MoinMoin: MoinMoinChat/Logs/moin-dev/2008-06-01

2008-06-01T01:19:35  <dreimark> re
2008-06-01T01:28:53  <dreimark> gn
2008-06-01T01:51:52  <PawelPacana> gn :)
2008-06-01T02:18:53  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3648:23851c20e53f 1.7/MoinMoin/ (3 files in 2 dirs): add ldap testing framework, add ldap_login tests
2008-06-01T02:19:02  <ThomasWaldmann> mmihaljevic:
2008-06-01T06:54:40  <gizmach> ThomasWaldmann: heh great :) I will take a look at it
2008-06-01T11:01:54  <zenhase> moin
2008-06-01T11:39:34  <ThomasWaldmann> moin
2008-06-01T11:54:20  <dreimark> moin
2008-06-01T12:00:45  <dreimark> 3 failing tests for me currently
2008-06-01T12:00:55  <dreimark> events/_tests/test_events.py:64
2008-06-01T12:02:05  <dreimark> and test_ldap_login.py[4] FFss
2008-06-01T12:04:24  <dreimark> bbl
2008-06-01T12:05:43  <ThomasWaldmann> i also saw the events test failing
2008-06-01T12:06:34  <ThomasWaldmann> the first 2 ldap tests work for me, but they maybe have some requirements not there on your machine
2008-06-01T12:07:56  <ThomasWaldmann> dreimark:
2008-06-01T12:08:35  <PawelPacana> moin
2008-06-01T12:11:00  <ThomasWaldmann> hi PawelPacana
2008-06-01T12:14:07  <TheSheep> moinmoin
2008-06-01T12:59:57  * ThomasWaldmann writes some environment check so we can skip the ldap tests if there is no ldap
2008-06-01T13:04:54  * waldi would just hardcode the answers used to test the implementation without relying on an actual server
2008-06-01T13:40:24  <ThomasWaldmann> i thought about that but finally (after talking to python-ldap author) decided it is more realistic with a real one
2008-06-01T13:41:41  <ThomasWaldmann> apparmor seems to be a bit paranoid
2008-06-01T13:44:01  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3649:df024fd0a129 1.7/MoinMoin/ (_tests/ldap_testbase.py auth/_tests/test_ldap_login.py): LDAP tests: add checks for the environment needed to run the tests. skip test with reason if environment is unsuitable.
2008-06-01T13:45:12  <ThomasWaldmann> dreimark: ^^ please run the tests again and see if it skips with some reasonable msg
2008-06-01T13:49:00  <waldi> ThomasWaldmann: which means that you test 3 things at once: the correct configuration of the ldap server, the correct configucation of the support and the own code
2008-06-01T13:58:29  <ThomasWaldmann> yes
2008-06-01T13:59:16  <ThomasWaldmann> but you can test stuff like failover realistically
2008-06-01T14:00:14  <ThomasWaldmann> and if you do your own api level server emulation, you test also multiple things at once
2008-06-01T14:01:24  <ThomasWaldmann> it is just your own simulation instead the real server and its config
2008-06-01T14:02:08  <ThomasWaldmann> so it doesn't mean that you have less problems, just different ones
2008-06-01T14:50:07  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3650:0b04ad4db631 1.7/MoinMoin/events/_tests/test_events.py: event tests: skip wrong test
2008-06-01T15:08:20  <dennda> mercurial pushes slooow today
2008-06-01T15:08:43  <dennda> I wonder if it crashed
2008-06-01T15:10:12  <CIA-50> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com default * 4019:2e4da6611325 1.8-storage-cdenter/ (129 files in 34 dirs): merging with 1.7-storage-hwendell (which has been merged with 1.7 in the interim)
2008-06-01T15:10:21  <dennda> ah :)
2008-06-01T15:10:53  <ThomasWaldmann> dennda: the problem often is rather cia being slow
2008-06-01T15:11:08  <dennda> ThomasWaldmann: no, it stuck at searching for changes
2008-06-01T15:11:13  <dennda> CIA was quite fast
2008-06-01T15:53:02  <dreimark> ThomasWaldmann: works now, I do wonder a bit about why we do get now the event test failure shown
2008-06-01T15:54:11  <ThomasWaldmann> i skipped the broken test
2008-06-01T15:54:43  <dreimark> I know but it should have been always showing it's brokeness, somthing has hided it
2008-06-01T15:55:07  <ThomasWaldmann> it was due to some other fixes i did 2 weeks ago or so
2008-06-01T15:55:29  <dreimark> ok
2008-06-01T15:55:45  * ThomasWaldmann .oO(ldap failover works :)
