_headervalue is broken
imo
That's the yucky part, that is not needed.
I just pushed a cleanup that should at least put us back to where we
were before the headertable change in terms of correctness. There are
very likely still some issues with search as Michael reported, but
hopefully no new issues.
I had the sql
Paul J Stevens wrote:
Jonathan Feally wrote:
The emailaddr column is used for sorting only. The headervalue column is
what is used for the where clauses.
Jon, in your latest code, the emailaddr/emailname columns are never used
again after insertion. We really shouldn't be storing
Thats ugly. I am having no issues on my FreeBSD box. Just run a fresh
clone a few minutes ago. I'm not sure where that confstatBD6747
directory came from. My guess is something is broken on your box. Did
you try it on another machine?
-Jon
On 5/30/2009 11:17 AM, Uwe Kiewel wrote:
Hi,
Sounds like we need to setup on the first x chars for indexing of those
columns. How many addresses were in the to: field? I'm not a postgres
guy, so I don't know how to do an index on the first x chars of the
field. From the error message below, it looks like we could use a size
of 2048.
a lot more brain power to program this once, but I think it's a
huge benefit over the currect (2.2.x) situation.
On Donnerstag 28 Mai 2009 Jonathan Feally wrote:
From the error message below, it looks like we could use a size
of 2048.
MySQL can only index 255 chars :-(
mfg zmi
Looks like I missed that use of the datefield. I'll get a patch over to
paul for commit later on today. Your attempt will work for the time
being, but the joins need to be done a bit different for performance
reasons.
-Jon
Максим Подоров wrote:
Paul, during IMAP search, SQLExceptions are
It is not a hugh deal as a bigint(21) column can hold something like
|9223372036854775807 messages. The cause of this is the way that
messages are delivered. They are first delivered to
_...@!internal_delivery_user!@__ to create all of the physical message and
cache entries. Then, for each
When you import a message with the script, is there a header being
inserted that matches your existing UIDL's? If you UIDL is being put in
the message headers, then we should be able to grab that out of the
header and update uidl all in sql. If that is the case, send me some
sample header
Check your mysqldump -t dbmail against the create_tables.sql to find any
differences that need correction. You could also sent that output to me
offlist so I can try to figure out where it went wrong.
Also, you are moving from which 2.2.x to 2.3-git (post 2.3.5)??
-Jon
Uwe Kiewel wrote:
idle_status = (yes/no) will change the behavior of what is sent to the
client when the idle timeout hits and the client has said idle. Normal
output is to only send information about the SELECTed mailbox. This
option will check all subscribed mailboxes, and send information about
any mailbox
Yeah - thats not the right flag. use -d for no data.
Thats what I get for using my memory instead of my man!
-Jon
Uwe Kiewel wrote:
Jonathan Feally schrieb:
Check your mysqldump -t dbmail against the create_tables.sql to find any
differences that need correction. You could also sent
I have been using trunk as is since you applied my last patch. I haven't
been able to catch the 100% load issue since then, but I have also been
restarting every 2 hours via cron just in case it happens when I'm out
and about. So far its working good with no issues on T-Bird.
-Jon
Paul J
Testing the lists. I haven't gotten anything since 3/19.
-Jon
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
___
Dbmail-dev mailing list
Dbmail-dev@dbmail.org
I have some performance issues on 5.1 when it comes to sorting and using
the indexes. Also had some queries just getting stuck processing. Stick
with 5.0.x for production use for a while.
I think 5.1.x and 6.x will start working better in the next couple
releases, but won't try it again until
I've found that with thunderbird, the select on the mailbox is not as
much of an issue as the list/lsub of all the mailboxes that Thunderbird
does on startup. My inbox usually hovers around 500 messages, but with
150-200 mailboxes, the list/lsub build a part of the cache out for those
boxes
I've found that with thunderbird, the select on the mailbox is not as
much of an issue as the list/lsub of all the mailboxes that Thunderbird
does on startup. My inbox usually hovers around 500 messages, but with
150-200 mailboxes, the list/lsub build a part of the cache out for those
boxes
Paul/Others,
Starting with the first wip: buffered reads commit, FreeBSD IMAP TLS is
broken. Thunderbird throws an error:
Error establishing an encrypted connection to server. Error Code: -12217
The certificate is being transmitted just fine, but after accepting the
cert, it blows up with that
I'm guessing you are speaking about limiting the connection throughput
to the clients? As it sits now, the daemons write out huge chunks of
data all at once towards the client socket. I don't see how we can limit
the speed of data written to the client with out some major reworking to
keep it
How many concurrent connections do you have when this happens? Perhaps
you are having a DOS attack?
