Re: EMBARRASSING TO THE LIST: Re: *WARNING* Your Email Account Will Be Closed

[Derrick J Brashear]

On Mon, 30 May 2005, Mark Nernberg wrote:


This, along with other similar such things, should be a total embarrassment
to the list administrators and operators.


I hear cmu is hiring if you'd like to fix it. Your email address suggests 
you're local.



Considering that Project Cyrus aims to produce "secure" software, the total
lack of even basic anti-virus screening on the listserve (or, if there is
screening, the total failure of it), reeks of laziness on the part of the
system operators.


No effort will be spent on majordomo. CMU is working on deploying mailman, 
and is short staffed enough that it makes more sense to work on that than 
on this.



As such, I now question the validity of the claims that Cyrus IMAPD and SASL
are relatively secure.  If the programmers are as lazy as the system
operators, then IMAPD and SASL must be full of holes.


If wishes were trees, trees would be falling.

And?

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: [WORKAROUND SUCCESS] Re: et_list & com-err vs imapd v2.2.12 on OSX 10.4.1

[Derrick J Brashear]

On Mon, 23 May 2005, OpenMacNews wrote:


they come with the release distro:



Fine, so just remove them, let them be regenerated, and move on with life.


huh?


Huh what, you said you did just this.


and shared with you the results of making 'it' work on OSX 10.4.1.

i've been trying to be helpful ... why the sudden attitude change?


There is none. That was a summary of (what I expect) needs to be done as a 
workaround.


The ultimate fix upstream is for these files to not be included, but I 
can't go back in time to fix 2.2.12.



---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: new install, error when connecting, "BYE Fatal Error: can't write proc file"

[Derrick J Brashear]

On Mon, 23 May 2005, Mark Nernberg wrote:


Ok.  I got the issues with compile_et & com_err worked out (by
specifying --with-com-err=/usr/local/dist/cyrus-imapd-2.2.10 in
./configure).

Now that its compiled, I need to test it (duh!).

When connecting via telnet to localhost 143, I get the following:

# telnet localhost 143
Trying 127.0.0.1...
Connected to localhost.
Escape character is '^]'.
* BYE Fatal error: can't write proc file
Connection closed by foreign host.


Having Googled this, I noted that there is a prior posting to the list
(from Dec.2004), but it went unanswered.


The directory where proc files are written (which is a subdirectory of 
configdirectory as specified in imapd.conf, named proc) is presumably not 
writable to the cyrus user. It should be owned by the cyrus user, not 
root, and of at least mode 755, but perhaps better mode 700.


---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: [WORKAROUND SUCCESS] Re: et_list & com-err vs imapd v2.2.12 on OSX 10.4.1

[Derrick J Brashear]

On Mon, 23 May 2005, OpenMacNews wrote:


they come with the release distro:


Fine, so just remove them, let them be regenerated, and move on with life.
The problem is that *our* compile_et doesn't match *your* libcom_err, in 
fact.


So this step:

manually recompile the .et files:

% cd $WORK
% find . -name "*.et" -print
./et/test1.et
./et/test2.et
./imap/imap_err.et
./imap/mupdate_err.et
./imap/nntp_err.et
./SIEVE/sieve_err.et

% cd $WORK/et && compile_et test1.et

[]

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: et_list & com-err vs imapd v2.2.12 on OSX 10.4.1

[Derrick J Brashear]

On Mon, 23 May 2005, OpenMacNews wrote:


hi derrick,

a bit more info/data ...

well it sure seems that com_err/et_list originate from / revolve around 
Kerberos.


Kerberos uses them, and they are from MIT, but there are non-Kerberos 
things which do.


Apple provides it with/due to Kerberos.


1st, after a fresh DL, we still have the 'suspect':

% grep et_list imap_err.c
struct et_list {


How were these generated?


and checking,

% grep et_list imap_err.c
% (<-- still EMPTY)


Which is good. But, did you rebuild the other et files also, or do they 
still have the references? And did you remake only the objects or did make 
regenerate the foo_err.c and foo_err.h files out from under you with the 
__et_list references again.



so, i'm not conviced that this is a Apple-specific issue ...


I never suggested it was.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: et_list & com-err vs imapd v2.2.12 on OSX 10.4.1

[Derrick J Brashear]

On Mon, 23 May 2005, OpenMacNews wrote:


  % which compile_et
  /usr/bin/compile_et

  % compile_et imap/imap_err.et
  Error Table "imap" has base -1904809472


Well, it writes out imap_err.h, imap_err.c, and the latter would have the 
reference.




  % ls -al /usr/ports/cyrus/cyrus-imapd-2.2.12/et/compile_et*
  -rw-r--r--  1 17985 staff 4039 May 23  2000
  /usr/ports/cyrus/cyrus-imapd-2.2.12/et/compile_et.1
  -rwxr-xr-x  1 17985 staff 1953 May 23  2000
  /usr/ports/cyrus/cyrus-imapd-2.2.12/et/compile_et.sh

nothin' else there ...


Wait. darwinports? The build directory is in a subdirectory.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: et_list & com-err vs imapd v2.2.12 on OSX 10.4.1

[Derrick J Brashear]

On Mon, 23 May 2005, OpenMacNews wrote:


hi again,


Well, so, if you manually use compile_et from /usr/bin to compile the et
file, does it emit a __et_list reference?


sorry, confused.

"the et_file ..." -- which file in the imapd distro?


Not the "et_file"; The et file. There are several. Ultimately you want 
them all, but for test purposes one will do

imap/imap_err.et imap/nntp_err.et
imap/mupdate_err.et


Otherwise, you have some other compile_et being used. Get rid of it.


"There can be only one ..."

all i've got is ...

  % locate compile_et
  /usr/bin/compile_et
  /usr/share/man/man1/compile_et.1
  /usr/ports/cyrus/cyrus-imapd-2.2.12/et/compile_et.1
  /usr/ports/cyrus/cyrus-imapd-2.2.12/et/compile_et.sh


And not, say, /usr/ports/cyrus/cyrus-imapd-2.2.12/et/compile_et (or 
wherver it's building)?

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: et_list & com-err vs imapd v2.2.12 on OSX 10.4.1

[Derrick J Brashear]

On Mon, 23 May 2005, OpenMacNews wrote:


  % make depend
  % make all

fails @:

  ...
  sievec.c
  gcc -L/usr/local/ssl/lib -L/usr/local/berkeley-db/lib
  -bind_at_load -ldl -L/usr/local/berkeley-db/lib -ldb
  -F/Library/Frameworks -framework SASL2 -o sievec sievec.o libsieve.a
  ../lib/libcyrus.a ../lib/libcyrus_min.a  libsieve.a
  -L/usr/local/cyrus-sasl/lib  -lsasl2  -lfl
  -L/usr/local/berkeley-db/lib  -ldb-4.3 -lssl -lcrypto  -lcom_err
  /usr/bin/ld: Undefined symbols:
  __et_list
  collect2: ld returned 1 exit status
  make[1]: *** [sievec] Error 1
  make: *** [all] Error 1


Ok,


which, as b4, i'd gather stems from:

  >> ... the fact that et_list is not defined in libcom_err/com_err.h, i.e.:
  >>
  >> % nm /usr/lib/libcom_err.dylib | grep et_list
  >> % (<-- empty)
  >>
  >> % grep et_list /usr/include/com_err.h
  >> % (<-- empty)

fyi, as i'm googling to unnderstand this, and figure out what - if anything - 
changes re: kerberos under Tiger -- i'm finding an old thread at:


Well, so, if you manually use compile_et from /usr/bin to compile the et 
file, does it emit a __et_list reference? Otherwise, you have some other 
compile_et being used. Get rid of it.


---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Beta release of Cyrus SASL 2.1.22

[Derrick J Brashear]

I'd like to announce the beta release of Cyrus SASL 2.1.22 on
ftp.andrew.cmu.edu. This version includes bug fixes for the build system 
and a fix for gcc 4.0.


Please send any feedback either to [EMAIL PROTECTED]
(public list) or to [EMAIL PROTECTED]

Download at:
ftp://ftp.andrew.cmu.edu/pub/cyrus-mail/BETA/cyrus-sasl-2.1.22.tar.gz

Thanks,
Derrick Brashear
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: compile problems with imap-2.2.10

[Derrick J Brashear]

On Mon, 23 May 2005, Mark J. Nernberg wrote:


From: Derrick J Brashear <[EMAIL PROTECTED]>
Date: Mon, 23 May 2005 13:27:15 -0400 (EDT)
To: 
Subject: Re: compile problems with imap-2.2.10

On Mon, 23 May 2005, Mark Nernberg wrote:


`initialize_imap_error_table_r':
/usr/local/warez/cyrus-imapd-2.2.10/imap/imap_err.c:68: undefined reference
to `initialize_error_table_r'
libimap.a(imap_err.o)(.text+0x14):/usr/local/warez/cyrus-imapd-2.2.10/imap/i
map_err.c:68: undefined reference to `initialize_error_table_r'


Wild guess, the compile_et and libcom_err are not from the same source.



Not being a programmer, would you please be so kind as to offer a suggestion
to how I might get around it (i.e. get them in the same source).


If you're linking against /usr/lib/libcom_err.{a or so} I'd expect the 
matching compile_et to be in /usr/bin


If you're using the internal to cyrus com_err (in the "et" directory) I'd 
expect to not use a /usr/lib/libcom_err or /usr/bin/compile_et.


---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: et_list & com-err vs imapd v2.2.12 on OSX 10.4.1

[Derrick J Brashear]

On Mon, 23 May 2005, OpenMacNews wrote:


  -L/usr/local/DarkMatter/berkeley-db/lib  -ldb-4.3 -lssl -lcrypto
  /usr/lib/libcom_err.a
  powerpc-apple-darwin8-gcc-4.0.0: /usr/lib/libcom_err.a: No such file or
  directory


I assume they provide /usr/lib/libcom_err.dylib. Edit the makefile to 
refer only to -lcom_err and not /usr/lib/libcom_err.a, and keep 
--with-com_err=/usr


---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: compile problems with imap-2.2.10

[Derrick J Brashear]

On Mon, 23 May 2005, Mark Nernberg wrote:


`initialize_imap_error_table_r':
/usr/local/warez/cyrus-imapd-2.2.10/imap/imap_err.c:68: undefined reference
to `initialize_error_table_r'
libimap.a(imap_err.o)(.text+0x14):/usr/local/warez/cyrus-imapd-2.2.10/imap/i
map_err.c:68: undefined reference to `initialize_error_table_r'


Wild guess, the compile_et and libcom_err are not from the same source.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Shared Mailboxes and Postfix

[Derrick J Brashear]

On Mon, 23 May 2005, Josh Whitver wrote:


Derrick J Brashear <[EMAIL PROTECTED]> on Monday, May 23, 2005 at 7:51 AM
-0600 wrote:

Yes. But you still need a postuser to inject them into the mail system
hinting that they are for delivery to shared folders.

post or bb are common choices.