2008-06-01T15:55:52  * dreimark tries to get closer to the wikisync problem
2008-06-01T16:01:04  <dreimark>     * FeatureRequests
2008-06-01T16:01:05  <dreimark>     * LdapMultipleAuth
2008-06-01T16:02:05  <dreimark> ThomasWaldmann: the reason you added to LdapMultipleAuth is necessary for every other auth mixing too
2008-06-01T16:03:17  <dreimark> but with ldap it becomes apparently
2008-06-01T16:05:29  <xorAxAx> hmm, my student hasnt been online for a few days
2008-06-01T16:12:54  <ThomasWaldmann> dreimark: yeah, sure
2008-06-01T16:13:48  <ThomasWaldmann> (at least when we query some external source)
2008-06-01T16:18:09  <dreimark> xorAxAx: it is weekend, last time I ve seen him was the 29th
2008-06-01T16:18:51  <xorAxAx> dreimark: but he has set a deadline for today :)
2008-06-01T16:24:22  <ThomasWaldmann> johill:             return ContinueLogin(user_obj, _('Missing password. Please enter user name and password.'))
2008-06-01T16:24:35  <ThomasWaldmann> shouldn't that be rather CancelLogin?
2008-06-01T16:58:15  <johill> ThomasWaldmann: hmm yeah looks like, where's that?
2008-06-01T16:58:33  <johill> ThomasWaldmann: actually, no, you might have another auth method like openid
2008-06-01T16:58:58  <xorAxAx> looks like a layering violation to me
2008-06-01T16:59:02  <johill> ThomasWaldmann: if all methods continue, you get the message, but if one of them is ok with the given info (say openid info was given and no username/password) then that method goes through
2008-06-01T16:59:12  <johill> xorAxAx is starting to sound like a broken record
2008-06-01T16:59:23  <johill> xorAxAx: go fix it
2008-06-01T16:59:24  <xorAxAx> to let an auth plugin stop the login process on its own
2008-06-01T16:59:30  <xorAxAx> johill: well ,its a design issue ...
2008-06-01T16:59:31  <johill> not at all
2008-06-01T16:59:36  <xorAxAx> why?
2008-06-01T16:59:56  <johill> not interested in a discussion about layering with you. go redesign it or forget it
2008-06-01T17:00:17  <xorAxAx> i dont want to discuss but just know the reason why you think its not
2008-06-01T17:00:39  <johill> because of negative login assertions
2008-06-01T17:00:48  <johill> when a user exists in the db but the password was wrong
2008-06-01T17:01:04  <xorAxAx> then you raise cancel? hmm
2008-06-01T17:01:40  <xorAxAx> makes sense if you want the auth providers to have distinct usernames
2008-06-01T17:01:44  <johill> like I said, I'm not interested in discussing the design. if you really cared, you'd have fixed it a long time ago because it was always pretty much broken. and if you don't please don't try to push me into redesigning everything the way you want it
2008-06-01T17:02:01  <johill> you do this for every nit
2008-06-01T17:02:13  <xorAxAx> i didnt say you should redesign it ...
2008-06-01T17:02:13  <johill> I just no longer care
2008-06-01T17:02:28  <xorAxAx> it just makes sense to be able to explain design decisions
2008-06-01T17:02:37  <johill> but you're implicitly accusing me of misdesigning it instead of asking about it
2008-06-01T17:02:40  <xorAxAx> otherwise they will be lost in the evolution of the system
2008-06-01T17:02:54  <johill> then go work on your social skills maybe and _ask_ instead of saying 'oh this sounds broken to me'
2008-06-01T17:02:56  <xorAxAx> yeah, i will word it differently next time :)
2008-06-01T17:07:15  <xorAxAx> johill: so you had 2 people not knowing the design concept in this case. i havent looked at the source code in this case but probably its not documented ... this is not an accusation :)
2008-06-01T17:07:38  <johill> yeah, maybe
2008-06-01T17:07:44  <johill> hey, it's documented better than the rest of moin ;)
2008-06-01T17:08:24  <xorAxAx> hehe
2008-06-01T17:10:18  <johill> in fact, it looks like I forgot to update the docs for the use of return objects
2008-06-01T17:10:33  <johill> ah, not exactly
2008-06-01T17:10:53  <johill> it's just that the helper classes could be described better
2008-06-01T17:13:31  <johill> do we have a 1.8 branch now?
