[Citadel Development] (no subject)
hm. somebody told me that our mamboo version is vulnerable for remote php code injection... ig, you should update that.
[Citadel Development] Re:Would like to port Citadel to the NSLU2 - ignore blank mesg
Dec 16 2005 4:41pm von michaelebrown Betreff: Would like to port Citadel to the NSLU2 - ignore blank mesg Anyone, I'd like to port Citadel to the NSLU2, but would like some guidance as how to start.In general i would say, no real 'porting' is needed, because of citadel runs for example on Sparc architecture too.You can read more about the NSLU2 at my website http://bytered.com/siteinfo.html And,at the NSLU2 Wiki site http://www.nslu2-linux.org/ The NSLU2 is a very small single board computer, 133MHz and 32MB of memory, sold by LinkSys as a network storage device. It can be overclockedto 266MHz, and you can also (if you're good at soldering) upgrade memory. The NSLU2 has been hacked for over a year now, and a huge number of applications have been ported to the NSLU2, and is soon(hopefully) to have a Debian OS port. My contributions have centered around the existing thttpd installation and configuration, and native development on the NSLU2.I personally like the openwrt packaging for embedded system. did you have a look at that? it aims to be debian alike. btw, The debian scripts you can find in the debian/ dirs are my work. I'm now shifting my attentionto mail, webmail, and other forms of community collaboration. I've been reviewing Postfix, qmail, Sendmail, Exim as potential SMTP agents, and I came across Citadel on WikiPedia. It's much more than Iwas searching for, but seems to solve mail transfer as well many other items on my shopping list. Most of the open solutions available entail a very long list of required supporting software. Forexample, before you can use SquirrelMail, you need to install Apache, Php, Sendmail (or other), and the list goes on... ending up with a huge list of processes. Hence, Citadel...Taken from the top of my athlon:The Citadel server itself: (data storage, services smtp, pop3, imap, citadel 25496 citadel 18 0 13852 4596 2784 S 0.0 0.9 3:44.34 citserver tcp 0 0 0.0.0.0:993 0.0.0.0:* ; LISTEN 25496/citserver tcp 0 0 0.0.0.0:995 bsp; 0.0.0.0:* LISTEN 25496/citserver tcp 0 0 0.0.0.0:587 0.0.0.0:* LISTEN 25496/citserver tcp 0 0 0.0.0.0:110 0.0.0.0:* LISTEN 25496/citserver tcp 0 0 0.0.0.0:143 0.0.0.0:* LISTEN 25496/citserver tcp 0 0 0.0.0.0:465 0.0.0.0:* LISTEN 25496/citserver tcp 0 0 0.0.0.0:504 0.0.0.0:* LISTEN 25496/citserver tcp 0 0 0.0.0.0:25 0.0.0.0:* LISTEN 25496/citserver Webcit: Provides a webserver, so you probably wouldn't need thttpd anymore. primary issue would be embedding the management scripts here. 5363 root 15 0 18760 2388 1396 S 0.0 0.5 0:00.10 webserverHere's a few questionsthat I'd appreciate some help with... The Citadel website mentioned a need for a minimum of 64MB for the software. How serious is this need? Can Citadel be configured to have less impact on memory?(Note: Some people have actually gotten Apache to run on the NSLU2, but I hear lots of concern of performance (totally expected), which is why I stayed with thttpd and it's small footprint.)citadel uses the Sleepy Cat database engine as backend storage. so it' memory hunger will depend on the number of users and data stored by them. TheNSLU2 does not have a video controller, and as such, is typically used in a server capacity. It appears that to me that Citadel could run on the NSLU2, and I would use the web interface fromFireFox on my PC to access Citadel on the NSLU2. Is this correct?yep. almost any configuration (aside of that needed to get the services up) may be changed from within webcit.Are there ways to build Citadel incrementally? And therefore provide a more usable footprint for this small machine. For example,only include the mail and chat facilities... ?you can abbandon calendaring features by not providing ical. allso you could allso drop ldap support, and localization to safe space. Many of the people that are now setting up their NSLU2s are more inexperienced than before, and they are severely confounded by an email installation.Citadel appears to be *much* easier to install and maintain. Am I correct in that? Does anyone have real-life experiences to share?yep. you can default configure citadel by environment variables to the setup program, so you can use values for the situation in your box apropriate, and do all the other stuff, domain name etc. from the web frontend.I couldn't find the Citadel sources, have they been moved? Arethere detailed instructions on how to checkout the source using Eclipse?svn is used to manage the citadel sources. see www.citadel.org the download section.Thanks to everyone for taking time to provide me with feedback. P.S. Citadel looks awesome!it is ;-)P.P.S. Anyoneknow of a challenging job in the northeast Atlanta area? I'm in the job market. Mike http://bytered.com/-- Imagination is more important than knowledge. Knowledge is limited.Imagination encircles the world. - Albert Einstein
[Citadel Development] (no subject)
i've added and commited the debian packaging scripts to rss2ctdl. i've allso added a shellscript to add another config to a system using the debian/citadel-rss2ctld.cron.d cronscript.
[Citadel Development] (no subject)
ah, and if everybody else could give me his (in case fresh created) sf.net account, i'll add it.i think it looks more representative if there are more members. And as the main reason for this action was getting attention we didn't get so far , please register. The more publicity we reach, the more interesting developers might stumble over citadel ;)
[Citadel Development] (no subject)
oO(one of ig's favourite hobbies... locking up CVS ;-) [EMAIL PROTECTED]:~/src/citcvs/webcit-6.21/intl$ cvs update -d cvs update: Updating . cvs update: failed to create lock directory for `/appl/cvs/webcit/intl' (/var/lock/cvs/webcit/intl/#cvs. lock): Permission denied cvs update: failed to obtain dir lock in repository `/appl/cvs/webcit/intl' cvs [update aborted]: read lock failed - giving up
[Citadel Development] (no subject)
ugh. something is broken with ssl in sid. Generating a certificate signing request. Program received signal SIGSEGV, Segmentation fault. [Switching to Thread -1211132224 (LWP 16371)] 0xb7ee7692 in BN_num_bits () from /usr/lib/i686/cmov/libcrypto.so.0.9.7 (gdb) bt #0 0xb7ee7692 in BN_num_bits () from /usr/lib/i686/cmov/libcrypto.so.0.9.7 #1 0x0008 in ?? () #2 0xb7f9afe4 in ?? () from /usr/lib/i686/cmov/libcrypto.so.0.9.7 #3 0x0811cdd8 in ?? () #4 0xb7f32539 in X509_ATTRIBUTE_create () from /usr/lib/i686/cmov/libcrypto.so.0.9.7 #5 0x0811cdd8 in ?? () #6 0xb7d650a4 in free () from /lib/tls/libc.so.6 Previous frame inner to this frame (corrupt stack?) (gdb)
[Citadel Development] (no subject)
ok, ig, this works. on index view mutt shows the sender of the message, after i've viewed it, anymore. shouldn't there be any further actions taken? i.e. mail to aide? or better an instant message (instead of the actual mail?)? besides, how do i best recover my database inconsistency? This starts struggling me as I rely on my mails :[ Cheers, Willi
[Citadel Development] Re:Cit livecd
as you notice you're using rpms, do you do a regular citadel installation, or do you use my aproach to lhfs'ifie citadel? if not, would you like to do so, and would you like to put up an apt/rpm repository? Will you share your specfiles? Would you attent citadel packages room? ;)
[Citadel Development] (no subject)
why not add a hidden post field to the editor tieing the message to a room?webcit could check if it is the same room as it was, and if not, zap into it, post the message and zap back...Allso a post field with the actual room-parameters would be great...that if you're viewing message 20 to 40 before entering the message the listview jumps back to that focus again.having to scroll back down, and switching back to the last view is anoying.the anoyance could be reduced by a similar dropdown at the top of the window.Willi
[Citadel Development] (no subject)
hy, i've tried building packages from latest svn, and it seems as if it doesn't build chkpwd anymore? the binary is missing from installation paths...is it still needed? i did make chkpwd, which started building it, but failed with a compie error: make chkpwd i486-linux-gnu-gcc -I/usr/sbin -Wall -g -O2 -Wall -Wcast-qual -Wcast-align -Wstrict-prototypes -D_REENTRANT -pthread -I. -DHAVE_ CONFIG_H -c chkpwd.c -o chkpwd.o i486-linux-gnu-gcc -I/usr/sbin -Wall -g -O2 -Wall -Wcast-qual -Wcast-align -Wstrict-prototypes -D_REENTRANT -pthread -I. -DHAVE_ CONFIG_H -c auth.c -o auth.o i486-linux-gnu-gcc chkpwd.o auth.o config.o -L/usr/sbin -o chkpwd auth.o: In function `validpw': citadel/auth.c:132: undefined reference to `crypt' collect2: ld returned 1 exit status make: *** [chkpwd] Error 1
[Citadel Development] Re:
Nov 1 2005 9:21am from [EMAIL PROTECTED] hy, i\'ve tried building packages from latest svn, and it seems as if ^^^ Did I miss something? When did we get to Disneyland?here near coogne we prefer the 'phantasialand' it's closer, not that expansive and you have some serious action in there.