Okay; I've got one set up, and it's called "Conference".  Will uppercase
characters cause problems?  I can change it to "conference" in imapd.conf if
that's all it'll take.


Unclear, it depends if the mta is passing in uppercase or not for 
instance.




The line "postuser: " in imapd.conf is all it takes to set that up,
right?  I'm not missing some important step somewhere?


That's it.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Autocreatequota and mailbox

[Derrick J Brashear]

On Sat, 21 May 2005, Greg A. Woods wrote:


[ On Thursday, May 19, 2005 at 16:49:53 (-0700), Wil Cooley wrote: ]

Subject: Re: Autocreatequota and mailbox

On Thu, 2005-05-19 at 17:20 -0400, Greg A. Woods wrote:


Does somebody know why this patch isn't
included in Cyrus? http://email.uoa.gr/projects/cyrus/autocreate/


I've been asking myself the same question for several years now, ever
since I created a very similar patch too.  :-)


Christos answered this very question a few weeks ago:

http://article.gmane.org/gmane.mail.imap.cyrus/17735


Yes, but that was not really a very good nor sufficient answer given
that the concept of automatically creating a missing INBOX on POST (or
login) has been around _FAR_ longer than the MURDER stuff has been.  I
sent my patch in back in the 1.x days, so there was more than ample time
to include the concept of doing this in the inter-server protocols.


I can't speak to previous maintainers, but the statement that when Murder 
is supportable by autocreate patches, they will be included, is correct.


The past is gone. We are where we are.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Shared Mailboxes and Postfix

[Derrick J Brashear]

On Sat, 21 May 2005, Josh Whitver wrote:


I saw it, but wasn't sure it was applicable.   I was told earlier that the
postuser ("Conference" in my case) didn't need to be an actual user on the mail
system.  I'm trying to clear up that confusion now, but am not getting a clear
answer.

Does the postuser need to be a true, actual user on the system for shared
mailboxes and bulletin boards to function?


No.


Is it possible to have shared
mailboxes/bulletin boards exist outside of the user mailbox heirarchy,


Yes. But you still need a postuser to inject them into the mail system 
hinting that they are for delivery to shared folders.


post or bb are common choices.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Cyrus SASL 2.1.21 released: addendum

[Derrick J Brashear]

On Mon, 16 May 2005, Derrick J Brashear wrote:
I'd like to announce the beta release of Cyrus SASL 2.1.21 on
ftp.andrew.cmu.edu. This version includes bug fixes most notably to the build
Correction, this is a real release. Apologies for the confusion.
Download at:
ftp://ftp.andrew.cmu.edu/pub/cyrus-mail/cyrus-sasl-2.1.21.tar.gz>
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Cyrus SASL 2.1.21 released

[Derrick J Brashear]

I'd like to announce the beta release of Cyrus SASL 2.1.21 on
ftp.andrew.cmu.edu. This version includes bug fixes most notably to the 
build system and log level setting, a rewrite of the Kerberos 4 and 
GSSAPI plugins to share more code from CMU School of Computer Science, 
and the ldapdb auxprop plugin from Howard Chu (as contributed by Igor 
Brezac)

Please send any feedback either to [EMAIL PROTECTED]
(public list) or to [EMAIL PROTECTED]
Download at:
ftp://ftp.andrew.cmu.edu/pub/cyrus-mail/cyrus-sasl-2.1.21.tar.gz
Thanks,
Derrick J Brashear
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Spam coming from list server??

[Derrick J Brashear]

On Sun, 15 May 2005, Patrick Gibson wrote:
Is there any particular reason why info-cyrus is not a closed list? It is 
unnecessary in my opinion to be receiving spam via a list that can easily 
restrict posting to subscribed members.
The spam can be (and on other lists has been) forged as from members. Then 
what?
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Searching multipart emails and using squatter

[Derrick J Brashear]

On Fri, 13 May 2005, Zachariah Mully wrote:
Due to XML object mime-type, SEARCH BODY requests don't work, as the
IMAP server has no idea what the heck it is, and therefore, the
webclient must fetch and retrieve the entire calendar mail store and
parse it to find relevant entries. This is, of course, miserably
inefficient and makes the webclient completely useless once the calendar
mail store has several hundred objects in it.
Hm. Bodes for an extension to deal with this, if (read further)
My question is, is it possible to get Cyrus to recognize the Kolab XML
object and index it so that the webclient can use SEARCH BODY requests
to fetch relevant calendar entries?
You have the source, everything is possible. Having something we can (and 
will) distribute is another thing. If it's well designed, certainly it's 
not right out.

separate cache of the data, but the developers of Kolab don't see a need
to expose the XML to the Cyrus' native caching and searching by changing
the mime-type to something like "text/xml" which Cyrus does understand.
So, perhaps they have a good reason for this; have you asked for comment 
from them?

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Error message opening imap boxes

[Derrick J Brashear]

On Tue, 10 May 2005, Andreas Haase wrote:
Hello,
after doing some (probably silly) stuff, I get error messages when some of my 
users are opening their imap boxes. The log file entries are:

--- snip ---
May 10 17:15:55 mail imapd[4925]: login: host.name[xxx.xxx.xxx.xxx] username 
plaintext
May 10 17:15:55 mail imapd[4925]: skiplist: invalid magic header: 
/var/lib/imap/user/u/username.seen
May 10 17:15:55 mail imapd[4925]: DBERROR: opening 
/var/lib/imap/user/u/username.seen: cyrusdb error
May 10 17:15:55 mail imapd[4925]: Could not open seen state for username 
(System I/O error)
May 10 17:15:55 mail imapd[4925]: open: user username opened INBOX
--- snap ---

So the user seems to get his mails correctly. Is there any way to get rid of 
the messages without deleting /var/lib/imap/user/u/username.seen? A 
reconstruct seems not to be the solution because after that the messages are 
still there.
Is it really a skiplist? What does the file look like?
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: stupid problems with virtual domains

[Derrick J Brashear]

On Mon, 9 May 2005 [EMAIL PROTECTED] wrote:
Consider / as mailbox delimiter, and thus:
unixhierarchysep: true
?
# cat /etc/imapd.conf
configdirectory: /srv/var/imap.cf
partition-default: /srv/var/spool/imap
admins: mailadmin
sasl_pwcheck_method: auxprop
sasl_auxprop_plugin: sasldb
virtualdomains: on
servername: mux
loginrealms: localhost mux mux.tld mux-it.tld
# grep -v '#' /etc/cyrus.conf
START {
 recover   cmd="/usr/cyrus/bin/ctl_cyrusdb -r"
}
SERVICES {
 imaps cmd="imapd -s" listen="imaps" prefork=0
 pop3s cmd="pop3d -s" listen="pop3s" prefork=0
 sieve cmd="timsieved" listen="sieve" prefork=0
 lmtpunix  cmd="lmtpd" listen="/var/spool/postfix/public/lmtp"
prefork=1
}
EVENTS {
 checkpointcmd="ctl_cyrusdb -c" period=30
 delprune  cmd="cyr_expire -E 3" at=0400
 tlsprune  cmd="tls_prune" at=0400
}
# sasldblistusers2
[EMAIL PROTECTED]: userPassword
[EMAIL PROTECTED]: userPassword
[EMAIL PROTECTED]: cmusaslsecretOTP
[EMAIL PROTECTED]: cmusaslsecretOTP
[EMAIL PROTECTED]: cmusaslsecretOTP
[EMAIL PROTECTED]: cmusaslsecretOTP
[EMAIL PROTECTED]: userPassword
[EMAIL PROTECTED]: userPassword
# grep ctl_cyrusdb /var/log/messages | awk -F: '{ print $3 " " $4 }'
03 mux master[3720]  about to exec /usr/cyrus/bin/ctl_cyrusdb
03 mux ctl_cyrusdb[3720]  checkpointing cyrus databases
03 mux ctl_cyrusdb[3720]  archiving database file
03 mux ctl_cyrusdb[3720]  archiving log file
03 mux ctl_cyrusdb[3720]  archiving database file
03 mux ctl_cyrusdb[3720]  archiving log file
03 mux ctl_cyrusdb[3720]  done checkpointing cyrus databases
# grep imap /var/log/messages | ridoftime
mux imap[3294]: skiplist: recovered /srv/var/imap.cf/user/b/mux.seen (0
records, 144 bytes) in 0 seconds
mux imap[3294]: seen_db: user mux opened /srv/var/imap.cf/user/b/mux.seen
mux imap[3294]: open: user mux opened INBO
--
+++ Neu: Echte DSL-Flatrates von GMX - Surfen ohne Limits +++
Always online ab 4,99 Euro/Monat: http://www.gmx.net/de/go/dsl
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: is this a DB lock problem?

[Derrick J Brashear]

On Thu, 5 May 2005, Phil Chambers wrote:
I am running cyrus-imap-2.2.8 and it has been running continuously for over a 
year
without a shutdown.  Today I suddenly have a problem which shows up with 
Mulberry.
Mulberry hangs at startup while trying to obtain a list of all users.
A protocol log on cyrus show that the the IMAP process has obtained a list of 
the
user's folders and then issued:
 A6 LIST "" "user.%"
command. The IMAP process has not responded and appears to be hanging and does 
not
time-out (at least, I have waited several minutes).
What does a syscall tracer show the imapd doing when this is happening?
Is it likely I have a genuine lock problem or could there be some other issue here?
I don't think you've given us enough information to know that.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Question about cyrus quotas - how it behaves

[Derrick J Brashear]

On Wed, 4 May 2005, David R Bosso wrote:
--On May 4, 2005 4:19:51 PM -0600 Sun Advocate Webmaster 
<[EMAIL PROTECTED]> wrote:

Just a question for all of you about quotas on cyrus imap.
As far as I see, there is no "grace" quota or warning system. Users not
watching the size of their mailboxes won't know they're near full until
they're actually full. Is this the case?
Check out quotawarn: and quotawarnkb: in man imapd.conf.  They result in IMAP 
ALERT messages being generated.
Note that these alerts are sent on SELECT, and it has been claimed that 
Outlook SELECTs a lot, making it unusable. I have not verified this.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: How can I increase cyrus loglevel?