2008-06-01T17:13:40  <johill> is anybody managing merges between 1.7/1.8?
2008-06-01T17:13:50  <johill> the 1.6/1.7 way with listing them all was quite a mess :/
2008-06-01T17:14:32  <CIA-50> Johannes Berg <johannes AT sipsolutions DOT net> default * 3651:ab3a5e89af14 1.7/MoinMoin/auth/__init__.py: add a note about auth vs. username not entered and similar
2008-06-01T17:14:33  <dreimark> we do have an 1.8 branch
2008-06-01T17:14:56  <johill> yeah, that was more rhetorical, I had seen the wiki update about it
2008-06-01T17:15:11  <johill> does anybody want to take responsibility for 1.7?
2008-06-01T17:16:41  <dreimark> I do debug currently wikisync, have not thought about the ongoing mess with two branches again, besides that I dislike it going the same way as 1.6/1.7
2008-06-01T17:16:54  <ThomasWaldmann> johill: i am currently working on ldap_login
2008-06-01T17:17:12  <johill> ThomasWaldmann: yeah, I saw your note that you were doing a testing framework
2008-06-01T17:17:33  <ThomasWaldmann> now I think it is even correct what you had, it just didn't match the old strict behaviour
2008-06-01T17:17:49  <xorAxAx> dreimark: what do you debug about it? the problem we talked about lately?
2008-06-01T17:17:52  <ThomasWaldmann> but I meanwhile found out that the old behaviour was a bit pointless
2008-06-01T17:18:03  <dreimark> xorAxAx: yes
2008-06-01T17:18:21  <CIA-50> Johannes Berg <johannes AT sipsolutions DOT net> default * 3644:5a760b6b6a97 1.8/MoinMoin/auth/__init__.py: add a note about auth vs. username not entered and similar
2008-06-01T17:18:41  <johill> ThomasWaldmann: I didn't even realise I changed behaviour
2008-06-01T17:19:00  <johill> but then, I never really understood the ldap code too well
2008-06-01T17:20:35  <johill> and pretty much failed to set up a test environment
2008-06-01T17:20:52  <johill> but as somebody recently blogged, apparently ldap admins need some initiation rituals and nothing is really documented ;)
2008-06-01T17:22:05  <ThomasWaldmann> i made it behave now like this:
2008-06-01T17:22:31  <ThomasWaldmann> the only issue that leads to CancelLogin is user found and wrong password
2008-06-01T17:23:02  <ThomasWaldmann> if user is not found, or server not reachable, or some exception, it gives subsequent auth list entries a chance
2008-06-01T17:23:58  <ThomasWaldmann> (and if you don't want other auth methods to get a chance, you just don't put them into the auth list)
2008-06-01T17:24:02  <dreimark> ThomasWaldmann: do you have removed the ldap default password?
2008-06-01T17:24:06  <johill> yeah, sounds sane; whether or not an auth method is allowed to do negative assertions is probably a matter of belief and probably _should_ be configurable, but I think we have too many config options already, and I think using multiple methods is pretty insane anyway unless it's like moin+openid or so
2008-06-01T17:24:16  <ThomasWaldmann> dreimark: not yet, but I'll look at it
2008-06-01T17:25:14  <dreimark> johill: do you think we should limit or do define pairs of auth's used together?