[Citadel Development] Re:
i'm missing the admin-name in in the environment docs...and DISABLE_OTHER_MTA should better be ACT_AS_MTA or that should be there too, and make it use port 25 or not...
[Citadel Development] (no subject)
so, after some party ;-) here's the backtrace:0xb7d0e4c0 in strcasecmp () from /lib/tls/libc.so.6 (gdb) bt #0 0xb7d0e4c0 in strcasecmp () from /lib/tls/libc.so.6
#1 0x0806d698 in print_value_of (keyname=0xb71944c8 ROOMPIC) at subst.c:169 #2 0x0806d97c in do_template (templatename=0x80a446c) at subst.c:236 #3 0x080579e5
in embed_room_banner (got=0xb7196104 200 Lobby|0|0|0|2|0|0|0|0|0|0|0|0|, navbar_style=1) at roomops.c:368 #4 0x0804e8bc in output_headers (do_httpheaders=1, do_htmlhead=1, do_room_banner=1,
unset_cookies=0, suppress_check=0, cache=0) at webcit.c:416 #5 0x08053e1f in display_main_menu () at mainmenu.c:15 #6 0x0805245d in session_loop (req=0x81717f0)
at webcit.c:1493 #7 0x0804cbf9 in context_loop (sock=9) at context_loop.c:409 #8 0x0804bf87 in worker_entry () at webserver.c:561 #9 0xb7ddccfd in start_thread () from
/lib/tls/libpthread.so.0 #10 0xb7d70f1e in clone () from /lib/tls/libc.so.6 this looks a bit strange to me: (gdb) frame 1 #1 0x0806d698 in print_value_of
(keyname=0xb71944c8 ROOMPIC) at subst.c:169 169 if (!strcasecmp(ptr-wcs_key,
keyname)) { (gdb) print ptr $1 = (struct wcsubst *) 0x110 (gdb) (gdb) print ptr-wcs_key Cannot access memory at address 0x118 indeed.i'll
let the debugger in that state, if there are further questions.Willi
[Citadel Development] Re:
Nov 5 2005 10:58pm from IGnatius T [EMAIL PROTECTED] 0xb7d0e4c0 in strcasecmp () from /lib/tls/libc.so.6 Ok, I've found and replaced some more absolute url's. Try it again. By the way, it looks like this technique is *not* going to be usable for GroupDAV or for the mailing list subscription stuff. Both of these modules require absolute knowledge of their URL paths...wouldn't it be a suitable way to add a 'basepath-extension' to the config done by setup?that the links realy go to /webcit ?that way would allso work for groupdav too, am i right? 'caus most implementation offer just a base url...
[Citadel Development] (no subject)
should webcit segfault on wrong urls? imho it should give a 404 in that case.
[Citadel Development] (no subject)
whats about citadel.control? the package install has problems with it disappearing, and citadel won't create it on start...would it pherhaps be better to create it if it doesn't exist on startup?also it tries to bind port 25 though i've configured act_as_mta=noNov 7 09:18:21 [citadel] Cannot open citadel.control: No such file or directory_ Nov 7 09:19:05 [citadel] citserver: Can't bind: Address already in use_ Nov 7 09:19:05 [citadel] TCP port 25: FAILED._
[Citadel Development] (no subject)
the new maildir view is cute.one thing that's not good is that the cursor changes to the 'place textcursor' cursor from the pointer.(the I-cursor)
[Citadel Development] (no subject)
another error:[Mon Nov 07 13:50:17 2005] [warn] proxy: Ignoring bogus HTTP header returned by /webcit/editpic (POST) appeared after i've uploaded my picture into my bio.
[Citadel Development] (no subject)
ig: one link is still absolute: the who is online icon:[Sat Nov 12 19:17:45 2005] [error] [client 84.175.98.200] File does not exist: /var/www/who_inner_html
[Citadel Development] (no subject)
i've researched the problem with the non existing citadel.control file.it is because of in serv_extensions.c initialize_server_extensions serv_inetcfg_init is called before serv_upgrade_init.serv_inetcfg_init calls get_mm, which is only enabled reading the control file, while serv_upgrade_init calls get_control, which creates the file allso.i've tried swapping them,seems to work, and control file is created.
[Citadel Development] (no subject)
as patch... patch Description: Binary data
[Citadel Development] (no subject)
whats the advanced-Network Config - Network run frequency (in seconds)?the Poll intervall of the nodes?(for the translation, sometimes one needs to understand ;-)
[Citadel Development] (no subject)
ig, there are still some problems with the who is online page. in suburl mode one needs to weaken up cookie policy. is there a meta refresh inside pointing directly to the webcit server?sometimes it gets to really funky happenings, suddenly the citadel login screen appears... ;) inside the center frame. The other stuff seems to work, the remote ip is shown.but what if something like this happenes? : X-Forwarded-For: 192.168.1.1, 192.168.2.1 i don't know how this exactly happenes... but just the first ip is shown...Viva Netcat ;-)
[Citadel Development] (no subject)
ig, please lets look to get it working, add missing strings afterwards, then let me give it s.b. to spellcheck it, and then release it...Allso i think i need to review it in a working environment, if it's consistent, as i don't allways know in which way the strings get glued together in a reallife environment.
[Citadel Development] Re:
Nov 15 2005 2:07am from IGnatius T [EMAIL PROTECTED] dothebart: sorry to say this, but I had to tend to an emergency at work tonight and didn't get to touch Citadel at all. (It's 2:00am here on the east coast of the US and I just got home a few minutes ago.) This means that putting the German translation online got pushed back another day or two. I promise I will make it a priority.np, it took me over a month due to being short of time too. I'm one of those firemen who solves urgent customer problems asap too ;)
[Citadel Development] (no subject)
this character issue is realy strange. I've put together some simple html file, read it from disk, and opened it in firefox. jam. added the charset header, umlauts.Added charset header to webcit, still jam.added an a uml ; to the translation file, the right umlaut is shown next to the others.strange things happen. perhaps one needs realy tell the localization which charset one is using?
