MoinMoin: MoinMoinChat/Logs/moin-dev/2009-07-28

2009-07-28T00:14:43  *** starshine_away has joined #moin-dev
2009-07-28T00:15:43  *** starshine_away is now known as starshine
2009-07-28T00:56:44  *** starshine is now known as starshine_away
2009-07-28T01:03:35  *** grzywacz has quit IRC
2009-07-28T01:10:01  *** hgg has joined #moin-dev
2009-07-28T01:34:25  *** hgg has quit IRC
2009-07-28T01:36:47  *** ser has quit IRC
2009-07-28T01:44:16  *** ser has joined #moin-dev
2009-07-28T02:01:03  *** SuBmUnDo_ has joined #moin-dev
2009-07-28T02:08:04  <CIA-43> Alexandre Martani <amartani AT gmail DOT com> default * 6397:6883c8853a04 2.0-storage-editor-amartani/MoinMoin/ (3 files in 2 dirs): Real-time editor: update mobwrite to r74
2009-07-28T02:08:38  *** amartani has quit IRC
2009-07-28T03:10:50  *** amartani has joined #moin-dev
2009-07-28T05:13:47  <CIA-43> Alexandre Martani <amartani AT gmail DOT com> default * 6398:d0b0dc41966d 2.0-storage-editor-amartani/MoinMoin/web/static/htdocs/applets/Mobwrite/mobwrite_core.js: Real-time editor: Fix mobwrite update.
2009-07-28T09:01:29  <dreimark> moin
2009-07-28T09:46:04  <dimazest> moin
2009-07-28T09:48:07  <ThomasWaldmann> MoinMoin
2009-07-28T09:49:26  <dreimark> dimazest: please check if we need a workaround for unicode too
2009-07-28T09:49:34  <dreimark> https://code.launchpad.net/~miracle2k/django-xappy/trunk
2009-07-28T09:49:56  <dreimark> http://bazaar.launchpad.net/~miracle2k/django-xappy/trunk/revision/38
2009-07-28T09:50:43  <dimazest> dreimark: i'll check that
2009-07-28T10:02:08  *** starshine_away has quit IRC
2009-07-28T10:02:08  *** CIA-43 has quit IRC
2009-07-28T10:02:08  *** xorAxAx has quit IRC
2009-07-28T10:02:08  *** vpv has quit IRC
2009-07-28T10:02:40  *** starshine_away has joined #moin-dev
2009-07-28T10:02:40  *** CIA-43 has joined #moin-dev
2009-07-28T10:02:40  *** xorAxAx has joined #moin-dev
2009-07-28T10:02:40  *** vpv has joined #moin-dev
2009-07-28T10:08:57  <ThomasWaldmann> that likely is for situations where some found word is shown within some context
2009-07-28T10:10:06  <ThomasWaldmann> xapian itself just knows bytes, so if you tell it to give context of +/- 50, it will maybe truncate within a utf-8 multibyte sequence
2009-07-28T10:11:01  <ThomasWaldmann> and of course if that happens, you can't decode that with 'strict'
2009-07-28T10:12:23  <ThomasWaldmann> and it depends a bit on the goals of the xappy api whether such workaround code should be done within xappy or not
2009-07-28T10:14:38  <dreimark> if we need it it should be discussed with richard (don't know this at the moment)
2009-07-28T10:16:16  <ThomasWaldmann> """All text passed to the interface is assumed to be UTF-8 encoded Unicode."""
2009-07-28T10:16:23  <ThomasWaldmann> from xappy docs
2009-07-28T10:16:58  <ThomasWaldmann> so, xappy is not a unicode api, it is based on str (bytes)
2009-07-28T10:17:27  <ThomasWaldmann> thus, this has to be done by code using xappy
2009-07-28T10:49:18  <ThomasWaldmann> bbl
2009-07-28T11:47:44  <dennda> "history-over-delete" == "history-about-deletions"?