2008-06-01T17:25:46  <johill> dreimark: I don't see much point in that either
2008-06-01T17:26:22  <johill> even ldap+moin+openid could make sense if the wiki is like 'company-website' and '(partially) open to public' at the same time
2008-06-01T17:27:15  <johill> in fact, anything +moin could make sense if you want to allow people to subscribe to changes no matter how the wiki is done
2008-06-01T17:27:31  <johill> though, it would be nice for that to have an ACL that says: "allow if auth method was LDAP"
2008-06-01T17:27:42  <dreimark> in the past I thought one could use any openid server but now I think you have to setup your own
2008-06-01T17:28:47  <dreimark> and if you want to clean up the user you may want to know from which auth a userprofile was created
2008-06-01T17:29:22  <dreimark> or if one ldap is missing you don't want the user be able to change the "moin" password without beeing able to sync it with ldap
2008-06-01T17:30:16  <dreimark> or http_auth or openid
2008-06-01T17:32:40  <johill> yeah that's still a sticky thing too
2008-06-01T17:32:55  <johill> changing passwords when the auth method is not moin
2008-06-01T17:33:04  <johill> it really should go via the auth method but having user storage there doesn't make it easy
2008-06-01T17:33:23  <johill> I'm somewhat hoping that ldap user groups stuff will work into a direction of storing more of that there and make that easier
2008-06-01T17:34:36  <johill> the current user.py code is really only suitable to moin auth
2008-06-01T17:34:39  <johill> it stores everything locally etc
2008-06-01T17:34:51  <johill> hence we need that hack with the 'user attrib override' stuff the auth method says
2008-06-01T17:35:07  <johill> it'd be better if all user management was done _via_ the auth method, which could of course hit local storage then
2008-06-01T17:35:11  <dreimark> if one adds two (ldap) servers it can be that a group name is used twice with different users
2008-06-01T17:35:26  <johill> that's possible, but a separate problem I think
2008-06-01T17:36:15  <dreimark> yeah, but it shows that we not only have the problem of two users using the same name
2008-06-01T17:36:56  <dreimark> but that is more relevant to gizmach
2008-06-01T17:40:20  <dreimark> so I would change this to  "allow if auth method was LDAP from server1"
2008-06-01T17:40:50  <dreimark> johill:
2008-06-01T17:46:20  <ThomasWaldmann> please have a look if you spot something incorrect:
2008-06-01T17:46:27  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3652:879d8bdf01a1 1.7/ (3 files in 3 dirs): ldap auth: cancel login only for invalid credentials, otherwise give subsequent auth list entries a chance, failover works, improved logging
2008-06-01T17:46:28  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3653:ba99c8023063 1.7/MoinMoin/auth/__init__.py: merge main
2008-06-01T17:53:54  <johill> dreimark: makes sense too
2008-06-01T18:02:26  <ThomasWaldmann> johill: is it intended that it is still MoinLogin, while everything else is FooAuth?
2008-06-01T18:02:39  <ThomasWaldmann> (could be MoinAuth)
2008-06-01T18:02:44  <johill> not really
2008-06-01T18:12:23  <ThomasWaldmann> ok, I'll change it for better consistency
2008-06-01T18:13:26  <johill> ok
2008-06-01T18:25:27  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3654:6060395dcdf1 1.7/ (5 files in 5 dirs): rename MoinLogin to MoinAuth for better consistency with other auth classes
2008-06-01T18:26:14  <ThomasWaldmann> (if somebody could change it on the master17, that would be appreciated :)
2008-06-01T18:29:33  <ThomasWaldmann> ok, i have a failing test now for the "NotStored" password
2008-06-01T18:50:13  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3655:2ef181f42a41 1.7/ (3 files in 3 dirs): fix MoinMoinBugs/DummyPasswordInAutoCreatedLdapUserProfiles and add a test for it
2008-06-01T18:50:31  <ThomasWaldmann> dreimark: ^^
2008-06-01T19:09:40  * ThomasWaldmann fixed master17 (MoinAuth)
2008-06-01T19:12:58  <ThomasWaldmann> ok, looks like rc2 is near :)
2008-06-01T19:24:03  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3656:97bb459d0873 1.7/MoinMoin/i18n/ (42 files): updated i18n
2008-06-01T19:32:58  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3657:c235cb93bea1 1.7/wiki/underlay.tar: updated underlay
2008-06-01T19:40:46  <vpv> "LDAP server failed" is not really clear for us translators, if it had some reason how it failed, it would be clearer, like "Connection to LDAP server failed" or something
2008-06-01T19:41:46  <ThomasWaldmann> yeah, it means that moin couldn't connect to it
2008-06-01T19:42:29  <ThomasWaldmann> could be also someone pulling some plug :)
2008-06-01T19:43:07  <ThomasWaldmann> but from moin's view, it has failed :)
2008-06-01T19:44:22  <dreimark> bb phonecall
2008-06-01T20:12:11  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3658:b3354cf420ff 1.7/MoinMoin/auth/_tests/test_ldap_login.py: ldap auth tests: fix failures due to order of test execution
2008-06-01T20:12:39  <ThomasWaldmann> dreimark: do they work for you now?