[Citadel Development] (no subject)
there seems to be some js error.the message selector doesn't work anymore:Error: msgomatic is not defined
[Citadel Development] specs...
Hy, I was wondering who did the rpm spec-files? I'd like to start over with that too.I think the .debs stabilize in feature completeness, and would like to see something similar for RPM based Distros too. I've contacted a Fedora Person who went over my way ;) and i'll start to get connected on that case soon.
[Citadel Development] (no subject)
I've added a gettext.c file, and 5 lines to context_loop.context_loop() scans now for the Accept-Language Header too, and if detected calls httplang_to_locale() which tries to detect the locale, and if it has a matching locale, set it apropriately.Though there are some things not yet ok for me, * !HELP! There seems to be a race condition about the malloc() i've used. is this wrong in that place? what else has to be done? (it appears only if another box requests the webcit page, not if you request it via loopback)* the available locales are not built from the files under po/; but a list in gettext.c* the weights aren't looked at at all.* I think there should be an en_US locale, that just consists of just the same strings as the c-files* shouldn't date and time settings (am,pm/0-24h) be picked there from too? (allso as the number formats?)
[Citadel Development] (no subject)
ready to test it ;) but i would need a smart way to migrate an existing moinmoin ;-) i think, we should have something of perl-migrating script for that. A perl implementation of the citadel api would be a cool thing, because of you could bind that up in irssi for chat-based allerting too ;)
[Citadel Development] (no subject)
hm, the error pages should allso have the expiry date of your birth, ig. i went into my citadel using one of the firefox history lines, which pointed me to nkrooms, which (as ok) gave me an error message.i hit 'sign-on' and logged in, clicket rooms icon, and was back out again...
[Citadel Development] (no subject)
did we yet try newsforge.com for press releases?
[Citadel Development] (no subject)
hm, should the sequence of the dialog be in german order too? dunno... usualy its day. month. year...so the selectors should have that sequence too?
[Citadel Development] (no subject)
ig, i've updated the po, and committed it, but no success... the newthinkers are moaning webcit is not colorfull enough ;-) I think one can change that in the stylesheet?
[Citadel Development] (no subject)
hmnz. that uml graph view is only available in c++ code with stacked objects :( a call tree would have been that neat :(
[Citadel Development] (no subject)
hmz, i can't seem to get kontact to cooperate with actual SVN.i enter the account, and the username and press 'update folder list', and webcit shows the following in the access log:2006/01/30 0:02:44.643836 [ 16] LKRA2006/01/30 0:02:44.706794 [ 16] GOTO Calendar2006/01/30 0:02:44.707365 [ 16] 08.Calendar 1 new of 1 total messages2006/01/30 0:02:44.710664 [ 16] MSGS ALL2006/01/30 0:02:44.713925 [ 16] MSG0 21|32006/01/30 0:02:44.713985 [ 16] CtdlOutputMsg() msgnum=21, mode=0, section=2006/01/30 0:02:44.714016 [ 16] CtdlFetchMessage(21, 0)2006/01/30 0:02:44.714123 [ 16] CtdlOutputPreLoadedMsg(TheMessage=not null, 0, 3, 1, 02006/01/30 0:02:44.721064 [ 16] GOTO Global Address Book2006/01/30 0:02:44.721977 [ 16] Global Address Book 0 new of 0 total messages2006/01/30 0:02:44.725079 [ 16] MSGS ALL2006/01/30 0:02:44.731218 [ 16] GOTO Contacts2006/01/30 0:02:44.731750 [ 16] 08.Contacts 0 new of 1 total messages2006/01/30 0:02:44.735078 [ 16] MSGS ALL2006/01/30 0:02:44.738348 [ 16] MSG0 20|32006/01/30 0:02:44.738411 [ 16] CtdlOutputMsg() msgnum=20, mode=0, section=2006/01/30 0:02:44.738441 [ 16] CtdlFetchMessage(20, 0)2006/01/30 0:02:44.738544 [ 16] CtdlOutputPreLoadedMsg(TheMessage=not null, 0, 3, 1, 02006/01/30 0:02:44.754555 [ 16] GOTO Tasks2006/01/30 0:02:44.755091 [ 16] 08.Tasks 1 new of 1 total messages2006/01/30 0:02:44.758354 [ 16] MSGS ALL2006/01/30 0:02:44.761612 [ 16] MSG0 19|32006/01/30 0:02:44.761673 [ 16] CtdlOutputMsg() msgnum=19, mode=0, section=2006/01/30 0:02:44.761703 [ 16] CtdlFetchMessage(19, 0)2006/01/30 0:02:44.761805 [ 16] CtdlOutputPreLoadedMsg(TheMessage=not null, 0, 3, 1, 0but the list of the available folders stays empty.it's a sid, KontactVersion 1.1.2I think i had a working feedback at the start of january...
[Citadel Development] (no subject)
the actual installation: [EMAIL PROTECTED]:~# netstat -lnp |grep webservertcp0 0 0.0.0.0:20000.0.0.0:* LISTEN 13218/webserver root 13218 0.0 0.7 20828 3672 ?Ss Jan26 0:05 /usr/sbin/webserver -f -p2000 uds /var/run/citadel/ 127.0.0.1 ssl.conf Description: Binary data
[Citadel Development] (no subject)
PROPFIND /groupdav/ HTTP/1.1 Host: 127.0.0.1:2000 User-Agent: Mozilla/5.0 (compatible; Konqueror/3.4; Linux) KHTML/3.4.3 (like Gecko) (Debian package 4:3.4.3-2) Pragma: no-cache Cache-control: no-cache Accept: text/html, image/jpeg, image/png, text/*, image/*, */* Accept-Encoding: x-gzip, x-deflate, gzip, deflate Accept-Charset: iso-8859-1, utf-8;q=0.5, *;q=0.5 Accept-Language: en Authorization: Basic d2dvOm9obXlnOHQ= Depth: 1 Content-Type: text/xml; charset=utf-8 Content-Length: 142 Max-Forwards: 10 Via: 1.1 tucholsky.ty.newthinking.de X-Forwarded-For: 192.168.2.234 X-Forwarded-Host: tucholsky.ty.newthinking.de X-Forwarded-Server: tucholsky.ty.newthinking.de
[Citadel Development] (no subject)
once again after tiny mce is booted a paste with newlines: PROPFIND /groupdav/ HTTP/1.1Host: 127.0.0.1:2000User-Agent: Mozilla/5.0 (compatible; Konqueror/3.4; Linux) KHTML/3.4.3 (like Gecko) (Debian package 4:3.4.3-2)Pragma: no-cacheCache-control: no-cacheAccept: text/html, image/jpeg, image/png, text/*, image/*, */*Accept-Encoding: x-gzip, x-deflate, gzip, deflateAccept-Charset: iso-8859-1, utf-8;q=0.5, *;q=0.5Accept-Language: enAuthorization: Basic d2dvOm9obXlnOHQ=Depth: 1Content-Type: text/xml; charset=utf-8Content-Length: 142Max-Forwards: 10Via: 1.1 tucholsky.ty.newthinking.deX-Forwarded-For: 192.168.2.234X-Forwarded-Host: tucholsky.ty.newthinking.deX-Forwarded-Server: tucholsky.ty.newthinking.de
[Citadel Development] (no subject)
somehow there seems to be a verry smart way to shoot yourself in the foot with kontact.for some reason it selected two calendars at one resource.then it would create the dates in one resource and display anothe resource.you can add your event, it gets displayed, kontact syncs it up to the server, and reads back the primary room, and *zap* no date in there, your date goes away!THIS IS AWFULL!