[Derrick J Brashear]

On Wed, 4 May 2005, Miguel Paolino wrote:
I've been searching the web and I can't find they way to increase the log
level of cyrus. How can I do this?
change syslog.conf to log local6.debug (or something higher than the 
default loglevel)

or do you mean something else?
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Cyrus Sasl Libary 2.1.20

[Derrick J Brashear]

On Tue, 3 May 2005, Stefan Berger wrote:
Hello!
is in the version 2.1.20 the crypt sql patch included?
no. and nothing has been submitted for 2.1.21 yet either, the patch 
provided for 2.1.20 had some logging changes which were basically not 
generally useful and so was not applicable as-is.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: [POLL] unexpunge behavior

[Derrick J Brashear]

On Tue, 3 May 2005, Ken Murchison wrote:
Igor Brezac wrote:
How does quota work?  I assume that a 'deleted message' (before it is 
cyr_expire'd) does not count against quota.
Correct.
And since the user can't get it back without intervention it's not 
secretly limitless quota, though it does make knowing how much disk space 
you've allocated less deterministic.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: intermittent mupdate failures.

[Derrick J Brashear]

On Mon, 2 May 2005, Darrell Fuhriman wrote:
But while that *works* I'm still failing to see how having it
fail when set to 3 can be considered anything other than a bug
and therefore worth addressing.
It's the difference between a bugfix and workaround.
The bug is that you can prefork mupdate 3 times at all. You should never 
have more than one running.


---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: cyrus-imapd/sasl on MacOS X 10.3.8: hang

[Derrick J Brashear]

On Mon, 2 May 2005, Charles Bailey wrote:
OK, to no-one's real suprise (including my own), if I rebuild Cyrus
SASL without GSSAPI, LDAP, or PAM support and run Cyrus IMAPD and
Sendmail against it, using sasldb for authentication, all is well.
(It's run clean for several days longer than any of the previous
attempts.)
I think we're on stable ground now, but I'm still uneasy walking away
from this, as it seems there's a simple-to-exploit DoS bug living in
OS X somewhere, which it seems unfriendly to ignore.I haven't
really had a chance to delve into the source and figure out why
saslauthd-pam/imapd (and possibly saslauthd-pam/sendmail) locks up the
system while things like sshd-pam do not, if in fact PAM's the
culprit.  Do any other obvious lines of investigation occur to folks
out there?  Is there a mechanism by which an outside user can file a
bug report with Apple?  (Perhaps the darwin-dev list works?)  Is there
a better forum in which to ask?
https://bugreport.apple.com/cgi-bin/WebObjects/RadarWeb.woa
You should be able to just join ADC for free to use this.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: intermittent mupdate failures.

[Derrick J Brashear]

On Thu, 28 Apr 2005, Darrell Fuhriman wrote:
Henrique de Moraes Holschuh <[EMAIL PROTECTED]> writes:
mupdate is self-forking. Start just one.
A reasonable suggestion, which does absolutely nothing to address
the issue of *why* such a behavior would happen.
It's not even worth addressing the issue unless it happens when you start 
a single mupdate. Does it?

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: mupdate exits with signal 11

[Derrick J Brashear]

On Thu, 28 Apr 2005, João Assad wrote:
Derrick J Brashear wrote:
If anyone else wants to try this (I'd guess auth_unix use in mupdate isn't 
common?)
I use saslauthd for authentication which then uses pam. Is there a way to get 
rid of auth_unix ?
sasl_pwcheck_method: saslauthd
sasl_mech_list: PLAIN
auth_unix's purpose is for groups, not for something pam will do for you. 
technically if you're not using unix groups for access control you don't 
need any of the group functionality. Really, though, I should figure out 
what the getgr*_r interface difference which is hitting you is and fix it.

Re: Message contains NUL characters ...

[Derrick J Brashear]

On Thu, 28 Apr 2005, Henrique de Moraes Holschuh wrote:
But we really need to do something about the also non-RFC compliant crap
that sends too-big lines as well, cyrus screws up when breaking those lines.
Is that bug completely fixed yet?
No. If you have a suggestion for how to do it without touching the 
prot_stdio layer that would make this much easier to know and limit (and 
test) impact of. Actually I think Ken had an idea, which if he did and he 
reminds me of I'll implement.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: mupdate exits with signal 11

[Derrick J Brashear]

If anyone else wants to try this (I'd guess auth_unix use in mupdate isn't 
common?)
http://www.contrib.andrew.cmu.edu/~shadow/auth_unix.c.diff
is a diff which probably works. I'm still unsure if there are platforms 
using the 3 argument version of getgrent_r() and it appears to be a 
non-standard function anyway, so this probably needs configure glue.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: mupdate exits with signal 11

[Derrick J Brashear]

And I guess this also:
--- auth_unix.c 16 Feb 2005 21:06:50 -  1.37.2.2
+++ auth_unix.c 26 Apr 2005 17:53:40 -
@@ -157,10 +157,11 @@
 size_t len;
 {
 static char retbuf[81];
-struct group *grp;
+struct group grp;
 char sawalpha;
 char *p;
 int username_tolower = 0;
+char buf[BUFLEN];
 if(!len) len = strlen(identifier);
 if(len >= sizeof(retbuf)) return NULL;
@@ -177,8 +178,8 @@
  */
 if (!strncmp(retbuf, "group:", 6)) {
-   grp = getgrnam(retbuf+6);
-   if (!grp) return 0;
+   if (getgrnam_r(retbuf + 6, &grp, buf, BUFLEN))
+   return 0;
strcpy(retbuf+6, grp->gr_name);
return retbuf;
 }

Index: auth_unix.c
===
RCS file: /afs/andrew.cmu.edu/system/cvs/src/cyrus/lib/auth_unix.c,v
retrieving revision 1.37.2.2
diff -u -r1.37.2.2 auth_unix.c
--- auth_unix.c 16 Feb 2005 21:06:50 -  1.37.2.2
+++ auth_unix.c 26 Apr 2005 17:43:56 -
@@ -221,9 +221,11 @@
static struct auth_state *mynewstate(const char *identifier)
{
struct auth_state *newstate;
-struct passwd *pwd;
-struct group *grp;
+struct passwd pwd, *pwdp;
char **mem;
+struct group grp, *grpp;
+char buf[BUFLEN], buf2[BUFLEN];
+int i;
identifier = mycanonifyid(identifier, 0);
if (!identifier) return 0;
@@ -238,22 +240,25 @@
if(!libcyrus_config_getswitch(CYRUSOPT_AUTH_UNIX_GROUP_ENABLE))
   return newstate;
-pwd = getpwnam(identifier);
+getpwnam_r(identifier, pwd, buf2, BUFLEN, &pwdp);
setgrent();
-while ((grp = getgrent())) {
-   for (mem = grp->gr_mem; *mem; mem++) {
-   if (!strcmp(*mem, identifier)) break;
-   }
-
-   if (*mem || (pwd && pwd->pw_gid == grp->gr_gid)) {
-   newstate->ngroups++;
-   newstate->group = (char **)xrealloc((char *)newstate->group,
-   newstate->ngroups * 
sizeof(char *));
-   newstate->group[newstate->ngroups-1] = xstrdup(grp->gr_name);
-   }
+while (1) {
+  i = getgrent_r(&grp, buf, BUFLEN, &grpp);
+  if (i)
+   break;
+  for (mem = grpp->gr_mem; *mem; mem++) {
+   if (!strcmp(*mem, identifier)) break;
+  }
+  if (*mem || (pwdp && pwdp->pw_gid == grpp->gr_gid)) {
+   newstate->ngroups++;
+   newstate->group = (char **)xrealloc((char *)newstate->group,
+   newstate->ngroups * sizeof(char 
*));
+   newstate->group[newstate->ngroups-1] = xstrdup(grpp->gr_name);
+  }
}
endgrent();
+
return newstate;
}


---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: mupdate exits with signal 11

[Derrick J Brashear]

On Tue, 26 Apr 2005, João Assad wrote:
sometimes mupdate exits with signal 11 . Im clueless why.
cyrus-2.2.12
gdb backtrace attached.
http://64.233.161.104/search?q=cache:CAPMOK5A4vwJ:asg.web.cmu.edu/archive/message.php%3Fmailbox%3Darchive.info-cyrus%26msg%3D34797+getgrent+auth_unix+cyrus&hl=en
It's the same problem.
I haven't tested this code yet, it was literally written right now while 
I'm eating my lunch:
Index: auth_unix.c
===
RCS file: /afs/andrew.cmu.edu/system/cvs/src/cyrus/lib/auth_unix.c,v
retrieving revision 1.37.2.2
diff -u -r1.37.2.2 auth_unix.c
--- auth_unix.c 16 Feb 2005 21:06:50 -  1.37.2.2
+++ auth_unix.c 26 Apr 2005 17:43:56 -
@@ -221,9 +221,11 @@
 static struct auth_state *mynewstate(const char *identifier)
 {
 struct auth_state *newstate;
-struct passwd *pwd;
-struct group *grp;
+struct passwd pwd, *pwdp;
 char **mem;
+struct group grp, *grpp;
+char buf[BUFLEN], buf2[BUFLEN];
+int i;

 identifier = mycanonifyid(identifier, 0);
 if (!identifier) return 0;
@@ -238,22 +240,25 @@
 if(!libcyrus_config_getswitch(CYRUSOPT_AUTH_UNIX_GROUP_ENABLE))
return newstate;
-pwd = getpwnam(identifier);
+getpwnam_r(identifier, pwd, buf2, BUFLEN, &pwdp);
 setgrent();
-while ((grp = getgrent())) {
-   for (mem = grp->gr_mem; *mem; mem++) {
-   if (!strcmp(*mem, identifier)) break;
-   }
-
-   if (*mem || (pwd && pwd->pw_gid == grp->gr_gid)) {
-   newstate->ngroups++;
-   newstate->group = (char **)xrealloc((char *)newstate->group,
-   newstate->ngroups * sizeof(char 
*));
-   newstate->group[newstate->ngroups-1] = xstrdup(grp->gr_name);
-   }
+while (1) {
+  i = getgrent_r(&grp, buf, BUFLEN, &grpp);
+  if (i)
+   break;
+  for (mem = grpp->gr_mem; *mem; mem++) {
+   if (!strcmp(*mem, identifier)) break;
+  }
+  if (*mem || (pwdp && pwdp->pw_gid == grpp->gr_gid)) {
+   newstate->ngroups++;
+   newstate->group = (char **)xrealloc((char *)newstate->group,
+   newstate->ngroups * sizeof(char *));
+   newstate->group[newstate->ngroups-1] = xstrdup(grpp->gr_name);
+  }
 }
 endgrent();
+
 return newstate;
 }

Re: DB Errors

[Derrick J Brashear]

On Tue, 26 Apr 2005, Daniel Hazelbaker wrote:
(Quotes regarding how to get the core dumps generated snipped)
Okay.  I finally got it to generate core dumps correctly.  Unfortunetly it 
did not help any.  Here is the backtrace:
(gdb) where
#0  0x0fb263b0 in strcmp () from /lib/tls/libc.so.6
#1  0x10005228 in ?? ()

Not very helpful.  So my only other solution will be to run lmtpd inside of 
gdb and track it down manually.  However I still cannot get lmtpd to run 
correctly using the -D command.  It always complains about the CYRUS_MASTER 
environment or whatever it is.  Anybody know how to properly run lmtpd under 
gdb so I can step through the functions until I figure out what is wrong?
Why not attach to it after it's running? But if you have no symbols, you 
have no symbols...