2008-06-01T20:13:58  * dreimark looks
2008-06-01T20:15:47  <dreimark> ThomasWaldmann: test_ldap_login.py[5] sssFF
2008-06-01T20:17:45  * ThomasWaldmann looks
2008-06-01T20:18:43  <ThomasWaldmann> dreimark: is this latest revision?
2008-06-01T20:19:18  <dreimark> yes
2008-06-01T20:19:48  <ThomasWaldmann> ok, please put the log into pastebin
2008-06-01T20:20:46  * ThomasWaldmann expected py.test.skip on module level to skip all tests, strange it does only skip the first 3
2008-06-01T20:21:16  <dreimark> sometimes it is somehow doing miracles
2008-06-01T20:22:03  <ThomasWaldmann> dreimark: is apparmor running on your machine?
2008-06-01T20:22:36  <dreimark> http://paste.pocoo.org/show/59161/
2008-06-01T20:22:41  * dreimark checks
2008-06-01T20:23:02  <dreimark> ThomasWaldmann: no
2008-06-01T20:23:18  <ThomasWaldmann> what is skip reason for first 3 tests?
2008-06-01T20:24:12  <ThomasWaldmann> obviously the last 2 fail because they can't contact the server
2008-06-01T20:24:15  <dreimark> reason: Skipped: Failed to start slapd process, please see your syslog / log files (and check if stopping apparmor helps, in case you use it).
2008-06-01T20:24:29  <ThomasWaldmann> ok, then do that :)
2008-06-01T20:24:48  <dreimark> in my case I do have it running /usr/sbin/slapd -g openldap -u openldap
2008-06-01T20:25:09  <ThomasWaldmann> that doesn't matter als long as it does not run on port 3890
2008-06-01T20:28:38  <ThomasWaldmann> can anyone else please run the tests? Having python-ldap, slapd and python 2.4 installed.
2008-06-01T20:29:00  <dreimark> gizmach:
2008-06-01T20:29:08  <ThomasWaldmann> (it just needs the slapd binary, no need to configure or run it on your own)
2008-06-01T20:36:37  <dreimark> looks like I have problems with ldap_env.start_slapd() returns False
2008-06-01T20:36:48  <dreimark> and slpad becomes  [slapd] <defunct>
2008-06-01T20:36:58  <ThomasWaldmann> then it has not started. what does syslog tell?
2008-06-01T20:38:34  <dreimark> looks like a wrong config call here desktop slapd:3890[7469]: lt_dlopenext failed: (back_bdb.la) file not found
2008-06-01T20:39:23  <dreimark> but if it is False it can be skipped too
2008-06-01T20:40:26  <ThomasWaldmann> yeah, i made a different startup function there
2008-06-01T20:41:44  <dreimark> btw. I am not sure if we should use test in the middle of e.g. ldap_testdata
2008-06-01T20:42:49  <ThomasWaldmann> it has to start with test iirc
2008-06-01T20:44:00  <ThomasWaldmann> btw, that's the berkeley DB backend, pretty standard for LDAP server
2008-06-01T20:44:36  <ThomasWaldmann> slapd -V tells what?
2008-06-01T20:45:48  <dreimark> $OpenLDAP: slapd 2.3.35 (Mar  5 2008 15:13:34)
2008-06-01T20:46:02  <dreimark> buildd@king:/build/buildd/openldap2.3-2.3.35/debian/build/servers/slapd
2008-06-01T20:46:28  * ThomasWaldmann has 2.4.7 on ubuntu
2008-06-01T20:47:38  <dreimark> that's the recent version for my kubuntu version, will do the upgrade when I do have some time
2008-06-01T20:47:48  <dreimark> and will switch to ubuntu
2008-06-01T20:48:18  <ThomasWaldmann> what does locate back_bdb tell?
2008-06-01T20:49:21  * dreimark has to update because it tells nothing 
2008-06-01T20:50:06  <ThomasWaldmann> look into /usr/lib/ldap/
2008-06-01T20:51:03  <dreimark> ls: /usr/lib/ldap/back_bdb: No such file or directory
2008-06-01T20:53:42  <waldi> it depends on the config if this module needs to be loaded or is built-in
2008-06-01T20:53:49  <dreimark> /usr/lib/ldap/back_bdb-2.3.so.0.2.23
2008-06-01T20:53:55  <dreimark> /usr/lib/ldap/back_bdb-2.3.so.0
2008-06-01T20:54:03  <dreimark> /usr/lib/ldap/back_bdb.la
2008-06-01T20:54:14  <dreimark> /usr/lib/ldap/back_bdb.so
2008-06-01T20:54:59  <dreimark> waldi: does it help if I do rename my slapd.conf ?