2009-07-28T11:48:57  <dreimark> I guess yes
2009-07-28T11:50:59  <dennda> Besides, viewing all trash items doesn't need to be "check for previous metadata" but rather "enumerate all trash items in trash backend"
2009-07-28T13:56:27  <ThomasWaldmann> dennda: i guess johill meant that you can track revision history older than deletion event(s)
2009-07-28T13:56:56  <ThomasWaldmann> currently with deletion just in metadata we have that
2009-07-28T13:57:06  <ThomasWaldmann> but with trashbin approach, we would not
2009-07-28T14:14:26  <dennda> yes. there you'd have to check the trashbin for items of interest
2009-07-28T14:16:53  <johill> if you o to the info action that works with deleted items while they are deleted
2009-07-28T14:17:17  <johill> thats what i was referring to
2009-07-28T14:17:29  <johill> (sorry typing on n810)
2009-07-28T14:23:28  <dennda> ah yes
2009-07-28T14:27:41  * ThomasWaldmann cancelled our ds5000 hetzner server (old moinmo.in)
2009-07-28T15:14:41  <_peck_> hoy
2009-07-28T15:15:28  <_peck_> i talked here about ldap groups about a week ago or 2
2009-07-28T15:16:04  <_peck_> sorry to be a bit long, i've got a question about groups implementation
2009-07-28T15:16:47  <_peck_> is it possible to retrieve the user object from its username to enable the ldap group to interact with the ldap auth
2009-07-28T15:17:43  <_peck_> the goal is to use the user dn instead of the uid for ldap users to avoid possible name collisions between ldap branches
2009-07-28T15:19:57  <_peck_> otherwise i can hope uid is really unique accross the ldap server but this may not be garanted
2009-07-28T16:52:02  <ThomasWaldmann> hi _peck_
2009-07-28T16:52:16  <ThomasWaldmann> dimazest: ^^
2009-07-28T16:52:51  <ThomasWaldmann> _peck_: you can create a User object from the username
2009-07-28T16:53:17  <_peck_> with which method ?
2009-07-28T16:53:22  <ThomasWaldmann> User()
2009-07-28T16:53:40  <ThomasWaldmann> (see ldap_login.py)
2009-07-28T16:54:16  <ThomasWaldmann> but the user object does not store details about ldap stuff
2009-07-28T16:54:52  <_peck_> but i would like to retrieve the user object nor create it
2009-07-28T16:55:19  <ThomasWaldmann> well, the current user is in request.user
2009-07-28T16:55:47  <_peck_> i could strore the dn from the auth modul in there
2009-07-28T16:56:25  <ThomasWaldmann> yes, if you just use u.dn = ... in ldap_login, that would work
2009-07-28T16:56:41  <_peck_> yes that's what i want to do
2009-07-28T16:56:51  <ThomasWaldmann> (it would be stored into the profile then, though, if you do not want that, it is u._dn or so)
2009-07-28T16:57:08  <dimazest> i cannot answer to question, it is more related to ldap code, not groups
2009-07-28T16:57:15  <_peck_> ok, then yes with an _
2009-07-28T16:57:20  <ThomasWaldmann> hmm
2009-07-28T16:57:27  <ThomasWaldmann> no, you need it to be stored
2009-07-28T16:57:44  <ThomasWaldmann> otherwise it will only be there in 1 request (when the user sends the login form)
2009-07-28T16:57:53  <_peck_> why ? is is for sso or sth like that ?
2009-07-28T16:58:03  <_peck_> i see
2009-07-28T16:58:10  <ThomasWaldmann> because ldap_login is just executed once
2009-07-28T16:58:20  <ThomasWaldmann> the session is kept by the session system
2009-07-28T16:58:24  <_peck_> what is there in the profile anyway ?
2009-07-28T16:58:33  <ThomasWaldmann> (you could also store it into the session)
2009-07-28T16:58:44  <_peck_> is it session related or is it in a database ?