[Citadel Development] (no subject)
i've added some code to do the brief calview. I've still got some problems with the colors of the table. please tell me if you like it that way as it is so far.
[Citadel Development] (no subject)
depends. whether you configure it with --auto-login afair.
[Citadel Development] (no subject)
hm, is that usual, that if you add a second calendar resource to kontact on the same citadel, and press folder update, it gets all folders double?but, ig, that change did the job for me. but first after i've removed my .kde directory. i doubt, that this will be a viable procedure for all others :(I'll have a look at that, when im over in berlin again next week.
[Citadel Development] (no subject)
hmz, that room with those dates imported from the zidestore via kontact keeps messing up webcit.is there away like ARTV EXPORT to dump one room? so i can give you that for debugging?actually this room contains the dates of happenings over the last year, so it is frequently used...
[Citadel Development] (no subject)
ig, its the 'raumbelegung ty' room. it doesn't happen allways, but verry often. I could search for the .ics file we dumped in. should i?
[Citadel Development] (no subject)
ok, here something about the starttls stuff not working for thunderbird.It claims that citadel doesn't send the TLS heades in the ehlo response, which is correct: [EMAIL PROTECTED]:~$ telnet tucholsky 25Trying 192.168.2.135...Connected to tucholsky.ty.newthinking.de.Escape character is '^]'.220 newthinking.de ESMTP Citadel server ready.ehlo jordy.ty.newthinking.de250-Hello jordy.ty.newthinking.de (192.168.2.240 [192.168.2.240])250-HELP250-SIZE 10485760250-AUTH LOGIN PLAIN250-AUTH=LOGIN PLAIN250 ENHANCEDSTATUSCODEShere a tls enabled sendmal as a comparison: [EMAIL PROTECTED]:~$ telnet smtp.dinoex.de 25Trying 80.237.237.4...Connected to smtp.dinoex.de.Escape character is '^]'.220 k1.dinoex.de ESMTP Sendmail 8.13.5/8.13.5; Wed, 1 Mar 2006 22:15:48 +0100 (CET); (No UCE/UBE) logging access from: [62.176.139.220](FAIL)[EMAIL PROTECTED]ehlo localhost250-k1.dinoex.de Hello [EMAIL PROTECTED], pleased to meet you250-ENHANCEDSTATUSCODES250-PIPELINING250-8BITMIME250-SIZE 4000250-DSN250-AUTH PLAIN LOGIN CRAM-MD5250-STARTTLS250-DELIVERBY250 HELPIn fact, do we offer TLS for clients? or should that be added?
[Citadel Development] (no subject)
ok, there is a way to build webcit on freebsd again:./configure --disable-nlsafter you've installed svn from ports, all other stuff needed will be there.though we still have a serious problem getting the internationalization work:it will search for locale_t, which on linux the glibc brings to you.We could use some help of more bsd'ish people to solve this issue.which headers of the system should provide this type? on linux it's /usr/inculde/locale.h, its glibc.i've found similar looking stuff in /usr/src/contrib/libstdc++ on freebsd, but i must admit, i'm not that familliar with all this bsdish mechanisms, i don't know how to install it ;-)saadly you have to bootstrap webcit on linux, because of the ports bring a verry poorly done installation of the autofoo shit, that falls appart on many points, regarding libtool, aclocal, autoreconf...I've made a monkey out of myself in #bsdforen.de trying to get some help on that, but instead of trying to get that propperly configured and installed on the system, i've experienced them spitting at these tools.My personal opinion is, that the bsd ports don't install it properly, and if the port maintainer would fix those issues, there would be much less trouble. just my 2ct.
[Citadel Development] (no subject)
hm, the autofoo is suggested to produce this file automagically from POTFILES.in i'll add this to svn.this will become an issue the next time we add a .c file to the process, either one has to modify both files by hand, or add it to the first, and use a newer gettext macro version to let it be made automaticI've added it, and now that configure runs nice again on uncensored (tested)I allso hope, that the numbers of issues arrizing decrease soon with that.in doubt we should perhaps put the things made by autogen.sh into svn too? this seems to be nessescarry for some systems...so all the Makefile.in/am etc...Do 13 Apr 2006 23:08:55 EDT von IGnatius T [EMAIL PROTECTED] mv: cannot stat `po/POTFILES': No such file or directory
[Citadel Development] (no subject)
in case of the startoff dialog, this must be handled by webcit. or an entry into the sent items with the subject Failure: failuremessage and the message text as usual could do the job. but loosing the message gives a bad feeling to the user.
[Citadel Development] (no subject)
Sun 23 Apr 2006 09:23:01 PM EDT from IGnatius T [EMAIL PROTECTED] Better spam filtering would be helpful, but we have to assume that no filter will ever be perfect. It's probably time to implement access controls on mail to rooms. I think we can safely assume that the following are always allowed: 1. Mail from non-SMTP sources (such as other Citadel nodes on the network) 2. Anything from an authenticated SMTP connection lmtp.unfiltered? And it's probably safe to accept mail from: 3. Someone whose address is subscribed to a mailing list room 4. Someone whose address is subscribed to a mailing list room in digest form So all we have to add is: 5. One extra type of listrecp that doesn't actually get anything delivered from them -- it merely identifies their email address to Citadel as someone who is allowed to post to a room via SMTP And possibly: 6. A global list of addresses (such as fleeb's cell phone) that can post to *any* room via SMTP. and there allso needs to be a way to have the old behaviour, if you want to use a room as 'shared inbox' for example stuff like service@ where people can publicaly mail to, and a team of users processes the mail. though this needs to be spam filtered too.
[Citadel Development] (no subject)
and there are three other things stretching my nerves, they are more tiny-mce related. * it's nice to have the original message indented, but tiny mce doesn't allow me (or i'm not clever enough) to interrupt the message, to add my own comments right the place where they are related to, so one mustn't read the mail all over (see my posting above where i've changed my text to bold, this should have zero indent) .* tiny mce takes to long to start. if i enter a text, and press backspace, firefox goes back to the room (which usually then has no unread messages)*if i remember not to press backspace while loading it, the curser ends up on position (0,0) even if i've typed some text before the js took over the textbox, it will remain at the start of the textbox, and i'll have to use the mouse to point it at the end. (i think the textbox sometimes doesn't even have the focus)are there some smart ways to preload it somehow? perhaps this could eliminate the later two issues...
[Citadel Development] (no subject)
hm, i think the user will confuse, if 'aliases' for i.e. rooms etc. show up as global vcard. it will allso trashn the global vcard for clients like kontact. On the other hand, using this mechanisms (that already exist) to make several aliases point to one destination is a smart idea.Perhaps a new room, that works as the global vcard room, but isn't visible to users and groupdav/imap? Access to this should be done in a stripped down vcard dialog somewhere in the config section... Should just aides be allowed to create global aliases, that don't point to a user account?and once you start doing this, can you pay attention to refusing to create name clashes on creating aliases?so that the user gets feedback ala 'mail to $foo gets already delivered to $bar' ?and once you have that, could you add this functionality to citmail? so that one can retrieve the user/room a specific mail would be delivered to?
