Aides can send an instant message to broadcast and everyone will receive
it.
(Hmm ... that's not in the docs is it)
Good question. I would imagine they'd receive broadcast messages, but
non-Aides wouldn't be allowed to send any messages at all to them.
** IMPORTANT **
We have shuffled the WebCit builds around. The branch formerly known as
webcit_before_automake is once again the trunk. Therefore, the
available webcit trees are:
/trunk/webcit (non-automake build)
/tags/webcit_automake (automake build)
Yeesh.Remember that big heinous memory leak in WebCit that I've been hunting, on and off, for a couple of months now?I've found it.I kept assuming that it had something to do with the way we were calling OpenSSL -- that we were using it in such a way that we were causing it to dereference
Interesting things are being tried this week.
Over at the Big Blue X we have a user who keeps a few mailboxes that have
tens of thousands of messages in them. He's been complaining about IMAP
performance, which baffled me because the problem operation in question
was a simple header
So I've been staring at a whiteboard for half an hour, trying to figure
out how we're going to implement the whole aliases / distribution lists /
virtual addresses thing.
We need to be able to define one or more destinations for an alias, and
those destinations can be local users, remote
Ok, if you want to try again, then go ahead ... but keep track of the
changes we're making in the trunk in case anything changes that would
break your branch later. I don't anticipate that being the case, though,
since pretty much everything we're doing at this point is up at the UI
layer
As for the aliases thing...
I'm inclined to go with a variation on LS's proposal. How about if we
boil it down to this:
1. A flag (or maybe a view setting) that says this room exists only as a
message distribution point (i.e. its netconfig file)
2. New field in either the
to that address. Example:
Citadel Development) @Server command: qdir [EMAIL PROTECTED]
200 IGnatius T Foobar @ uncnsrd
Citadel Development) @Server command: qdir [EMAIL PROTECTED]
200 IGnatius T Foobar @ uncnsrd
If you think it's malfunctioning, you can always run an IGAB command to
rebuild your
The difference between Zimbra AJAX and Citadel AJAX is that Zimbra is
designed from the ground up to have a never reload the whole page
approach, while Citadel (WebCit, actually) uses the AJAX model for a
performance boost in places where we think it makes sense.
Both are valid
Thanks, I'll go apply that now.
Oh crap, I think I know what that's about.
Ok, I think it's time to publish a set of 6.82 tarballs. Does anyone have
any issues which we need to cover first?
Two new languages, more improvement in IMAP performance, and some fairly
significant bugfixes ... we've got to get this stuff out there.
Plus it gives us an excuse to
Better yet, if malloc.h is no longer needed on Linux *or* FreeBSD then
we can just pull it altogether.
Ok, the patches are applied and the new functionality works. Thank you!
Please double check to make sure I got your name correct in the credits.
So ... what areas of the Citadel system would you find interesting to
work with? I don't like to just hand out programming tasks
Question for IO:
citadel_ipc.c, line 2666:
if (errval == SSL_ERROR_ZERO_RETURN ||
errval == SSL_ERROR_SSL) {
serv_read(ipc, buf[len], bytes - len);
return;
Nice to hear from you again, Seg. Support will be provided for whatever
you need to to to make up those Fedora modules. Then we need to make some
*noise* on the Fedora list!
I'm going to have limited access to the computer this coming week, so I
will only be here sporadically. I'll
That's cool.
Wow, those folks are picky.
I have been told that the new version of the connector is in active
development. Hopefully we will have a working version, or at least a
beta, this year. I'm looking forward to it too.
Ok, I've made some progress on this, and also uncovered another bug.
When ical_dezonify() encounters a timestamp that is not UTC but has no
timezone property, it now converts from the default timezone to UTC.
Right now the default timezone is hardcoded as America/New_York but I
did
Blah. I spent several hours last night trying to come up with a way to
get libical to know the timezone this computer is currently in but could
not come up with anything. I would appreciate any help anyone could
offer. This would be a nice little bit of code that doesn't require
in-depth
And how the heck do they wind up in this room?
Spam arrives here because it's mailed to the room's email address, which
spammers have gotten hold of via an archive somewhere.
In the not too distant future we're going to have to find a way to
authenticate messages posted via
It is. SpamAssassin is far from perfect.
How do you feed spam to it? More importantly, how do you feed ham to it?
Matt: I like what I see. Clearly you have better web-fu than I do. :)
Would you be willing to start folding your changes into the main SVN
tree?
I'm quite interested in this lightweight XML parser. I don't plan to
begin parsing PROPFIND requests unless I find that the current
Development notes...