2009-07-28T16:58:46  <ThomasWaldmann> settings, email, ...
2009-07-28T16:59:10  <ThomasWaldmann> for moin < 2.0, user profiles are in data/user/*
2009-07-28T16:59:35  <_peck_> it may be peferable to make it hidden
2009-07-28T17:00:24  <ThomasWaldmann> so, what's your plan? a lazy group backend for ldap?
2009-07-28T17:01:00  <_peck_> yes
2009-07-28T17:01:38  <ThomasWaldmann> ok, so the problem to solve is <u> in <g>
2009-07-28T17:02:18  <ThomasWaldmann> <u> is not necessarily the current user (request.user) - e.g. if we want to check whether to send a email notify about a page change and whether some user may read that stuff
2009-07-28T17:02:41  <ThomasWaldmann> thus, if you need the dn, you must store it in user profile
2009-07-28T17:02:52  <_peck_> ok
2009-07-28T17:03:24  <_peck_> and will this cover all the cases ?
2009-07-28T17:03:36  <ThomasWaldmann> then you'll have <u>  --> User(...<u>...)  --> dn  --> ldap lookup  --> result
2009-07-28T17:04:17  <ThomasWaldmann> (and quite some configuration stuff to make this flexible enough to cope with everyone's ldap setup)
2009-07-28T17:05:03  <ThomasWaldmann> if it is stored in profile, you can have access to it at any time
2009-07-28T17:05:09  <_peck_> is that you who wrote the ldap_auth stuff ?
2009-07-28T17:05:51  <_peck_> i may have to rewrite it a bit for this
2009-07-28T17:06:23  <ThomasWaldmann> yes and Nick Philips
2009-07-28T17:06:28  <_peck_> is the profile fully visible from the interface or only an object stuff ?
2009-07-28T17:06:52  <ThomasWaldmann> the ui just shows some of the stuff
2009-07-28T17:07:09  <ThomasWaldmann> it won't show the dn unless you make it do that
2009-07-28T17:07:20  <_peck_> ok, that's good
2009-07-28T17:08:14  <ThomasWaldmann> (I must add that I don't use ldap_login myself, but I work on it from time to time when doing support for $SOMEBIGCOMPANY)
2009-07-28T17:09:12  <_peck_> :)
2009-07-28T17:09:52  <ThomasWaldmann> what we need to do later, if to rip out the ldap connection into an own object
2009-07-28T17:10:25  <ThomasWaldmann> and give that object to both ldap_login and ldap_groups
2009-07-28T17:10:30  <_peck_> that's what i started with for groups
2009-07-28T17:11:30  <ThomasWaldmann> ah, good
2009-07-28T17:12:19  <ThomasWaldmann> _peck_: do you have multiple ldap servers?
2009-07-28T17:12:24  <ThomasWaldmann> like pdc and bdc?
2009-07-28T17:12:25  <_peck_> no
2009-07-28T17:12:39  <_peck_> but i might have in the future, i don't know yet
2009-07-28T17:14:11  <_peck_> to configure a group manager, I see in the wiki that you override the groups method in Config
2009-07-28T17:14:22  <_peck_> but i found this in the svn : group_manager_init = lambda self, request: WikiGroups(request)
2009-07-28T17:14:53  <ThomasWaldmann> you mean in hg :)
2009-07-28T17:15:01  <_peck_> yes of course :)
2009-07-28T17:15:23  <ThomasWaldmann> it used to be named "group_manager_init", but we decided this is a bit too loooong, so it is just "groups" now.
2009-07-28T17:15:40  <_peck_> ok
2009-07-28T17:16:16  <ThomasWaldmann> btw, did you see the groups stuff was merged into moin/1.9 main repo?
2009-07-28T17:16:51  <ThomasWaldmann> (and dicts also of course. and openid.)