[Citadel Development] (no subject)
hm, is there a way to make citadel have a default charset for mail import, if the headers just say 8 bit, and nothing more?the headers look like that:Content-Disposition: inline Content-Transfer-Encoding: 8bit Content-Type: text/plain MIME-Version: 1.0 Precedence: bulk X-Mailer: MIME::Lite 3.01 (F2.71; A1.58; B2.12; Q2.03) and i think the usual default for me would be to treat it as iso 8859-1, which it should be for all mail with this kind of headers.its just right for me, for the geographical position of my installation, so i think it should be configurable somewhere sitewide...or did i miss something? i've changed my personal charset to that, no change.
[Citadel Development] (no subject)
hm, i've added a domain 'foo3' next to foo under advanced-configure domain and internet mail- upper left dialog box.my setup is a postfix frontend, that sends mail over via lmtp.sock.i've added a user test with matching 'aliases' in his vcard in the second domain. now i send mail to [EMAIL PROTECTED], postfix will pass it through, citadel delivers it, everybody happy.then i send a mail to [EMAIL PROTECTED], postfix will allso pass it through, citadel accepts it, says it can't deliver it localy, and retransmit it to postfix, that again delivers it to citadels lmtp.sock that transmits it over to postfix again...same happenes if i try to send a mail from webcit local to [EMAIL PROTECTED]I've shut down citadel server inbetween with 'sendcommand down' and restarted it, no change.what did i miss?
[Citadel Development] (no subject)
if you're more into the text-client you could teach it to view all the vcard/ical stuff in a proper manner. this aplies to calendar/adressbook/todo lists too. perhaps somethin like the bsdmainutils:cal view would be cool for that i think. it hilights the actual day on more modern debian systems. doing this with the commandline client would be neat!speaking about the auto purger... it has aten our citadel documentation room too. this shouldn't happen. there should be an archiving method for public room.Disk space isn't that expansive anymore than 10 years ago. pipe it through bzip2, put it into purged/roomname/datebz2, at least for public rooms. we do need a history. else we're like that guy in the william gibbson book, that drops all events longer ago than five minutes. was it mona lisa overdrive? biochips?
[Citadel Development] (no subject)
hm, this is strange, i can only see it call it too.
[Citadel Development] (no subject)
how about using colinux images for windows, or vmware images for vmware player, to enable the average (p/b)auer-user to launch a preinstalled citadel version out of the box, to easy check it out? In that case the master user and his passvoid could be anounced on the citadel login screen, so we perhaps can more specificially advise people how to watch it starting because of we know exactly which os it is?* bauer is the german word for farmer ;-)
[Citadel Development] postfix user mapping
i've digged a bit deeper into that problem, and consulting some people more closer to the source (experienced with postfix and their ml and so on) i was told, that there is a tcp - user dictionary, which would be the prefered way to make postfix backcheck if citadel knows the user. see http://www.postfix.org/tcp_table.5.htmlfor details. As the qdir command allmost does the job, i had a look at that.it is registered via the plugin api, that just specifies 4 character commands, so just hooking in another command with get is not possible. I've patched citadel.c to make it know the get cmd, which breaks well... clean code? it would do the job i think, but please take a look.
Index: citserver.c
===
--- citserver.c (revision 4564)
+++ citserver.c (working copy)
@@ -1309,6 +1309,14 @@
cmd_isme(cmdbuf[5]);
}
+ else if (!strncasecmp(cmdbuf, GET, 3)) {
+ // ok, as external plugin we don't have a chance,
+ // because of we don't have 4 chars. we hack this
+// to make it fit together.
+ extern void cmd_get(char *argbuf);
+ cmd_get(cmdbuf[4]);
+ }
+
else if (!DLoader_Exec_Cmd(cmdbuf)) {
cprintf(%d Unrecognized or unsupported command.\n,
ERROR + CMD_NOT_SUPPORTED);
Index: serv_vcard.c
===
--- serv_vcard.c (revision 4564)
+++ serv_vcard.c (working copy)
@@ -876,7 +876,27 @@
cprintf(%d %s\n, CIT_OK, citadel_addr);
}
+/*
+ * Query Directory, in fact an alias to match postfix tcp auth.
+ */
+void cmd_get(char *argbuf) {
+ char citadel_addr[256];
+ char internet_addr[256];
+ if (CtdlAccessCheck(ac_logged_in)) return;
+
+ extract_token(internet_addr, argbuf, 0, '|', sizeof internet_addr);
+
+ if (CtdlDirectoryLookup(citadel_addr, internet_addr, sizeof citadel_addr) != 0) {
+ cprintf(500 %d %s was not found.\n,
+ ERROR + NO_SUCH_USER, internet_addr);
+ return;
+ }
+
+ cprintf(200 %d %s\n, CIT_OK, citadel_addr);
+}
+
+
/*
* We don't know if the Contacts room exists so we just create it at login
*/
@@ -1112,6 +1132,7 @@
CtdlRegisterProtoHook(cmd_greg, GREG, Get registration info);
CtdlRegisterProtoHook(cmd_igab, IGAB,
Initialize Global Address Book);
+// CtdlRegisterProtoHook(cmd_get, GET , Alias for Query Directory; Postfix tcptable compatibilicty version);
CtdlRegisterProtoHook(cmd_qdir, QDIR, Query Directory);
CtdlRegisterUserHook(vcard_newuser, EVT_NEWUSER);
CtdlRegisterUserHook(vcard_purge, EVT_PURGEUSER);
[Citadel Development] (no subject)
would you make this a compiletime, starttime, or webcit configurable option? btw, using usermaps with different targets could fork spam-default victim users directly to a spamassesin, even before citadel gains it, i think. The firewall issue is true, it would be better that way. where would one plug that in?
[Citadel Development] (no subject)
hm, i think, setup should ask / handle the /usr/sbin/sendmail issue... move the existing file to /usr/sbin/sendmail.distorg, and replace it with a link to citmail, and set the flags in an appropriate manner. for shure this needs to be disabled/selectable for setup calls from the distro packages.