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: User-deleted folders still on disk?

[Derrick J Brashear]

On Tue, 26 Apr 2005, Christophe Boyanique wrote:
Derrick J Brashear a écrit :
Well, a while ago there was a discussion about this related to NFS, and 
it's quite possibly the same problem, though it's much less an issue if 
you're not show your users du output.

http://64.233.187.104/search?q=cache:9aAtHxcLfIUJ:asg.web.cmu.edu/archive/message.php%3Fmailbox%3Darchive.info-cyrus%26msg%3D34476+cyrus+nfs+rmdir+site:asg.web.cmu.edu&hl=en 
So it seems that someone read that mail :)
I recall but am too lazy to check that he posted the patch after I 
suggested snooping nfs protocol traffic to determine if the directory was 
empty. So, well, I think there was even evidence of it at the time;-)

Is there a way to integrate this patch at least in Cyrus 2.1 and 2.2 ?
2.2: probably. it's actually uncommitted in my sandbox now. Ken's point 
about "can you remove your cwd" remains.
2.1: it's not a security fix.

Re: Problem with POP, STLS and Pegasus Mail?

[Derrick J Brashear]

On Tue, 26 Apr 2005, Sebastian Hagedorn wrote:
Well, I have meanwhile tried it myself and have found that it definitely 
*does* work for me. The user told me that a virus scanner is involved. I'm 
guessing that it's a POP proxy server and that *it* doesn't support the STLS 
command. The user claims that it doesn't even work without the virus scanner 
and that it *did* work with our old UW ipop3d server, but at this point I'm 
not sure I believe him ;-)
I don't. A proxy was my second guess, the non-English text had to come 
from somewhere and it's not coming from us.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Problem with POP, STLS and Pegasus Mail?

[Derrick J Brashear]

On Mon, 25 Apr 2005, Sebastian Hagedorn wrote:
[*] Connection established to 134.95.19.46
0132 +OK cyrus.rrz.uni-koeln.de Cyrus POP3 v2.2.12-Invoca-RPM-2.2.12-
1.ZAIK server ready <[EMAIL PROTECTED]>
<< 0006 STLS
0034 -ERR Nicht unterstützter Befehl.
<< 0014 USER xxx
0055 -ERR [AUTH] USER command only available under a layer
Now I have no idea why the error message following STLS is in German. It says 
"command not supported". When I try this with telnet, it looks fine:
Well, at least the other matches our source:
/* possibly disallow USER */
if (!(kflag || popd_starttls_done ||
  config_getswitch(IMAPOPT_ALLOWPLAINTEXT))) {
prot_printf(popd_out,
"-ERR [AUTH] USER command only available under a 
layer\r\n");
return;
}
Perchance if you tcpdump you see the client lying and not bothering to 
send the STLS at all?

Re: Sieve

[Derrick J Brashear]

On Mon, 25 Apr 2005, Derrick J Brashear wrote:
You mean a script that uses the perl modules? We have one around which I just 
had some fun debugging (there's a debug in the "new protocol" support which 
we found Thursday, it will be fixed in 2.2.13). I'll find it.
http://www.contrib.andrew.cmu.edu/~shadow/sieve.tar.gz
are several scripts we have that we run for our users.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Sieve

[Derrick J Brashear]

On Mon, 25 Apr 2005, NM Public wrote:
* I use Tuffmail.com and they support the ManageSIEVE protocol.
 I'd like to be able to update my Sieve greenlist (aka
 whitelist) from Pine. E.g., In Pine I pipe a message to a
 script and the script extracts the From: header and uses
 ManageSIEVE to update my remote Sieve script. Where can I find
 a ManageSIEVE script that will give me clues about how to do
 this? I run Pine on Mac OS X, FreeBSD, and Debian Linux so a
 Unix-type shell script would be great. BTW, I know how to
 extract the address from the From: header because I've been
 doing this for years for my Procmail recipes.
You mean a script that uses the perl modules? We have one around which I 
just had some fun debugging (there's a debug in the "new protocol" support 
which we found Thursday, it will be fixed in 2.2.13). I'll find it.


* Is it possible for Sieve to use "include" files so I can
 compartmentalize the various sections of my Sieve script? This
 seems like it would be especially useful in ensuring that I
 don't corrupt my entire Sieve script using the ManageSIEVE
 script that I describe above.
No (well, yes, but only starting in 2.3 versions)
* Are there Sieve syntax checking tools that will tell me if I
 have specified exactly the right "require" items?
I know of none.
 If so, where
 can I find them? Does it make much of difference if I specify
 more than is needed.
No
* I've read some Sieve examples on the Web and I don't understand
 why some people use 'header :contains' when 'address :all
 :comparator "i;ascii-casemap" :contains' seems to be more
 correct. Why would someone use 'header' when they are looking
 for an 'address'? Is this an example of cluelessness on the
 part of the Sieve scripter or is there something I'm not
 getting?
Sieve is an evolving language. Some syntax didn't exist at the beginning.
Of course, if you claim to prefer minimalism, why would you explicitly 
specify the default comparator? ;-)

* I've seen some recipes that have this:
   fileinto "foo"
   stop;
 Is the 'stop;' redundant here? In procmail, delivery means stop
 (unless the 'c' flag is used). What's the story in Sieve?
It means no later rules will apply (so you won't have multiple copies 
filed)
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: User-deleted folders still on disk?

[Derrick J Brashear]

On Mon, 25 Apr 2005, Jeff Blaine wrote:
Using cyrus-2.2.12
We're experiencing a situation where user-deleted folders still
remain on disk.
Running 'reconstruct -r user.USERNAME' does not address the
problem.
Is this expected behavior?  If so, I'm curious to hear the
reasoning.
Well, a while ago there was a discussion about this related to NFS, and 
it's quite possibly the same problem, though it's much less an issue if 
you're not show your users du output.

http://64.233.187.104/search?q=cache:9aAtHxcLfIUJ:asg.web.cmu.edu/archive/message.php%3Fmailbox%3Darchive.info-cyrus%26msg%3D34476+cyrus+nfs+rmdir+site:asg.web.cmu.edu&hl=en
Of course solving it in 2.3 is a little harder but this will probably do 
it:
Index: mailbox.c
===
RCS file: /afs/andrew.cmu.edu/system/cvs/src/cyrus/imap/mailbox.c,v
retrieving revision 1.147.2.28
diff -u -r1.147.2.28 mailbox.c
--- mailbox.c	7 Apr 2005 00:27:09 -	1.147.2.28
+++ mailbox.c	25 Apr 2005 16:39:25 -
@@ -2606,7 +2606,7 @@
 int r, rquota = 0;
 DIR *dirp;
 struct dirent *f;
-char buf[MAX_MAILBOX_PATH+1], *path;
+char buf[MAX_MAILBOX_PATH+1], *path, *lpath, *mpath = NULL;
 char *tail;
 struct txn *tid = NULL;

@@ -2640,7 +2640,11 @@
 }
 /* remove data (message file) directory */
-path = mailbox->path;
+lpath = strdup(mailbox->path);
+if (mailbox->mpath) 
+  mpath = strdup(mailbox->mpath);
+mailbox_close(mailbox);
+path = lpath;

 do {
/* remove all files in directory */
@@ -2686,10 +2690,11 @@
} while (rmdir(buf) == 0 && (tail = strrchr(buf, '/')));
/* remove metadata directory (if exists) */
-} while (mailbox->mpath && (path != mailbox->mpath) &&
-(path = mailbox->mpath));
+} while (mpath && (path != mpath) &&
+(path = mpath));
+free(lpath);
+if (mpath) free(mpath);
-mailbox_close(mailbox);
 return 0;
 }
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: DB Errors

[Derrick J Brashear]

On Sun, 24 Apr 2005, Daniel Hazelbaker wrote:
Okay.  I'm hitting the annoyed state with this problem because it is feeling 
close to being solved, but not quite!  The cwd of lmtpd process (and cyrus 
itself) is the root directory.  I did a quick test of setting / to mode 777 
(ya I know) and waited for an abnormal exit. Checked and still no core file. 
Nor can I get the debug_command option in lmtpd to work to see what is going 
on when it crashes.

I have, however, been able to track it down to crashing when I try to deliver 
to a non-existant mailbox.  For example, 'deliver user.blahblahblah' causes 
it to crash in a BUSY state, same goes for 'deliver blah'. So I am closer, 
but still can't figure out exactly why it is crashing, or not core dumping 
for that matter.  Short of recompiling lmtpd to not fork, is there something 
I am missing?  Here is my debug_command:
/usr/bin/gdb /usr/lib/cyrus-imapd/%s %d
Well, if you can crash it, don't worry about the core, and just get a 
backtrace in gdb.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

RE: Message contains NUL characters ...

[Derrick J Brashear]

On Mon, 25 Apr 2005, Michael Sims wrote:
Marc G. Fournier wrote:
Apr 24 15:29:02 hub postfix/lmtp[54800]: 0367A1298D0:
to=<[EMAIL PROTECTED]>, relay=/var/run/socket/lmtp[/var/run/socket/lmtp],
delay=273190, status=bounced (host
/var/run/socket/lmtp[/var/run/socket/lmtp] said: 554 5.6.0 Message
contains NUL characters (in reply to end of DATA command))
Hi,
This is a FAQ (although it appears not to be in the actual FAQ or the
wiki...). Basically you want to configure your MTA so that NULs are stripped
before the messages containing them are passed to LMTP.  Check the archives
(watch for wrapping on this URL):
http://marc.theaimsgroup.com/?l=info-cyrus&w=2&r=1&s=contains+nul+characters
+postfix&q=b
As for why this suddenly became a problem when it wasn't before, I have no
idea; maybe someone else can shed some light on that...
Well, there's also the possibility as we discussed in October that your 
mta is passing (too) long (to be compliant) lines to cyrus, and the logic 
doesn't distinguish and prints the same error.

The most recent patch we got messes with the prot_stdio code which is 
really not the most desirable way to deal. We'll come up with another 
way.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Sieve

[Derrick J Brashear]

On Mon, 25 Apr 2005, NM Public wrote:
On 2005-04-24 in , Forrest 
Aldrich wrote:
Either way, it would be really nice to have this at some point - a
mailing list would be a good ajunct.
Here's another vote for a Sieve user mailing list, preferably
linked to a Gmane newsgroup.
Anyone can link a list to a gmane newsgroup though, or am i wrong? (I 
understand the value of this, though, so you'll get no argument here)
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: DB Errors

[Derrick J Brashear]

On Sun, 24 Apr 2005, Daniel Hazelbaker wrote:
Hmm. One problem.  I can't find any core files. :)  As best I can tell they 
aren't disabled anywhere, (/proc/sys/kernel has valid core_pattern and 
core_uses_pid values ("core" and "1" respectively). ulimit shows core size of 
0 (I'm assuming this is unlimited).  My /etc/security/limits.conf is 
essentially empty.  Is there somewhere else I should check for 
enabling/disabling core files?
current working directory of lmtp writable to the uid that it's running 
as?