2009-07-28T17:19:04  <_peck_> no
2009-07-28T17:19:17  <_peck_> i'm still working with 1.9-groups-dmilajevs
2009-07-28T17:25:36  <_peck_> what i a dict for moinmoin exactly ?
2009-07-28T17:26:26  <ThomasWaldmann> the same as in python :)
2009-07-28T17:26:44  <ThomasWaldmann> but keys and values are just unicode objects
2009-07-28T17:27:17  <_peck_> and what does wmoinmoin store with dict backends ?
2009-07-28T17:27:33  <_peck_> pages ? anything ?
2009-07-28T17:27:48  <ThomasWaldmann> not much. you can use it to extend i18n and for @FOO@ stuff.
2009-07-28T17:28:15  <_peck_> ok
2009-07-28T17:33:17  <_peck_> how does an acl detect that it applies to a group ?
2009-07-28T17:33:41  <_peck_> if it's only with the Group suffix, my ldap groups don't have them
2009-07-28T17:34:20  <ThomasWaldmann> you need to have a transformation in your stuff then
2009-07-28T17:34:46  <ThomasWaldmann> so that it e.g. strips the Group (or whatever else) suffix
2009-07-28T17:35:26  <ThomasWaldmann> (use page_group_regex to extract the relevant stuff)
2009-07-28T17:36:06  <ThomasWaldmann> or rather page_group_regexact even
2009-07-28T17:36:33  <ThomasWaldmann> for: FooGroup -> Foo
2009-07-28T17:37:20  <_peck_> ok
2009-07-28T17:45:13  <_peck_> last question, how do i enable logging and change log level ?
2009-07-28T17:46:53  *** johill has left #moin-dev
2009-07-28T17:49:07  <ThomasWaldmann> _peck_: using wikiserver.py?
2009-07-28T17:50:29  <_peck_> wikiconfig.py
2009-07-28T17:50:44  <ThomasWaldmann> i mean "which server are you using"?
2009-07-28T17:50:52  <_peck_> i found the log.load_config command but what is the config file format ?
2009-07-28T17:51:14  <_peck_> apache an cgi for the tests
2009-07-28T17:51:15  <ThomasWaldmann> see wiki/config/logging for samples. there is a ready to use one for auth logging
2009-07-28T18:40:14  <_peck_> \o/ my first working ldap group
2009-07-28T18:42:58  <_peck_> well, not really working, the dn is faked
2009-07-28T18:43:30  <_peck_> how do i retrieve a user from the profile list ?
2009-07-28T18:45:51  <dimazest> ThomasWaldmann: dreimark today i've studied Moin.search code, i think it is too complicated
2009-07-28T18:47:20  <dimazest> the main reason is that xappy provides simplier api, there is no need to specify terms, probably, stemming is done internally, there is query parser
2009-07-28T18:47:43  <dimazest> it can return maching results and highlight them
2009-07-28T18:48:27  <dimazest> so, my proposal is to make search code modular, and have 2 backends: for moin search and for xappy search
2009-07-28T18:49:19  <dimazest> then it should be easier to use other searchers, libraries in the future
2009-07-28T18:49:36  <dreimark> dimazest: if one enters a regex while xapian is enabled it needs to do the slow moin search
2009-07-28T18:50:23  <dimazest> but this can be decided befor backend is applied
2009-07-28T18:50:50  <dreimark> yes, I just want to mention it that it works currently this way
2009-07-28T18:51:21  <dreimark> I think it is a good idea to make it modular
2009-07-28T18:51:56  <dimazest> and i think in the end it will be simplier
2009-07-28T18:52:09  <dimazest> (i mean the code)
2009-07-28T18:54:03  <dreimark> probably yes. also I think it is good to use the functionality of xappy/xapian instead of doint i t by much own code when it is only the same result
2009-07-28T18:56:05  <dreimark> dimazest: is the sort issue fixed by doing this automaticly?