[Citadel Development] postfix tcp dict
If you want to plug citadel behind a frontend postfix for whatever reason (smothter spamassesin integration, improved security (maybe postfix has been audited more often than citadel...? ))So far you would have run into a major drawback: how would postfix know, whether it should deliver a mail, or reject it due to being sent to a non existant user? At present times spammer tend to do dictionary attacks, and use wrong from addresses. If postfix accepts the mail, and gets a reject by postfix, it will try to notify the sender that it wasn't able to deliver it. in most cases, non existant senders, which will cause double-bounces filling up your aide folder.So here's the solution to the case:Postfix supports non local (unix users) delivery tables. You can do user lookup for example in a mysql table, or postgres, dbm, ldap and, since later versions of postfix (2.2 and later, earlier in patches, for example debian sarge; search for dict_tcp.so) via TCP lookup.Now citadel offers such a service. To configure it, follow these steps:Upgrade to latest citadel and webcit(6.84 and later).Configure Administration - Edit site-wide configuration - Network - Postfix TCP Dictionary Portto a port of your choice. Restart the citadel system to make it open the port using sendcommand down as root on your shell.To check back if it worked out, do netstat -lnp if citserver provides your port.Now postfix. Check if /etc/postfix/dynamicmaps.cf contains something like tcp /usr/lib/postfix/dict_tcp.sodict_tcp_openTake the appended main.cf, and replace 777 with your port (or add the matching lines to your config). Make postfix reload its config using postfix reload as root on your shell.Use tail -f /var/log/mail or tail -f /var/log/mail/current depending on your syslog facility to check the effort:Aug 27 23:29:00 [postfix/postfix-script] refreshing the Postfix mail systemAug 27 23:29:00 [postfix/master] reload configurationor if you do postfix stop; postfix start :Aug 28 22:32:30 [postfix/postfix-script] stopping the Postfix mail systemAug 28 22:32:30 [postfix/master] terminating on signal 15Aug 28 22:32:33 [postfix/postfix-script] starting the Postfix mail systemAug 28 22:32:33 [postfix/master] daemon started -- version 2.1.5and use another mail account to verify that it works:sending Mail to an existing user should look like that:Aug 29 00:34:41 [postfix/smtpd] connect from yourtestmta.org[testmta ip]Aug 29 00:34:41 [postfix/smtpd] E69F29ED1: client=yourtestmta.org[testmta ip]Aug 29 00:34:41 [postfix/cleanup] E69F29ED1: message-id=[EMAIL PROTECTED]Aug 29 00:34:41 [postfix/qmgr] E69F29ED1: from=, size=838, nrcpt=1 (queue active)Aug 29 00:34:41 [postfix/smtpd] disconnect from yourtestmta.org[yourtestmta ip]Aug 29 00:34:42 [citadel] 8137: from=, nrcpts=1, relay= [], stat=250 2.0.0 Message accepted.__Aug 29 00:34:42 [postfix/lmtp] E69F29ED1: to=, relay=/var/run/citadel/lmtp.socket[/var/run/citadel/lmtp.socket], delay=1, status=sent (250 2.0.0 Message accepted.)Aug 29 00:34:42 [postfix/qmgr] E69F29ED1: removedif you view citadel logging:Aug 29 00:34:32 [citadel] : get [EMAIL PROTECTED]Aug 29 00:34:32 [citadel] sending 200 OK [EMAIL PROTECTED]and the usual delivery and sorting afterwards.now the non match case:Aug 29 00:33:40 [postfix/smtpd] connect from unknown[testmta ip]Aug 29 00:33:41 [postfix/smtpd] NOQUEUE: reject: RCPT from unknown[testmta ip]: 550 : Recipient address rejected: User unknown i main.cf Description: Binary data
[Citadel Development] from:
i think we should have a way to preconfigure the sitewide from. there are many company defaults, so you should be able to say i want [EMAIL PROTECTED] or [EMAIL PROTECTED], and i think we should add the real name, as there exists a format for that to like CName Name [EMAIL PROTECTED] as a site wide preconfig. This all would just threatn the user to fill his vcard, which is a good thing [tm]
[Citadel Development] (no subject)
yep, forgot that :(and yes, i'd like to finish that. I'm just testing with the test sample of php-sieve.http://sieve-php.sourceforge.net/with the test.php i got the handshake working.I think i need some further assistance on the rest of these, but i'm willing to finish that.Pherhaps you could look at the unfinished code, and see if you get some major hickups with that...further i think, when the php-sieve thingie is at least working, we can start with the server implementation. when this work, we should try to replicate it's functionality with webcit.(though webcit wouldn't talk to citadel with the managesieve ;-)
[Citadel Development] (no subject)
IDEA! this mechanism could be used to have a maildir for a learning spam/bayanese/foo. who of you actually does stuff like that already? any ideas how this could be used for per user writing? could this be used to be plugged in front of the imap server code to write maildirs for spam tools?
[Citadel Development] tb groupdav...
hm, fails on me. maybe this is a problem with the https apache wrapper. Ig, can we have a solution for that? Something like a enable-able option to fixiate the groupdav url to an https host? attachment: groupdav_tb_error.jpg
[Citadel Development] (no subject)
Loanshark, you could probably use that for rss-bookmarks too. the rss headers just need to look a bit different.
[Citadel Development] (no subject)
if i would, would i waste my time? do i sound as if i would?
[Citadel Development] (no subject)
How about adding the article on citadel.org asking the users to give feedback to easyinstall as a mail to room_aide with a from to reply to citadel.org? there could be a link to uncensored too...
[Citadel Development] welcome mail...
I've created a welcome mail, that could be injected by easy install into the lobby. I think it would be nice to get a warm welcome with some information when one enters his fresh citadel. Welcome to your new Citadel installation! Congratulations! you've just completed your Citadel Installation. Your Citadel system provides you with the following Protocols to communicate with your Client Software: Webmail via http and https. All configuration Tasks you didn't do yet, will most probably be doable in Webcit. Imap Directory Services. Access your Inbox and other Rooms easily via Imap. POP3 Mail Delivery. If you'd like to poll mails over to your Client, use Pop3. GroupDAV Callendaring and Addressbook Access. Connect your Callendaring Client to a central Storage, and share callendars with others. ICal Access to your personal Callendar. SMTP / MSA access with TLS support. Receive mails from others or relay your Client via your Citadel. If you need advice howto configure your favorite Mailclient to interact with your Citadel, have a look at the ClientMatrix . If you don't find your client there, please sign into Uncensored! and tell us about it in the Citadel Documentation Room! If you like to provide screenshots on how you did the configuration for documentation purposes, we'll appreciate them. If the FAQ for more general purposes, or the Citadel Administration Manual , join us at Uncensored! and post your message into the Citadel Support Room, you will get help there. You might allso join us on IRC (irc.citadel.org, or freenode network) in #Citadel. If you 'd like to make Webcit speak your Tongue, and add a translation, tell us at Uncensored in the Citadel Development Room. If You like Citadel, Please let us know. Just reply to this email, or log into Uncensored and drop us a note! Feel free to join the Citadel Group! Your Citadel Group Read more about Citadel on Wikipedia
[Citadel Development] (no subject)
Here we are! http://citadel.org/index.php?option=com_contenttask=viewid=122Itemid=47clientmatrix incooperated into the citadel webpage.
[Citadel Development] (no subject)
looks as if it was time to add such a riddle picture to the login screen. these are citadel authenticated spams, right? or do they come in via the ml subscription interface?
[Citadel Development] (no subject)
perhaps we should have a closer look at how the squirrel mail sieve plugin works, and get some inspiration from it.
[Citadel Development] calendar buble leaves browserwindow
i've got a bug report regarding our calendar webcit overview. It looks as if sometimes the bubble-hints in the left column leave the screen.
[Citadel Development] spam capability...
i've found that in the debian build directory. perhaps this might help? i didn't do any implementation capabilities yet, but perhaps you might read more out of this mail with patches ig? sieve-spamasssassin Description: Binary data
[Citadel Development] (no subject)
ok, here they are. but, ig, the putscript callback doesn't do what i want it to do... but, perhaps i should have a closer look, or, can you give me a hint, ig?
[Citadel Development] (no subject)
regarding the pickyness of sieveshell, should i strip of traling white spaces and newlines on adding a script? as the c library uses a flex parser, i can't debug what it exactly complains about...
[Citadel Development] (no subject)
today i've gottn hold of two interesting firefox extensions to explore webinterfaces and doms etc. have a look at them:Analyze the http stuff, and manipulate headers...https://addons.mozillaorg/firefox/966/Dom and CSS Analysis:https://addons.mozilla.org/firefox/60/
[Citadel Development] (no subject)
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.