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: DB Errors

[Derrick J Brashear]

On Sun, 24 Apr 2005, Daniel Hazelbaker wrote:
Okay.  Got the log information and db_stat run today. From what I can tell 
though, lmtpunix keeps having SEGV errors, but no indication of why.  the 
pop3d process failed a few times, but not with SEGV - it had 
tls_start_servertls() errors.
Ok, so how about cores? You should be able to get a backtrace, and we can 
try to fix the problem.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Sieve

[Derrick J Brashear]

On Sun, 24 Apr 2005, Forrest Aldrich wrote:
Hi Derrick,
Does there exist an outline of requirements for a technical writer - ie:
what needs to be accomplished etc.
You mean for Cyrus or generally? Realize if Computing Services does manage 
to hire one we'll want them to write much more than just Cyrus 
documentation.

As for procmail, I've read some examples of getting this working with
Cyrus - but it seems to require a login account, which I believe defeats
our purpose of Cyrus and a closed system.   ?
It requires an /etc/passwd entry. Said account doesn't need a real shell, 
or to log in.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Sieve

[Derrick J Brashear]

On Sat, 23 Apr 2005, Forrest Aldrich wrote:
The Sieve web page states:
   At present, there is no mailing list or newsgroup for end users of
   Sieve.
I'm curious, how is it expected Sieve will be more widely adopted with
1) limited documentation (ie: not end-user friendly) and 2) no formal
resources to discuss the protocol/language.
"It was like that when I found it", so I won't take the blame for the 
current situation. The thing that kills us is we've never been able to 
keep a technical writer long enough to get good documentation for 
anything. It's been on my "sucks" list since at least 2000, and it's an 
issue for internal software and not just Cyrus.

A mailing list can be gotten trivially though but that's not really all it 
takes.

I've spent a significant amount of time perusing the net for resources -
I found Websieve, great - but that doesn't help me educate end users on
how to use this.
This probably sounds more accusatory than I mean it to, but if your users 
figured out how to use procmail I'd be shocked if they couldn't use sieve; 
In the pre-sieve days I used procmail on my own system, and sealed system 
was not a problem (that's what distributed filesystems are for); It would 
certainly be possible if not necessarily simplest to continue using 
procmail.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: crypto mailbox

[Derrick J Brashear]

On Fri, 22 Apr 2005, Cyrus Daboo wrote:
Hi Derrick,
[]
Even with that there is still a 'window of opportunity' in which the clear 
text message data exists somewhere on the machine (in a temp file, resident 
in memory etc) and root will likely be able to get to it. The bottom line is
To be honest my target was more along the line of Sam Weiler's worries, 
namely, for people who are paranoid a warrant will show up and their 
machine will walk away. Presumably there's still the "wiretap" equivalent 
problem (someone gets a copy of the data before it's encrypted) but if it 
went over the wire in the clear you already have that problem. This only 
precludes the bits from being available at a later time, not any time up 
until they're encrypted.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

RE: crypto mailbox

[Derrick J Brashear]

On Fri, 22 Apr 2005, Jason Huddleston wrote:
Why don't they use PGP for messages of this importance???
I can't control the sender's use of PGP.
(the excerpted bit from me which your mail client did something weird with 
instead of quoting follows:)

To be honest my target was more along the line of Sam Weiler's worries,
namely, for people who are paranoid a warrant will show up and their
machine will walk away. Presumably there's still the "wiretap" equivalent
problem (someone gets a copy of the data before it's encrypted) but if it
went over the wire in the clear you already have that problem. This only
precludes the bits from being available at a later time, not any time up
until they're encrypted.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: DB Errors

[Derrick J Brashear]

On Fri, 22 Apr 2005, Daniel Hazelbaker wrote:
Thanks, I'll take a look and check on that.  While I am doing so, here is the 
actual output from my log files showing the errors (grepped for DBERROR).  We 
only have 74 mailboxes and my mail system is set to only deliver 5 messages 
at a time.  Could this be a "locker" leak or something I am running into?
Nothing is dying uncleanly and orphaning locks?
cd into (confdir)/db (the directory with the __db.001 files) and run 
db_stat -Cl, then put the file preferably somewhere other than in what 
will be a huge mail if you send it to the list;-)

Apr 22 08:38:49 kempis lmtpunix[14284]: DBERROR db4: 1328 lockers
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: crypto mailbox

[Derrick J Brashear]

On Thu, 21 Apr 2005, Natalino Picone wrote:
Hi all,
i was wondering if there is a way to crypt the user mail spool in order to 
avoid that other users (root included) can read the user mailbox content. 
Something like gpg encryption on the mail file. Any idea ?
I had an idea for this where basically you'd put a public key into an 
annotation on a mailbox, and all mail payloads (not the headers) would be 
encrypted. It breaks things like search, but has the benefit you want. But 
it's low on the list of things to do. You could do similar with a sendmail 
milter or somesuch.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: change default mailbox names

[Derrick J Brashear]

On Wed, 20 Apr 2005, Nick Trenary wrote:
There answer may be "no" but unless you give an example of what you mean 
it's hard to know what you're really asking for.
The default mailbox names created are Inbox, Deleted Messages, Drafts, and 
Sent Messages.  I want to shorten the names to something like Inbox, Trash, 
Drafts and Out(box) respectively if I can.
inbox is a synonym for user.(you).
the rest are presumably being created by the client. if you want the 
client to create something else, well, you need to pervert the client to 
do so. what client?

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: change default mailbox names

[Derrick J Brashear]

On Tue, 19 Apr 2005, Nick Trenary wrote:
Is there a way to change the default mailbox names that are created in Cyrus? 
I'm using Cyrus on OS X.3 and cannot a config file that allows you to change 
the names.  Can someone point me in the right direction?
There answer may be "no" but unless you give an example of what you mean 
it's hard to know what you're really asking for.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Problem : unable to init duplicate delivery database ??

[Derrick J Brashear]

On Fri, 15 Apr 2005, Martine James (02 23 23 71 31) wrote:
I installed Cyrus one year ago and it used to work nicely with option 
"duplicatesuppression" as set by default (YES)
Things went wrong four days ago with this message:
Apr 14 12:00:04 machine.etu.univ-rennes1.fr pop3[7118]: [ID 729713 
local6.error] opening /var/imap/tls_sessions.db: Not enough space
Apr 14 12:14:11 machine.etu.univ-rennes1.fr lmtpunix[10565]: [ID 729713 
local6.error] DBERROR: opening /var/imap/deliver.db: Not enough space
???
Indeed, I checked ( twice) that no  partitions was full. I checked the
number of inodes, number of files descriptors. everything's ok.
Berkeley db, yes? You apparently had stuff crashing and leaking locks at 
some point.

I tried to restart cyrus. "ctl_cyrudb -r" command never ends.
Trace with strace. Does it keep trying to open file descriptors, never 
closing them?

You can try simply removing the duplicates database entirely and letting 
it get reinit'd. I can suggest a patch but this is simpler and faster.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: segfaults when executing quota command

[Derrick J Brashear]

On Thu, 14 Apr 2005, Craig White wrote:
NOTE: I would bet that the /usr/local/include/glob.h is the issue - that
would be on both systems exhibiting the problem and are apparently put
there by my compiling heimdal from source (did it on both systems)
Now, how can I fix this?
I should have guessed it was the roken glob.h issue. Sigh.
Well, either link libroken in (and thus use the roken glob) or remove 
/usr/local/include/glob.h

The former is presumably
setenv LDFLAGS -lroken; ./configure
the latter is obvious
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: segfaults when executing quota command

[Derrick J Brashear]

Starting program: /usr/lib/cyrus-imapd/quota -f
[Thread debugging using libthread_db enabled]
[New Thread -1218568064 (LWP 23872)]
/var/lib/imap/quota/q/
/var/lib/imap/quota/?/*
Ok, let's take a wild stab at this. I see glob() is being called from 
libc. What glob.h header is being found during compile?

find / -name glob.h -print (unless you have a network filesystem mounted, 
and then, I guess I need to see the output of df)

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: segfaults when executing quota command

[Derrick J Brashear]

On Wed, 13 Apr 2005, Craig White wrote:
what I don't know - and pardon my ignorance here...which binary file am
I supposed to copy over?
Presumably the one called quota, since that's what you're running... 
right?
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: segfaults when executing quota command

[Derrick J Brashear]

On Wed, 13 Apr 2005, Craig White wrote:
difference in size worries me...
# ls -l /usr/lib/cyrus-imapd/quota
-rwxr-xr-x1 root root   719864 \
Jan 15 11:33 /usr/lib/cyrus-imapd/quota
# ls -l /usr/src/redhat/BUILD/cyrus-imapd-2.2.12/imap/quota
-rwxr-xr-x1 root root  1295663 \
Apr 13 19:55 /usr/src/redhat/BUILD/cyrus-imapd-2.2.12/imap/quota
that's a lot of difference - is this the file?
if you built with symbols (-g) the new one better be bigger
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: segfaults when executing quota command

[Derrick J Brashear]

On Wed, 13 Apr 2005, Craig White wrote:
rpmbuild -bb /usr/src/redhat/SPECS/cyrus-imap.spec
which definitely worked
If that makes sense, I will uninstall the previously installed cyrus-
imapd rpms and install these new ones and then re-run the 'quota / gdb'
trace and hopefully this will give us more info to go on.
Actually you should be able to run the new quota binary against your 
existing installed binary set just for testing for the moment.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: mail to multiple users

[Derrick J Brashear]

On Thu, 14 Apr 2005, Helmut Weigel wrote:
Hi all!
I 've got a question about how messages are stored on the system.
Think of the following situation:
I have a huge number of users in my mailstore, lets say 5.
And now i want to send one single email, maybe with a fully featured 150 
Pages colored PDF Attachment, to them all.

Can you see any chance, that this message is just stored once in the 
mailstore and there are 5 references /links to that email?

What would the MTA have to tell cyrus to do that, if even possible?
I think a way around this are public imap folders. But it would be great if 
the first szenario would be possible.

Can anybody put a little light on this?
http://asg.web.cmu.edu/cyrus/download/imapd/overview.html#singleinstance
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: segfaults when executing quota command

[Derrick J Brashear]

On Tue, 12 Apr 2005, Craig White wrote:
Not especially. Having symbols would help. It's going to be something in
foreach() in lib/cyrusdb_quotalegacy.c, though, since that's the only
place you'd be using the system glob().
I suppose, for simplicity, you can just add a
printf("%s\n", quota_path);
before and after
 /* strip off the qr specific path and replace with pattern */
 sprintf(strstr(quota_path, FNAME_QUOTADIR) + strlen(FNAME_QUOTADIR),
 "?/%s*", prefix);
And see what 2 lines you get just before the crash, as a way of cheating.