2009-07-28T18:56:16  <dreimark> or is that problem then forwarded to xappy?
2009-07-28T18:56:58  <dimazest> dreimark: do not know
2009-07-28T18:57:03  <dimazest> let me see
2009-07-28T19:02:32  <dimazest> dreimark: http://paste.pocoo.org/show/131291/
2009-07-28T19:05:47  <dreimark> hmm that looks like a title search but not like a text search or?
2009-07-28T19:07:29  <dreimark> dimazest: it is now aplhabetically is that also given on a text search?
2009-07-28T19:10:58  <dreimark> and could the order be controlled
2009-07-28T19:11:46  <dimazest> http://paste.pocoo.org/show/131297/
2009-07-28T19:12:00  <dimazest> dreimark: you can specify sort order
2009-07-28T19:12:01  <ThomasWaldmann> dimazest: did you look at johill's search stuff in storage repo?
2009-07-28T19:12:51  <dimazest> ThomasWaldmann: there are not that much changes comparing with 1.9
2009-07-28T19:14:02  <dimazest> dreimark: http://paste.pocoo.org/show/131298/ sorting by title
2009-07-28T19:17:33  <dreimark> ok
2009-07-28T19:17:51  <dreimark> could you say something about the speed compared with xapwrap?
2009-07-28T19:19:20  <ThomasWaldmann> dimazest: iirc it is search/term.py
2009-07-28T19:19:53  <ThomasWaldmann> dreimark: as most of the work is done in xapian C libs, the speed will be pretty much the same
2009-07-28T19:20:30  <ThomasWaldmann> except if we optimize OUR stuff to search more intelligently :)
2009-07-28T19:21:40  <dimazest> ThomasWaldmann: i neeed to look at that code
2009-07-28T19:22:27  <_peck_> ThomasWaldmann: do you know how do i retrieve a user profile ?
2009-07-28T19:22:44  <ThomasWaldmann> _peck_: User()
2009-07-28T19:23:01  <_peck_> but this doesn'y bring me back my dn attribute
2009-07-28T19:23:04  <ThomasWaldmann> see ldap_login.py
2009-07-28T19:23:12  <ThomasWaldmann> it should
2009-07-28T19:23:22  <_peck_> i lookes at the wiki/user file and it is stored
2009-07-28T19:23:37  <_peck_> i can't find in the User code where they are restored
2009-07-28T19:23:55  <ThomasWaldmann> it (re)stores everything except some special stuff
2009-07-28T19:24:16  <_peck_> all i found is that is restores a static list or members
2009-07-28T19:24:54  <ThomasWaldmann> _peck_: you did set u.dn = "foo" ?
2009-07-28T19:25:00  <_peck_> yes
2009-07-28T19:25:00  <dimazest> ThomasWaldmann: dreimark what do you think about making search code modular?
2009-07-28T19:25:22  <_peck_> and i've got a dn=foo in the wiki/users/xxx file
2009-07-28T19:25:27  <ThomasWaldmann> dimazest: what we have tried to be somewhat modular :)
2009-07-28T19:25:42  <ThomasWaldmann> but I guess it can be improved :)
2009-07-28T19:25:59  <ThomasWaldmann> _peck_: ok, so saving works
2009-07-28T19:26:08  <ThomasWaldmann> restoring doesn't?
2009-07-28T19:26:11  <_peck_> yes
2009-07-28T19:27:40  <_peck_> but i couldn't find a reference to __filename or to _cfg.user_dir in the User code except in save()
2009-07-28T19:27:43  <dreimark> dimazest: ok from me. but please document changes important for developers
2009-07-28T19:27:58  <ThomasWaldmann> dimazest: btw, keep your remaining time in mind. rewriting that whole search stuff is maybe a bit much.