2008-06-01T20:55:18  <dreimark> it has moduleload      back_bdb
2008-06-01T20:56:03  <dreimark> I do remove my changes of a definition for a test
2008-06-01T20:57:13  <ThomasWaldmann> dreimark: try if it works if you strip the .la from the config moin uses
2008-06-01T21:03:13  * dreimark reboots
2008-06-01T21:10:11  <dreimark> does not help, SERVER_DOWN: {'desc': "Can't contact LDAP server"}
2008-06-01T21:11:22  <ThomasWaldmann> ok, I'll add the same skip as for the single ldap server tests
2008-06-01T21:11:58  <ThomasWaldmann> syslog?
2008-06-01T21:13:09  <dreimark> http://paste.pocoo.org/show/59198/
2008-06-01T21:13:47  <dreimark> try to get why my local ldap works
2008-06-01T21:19:01  <ThomasWaldmann> dreimark: looks like it doesn't find its own stuff
2008-06-01T21:44:08  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3659:2d5a325ae247 1.7/MoinMoin/ (_tests/ldap_testbase.py auth/_tests/test_ldap_login.py): ldap tests: rename base class so py.test doesn't find it, skip tests if slapd is not started
2008-06-01T21:44:13  <ThomasWaldmann> dreimark: try that
2008-06-01T21:44:33  <ThomasWaldmann> at least you should get all skipped now
2008-06-01T22:04:18  <dreimark> ThomasWaldmann: is thge meaning of threads in your config "The tool-threads parameter sets the actual amount of cpu's that is used for indexing."
2008-06-01T22:04:53  <dreimark> in mine config it is named tool-threads. but this is not the reason for the current problem
2008-06-01T22:05:01  <dreimark> now all tests are skipped
2008-06-01T22:05:39  <dreimark> I think we should be better conservative with nCPUs = 1 :)
2008-06-01T22:07:08  <ThomasWaldmann> no, threads is not tool-threads
2008-06-01T22:07:48  <ThomasWaldmann> (and the default is 16 iirc)
2008-06-01T22:07:59  <ThomasWaldmann> dreimark:
2008-06-01T22:34:42  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3660:9901ffff5280 1.7/ (4 files in 3 dirs): 1.7.0rc2 release (bump version number, update CHANGES, ...)
2008-06-01T22:34:44  <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3661:ae77b776457b 1.7/.hgtags: tagged release 1.7.0rc2
2008-06-01T22:43:05  <dreimark> ThomasWaldmann: you were right it is only a path problem
2008-06-01T22:43:09  <dreimark> http://paste.pocoo.org/show/59257/
2008-06-01T22:43:12  <dreimark> works for me
2008-06-01T22:47:59  <ThomasWaldmann> dreimark: waldi: shouldn't a working default of this be compiled into the binary?
2008-06-01T22:49:03  <waldi> for what?
2008-06-01T22:53:41  <dreimark> ThomasWaldmann: http://www.installationwiki.org/Ldap#Basics
2008-06-01T22:53:48  <dreimark> The moduleload directive instructs OpenLDAP to load a particular module. The directive takes either the file name of the module to be loaded (such as back_hdb) or a full path (beginning with /) to a module file. If just the name is specified, SLAPD will look in the directories specified in modulepath. If the entire path is specified, it will attempt to load from exactly that path (it will not use modulepath at all).
2008-06-01T22:54:58  <dreimark> it would be nicer it would know where to search itselfs if nothing is entered
2008-06-01T22:56:11  <ThomasWaldmann> waldi: path for loading its backend modules
2008-06-01T22:59:48  * ThomasWaldmann thinks about switching moinmo.in to 1.7.current next weekend.
2008-06-01T23:00:38  <dreimark> ThomasWaldmann: did you run in your version slaptest -v -f on slapd.conf?
2008-06-01T23:01:33  <dreimark>  -u     enable dryrun mode
2008-06-01T23:03:20  <ThomasWaldmann> no, it worked :)
2008-06-01T23:03:56  <ThomasWaldmann> (and if i put that path into there, it would be broken on installations using some other path)
2008-06-01T23:04:44  <dreimark> I know (we have then to add many pathes)
2008-06-01T23:15:36  <dreimark> (well it is a unit test for a config)