OK - excuse my ignorance here please...
No, excuse mine, didn't even occur to me you weren't building from source 
directly.

I used invoca.ch srpms and did rpmbuild from them and never actually
compiled from source so I suspect that is why I can't locate it.
Well, I guess rpmbuild cleaned up after itself.
This is a whiteboxlinux 3 (RHEL 3 clone) - and I would take direction
here...should I install some symbol package for debugging
Would be nice but I wouldn't hold my breath for it. Maybe someone else 
here knows.

or untar the tarball - and compile after your suggested changes ?

Which would be the
better course? I actually have this problem on my system (whiteboxlinux)
and a client system (RHEL 3 ES) but I would rather fool with this on my
system.
Well, the first would be better, but I bet you can't have it.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: DB4 Locked

[Derrick J Brashear]

On Tue, 12 Apr 2005, Justin wrote:
lmtp[9563]: DBERROR db4: 82 lockers
IIRC, this harmless? Can someone please ack/nack on this?
This is fine as long as the number of lockers does not continually
increase.

...And if it does continually increase?
Do you have imapds or lmtpds (or something else which holds locks) 
crashing and orphaning their locks?

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: segfaults when executing quota command

[Derrick J Brashear]

On Mon, 11 Apr 2005, Craig White wrote:
thanks - I know not of these things - appreciate your patience with me.
(gdb) where
#0  0x00be1513 in _int_realloc () from /lib/tls/libc.so.6
#1  0x00be0156 in realloc () from /lib/tls/libc.so.6
#2  0x00c1b586 in glob_in_dir () from /lib/tls/libc.so.6
#3  0x00c1a7fc in glob () from /lib/tls/libc.so.6
#4  0x0808328e in mboxlist_findsub_alt ()
#5  0x0804baa0 in ?? ()
#6  0x0811b1e8 in ?? ()
#7  0xbfff8580 in ?? ()
#8  0x in ?? ()
indeed, it was not the only stack. There was no rest of the backtrace
unless you are asking for the stuff that came 'before' I executed '(gdb)
run -f' which I didn't include because I didn't think that it had
meaningful information but for completeness...
no, you answered what i wanted to know, above. I need to look at the regex
code in mboxlist_findsub_alt

any progress here?
Not especially. Having symbols would help. It's going to be something in 
foreach() in lib/cyrusdb_quotalegacy.c, though, since that's the only 
place you'd be using the system glob().

I suppose, for simplicity, you can just add a
printf("%s\n", quota_path);
before and after
/* strip off the qr specific path and replace with pattern */
sprintf(strstr(quota_path, FNAME_QUOTADIR) + strlen(FNAME_QUOTADIR),
"?/%s*", prefix);
And see what 2 lines you get just before the crash, as a way of cheating.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: mupdate thread died with SIGSEGV

[Derrick J Brashear]

On Sun, 10 Apr 2005, João Assad wrote:
This time I have strace and gdb logs. Im sending attached the last 1000 lines 
of the mupdate process strace, the last 1000 of the mupdate thread strace and 
the gdb trace I managed to extract from the core dump.
So, I assume it was thread 33, the only one not in int80, which crashed.
Thread 33 (process 15913):
#0  0x00534478 in strcmp () from /lib/tls/libc.so.6
No symbol table info available.
#1  0x08075b07 in auth_newstate (identifier=0x810de40 "cyrus") at 
auth_unix.c:24
8
newstate = (struct auth_state *) 0x3834f5d8
pwd = (struct passwd *) 0x5e7fb8
grp = (struct group *) 0x5e7f08
mem = (char **) 0x8122494
#2  0x0805f8d9 in mysasl_proxy_policy (conn=0x8d07230, context=0x0, 
requested_user=0x8d07b40 "cyrus", rlen=5,

if (!strcmp(*mem, identifier)) break;
where *mem came from:
while ((grp = getgrent())) {
for (mem = grp->gr_mem; *mem; mem++) {
Of course in another thread I see:
#12 0x0059ac31 in __nss_getent () from /lib/tls/libc.so.6
No symbol table info available.
#13 0x00551f51 in getgrent () from /lib/tls/libc.so.6
No symbol table info available.
#14 0x08075ad6 in auth_newstate (identifier=0x810de40 "cyrus") at 
auth_unix.c:24
6
newstate = (struct auth_state *) 0x383bd950

Which screams that we walked on some data in a non-reentrant interface.
I need to think about the right way to fix it.

Re: Cyrus Imap and crypt SQL Passwords?

[Derrick J Brashear]

On Mon, 11 Apr 2005, Berger, Stefan (IT- Management) wrote:
Hi,
is there a possibility for use crypt mysql passwords with cyrus imap? Is
there a patch available or any other solution? I've patched the cyrus sasl
2.1.19  and this works fine.
Not for any scheme which is hash based, only schemes which take a password 
(e.g. login). Otherwise the one-way crypt is not reversible and you have 
no way to transform the password.

The archives have a pointer to a patch, if this is what you want to do.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: cyrus-murder problems with database corruption in the frontend/master

[Derrick J Brashear]

On Fri, 8 Apr 2005, João Assad wrote:
I could do a strace -f wich would dump all the traces from all the threads 
into a single file... but its a nightmare to read it.
by reading some strace output here I've noticed mmaps complaining about 
ENOMEM way before the mmap inside map_refresh goes crazy.
then It came to me that cyrus only do mmap inside map_refresh and It seemed 
to me that it was tcp_wrappers mmap that was causing the rist ENOMEM

Im trying to recompile without tcp_wrappers and check if I can clean strace 
output a bit.

Actually its not tcp_wrappers.
When I start importing , I get 723 messeages like the ones below:
16677 22:43:37.184060 rt_sigaction(SIGPIPE, {SIG_IGN}, NULL, 8) = 0
16677 22:43:37.184221 mmap2(NULL, 10489856, PROT_READ|PROT_WRITE, 
MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = -1 ENOMEM (Cannot allocate memory)
[]
So, prior to this presumably you've mmap2()'d some memory, have there been 
any munmaps? 
cyrus-fe1 cyrus/mupdate[16671]: could not start a new worker thread (not 
fatal)
cyrus-fe1 last message repeated 722 times

Re: cyrus-murder problems with database corruption in the frontend/master

[Derrick J Brashear]

On Fri, 8 Apr 2005, João Assad wrote:
João Assad wrote:
Managed to get a backtrace using debug_command ( thanks for this nifty 
feature Henrique de Moraes )
2 gdb backtraces from the production server.
curiously, the strace output isn't showing an mmap() call fail, that I 
see, before the error shows up.

Re: cyrus-murder problems with database corruption in the frontend/master

[Derrick J Brashear]

On Thu, 7 Apr 2005, João Assad wrote:
Ok I got a backtrace ( I think ) . I dont really know how to use gdb
did you compile without giving gcc the -g option? Probably. Having 
unstripped binaries with useful symbols would probably make for a more 
useful backtrace. (at least i hope so)

Re: cyrus-murder problems with database corruption in the frontend/master

[Derrick J Brashear]

On Wed, 6 Apr 2005, João Assad wrote:

and then give us a backtrace from the core which you will then get?

After doing that, the mupdate process now exits with signal 11 as expected.
OTOH the core isnt getting dumped to disk for some reason...
The only internal resource limit play happens for fds, not coredumps, so 
we haven't disabled it, if you can't get a core, I'm somewhat lost as to 
suggestions.

Re: cyrus-murder problems with database corruption in the frontend/master

[Derrick J Brashear]

On Wed, 6 Apr 2005, João Assad wrote:
cyrus/mupdate[12614]: IOERROR: mapping /var/lib/imap/mailboxes.db file: 
Cannot allocate memory

Resource limited memory, or are you really running out of memory?
Letting processes continue running in the face of an mmap failure needs to 
be re-examined I guess.
[]
I've noticed that once I get the mmap error, the system will still run 
without spitting db errors for anywhere from a few mins to a few hours. Also, 
I never get more than one mmap error before the db becomes unnusable.
that's not shocking.  i'd still like to know why mmap fails, but, failing 
that, and since you're already getting corruption, can you in mupdate.c 
replace fatal with a function which does:

void fatal(const char *s, int code)
{
  assert(0);
  exit(code); /* shut up GCC */
}
and then give us a backtrace from the core which you will then get?

Re: bugzilla feedback

[Derrick J Brashear]

On Wed, 6 Apr 2005, Patrick Welche wrote:
Is there any way to get feedback on bugs? (eg 2588)
Wait until someone (me, probably) has time to deal with it. Really, 
though, I expect before the end of the week to re-review all the sasl bugs 
to see what can/needs to be resolved so we can have another release.

Realistically, it looks fine, but there wasn't really much to say about it 
before it was committed.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: segfaults when executing quota command

[Derrick J Brashear]

On Sun, 3 Apr 2005, Craig White wrote:
(gdb) run -f
Starting program: /usr/lib/cyrus-imapd/quota -f
(no debugging symbols found)...(no debugging symbols found)...(no
debugging symbols found)...(no debugging symbols found)...(no debugging
symbols found)...(no debugging symbols found)...(no debugging symbols
found)...(no debugging symbols found)...(no debugging symbols
found)...(no debugging symbols found)...(no debugging symbols
found)...(no debugging symbols found)...(no debugging symbols
found)...(no debugging symbols found)...[Thread debugging using
libthread_db enabled]
[New Thread -1218568064 (LWP 7193)]
(no debugging symbols found)...
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1218568064 (LWP 7193)]
0x00d1c513 in _int_realloc () from /lib/tls/libc.so.6
If you have one stack frame only, you have some serious problems. I doubt
you do. What's the rest of the backtrace?
(gdb's "where" commands)

thanks - I know not of these things - appreciate your patience with me.
(gdb) where
#0  0x00be1513 in _int_realloc () from /lib/tls/libc.so.6
#1  0x00be0156 in realloc () from /lib/tls/libc.so.6
#2  0x00c1b586 in glob_in_dir () from /lib/tls/libc.so.6
#3  0x00c1a7fc in glob () from /lib/tls/libc.so.6
#4  0x0808328e in mboxlist_findsub_alt ()
#5  0x0804baa0 in ?? ()
#6  0x0811b1e8 in ?? ()
#7  0xbfff8580 in ?? ()
#8  0x in ?? ()
indeed, it was not the only stack. There was no rest of the backtrace
unless you are asking for the stuff that came 'before' I executed '(gdb)
run -f' which I didn't include because I didn't think that it had
meaningful information but for completeness...
no, you answered what i wanted to know, above. I need to look at the regex 
code in mboxlist_findsub_alt
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: segfaults when executing quota command