2009-07-28T19:28:00  <_peck_> so i don't knwo how it can restore data :)
2009-07-28T19:28:11  <ThomasWaldmann> but I have nothing against improving it :)
2009-07-28T19:28:47  <ThomasWaldmann> _peck_: you create a User instance
2009-07-28T19:28:55  <ThomasWaldmann> like the auth code does
2009-07-28T19:29:12  <ThomasWaldmann> then you should have your attribute back as it was
2009-07-28T19:30:04  <_peck_> but where in the __init__ code is that done ?
2009-07-28T19:30:18  <ThomasWaldmann> (create a User instance != create a user profile on disk)
2009-07-28T19:30:19  <_peck_> as i don't get the attribute i look at the source
2009-07-28T19:30:35  <_peck_> is it possible that i have a tooold version ?
2009-07-28T19:30:50  <ThomasWaldmann> no, that's like that since quite a while
2009-07-28T19:30:54  <dreimark> _peck_: which do u use?
2009-07-28T19:30:56  <dimazest> ThomasWaldmann: i aware about time, but if i just change code it can take the same time, and it won't be good xappy usage
2009-07-28T19:31:24  <ThomasWaldmann> we prefer good xappy usage. :)
2009-07-28T19:31:31  <dreimark> yeah !!! :)
2009-07-28T19:31:32  <dimazest> ThomasWaldmann: and i can continue working on it after soc
2009-07-28T19:32:04  <dreimark> that is most welcome
2009-07-28T19:32:05  <ThomasWaldmann> maybe just have it in a not to desolate state when final evals come
2009-07-28T19:33:03  <_peck_> i use this : changeset:   4848:85c15a83b7f2
2009-07-28T19:33:11  <_peck_> summary:     Groups2009: Configuration snippets for dicts.
2009-07-28T19:33:16  <_peck_> date:        Fri Jul 17 12:40:57 2009 +0200
2009-07-28T19:33:51  <ThomasWaldmann> _peck_: freshness of the code is no problem. did you commit any own changesets already to your repo?
2009-07-28T19:34:00  <_peck_> no
2009-07-28T19:34:23  <ThomasWaldmann> then maybe create a diff of what you have (hg diff >whatihave.diff)
2009-07-28T19:34:31  <ThomasWaldmann> clone the moin/1.9 repo
2009-07-28T19:34:40  <ThomasWaldmann> and reapply your diff to that
2009-07-28T19:34:55  <ThomasWaldmann> also copy all files you have in your repo that are not revision controlled
2009-07-28T19:35:12  <ThomasWaldmann> then you at least work with the main repo
2009-07-28T19:35:37  <ThomasWaldmann> but as I said, that is not related to your user profile problem, just a general idea
2009-07-28T19:36:20  <_peck_> ok
2009-07-28T19:36:28  <_peck_> i will have to go for now
2009-07-28T19:36:45  <dimazest> ThomasWaldmann: i can only promise that i will work on it after soc
2009-07-28T19:36:52  <_peck_> but i still don't see anywhere in the code the equivalent of save to restore attributes
2009-07-28T19:38:04  <dreimark> dimazest: final evaluations begin August 17. looks like 20 days
2009-07-28T19:38:31  <dreimark> I think you will have much of the stuff in that time.
2009-07-28T19:39:16  <dimazest> dreimark: i think so
2009-07-28T19:39:19  <dreimark> at least there would be some promising results / tests at that time
2009-07-28T19:39:36  <ThomasWaldmann> dimazest: we are obliged to judge you on the work being done in the soc timeframe
2009-07-28T19:39:59  <ThomasWaldmann> so working on it afterwards is fine and very welcome, but doesn't influence finals
2009-07-28T19:41:46  <dimazest> ThomasWaldmann: i know, by the final term there should be working code
2009-07-28T19:41:56  <dimazest> may be not completely, but mainly
2009-07-28T19:42:02  <ThomasWaldmann> cool :)
2009-07-28T19:49:55  <ThomasWaldmann> brb
2009-07-28T19:49:56  <dreimark> dimazest: we drop the work on "Add a plugin which visualizes access rights of each user for different settings"
2009-07-28T19:58:40  <dimazest> dreimark: yes
2009-07-28T19:59:13  <dimazest> we discussed that search has higher priority
2009-07-28T20:00:05  <dimazest> dreimark: is there  any documentation about search?