[Citadel Development] Re:SVN commit log: revision 4715
You should please do something like that in you citadel commandline client rewrite too, io-error...Wed 29 Nov 2006 05:35:20 PM EST from Citadel commit log in 00.Sent Items Subject: SVN commit log: revision 4715r4715 | dothebart | 2006-11-29 17:35:20 -0500 (Wed, 29 Nov 2006) | 2 linesChanged paths: M /trunk/webcit/messages.c* add x-mailer header for webcit.
[Citadel Development] (no subject)
i had a closer look at this file from apache:httpd-2.0.55$ less srclib/apr/threadproc/unix/procsup.cthe improvement is, that there is no more tty in ps -xjafwhich is one step forward in the right direction, but it won't make it with the postinst script yet :(
[Citadel Development] (no subject)
CitStartedFromConsoleWillExitAfterParsingCommandlineArgs |CitFirstForkedWillLoopArroundForkWaitPid |CitServerThatDoesTheJobAndMightCrasho k, i've found the issue. one had to finish off debconf before starting the deamon:http://www.calivia.com/bk/debian/debconfas formorer pointed me to. Many Thanks!I've attached my sysdep.c, that does all the forking now.Ig, Should i commit it that way and remove ctdlsvc?
/*
* $Id: sysdep.c 4763 2006-12-20 19:04:43Z ajc $
*
* Citadel system dependent stuff.
* See copyright.txt for copyright information.
*
* Here's where we (hopefully) have most parts of the Citadel server that
* would need to be altered to run the server in a non-POSIX environment.
*
* If we ever port to a different platform and either have multiple
* variants of this file or simply load it up with #ifdefs.
*
*/
#include sysdep.h
#include stdlib.h
#include unistd.h
#include stdio.h
#include fcntl.h
#include ctype.h
#include signal.h
#include sys/types.h
#include sys/stat.h
#include sys/wait.h
#include sys/socket.h
#include sys/syslog.h
#if TIME_WITH_SYS_TIME
# include sys/time.h
# include time.h
#else
# if HAVE_SYS_TIME_H
# include sys/time.h
# else
# include time.h
# endif
#endif
#include limits.h
#include sys/resource.h
#include netinet/in.h
#include netinet/tcp.h
#include arpa/inet.h
#include netdb.h
#include sys/un.h
#include string.h
#include pwd.h
#include errno.h
#include stdarg.h
#include grp.h
#ifdef HAVE_PTHREAD_H
#include pthread.h
#endif
#include citadel.h
#include server.h
#include serv_extensions.h
#include sysdep_decls.h
#include citserver.h
#include support.h
#include config.h
#include database.h
#include housekeeping.h
#include tools.h
#include serv_crypto.h
#include serv_fulltext.h
#ifdef HAVE_SYS_SELECT_H
#include sys/select.h
#endif
#ifndef HAVE_SNPRINTF
#include snprintf.h
#endif
#ifdef DEBUG_MEMORY_LEAKS
struct igheap {
struct igheap *next;
char file[32];
int line;
void *block;
};
struct igheap *igheap = NULL;
#endif
pthread_mutex_t Critters[MAX_SEMAPHORES]; /* Things needing locking */
pthread_key_t MyConKey;/* TSD key for MyContext() */
int verbosity = DEFAULT_VERBOSITY; /* Logging level */
struct CitContext masterCC;
time_t last_purge = 0;/* Last dead session purge */
static int num_threads = 0; /* Current number of threads */
int num_sessions = 0;/* Current number of sessions */
pthread_t indexer_thread_tid;
pthread_t checkpoint_thread_tid;
int syslog_facility = LOG_DAEMON;
int enable_syslog = 0;
extern int running_as_daemon;
/*
* lprintf() ... Write logging information
*/
void lprintf(enum LogLevel loglevel, const char *format, ...) {
va_list arg_ptr;
if (enable_syslog) {
va_start(arg_ptr, format);
vsyslog((syslog_facility | loglevel), format, arg_ptr);
va_end(arg_ptr);
}
/* stderr output code */
if (enable_syslog || running_as_daemon) return;
/* if we run in forground and syslog is disabled, log to terminal */
if (loglevel = verbosity) {
struct timeval tv;
struct tm tim;
time_t unixtime;
gettimeofday(tv, NULL);
/* Promote to time_t; types differ on some OSes (like darwin) */
unixtime = tv.tv_sec;
localtime_r(unixtime, tim);
if (CC-cs_pid != 0) {
fprintf(stderr,
%04d/%02d/%02d %2d:%02d:%02d.%06ld [%3d] ,
tim.tm_year + 1900, tim.tm_mon + 1,
tim.tm_mday, tim.tm_hour, tim.tm_min,
tim.tm_sec, (long)tv.tv_usec,
CC-cs_pid);
} else {
fprintf(stderr,
%04d/%02d/%02d %2d:%02d:%02d.%06ld ,
tim.tm_year + 1900, tim.tm_mon + 1,
tim.tm_mday, tim.tm_hour, tim.tm_min,
tim.tm_sec, (long)tv.tv_usec);
}
va_start(arg_ptr, format);
vfprintf(stderr, format, arg_ptr);
va_end(arg_ptr);
fflush(stderr);
}
}
/*
* Signal handler to shut down the server.
*/
volatile int time_to_die = 0;
volatile int shutdown_and_halt = 0;
static RETSIGTYPE signal_cleanup(int signum) {
lprintf(CTDL_DEBUG, Caught signal %d; shutting down.\n, signum);
time_to_die = 1;
master_cleanup(signum);
}
/*
* Some initialization stuff...
*/
void init_sysdep(void) {
int i;
sigset_t set;
/* Avoid vulnerabilities related to FD_SETSIZE if we can. */
#ifdef FD_SETSIZE
#ifdef RLIMIT_NOFILE
struct rlimit rl;
getrlimit(RLIMIT_NOFILE, rl);
rl.rlim_cur = FD_SETSIZE;
rl.rlim_max = FD_SETSIZE;
setrlimit(RLIMIT_NOFILE, rl);
#endif
#endif
/* If we've got OpenSSL, we're going to use it. */
#ifdef HAVE_OPENSSL
init_ssl();
#endif
/* Set up a bunch of semaphores to be used for critical sections */
for (i=0; iMAX_SEMAPHORES; ++i) {
pthread_mutex_init(Critters[i], NULL);
}
/*
* Set up a place to put thread-specific data.
* We only need a single pointer per thread - it points to the
* CitContext structure (in the ContextList linked list) of the
* session to which the calling thread is currently bound.
*/
if (pthread_key_create(MyConKey, NULL) != 0) {
lprintf(CTDL_CRIT, Can't create TSD key: %s\n,
strerror(errno));
}
[Citadel Development] (no subject)
Fri 19 Jan 2007 07:25:19 AM EST from [EMAIL PROTECTED] (comalies) ... removing autoheader cache files ... ... running autoheader ... autoheader: error: AC_CONFIG_HEADERS not found in configure.acI checked out a fresh copy to be sure, still happens..we don't use config headers in webcit. plus the new configure gets created. so it should do no harm.
[Citadel Development] Re:First Citadel CSS Guide - login screen
thierry, maybe you want to sign into the wiki, and do the documentation there?we're using dokuwiki, if you want to test somewhere else first.you could do stuff like code html to make the wiki display syntax hilighted html code...join us on freenode #citadel, or contact me for an account.