[Derrick J Brashear]

On Sun, 3 Apr 2005, Craig White wrote:
(gdb) run -f
Starting program: /usr/lib/cyrus-imapd/quota -f
(no debugging symbols found)...(no debugging symbols found)...(no
debugging symbols found)...(no debugging symbols found)...(no debugging
symbols found)...(no debugging symbols found)...(no debugging symbols
found)...(no debugging symbols found)...(no debugging symbols
found)...(no debugging symbols found)...(no debugging symbols
found)...(no debugging symbols found)...(no debugging symbols
found)...(no debugging symbols found)...[Thread debugging using
libthread_db enabled]
[New Thread -1218568064 (LWP 7193)]
(no debugging symbols found)...
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1218568064 (LWP 7193)]
0x00d1c513 in _int_realloc () from /lib/tls/libc.so.6
If you have one stack frame only, you have some serious problems. I doubt 
you do. What's the rest of the backtrace?
(gdb's "where" commands)
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: segfaults when executing quota command

[Derrick J Brashear]

I suspect that this is caused by my setup.
Quite possibly, I'd guess we'd have seen it before if it were a general 
issue, but...

number of things including openldap, cyrus-sasl, heimdal-kerberos,
openssl and what I perceive to the problem, sleepycat db-4.2.52 (+ a
couple of patches for openldap transactions).
Ok, but,
# rpm -q db4
db4-4.1.25-8
if your LD_RUN_PATH wasn't set usefull you could be linked against one set 
of libraries at link time but finding the other at runtime, depending on 
the names of the libraries.

anyway...
since any and all attempts to run quota segfault, I have included the
end of the stream of an strace on the command which may help someone
tell me what I need to do
I'd suggest (assumes csh, since guess what I use;-)
1)  strace su - cyrus -c '/usr/lib/cyrus-imapd/quota -f'|& grep ^open
should tell us exactly which shared libraries are being used and for that 
matter what else is opened.

2) su - cyrus -c 'gdb /usr/lib/cyrus-imapd/quota' and then at the gdb 
prompt,
run -f

and see if you get at least function names in the backtrace.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: cyrus folders synchronization

[Derrick J Brashear]

There is no sasl question here. cyrus-sasl list pruned...
On Sun, 3 Apr 2005, sam wun wrote:
Hi,
I just finished building postfix with cyrus,openldap2.2 and db4.2, however I 
can't get isync built with db4.2.
probably out of scope, but what problem are you having?
I want to use isync to synchronize cyrus dirs to a backup server. if isync is 
not good to perform this task, what other method I can use to handle cyrus 
folders synchronization with a remote backup server?
it's probably not the most efficient way to do it, but it will work.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: spamassassin - sa-learn and public

[Derrick J Brashear]

On Fri, 1 Apr 2005, Craig White wrote:
On Fri, 2005-04-01 at 10:37 -0800, Norman Zhang wrote:
Craig White wrote:
Now that I have a functional bayes system for spamassassin...does it
make sense to have a subfolder in the 'public' mailbox for people to
drag their low scoring spam into so sa-learn can add it to the bayes db
by cron based scripts?
Is there a better way?
If 'anyone' has lrs rights on public but lrsiwpcda rights on
public.SPAM-learn would they be able to do this?
I've googled the following thread earlier. Perhaps it is helpful.
http://marc.theaimsgroup.com/?t=11088543253&r=1&w=2

it's interesting and I have done that in the past with uw-imap but I
never had any shared folders in uw-imap.
I can't see the logic of automatically taking people's inbox
automatically as ham...
My thought would be:
1) set up per-user bayesian filters
2) set up per-user spamtrain and hamtrain folders (someuser lrs, and 
authenticate a cron job as someuser)
3) cron job reads your spam and ham hints, updates your bayesian filter.
(and if you want to give it more access, flushes the folder)

But I suspect this doesn't sufficiently meet your needs.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: IMAP session minor issue - clarification needed

[Derrick J Brashear]

On Fri, 1 Apr 2005, Simon Matter wrote:
Nikola Milutinovic wrote:
Hi all.
I have Cyrus IMAP 2.2.10 on 12 servers. I have noticed a strange issue
when my users use IMAP service. Maybe it is a feature...
I could try to use POPalarm, since both POP3 and IMAP4 are enabled,
but that is an ugly workaround.
Hi,
Well, im using IMAP checker (http://www.m-tecs.net/files/imapchk.exe) -
freeware and it works great.
Server is 2.2.12. Clients are outlook or thunderbird.
Wait a minute, is this plain Cyrus IMAP 2.2.12 or a patched one? There is
an OE seen state patch floating around which I've included in my rpms and
IIRC it just fixes the problem discussed here. IIRC the patch is really a
hack but at least it does work.
that would presumably be this:
http://www-uxsup.csx.cam.ac.uk/~dpc22/cyrus/patches/2.1.16/OutLookExpress-seenstate.patch
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: krb.equiv not working

[Derrick J Brashear]

On Fri, 1 Apr 2005, Giri Raichur wrote:
Derrick,
Thanks for your answer. Is there something similar for GSSAPI i.e
mapping kerberos principal to mailbox name?
Not within Cyrus. You could certainly set
loginuseacl: t
and put the real username on the acl of the mailbox for the username you'd 
like them to log in as.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: krb.equiv not working

[Derrick J Brashear]

On Thu, 31 Mar 2005, Derrick J Brashear wrote:
On Wed, 30 Mar 2005, Giri Raichur wrote:
Derrick,
[EMAIL PROTECTED] root]# klist
Ticket cache: FILE:/tmp/krb5cc_0_ftp582
Default principal: [EMAIL PROTECTED]
ok, so i guess the other thing i should ask is what is the local realm?
Actually, the answer is krb.equiv only applies to kerberos 4 and not 
gssapi.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: cyrus-murder problems with database corruption in the frontend/master

[Derrick J Brashear]

On Thu, 31 Mar 2005, João Assad wrote:
cyrus/mupdate[12614]: IOERROR: mapping /var/lib/imap/mailboxes.db file: 
Cannot allocate memory
Resource limited memory, or are you really running out of memory?
Letting processes continue running in the face of an mmap failure needs to 
be re-examined I guess.

Re: cyrus-murder problems with database corruption in the frontend/master

[Derrick J Brashear]

On Thu, 31 Mar 2005, João Assad wrote:
My db just got corrupted again 2 hours ago. seems like moving 200k mailboxes 
between backends really speed it up.
I have my corrupted mailboxes.db and I can send it to anyone interested in 
taking a look.
how about a url for it?

Re: krb.equiv not working

[Derrick J Brashear]

On Wed, 30 Mar 2005, Giri Raichur wrote:
Derrick,
[EMAIL PROTECTED] root]# klist
Ticket cache: FILE:/tmp/krb5cc_0_ftp582
Default principal: [EMAIL PROTECTED]
ok, so i guess the other thing i should ask is what is the local realm?
List Archives/Info: 
http://asg.web.cmu.edu/cyrus/mailing-list.html
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: 
http://asg.web.cmu.edu/cyrus/mailing-list.html


---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: cyrus-murder problems with database corruption in the frontend/master

[Derrick J Brashear]

Well, you have skiplist corruption, but there's 
not really anything in your report which is 
helpful at suggesting why you do, or helping to 
reproduce it so (if it is a bug) it can be 
tracked and killed.

Even posting your corrupted skiplist would be 
more useful.
Yeah I know the info I gave isn't very helpfull.. 
the problem is that the corruption is random.. 
sometimes it takes 2 days to happen, sometimes a 
week. Any sujestions on how to debug and try catch 
the problem ?

Anyway, I've droped the fedora package and 
compiled the 2.2.12 source with only the features 
I need... lets see if it happens again.
The copout would be to use some other database type than skiplist, but you 
may not want to do that, really. I can't offhand think of anything useful 
to suggest that's likely to be effective, but I supose if your mupdate 
connections are cleartext you could tcpdump to capture the traffic too and 
perhaps correlate time, or replay to reproduce. But maybe I can come up 
with a more useful idea.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: cyrus-murder problems with database corruption in the frontend/master

[Derrick J Brashear]

On Tue, 29 Mar 2005, João Assad wrote:
Come on guys, someone must have at least an idea I can try.
Anything will help, maybe Im missing something obvious.
Well, you have skiplist corruption, but there's not really anything in 
your report which is helpful at suggesting why you do, or helping to 
reproduce it so (if it is a bug) it can be tracked and killed.

Even posting your corrupted skiplist would be more useful.
 cyrus/mupdate[1433]: DBERROR: skiplist recovery
 /var/lib/imap/mailboxes.db: 141BAD94 should be ADD or DELETE
 cyrus/mupdate[1433]: DBERROR: error updating database user.1215407: cyrusdb 
error

Re: krb.equiv not working

[Derrick J Brashear]

On Wed, 30 Mar 2005, Giri Raichur wrote:
Hello list,
I am trying to use krb.equiv to map kerberos principal names to mailbox
names when the names are not the same.
[EMAIL PROTECTED] has a valid kerberos ticket and
I have an entry [EMAIL PROTECTED] mapping to gzzr in krb.equiv but I see
the following in the logs:
What does klist say, and what does cat /etc/krb.equiv say?
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Erro compiling -- undefined reference to netsnmp_enable_subagent

[Derrick J Brashear]

On Mon, 28 Mar 2005, Javier Ramos wrote:
Hi all,
I am getting the following error compiling under Red Hat Enterprise Linux 
3.0, and cannot figure where the problem is... can anyone give any clues?
Thanks a lot for your help...
Regards,

Javier
gcc -c -I.. -I./../lib -I./../imap  -I/usr/include/rpm 
-I/usr/include/net-snmp -I/usr/include/et -I/usr/kerberos/include 
-I/usr/include  -DHAVE_CONFIG_H  -O \
cyrusMasterMIB.c
gcc -s  -o master master.o masterconf.o cyrusMasterMIB.o ../lib/lock_fcntl.o 
../lib/libcyrus_min.a -L/usr/lib -lnetsnmpagent -lnetsnmpmibs 
-lnetsnmphelpers -lnetsnmp -ldl -lrpm -lrpmio -lpopt -lbz2 -lz -lcrypto -lelf 
-lm -L/usr/lib/lib -lwrap -lresolv -lssl -lcrypto   -lfl -L/usr/lib -lnetsnmp 
-lcrypto -lelf -lm   -ldb-4.1 -lcom_err
/usr/bin/ld: warning: libcom_err.so.3, needed by /usr/lib/libcrypto.so, may 
conflict with libcom_err.so.2
master.o(.text+0x3135): In function `main':
: undefined reference to `netsnmp_enable_subagent'
The version of the netsnmp package you have is either too old or too new, 
I guess, to have that symbol. You can disable snmp at configure time if 
you don't want to install another version.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: acls on shared folders