2009-07-28T20:02:34  <dreimark> HelpOnSearching
2009-07-28T20:06:10  <dimazest> ok
2009-07-28T20:27:42  *** grzywacz has joined #moin-dev
2009-07-28T21:38:17  *** thw has joined #moin-dev
2009-07-28T21:41:45  <thw> re
2009-07-28T21:53:37  <_peck_> i'm back
2009-07-28T21:53:45  <_peck_> i migrated to 1.9
2009-07-28T21:54:28  <thw> ah, good :)
2009-07-28T21:54:45  * thw migrated to a web irc client :P
2009-07-28T21:54:53  <_peck_> i found the :)
2009-07-28T21:55:29  <_peck_> the problem is that ldap authentication doesn't work anymore :-|
2009-07-28T21:57:08  <thw> enable debug log and look into the log
2009-07-28T22:00:41  <_peck_> that's what i did and : after auth: user is invalid
2009-07-28T22:04:38  <thw> autocreate=True
2009-07-28T22:10:43  <_peck_> :) yep
2009-07-28T22:11:21  <_peck_> my previous config didn't have it
2009-07-28T22:11:28  <_peck_> maybe the default was different
2009-07-28T22:12:00  <dreimark> if you have had the user profile you don't need autocreate
2009-07-28T22:12:22  <dreimark> moins default hasn't changed
2009-07-28T22:15:24  <_peck_> not better for the dn attribute but i found where it should be loaded
2009-07-28T22:25:34  <_peck_> found : valid, changed = self._validatePassword(user_data, password)
2009-07-28T22:25:53  <_peck_> it returns false for a user from ldap
2009-07-28T22:31:19  <thw> _peck_: is this about ldap_login.py?
2009-07-28T22:31:49  <_peck_> this is in load_from_id() in User
2009-07-28T22:32:21  <thw> called from ldap_login?
2009-07-28T22:33:57  <_peck_> found, it's called from my ldap_group code, I create a new User() but dont provide a user
2009-07-28T22:35:10  <_peck_> a *password*
2009-07-28T22:35:55  <thw> you don't need a password
2009-07-28T22:36:07  <thw> look like ldap_login does it
2009-07-28T22:37:03  <_peck_> ok it seems that i should provide an id
2009-07-28T22:37:17  <_peck_> otherwise i go here :  # no password given should fail
2009-07-28T22:38:19  <thw> look like ldap_login does it
2009-07-28T22:39:02  <_peck_> aaaahhhh i used name instead of auth_username
2009-07-28T22:43:33  * dreimark remembers that docstring too 
2009-07-28T22:43:42  <dreimark> we wanted to change that
2009-07-28T22:44:13  <_peck_> ok now my group implementation returns True
2009-07-28T22:47:07  <_peck_> tomorrow i will make this clean
2009-07-28T22:51:00  <thw> tomorrow it can also return False ;)
2009-07-28T22:57:42  <_peck_> hehe
2009-07-28T22:58:26  <thw> btw, that code could be quite interesting for other users
2009-07-28T22:58:43  <thw> so, if you have cleaned it up, maybe publish a diff for review
2009-07-28T22:59:19  <_peck_> of course
2009-07-28T22:59:33  <_peck_> which way do you prefer ?
2009-07-28T22:59:41  <_peck_> to a mailing-list ?
2009-07-28T23:00:52  <dreimark> _peck_: moin wiki as a Feature Request ?
2009-07-28T23:01:35  <_peck_> we can join a patch to a feature request ??