[Citadel Development] (no subject)
hm, on the multiple login stuff... it looks to me as the update poller is guilty for that. i've been talking to such a person lately, he wasn't aware that he is doing harm to the service; he justs did a simple login with webcit.
[Citadel Development] Re:SVN commit log: revision 5140
So 06 Mai 2007 10:48:53 EDT von [EMAIL PROTECTED] Betreff: Re:SVN commit log: revision 5140 sam 05 mai 2007 13:35:21 EDT de [EMAIL PROTECTED] Objet : Re:SVN commit log: revision 5140 thierry, once you're doing the editor window...could you have a look that the file buttons and the post buttons are devided to stay on the left / right? OKI suggest to remove the Post and Cancel buttons at the bottom to keep alone the upload files section.Why the buttons of the editor are below ? h. yes you're right since there is no resize button like we have in the wikki page (which i would apreciate to... editing in a stamp isn't that fun) it doesn't make sense to have them above and under...
[Citadel Development] (no subject)
cvs couldn't delete dirs.. ATTIC... i hated that one verry much.moving files wasn't nice too... Using a webserver for readonly access is preferable imho...svn isn't perfect for sure.. but how was that saying?Perfection is the enemy of the good?
[Citadel Development] edit view feedback
thierry, i like it as being a good stsart... but some more input. that 'in Mail (date) wraps arround, i don'th think it's good in there..maybe it could find some place in the header line? together with that writing icon?maybe igs cool new address thang could be put there too?on the icon... maybe the peoples could jump out of a drawer? maybe an arrow aside them?on the 'send message' button... i don't like it above the editing. maybe it could put in some icon in the gray zone too?and on the botom line... if we put it horizontally like you did it now, i'd put a tiny horizontal separator line between it, and center the items. maybe a thin vertical separator line between the attachment selector and the Attach file writing would be nice to group the items logically?On tinymce.. It seems as if the new version i've tested seems to have two advances usefull for us. It is compatible with opera, and it has the ability to add a vertical resize thumb in an aditional bottomline. I've plaied arround with it a bit, and found that combination: tinyMCE.init({ mode : \textareas\, width : \100%%\, browsers : \msie,gecko\,theme : \advanced\, plugins : \iespell\,theme_advanced_buttons1 : \bold, italic, underline, strikethrough, justifyleft, justifycenter, justifyright, justifyfull, bullist, numlist, cut, copy, paste, link, image, help, forecolor, iespell, code\,theme_advanced_buttons2 : \\,theme_advanced_buttons3 : \\,theme_advanced_path_location : \bottom\,\n theme_advanced_resizing : true,\n// theme_advanced_resize_horizontal : false,\n content_css : \static/webcit-tinymcecss\ });maybe we should put tinymces icons to the top then? I really like it to be vertical enlargable so i can see more message lines at once, if i like to.and... other input boxes, like the subject: one, could it use the same font and size as tiny mce does? so we get a little more space out there? and have it being more flat?
[Citadel Development] imap trouble with debian sid mutt
I've done some chatting with a mutt guy on that. Ig, maybe you could have a look at it? It's with my latest stable citadel. 11:23 dothebart: hm, mutt in latest version seems to have trouble with pipelining according to the guys on mutt 11:23 dothebart: http://tar.gzipped.org/muttdebug.0 11:27 dothebart: 11:27 pdmef: Hmm. The a0002 and a0003 are send in a batch, the server answers for a0002 but not for a0003 11:28 dothebart: 11:28 pdmef: Is that all? If so, I think my guess was right if there's no more debugfile contents... :( 11:33 dothebart: 11:30 dothebart: whats it expecting there? should the server behave different? 11:33 dothebart: 11:33 pdmef: Each command gets a (session wide) unique ID so that the client can send in any number commands (well, not really any but some) and the server puts the same ID in its response so that the client knows for what command it's the answer 11:34 dothebart: 11:33 pdmef: In your case mutt issues MYRIGHTS and SELECT, the server only responds for MYRIGHTS and silenty drops SELECT while mutt waits for the SELECT answer... 11:34 dothebart: 11:33 pdmef: Well, I think, I'm no IMAP expert
[Citadel Development] Re: SVN commit log: revision 5168
Mon 21 May 2007 09:20:17 AM EDT from IGnatius T [EMAIL PROTECTED] Subject: Re: SVN commit log: revision 5168 There is no such thing as reply all or forward in a bulletin board conversation. A reply is posted to the same bulletin boardBut, as bulletin boards and mailinglist merge almost transparently in citadel, shouldn't that go away too? if i've read a discussion, and want to tell a friend about that, wouldn't 'forward' be the right thing to do? (which i actualy are enabled to if i do imap...)
[Citadel Development] who's online button...
hm, maybe we could do such a bubble thing there too instead of putting that into the bar?plus, that list should be uniq in any case, so multiple logins of the same user don't appear...
[Citadel Development] (no subject)
no success on the kontact side. if we change the content type for the outlook connector, it will break kontact imap calendal access. Ig, can you have a word on that?
[Citadel Development] (no subject)
thierry, actualy my name extends the bar in the new 'who is online' button.either it must cut the nick or wrap it... but it goes over the bar.If the 'calendar' for example is displayed, it is cut..
[Citadel Development] (no subject)
thierry, uncensored moved to a new box. you need to ssh to it once, and erase the old key in .ssh/known_hosts ; ssh to it again, say yes, and then your svn will work again.
[Citadel Development] (no subject)
hm, the scrollwheel stopped working here :( just the submit buttons are blinking...
[Citadel Development] browser
nope, it was an iceweasel 2.0... so allmost a firefox...
[Citadel Development] (no subject)
hm, funny, it works with this etch iceweasel...
[Citadel Development] Re:SVN commit log: revision 5219
Mi 13 Jun 2007 17:16:16 EDT von Citadel commit log in 00.Sent Items Betreff: SVN commit log: revision 5219r5219 | dothebart | 2007-06-13 17:16:15 -0400 (Wed, 13 Jun 2007) | 2 linesChanged paths: M /trunk/ctdlphp/ctdlprotocol.php M /trunk/ctdlphp/myinfo.php* put the vcard we just read back in.Did i do that right? at least my preferences look the same after the next request...
[Citadel Development] (no subject)
Sa 23 Jun 2007 03:28:57 EDT von [EMAIL PROTECTED] ven 22 jun 2007 17:33:01 EDT de IGnatius T [EMAIL PROTECTED] I really like the project statistics, metrics, and graphs that Ohloh generates. It does all of that automatically by fetching your entire SVN repo and analyzing the code. Pretty cool.Yes. I suggested an improvement with Ohloh guys : My Ohloh registration name is jeau. I could not use thierry. Maybe this id was already taken. It wasn't a problem but the google map of Citadel project show my Ohloh id, not my commit id. It would be great to show ids like this : thierry (jeau) thierry, under 'my profile' there is in the center of a screen an button Experience: Newyou add citadel there, and put in your committer name.then you will show up on the developers map.for some reason ig has dissapeared again, fancy.
[Citadel Development] Re:SVN commit log: revision 5249
Ig, is that position in the dialog right? The wording and spelling ok? if, we can add it to the po's
[Citadel Development] Re:SVN commit log: revision 5248
is that flag setting as you thought it to be? or should it be different? Else i need to add it to the documentation...On webcit... if this is proper, it would be nice if s.b. could please add some js that steps on the users toes, if a subject is recommended, and he didn't add one... TIA.These things are missing from the text client so far.