[Derrick J Brashear]

On Sat, 26 Mar 2005, seph wrote:
wouldn't it let cyrus deliver to shared boxes without having to give
anyone p? Am I missing something obvious?
Yes, but what is that saving you from?
users shooting themselves in the foot. In my experience, users viewing
shared folders through an MTA quite often do things they oughtn't:
drag messages in, bypassing the distribution list; accidentally delete
the folder... Not giving them them permission beyond lrs saves a slew
of problems.
Ok, then it makes sense that you want this.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: acls on shared folders

[Derrick J Brashear]

On Fri, 25 Mar 2005, seph wrote:
Well, it hardcodes "postman" in yet another place but that's about the
only complaint.
But what does doing this buy you?
wouldn't it let cyrus deliver to shared boxes without having to give
anyone p? Am I missing something obvious?
Yes, but what is that saving you from?
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: parse_address() and restricting SIEVE vacation replies

[Derrick J Brashear]

On Fri, 25 Mar 2005, James Ralston wrote:
Because of SpamCop's attitude towards autoresponders*, I wish to patch
the implementation of the SIEVE vacation extension so that it will
never respond to remote addresses.
Wow, that's pretty useless:
"Solution: Do not use these systems. Inform your normal corespondents of 
your absence before you depart. Or let a co-worker answer your email in 
your absence. Publish FAQ information on a web-site. If you wish to 
dispense information via email, it's easy to reject a message while 
referring the sender to a FAQ web-page. Using sendmail, this is done in 
the access.db table like so:

to:[EMAIL PROTECTED] 550 Old address no longer valid, please see: 
http://www.example.com/NewAddressInfo.html";

So according to them I should either let someone else read my mail, or 
reject my mail... by giving users the ability to load addresses into 
sendmail config files!

From looking at sieve/bc_eval.c, this actually seems pretty
straightforward to do; I can just add another test case to
shouldRespond().
There's a comment fairly early in shouldRespond:
   /* Note: the domain-part of all addresses are canonicalized */
Does this mean that parse_address() is always going to return an
addr-spec, as per RFC2822 section 3.4.1?
As opposed to what?
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: acls on shared folders

[Derrick J Brashear]

On Fri, 25 Mar 2005, seph wrote:
I'd like to use a couple shared imap folders to archive some of our
internal lists. I can't figure out how to make the acl do what I
want. Is there any way to avoid giving anyone the post right?
This seems to have come up countless times before, and I haven't ever
seen a conclusive answer. My lmtpd logs that it's pre-auth'ed, but I
can't figure out how to grant it rights. I found this patch
http://asg.web.cmu.edu/archive/message.php?mailbox=archive.cyrus-devel&msg=745
but it doesn't seem to have been applied to the current release. Is
there some reason that approach is wrong?
Well, it hardcodes "postman" in yet another place but that's about the 
only complaint.

But what does doing this buy you?
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Cyrus, Sieve, and Evolution.

[Derrick J Brashear]

On Thu, 24 Mar 2005, Brian Dockter wrote:
I have Cyrus 2.1.17 set up on a Debian Sarge box. I use Evolution 2.0.3
to read my email via IMAP. Recently, I have started to use Sieve to sort
my incoming email. One problem I've noticed is when I click on a new
message (not all new messages, typically just the most recent),
Evolution displays an empty message body. If I change to a different
message and then go back the first message, the message is displayed
correctly. Could this be a Cyrus issue or should I go bug the people
over at Evolution?
It could be a Cyrus issue but I'm guessing you're getting a body displayed 
before the client finishes fetching the body. You could possibly enable 
telemetry logging on the server to see when the body is fetched.


---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Can someone reproduce this possible bug?

[Derrick J Brashear]

On Thu, 24 Mar 2005, Eddy Beliveau wrote:
Hi!
When using an accentuated character on a subject line header,
your client will modify the Subject line as in:
Subject: =?iso-8859-1?Q?=5BMy=5Ffilter=5D_Ol=E0!?=
Maybe the filter mechanism see a difference between "=5BMy=5Ffilter=5D" and 
"[My filter]"

Can you try to change your rule as:
if header :contains "Subject" "=5BMy=5Ffilter=5D" {
fileinto "INBOX.myfilter";
}
Right now, sieve won't decode =?iso-8859-1?Q?=5BMy=5Ffilter=5D_Ol=E0!?= to 
[My filter], basically we need a canonical version to do matching on, but 
the canonical version used for SEARCH is not suitable; Worse, since sieve 
does regex matching it's conceivable that after rewriting to a canonical 
form if we're not careful we may also need to rewrite the sieve script's 
regex to still match on changes we made, and I'd like to avoid that.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Quota warnings with Outlook

[Derrick J Brashear]

On Mon, 21 Mar 2005, Gary Mills wrote:
We've had a few complaints from people who are using MS Outlook as a
mail reader about pop-up messages for quota warnings.  Apparently,
once usage exceeds 90%, the pop-up messages appear so frequently that
the application becomes almost unusable.  Is there anything that can
be done, either in Outlook or Cyrus, to make this feature less
disruptive?
Does it select the mailbox repeatedly? Or does it, upon seeing the alert, 
reselect the mailbox more rapidly than it would otherwise?

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Cyrus strikes after SuSE Online Update

[Derrick J Brashear]

On Mon, 21 Mar 2005, Thomas Schilling wrote:
Hi Folks.
After Update my 9.1 pro. System, cyrus break down
- messages ---
Mar 19 12:14:17 hal2 ctl_cyrusdb[1711]: DBERROR db4: Berkeley DB library
configured to support only DB_PRIVATE environments


I've google around since 2 Days but I cant't find a solution.
30 seconds with google suggests this:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=91933
Anyway, what kernel version do you have, and was it the only choice? I 
would expect if that's the problem, you'd have a choice.

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: weird sieve issues

[Derrick J Brashear]

On Wed, 16 Mar 2005, Cory D. Wiles wrote:
Programs Installed:
Cyrus IMAPd 2.2.12
Cyrus SASL 2.1.20
I have been trying to get sieve running all day, but to no avail. If I
run the following command:
shell>$sieveshell --user=cwiles -a cyrus localhost
I get the following error:
connecting to localhost
Bad protocol from MANAGESIEVE server: EOL2
When I looked at my imadp.log I noticed that I was getting the following
error:
IOERROR: not a sieve bytecode file /usr/local/sieve/c/cwiles/defaultbc
Said file gets created when you enable a script as default. Don't 
precreate it unless you're putting in a script for all your users in 
advance, and if so put the script there also.

When I removed the defaultbc file I now get:
IOERROR: fstating sieve script /usr/local/sieve/c/cwiles/defaultbc: No
such file or directory
find /usr/local/sieve -print|xargs chown cyrus
and see if that fixes it? I can't look at source right this second but I'm 
guessing the ownership is wrong...


If anyone has any insight as to what I am  missing then i would really
appreciate it.  I have been googling all day long and I am out of ideas.
Thanks,
Cory
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Problems building cyrus-imapd

[Derrick J Brashear]

On Tue, 15 Mar 2005, Robert wrote:
Hi.  I am trying to build cyrus-imapd 2.2.12 and I am
having a bit of difficulty getting it to build.  The
specific error I get is:
--
gcc -L/usr/local/ssl/lib -Wl,-rpath,/usr/local/ssl/lib
-L/usr/local/BerkeleyDB.4.2/lib
-Wl,-rpath,/usr/local/BerkeleyDB.4.2/lib
-L/usr/local/lib -L/usr/local/BerkeleyDB.4.2/lib
-L/usr/local/ssl/lib -L/usr/local/lib -o sievec
sievec.o libsieve.a ../lib/libcyrus.a
../lib/libcyrus_min.a  libsieve.a -L/usr/local/lib/lib
-Wl,-rpath,/usr/local/lib/lib -lsasl2
-L/usr/local/lib -lgssapi_krb5 -lkrb5 -lk5crypto
-lcom_err -lresolv -lresolv  -lfl  -lresolv
-L/usr/local/BerkeleyDB.4.2/lib
-Wl,-rpath,/usr/local/BerkeleyDB.4.2/lib -ldb-4.2
-lssl -lcrypto  -lcom_err
libsieve.a(sieve_err.o): In function
`initialize_siev_error_table':
/vol/source/cyrus-imapd-2.2.12/sieve/sieve_err.c:40:
undefined reference to `_et_list'
What com_err.h and compile_et binary were used during compile? What 
com_err library?

probably
which compile_et
and
find /usr -name com_err.h -print
and
find /usr -name libcom_err.a
I bet you need comerr-devel rpm or somesuch.
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: CVSweb problem

[Derrick J Brashear]

I'll fix the commit script to send good links. Ken pointed it out but I've 
been sick the last couple days.

On Wed, 16 Mar 2005, Christos Soulios wrote:
Hi all
 There seems to be a problem with cvs web, because during the last month
every link to a diff from the cyrus-cvs list is broken.
Could it be fixed because this feature is very useful for monitoring
changes in the Cyrus souces.
Regards,
Christos
--
Did you visit http://email.uoa.gr?
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Problem while deleting mailbox on a private spool on NFS

[Derrick J Brashear]

On Tue, 15 Mar 2005, Christophe Boyanique wrote:

When I delete a mailbox (via cyradm for example), 
the mailbox directory is not deleted on the server 
with the NFS spool; but the content of the mailbox 
is deleted (including the cyrus.header, cache et 
index).
I suspect if you watch nfs network traffic you see a failed attempt to 
delete the directory, the problem is likely that the nfs server for some 
reason believes a file in the directory is still existant or possibly even 
open and so is unwilling to delete. Is anything in your log?

---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: IMAP Compile Error: initialize_mupd_error_table_r

[Derrick J Brashear]

On Sat, 12 Mar 2005, John Lane wrote:
Hello,
I am trying to build cyrus-imapd-2.2.12 but I get a compile error as shown
below. This same error has been reported before and I found two references
to it in the mailing list archives (#26174, #30497), both without replies.
If, after receiving the error, I comment out line 20 from imap/mupdate_err.h
I can finish the build off successfully.
I'm going to proceed on this basis but, please, can someone please explain
the correct solution to this problem,
Many thanks,
John Lane
gcc -c -I.. -I./../lib  -I./../sieve-DHAVE_CONFIG_H  -g -O2 \
mupdate-client.c
In file included from mupdate-client.h:47,
from mupdate.h:56,
from mupdate-client.c:75:
mupdate_err.h:20: warning: `struct et_list' declared inside parameter list
struct et_list isn't being defined. Normally it comes from your com_err.h 
header (probably /usr/include or /usr/local/include). I assume it's not 
defined there.

So I'd guess you have a com_err.h header being used which doesn't match 
the compile_et that's being used.

What platform?
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html