As some of you are aware, our QA department is always complaining about
poor performance moving messages in or out of very large folders (mainly
because our QA department treats his Trash folder as an archive of every
message I've ever received -- but he provides
I saw that mentioned by an Apple engineer on the CalDAV mailing list.
Feh. More f***ing CalDAV.
Anyway, if by write support for ical you mean the ability to publish
your calendar using straight iCal-over-webdav, then the answer is that
it's already there. The subscribe URL's are also
I think we should implement a separate module for this, which listens on a
different port, and is disabled by default. Anyone who wants to use it
can enable the service by setting the port number. This would allow
firewall rules to be constructed.
Ok, I've posted your Postfix TCP Dictionary document into the knowledge
base on www.citadel.org
dothebart: where is serv_managesieve.c ??
The current build appears to require it, but I don't see that file
anywhere. Did you forget to check it in?
Also ... I'm going to play with libsieve a bit. Right now I'm going to
work on the configure script stuff required to look for
Nope, I'm getting the same error, straight across the local network
without any encryption. I've reported it to the author.
Enough already. I do think it's a good idea, and we should put it in the
plan. I am envisioning a future version of the Citadel system in which
the functionality currently provided by rss2ctdl is embedded directly into
the server. This would of course require that we make a decision about
Thanks IO, now it would have been nice if you told me how to do that ...
hmm ... ok, after RTFM it seems I have to add -v to the svn log
command in the post-commit hook. Let's try it...
matt: sure, why not? Go ahead and submit it.
Great patches, thanks for submitting them!
i want it to do... but, perhaps i should have a closer look, or, can
you give me a hint, ig?
msiv_putscript() only alters the Sieve configuration in memory.
You have to do msiv_load() first, then msiv_putscript(), then
msiv_store().
Look at how cmd_msiv() does it.
spaces and newlines shouldn't really matter; they're just extra
characters.
HA. That's EXACTLY how I had planned to do it. :)
The queue view and the optimized mime downloading are things I've been
meaning to put in for a while now, and they didn't take long. With those
distractions out of the way, now it's time to work on a Sieve configurator
in WebCit.
And yes, it was Matt's post to the Programming room that got me thinking
about pluggable mail filtering. Maildrop would be pretty cool, but
Milter would be downright excellent.
Ok, here's where we are right now and where I see us going for the
remainder of this year:
1. Finish all the Sieve stuff
* dothebart has to finish his 'managesieve' implementation
* I have to finish the WebCit rules editor
* After we complete these, we will
Anything which speaks SOAP or any DAV variant is going to have to speak to
WebCit. Once we decide on an XML parser, it will be linked into both the
Citadel server *and* WebCit.
I'm leaning towards Expat. NunniCMAX is nice and small, but it doesn't
support namespaces. libxml2 is bloated, and too closely associated with
the Bad People.
Access to OS X is somewhat limited for most of us. I'll try to provide
some insight but I'm afraid I probably won't have much to offer.
matt: I'm open to ideas, if we need to come up with some sort of
event-driven push. I think it might be a great idea to dovetail Sieve
into push email somehow.
Status update: as you've probably noticed, Uncensored is running
something called Citadel 6.90 (beta). This is basically SVN
oops :)
Uh oh.
Ubuntu has
done away with inittab, and now uses upstart to bring up the
system. However, there is still the /etc/rc[0-3S].d directories for
the runlevels and /etc/init.d for the init scripts, so installs that
expect to see a SysV init environment have no problem.
You know what this
Ok, that ought to do it, at least for systems with a SysV init style
directory layout. Let's get some testing in on this, and if it's good,
we'll do the same for the Citadel server.
The other thing we could do would be to detect host systems that have a
BSD style startup (FreeBSD,
btw, messages... i'd vote for marking messages in the 'sent items'
room as 'sent'. having another folder in webcit in highlight state
keeps anoying me. Plus i think at least after sending pages out
successfull in webcit these should be marked sent too imho.
It is rather
Currently, the manual is part of the software distribution, not part of
the web site. That's just because of the history of its development.
It's good because they get some docs when they download Citadel, but it's
bad because the manual isn't searchable from the web site. I'm open to
Just as an FYI ... even though the officially sponsored Bad People have
called it quits, there are still a number of community Bad People who
intend to continue their Citadel ripoff project. This means we *do* have
to work hard to get Citadel out in front of people, and continue to
Excellent. Thank you.
I was confused for a moment there, when fr.po gave me a bunch of invalid
multibyte sequence errors, but then I realized that you had saved it as
ISO-8859-1 instead of UTF-8. A quick run of iconv later, we were up and
running in French.