2009-07-28T23:01:57  <_peck_> btw where in the tree should i put the ldap_server implementation ?
2009-07-28T23:01:59  <_peck_> util ?
2009-07-28T23:02:40  <thw> http://moinmo.in/Groups2009 < ldap groups are mentioned there
2009-07-28T23:03:00  <thw> (and it is a general todo, not a "summer of code soctodo"
2009-07-28T23:03:02  <thw> )
2009-07-28T23:03:39  <thw> _peck_: you mean the connection class?
2009-07-28T23:03:45  <_peck_> yes
2009-07-28T23:05:11  <ronny> sup
2009-07-28T23:05:43  <_peck_> support
2009-07-28T23:05:54  <ronny> i could use some infos on how moin uses backends; just to get some ideas for the anyvc commit builder apis
2009-07-28T23:05:59  <thw> _peck_: no, util or just in ldap_login
2009-07-28T23:06:12  <thw> (support are 3rd party packages)
2009-07-28T23:06:13  <ronny> (1.9 storrage)
2009-07-28T23:06:41  <thw> ronny: look at 2.0-storage, 1.9-storage is outdated
2009-07-28T23:06:52  <ronny> ok
2009-07-28T23:08:22  <ronny> can anyone throw a few quick pointers for code i could/should read
2009-07-28T23:08:47  <dennda> MoinMoin.storage
2009-07-28T23:09:09  <dennda> (surprise!)
2009-07-28T23:09:20  <thw> :)
2009-07-28T23:09:37  <ronny> oh my, how unexpected
2009-07-28T23:13:31  * ronny reads the mercurial backend
2009-07-28T23:13:56  <dennda> storage api definition is in MoinMoin.storage.__init__
2009-07-28T23:14:39  <ronny> dennda: i want to get how vcs based backends work internally, not the outside exposure
2009-07-28T23:14:55  <dennda> well then, hg backend
2009-07-28T23:14:59  <ronny> i"ll get at that later when anyvc is cappable of being used for moin
2009-07-28T23:15:24  <ronny> currently my commit builder is rather retarded
2009-07-28T23:16:30  <ronny> it works, for making one file :P
2009-07-28T23:16:52  <ronny> in the first revision
2009-07-28T23:17:01  <ronny> then it starts to fsck up
2009-07-28T23:18:01  <dreimark> http://google-opensource.blogspot.com/2009/07/midterm-report-on-google-summer-of-code.html
2009-07-28T23:21:19  <thw> "fail" is missing X)
2009-07-28T23:30:19  <dreimark> 7%
2009-07-28T23:31:31  <dreimark> is PIL slower on windows ?
2009-07-28T23:32:13  <ronny> hmm
2009-07-28T23:32:33  <ronny> the storrage api seems to be more pain than it should
2009-07-28T23:33:15  <dennda> because...?
2009-07-28T23:35:38  <ronny> dennda: it seems to spil out many things with more code than necessary, i'll figure a more concise way once i have anyvc ready
2009-07-28T23:36:07  <dennda> e.g....?
2009-07-28T23:37:12  <ronny> dennda: i have to understand metadata a bit more first
2009-07-28T23:37:39  <ronny> hmm, ok
2009-07-28T23:38:11  <ronny> its bascially a filesystem with pickled metadata
2009-07-28T23:38:36  <dennda> dunno how hg backend works internally
2009-07-28T23:38:56  <dreimark> gn
2009-07-28T23:39:15  <ronny> dennda: it seems kinda bad
2009-07-28T23:39:24  <dennda> not my fault %-)
2009-07-28T23:39:25  <ronny> it locs single files, then commits them
2009-07-28T23:40:02  <ronny> and the metadata store locks weird, too
2009-07-28T23:40:26  <ronny> (given that hg has an own nice way to deal with per file metadata since years
2009-07-28T23:41:27  * dennda has no idea
2009-07-28T23:41:39  <dennda> not my area of expertise