-Jon
Jorge Bastos wrote:
Paul, continuing this problem, it still happens.
--
Dec 13 15:22:59 lira dbmail/imap4d[21103]: [0x87aa260] Error:[message]
_set_content_from_stream(+783): opening
Try:
autoreconf -i
./configure
[g]make
[g]make install
You may need to add CFLAGS, LDFLAGS, and/or --with-sieve to your
configure line depending on your system
My FreeBSD system uses
./configure CFLAGS=-I/usr/local/include LDFLAGS=-L/usr/local/lib
--with-sieve
-Jon
Uwe Kiewel wrote:
I'm
Is anyone experiencing an issue with the internal date being returned
wrong for messages on IMAP?
I am running FreeBSD, MySQL 5.1 delivered via postfix LMTP. Date on db
row is corrected for TZ from -0800 (local machine time and zone) to
+, but when recalled it is taken as is and the -0800
this issue with the 2.2.x series with the same pda
phone, same database settings, same host settings. Only thing changed
was the dbmail.
-Jon
Paul J Stevens wrote:
Jonathan Feally wrote:
Is anyone experiencing an issue with the internal date being returned
wrong for messages on IMAP?
I am
v1.0.0
Sender: [EMAIL PROTECTED]
Errors-to: [EMAIL PROTECTED]
X-original-sender: [EMAIL PROTECTED]
Precedence: bulk
X-list: dovecot
X-UID: 3
Status: O
Jonathan Feally wrote:
Ok - so what I am seeing is with message insertion via lmtpd care of
postfx.
Sample e-mail is delivered at 2008-12-01 00:00
I've tested this patch on FreeBSD and adjusted the length of the pw
variable 200-130. This also requires an adjustment to the
dbmail_users.passwd column to accept the longer hash result. All seems
well and working. Tested both pop3 and imapd
Paul,
I've commited this to my ready branch and
Jorge, please do let us know how it is working, along with the selective
logging options now available. See the new dist/dbmail.conf on those new
options and my last email to -dev.
-Jon
-Original Message-
From: Jorge Bastos [EMAIL PROTECTED]
Sent: Sunday, October 05, 2008 8:55 AM
To:
production server, if it's working OK
-Original Message-
From: [EMAIL PROTECTED] [mailto:dbmail-dev-
[EMAIL PROTECTED] On Behalf Of Jonathan Feally
Sent: domingo, 5 de Outubro de 2008 21:14
To: DBMAIL Developers Mailinglist
Subject: RE: [Dbmail-dev] last SVN
Jorge, please do let us
Paul and myself have been working on the logging of dbmail.
Selective Logging:
We have moved in favor of a selective logging level system that better
mirrors syslog.
Differences:
Old:
6 Levels 0-5, each level including
Syslog logging only logged to levels alert or notice depending on
I was having an issue coping messages from another account into 2.3.x
where large messages caused a seg fault 11. I was able to narrow it down
to the creation of the buff char variable in dbmail-imapsession.c being
initialized as the size of the message from the append {size}
information.
to test and let you know.
-Original Message-
From: [EMAIL PROTECTED] [mailto:dbmail-dev-
[EMAIL PROTECTED] On Behalf Of Jonathan Feally
Sent: terça-feira, 19 de Agosto de 2008 19:31
To: DBMAIL Developers Mailinglist
Subject: [Dbmail-dev] IMAP append seq fault 11 problem with 2.3.x
I don't think so. I compiled just last night with all 3 drivers, ldap
and sieve and it seems to be ok.
-Jon
Keith Hopkins wrote:
Hello,
Is there any issues with the resulting dbmail binary if it built with more
then one database type specified? (assuming all libraries are available to
The attached patch fixes an issue I was having with my system.
Apparently the g_strncasecmp call was broken with glib 2.12.9. The call
seemed to think that an empty buffer still matched DONE. That caused
all my connection to go into a bogus state.
Upgrading to glib 2.12.12 (current version
Just to be clear - are you taking the headers of the mime part and
keeping them seperate of the body of said part? My original post to keep
them seperate was to allow for the attached file be renamed but not
create duplicates, and also some mail clients may create the headers of
the mime part
74 matches
Mail list logo