I did notice that I missed one string. The Add
Weird ... somehow it's already in there, even though I had forgotten to
localize the string. It's translated as Ajouter des usagers in French.
I had noticed that some strings were starting to appear even though I
didn't remember receiving translations for them. Is it possible that
matt: thanks for the excellent work -- it looks like stuff we REALLY could
make use of. But we are in feature freeze, so if I could ask you to hold on
to those patches for just another week or two I'd appreciate it. We'll get
all of your changes in immediately afterwards.
Do you guy's have contact with Omar at all? Is libical dead again? Is
there a new maintainer?
I've sent him a few emails but I seem to have lost touch with him.
He was given admin access to the SourceForge site, and we all expected
that he was going to upload his code there,
I'm inclined to go with IO's suggested action. setup does far more than
just write values to citadel.config. We've provided a number of ways to
feed information to it in batch mode and then run it quietly. This is the
method packagers should use. I can provide more specifics when you need
IG, can you point me to a list of new features in 7.00 (since, say,
6.00)?
6.xx was such a long cycle that it'll be a long list. 6.00 was released in
late 2002 and had a barely-working IMAP server.
So here are some highlights, in no particular order. Cull this list in
whatever way your
Let's plan for the 7.00 release to occur early next week. I'm actually
ready now, but I'd rather not have lots of people installing or updating
over the weekend, in case the need arises for last-minute fixes.
1. we don't close stin/err/out correctly i think, for that reason the
init script won't detach the console.
2. the pid?
Are your scripts calling ctdlsvc? ctdlsvc specifically addresses both of
those issues. It closes stdin/stdout/stderr so that you can detach -- I know
this works because
Commentary on comments (and code...) in routines2.c ...
/* This appears completely unnecessary; why was it here? -IO
r = CtdlIPCGetRoomNetworkConfig(ipc, listing, buf);
if (r / 100 == 1) {
while(listing strlen(listing)) {
Possibly.
Well, I also had to change a !strcmp() to a strcmp() in order to make it
work. So maybe that had something to do with it as well. It's probably
one of those little things that got b0rked when we moved to the new client
protocol API.
I tested with the Citadel Support room, because it's
7.00 is now on the wire. w00t
we have one final gap, the postinst script will block if it starts
citadel.
Ok, here's what we need to do:
Do the postinst, and get it to block.
While it's blocking, do a dump of citserver's open files, using either
lsof -p citserver_pid or cat /proc/citserver_pid/fd.
MATT: feature freeze is now over. Please re-submit any patches you've
got ready.
Thanks!
Have it scan for IP addresses in the previous init file (and/or
inittab), as well as other parameters, like -x, -t and -l.
Good idea ... we'll carry over as much as we can.
Are we going to roll out 7.03 with the client fix for my .ASG crash?
It's got what *may* be a fix for it, but I haven't been able to come up
with a definite reason for the crash.
Wireless in general is interesting. I often wonder whether the text-mode
BBS client could end up finding new uses as a minimalist UI for things
like mobile devices (or even ham radio terminals, as was discussed in
another room).
I think we're *just* starting to see some of the buzz that
A little troubleshooting last night with a Citadel user running Gentoo
revealed that our PAM configuration file was a little less than portable:
#%PAM-1.0
#
auth required /lib/security/pam_listfile.so item=user sense=deny
file=/etc/ftpusers onerr=succeed
auth required
Yes, I'm going to implement the IMAP ACL extension. We're going to need
it in order to support any of the Outlook connector products that we might
have the opportunity to use in the future.
I've just about run out of patience with Bynari. Their connector would
have been easy to
Odd. Right after we've discussed connectors, I received an unsolicited
email from Hyun Kim, the president of Bynari. She said that the
iCal/vCard connector is delayed because they signed a very large deal with
a partner who needed support for Kolab, so they added XML and
annnotatemore to the
Well, I just got off the phone with the folks at Bynari. Here are some
highlights from the conversation:
* iCal/vCard support in their Outlook connector is mostly completed.
They believe they can offer us a beta test by February or March.
* They are still willing to provide a
thierry: when I tried to apply your patch, it seemed to think that your
patch had already been applied. Have you send this to me before?
thierry: even the new version of your patch is mostly made up of changes
that have already been applied (except for the room banner update, of
course).
I want to apply your changes, but you have to post patches against the
current svn. Please update your tree to the latest svn, then
Il don't undersand, i applied this last patch on the current svn and
it's work
Ok, I don't know what I was doing wrong, but I tried again and it worked.
Strange.
I've found the cause of the slow TCP IPC connections bug on OS X,
namely buffering output on the socket will cause the system to wait
too long to spit out the results. Changes: * Check for
matt: I've applied this patch, but I had to make a change, otherwise it
broke the
thierry: I like your plan for a theme engine. I like it a lot, in fact.
What do we need to do first?
Okay.
I continue to move what which is hardcoded. Then we will see for
themes.
thierry: your latest patch has been applied. Would you like to have
write access to svn in order to continue the process without having to
submit patches?
I've released the Citadel email connector for Funambol:
http://bionicmessage.net/?q=node/11 This is a custom-built connector
which interacts much more nicely with Citadel (using the native
Citadel protocol)
That's *great* news!
I wish I had a mobile device that worked with Funambol so I
We're going to use Expat.
Re-bootstrap your build and start over ... the build has changed a bit and
you've probably got an old configure script.
Citadel now talks to Funambol, Funambol contacts phone, phone
downloads email, everyones happy :) Attached patches implement the
code required to tell Funambol we've got mail, add the config options
necessary, and add a new pane to the Webcit site config dialog.
Ok, I like the
Ok, if there are follow-up patches on the way, then I'll apply the first
patch set when I have some hacking time tomorrow.
Freakdog: yes, that's exactly what it's for. Citadel will be one of the
first (if not *the* first) open source groupware servers to do push.
I'll submit a second round to add the settings to the text client
and vandelay, as for queuing ping events, copying each message marked
for notification to a spool out room and doing it SMTP queue style
should do it..
Hey matt, I just thought of something that could make this
hm, i keep on getting these warnings over and over.
Ig, could you have a look at that?
I checked out all of them (except for the ones inside zlib), and changed
some code to initialize more of the variables during declaration. But I
don't think it'll make Valgrind any quieter.
I have the same error as mentioned earlier when trying to use apt.
(BTW: After we get this fixed , add a note after editing sources.list
to apt-get update)
The links for apt are somehow circular.
Fixing the packages must be an immediate priority.
If people are
IG, can you repatch msgbase.c and sysconfig.h using
revised_push_patch.diff. The text client patch in routines2.c doesn't
matt: I don't seem to have a revised_push_patch.diff. Can you post it
please?
Thanks. It's been applied. There were some compiler warnings which I
silenced, but there are a few left you might want to have a look at.
Hey everyone ... just wanted to drop a note to say that I've been somewhat
scarce over the last couple of days because of a big rush project at work.
Things should calm down by later this week and I'll jump back into
Citadel stuff. Thanks for your patience.
Actually, that makes perfect sense. The call to Nifty Corners is made at
the end of a page load. Since the PeriodicalUpdater is not a page load,
the call isn't made.
thierry, you're an excellent web developer. It's such a pleasure to have
you working with us.
Not familiar with firebug, but I'll check it out. Thanks!
In case anyone is interested in why Citadel crashed a couple of times
on Thursday/Friday after a user named bandulik logged in several hundred
times ... it was, quite simply, that we ran out of available file
descriptors.
The new feature I checked in tonight is intended to handle a type of spam
attack that we've been suffering from over the last month or two. The
spammer uses a botnet to bombard our SMTP port with about 50 simultaneous
connections from a lot of different addresses. I don't know why they
think
h0h0h0 ... it's working already.
2007/02/04 7:44:13.612356 [ 23] Session started.
2007/02/04 7:44:13.760662 [ 23] RBL: Blocked - see
http://www.spamcop.net/bl.shtml?195.131.139.216
2007/02/04 7:44:14.166325 Purging session 23
2007/02/04 7:44:14.166486 RemoveContext() session 23
2007/02/04
By the way ... the qpdecoder still isn't right. It's inserting spaces
where there are linebreaks.
During a meeting with Ragnar (the Citadel project's QA lead), he pointed
out to me that there is currently a rapid surge of interest in VMware
Virtual Appliances. Read more at http://www.vmware.com/vmtn/appliances/
Basically, you build a VM containing a bare-bones version of Linux, plus
Most of the appliance images are about that size.
One thing to remember is that this is primarily a marketing tool. VMware
already has the infrastructure in place to promote our appliance. If you
look at their marketplace you can see that it's growing quickly. There
are a lot of
Ok folks, this is important.
I need *everyone* to submit a Citadel-related story to Slashdot, either
very late this evening or tomorrow morning.
It can be about anything as long as it has to do with Citadel. Something
about the new Virtual Appliance would be great, or perhaps the
Hmmm ... I wonder if there would be any value in allowing Citadel to be an
OpenID provider, even though our multiprotocol design prevents us from
being able to have Citadel be an OpenID consumer?
101 - 200 of 1678 matches
Mail list logo
