MoinMoin: MoinMoinChat/Logs/moin-dev/2009-06-11

2009-06-11T00:17:38  <ThomasWaldmann> dennda:     1.74 +                # The item returned needs to be wrapped because otherwise the
2009-06-11T00:17:41  <ThomasWaldmann>     1.75 +                # item's methods (like create_revision) wouldn't be wrapped.
2009-06-11T00:19:00  <dennda> proposal for better wording?
2009-06-11T00:29:56  <CIA-19> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5931:e331daf04c11 2.0-storage/MoinMoin/storage/ (__init__.py backends/__init__.py backends/acl.py): storage: Make MoinMoin.storage.backends.acl ready for epydoc
2009-06-11T00:56:47  *** grzywacz has quit IRC
2009-06-11T05:32:11  <CIA-19> Alexandre Martani <amartani AT gmail DOT com> default * 5929:0587878a350f 2.0-storage-editor-amartani/contrib/GoogleMobwrite/ (mobwrite_core.py mobwrite_daemon.py): Real-time editor: update mobwrite to r69.
2009-06-11T05:34:13  *** amartani has quit IRC
2009-06-11T09:03:01  <ThomasWaldmann> moin
2009-06-11T09:44:11  * ThomasWaldmann reads storage docstrings
2009-06-11T10:03:19  <ThomasWaldmann> http://www.angryflower.com/bobsqu.gif < please everybody read this
2009-06-11T10:04:16  <dennda> hehe :D
2009-06-11T10:04:26  <dennda> I assume I did it wrong sometimes last year
2009-06-11T10:04:34  <dennda> recently read up on it again
2009-06-11T10:07:38  * ThomasWaldmann is creating a diff for storage/__init__.py
2009-06-11T10:09:59  <dennda> .oO(Do I really want to see?) :D
2009-06-11T11:03:48  *** ChanServ has quit IRC
2009-06-11T11:04:15  *** ChanServ has joined #moin-dev
2009-06-11T11:04:15  *** irc.freenode.net sets mode: +o ChanServ
2009-06-11T11:10:04  *** ChanServ has quit IRC
2009-06-11T11:10:12  *** ChanServ has joined #moin-dev
2009-06-11T11:10:12  *** irc.freenode.net sets mode: +o ChanServ
2009-06-11T12:21:40  *** grzywacz has joined #moin-dev
2009-06-11T12:46:49  <dimazest> dreimark: i've updated http://moinmo.in/Groups2009 could you check Group, Backend, wiki_group and config_group description
2009-06-11T12:51:07  <ThomasWaldmann> dimazest:     2.60 +        for (group, members) in self.expanded_groups.iteritems():  --> no parentheses needed
2009-06-11T12:51:25  <dimazest> really?
2009-06-11T12:51:41  <dennda> the former aren't
2009-06-11T12:51:48  <dennda> for group, members in ...
2009-06-11T12:52:09  <dimazest> a these, i thought about iteritems()
2009-06-11T12:52:10  <dennda> you still need them to call the iteritems method, of course
2009-06-11T12:52:39  <dimazest> ThomasWaldmann: which changeset is it?
2009-06-11T12:53:47  <ThomasWaldmann> it is in the tests
2009-06-11T12:53:51  <ThomasWaldmann> +            assert len(returned_members) == len(members) 2.74 +            for member in members: 2.75 +                assert member in returned_members
2009-06-11T12:54:11  <ThomasWaldmann> that looks like being easier with a set comparison
2009-06-11T12:55:17  <ThomasWaldmann> test_groups and expanded_groups could also use sets internally
2009-06-11T12:58:37  <ThomasWaldmann> dimazest: btw, the tuple operator is ",", not "()"
2009-06-11T12:59:15  <dimazest> not necesserely because dublicates can be passed, and backend must cope with them
2009-06-11T13:00:39  <ThomasWaldmann> ah, ok. do you test for that?
2009-06-11T13:01:25  <dimazest> no, but i think i should. something like test_groups = {u'EditorGroup': [u'AdminGroup', u'John', u'JoeDoe', u'Editor1', u'John'],
2009-06-11T13:03:44  <dimazest> and i think, it too early to think about mapping, because wiki_page and config backends do not use it
2009-06-11T13:06:47  <ThomasWaldmann> btw, _iter seems to be unnecessary double-checking
2009-06-11T13:06:50  <ThomasWaldmann> +doing
2009-06-11T13:07:41  <ThomasWaldmann> backends/__init__.py 90 and 92 seem to be superfluous
2009-06-11T13:09:11  <ThomasWaldmann> and 84 should be before 83
2009-06-11T13:11:35  <dimazest> http://paste.pocoo.org/show/122392/
2009-06-11T13:12:52  <ThomasWaldmann> no
2009-06-11T13:13:18  <ThomasWaldmann> indentation of 90 is wrong
2009-06-11T13:14:53  <ThomasWaldmann> backend test mapping: would be fine, but please more realistic test case
2009-06-11T13:15:45  <ThomasWaldmann> (and yes, config and wiki backend does not really need it)
2009-06-11T13:16:04  <dimazest> ThomasWaldmann: i want to do it later, because now it is nothing to test, and realistic test will be with sql backend
2009-06-11T13:16:28  <ThomasWaldmann> in fact, maybe the mapping stuff should be removed and be a private part of the child class
2009-06-11T13:16:46  <dimazest> ThomasWaldmann: that's what i want to do
2009-06-11T13:16:50  <ThomasWaldmann> (if needed there)
2009-06-11T13:16:58  <ThomasWaldmann> ok, do it :)
2009-06-11T13:17:16  <ThomasWaldmann> did you run the tests with the changesets you pasted?
2009-06-11T13:17:46  <dimazest> yes
2009-06-11T13:18:50  <ThomasWaldmann> just to check if I am right :)
2009-06-11T13:20:00  <ThomasWaldmann> btw, check line spacing to conform with pep8
2009-06-11T13:20:36  <ThomasWaldmann> (not sure whether our pep8 checker finds that, but methods usually have 1 empty line between, classes and toplevel functions usually have 2)
2009-06-11T13:21:05  <dimazest> i try to folloow this
2009-06-11T13:27:43  <ThomasWaldmann> after you committed that, please change the GroupManager with *backends
2009-06-11T13:27:58  <ThomasWaldmann> that'll make configuration easier
2009-06-11T13:28:25  <dimazest> ok
2009-06-11T13:39:15  <ThomasWaldmann> The wiki group backend enables you to define groups in a configuration file.
2009-06-11T13:39:26  <dimazest> gmm
2009-06-11T13:39:29  <ThomasWaldmann> that is from the config group backend, please fix.
2009-06-11T13:42:19  <ThomasWaldmann> about the config group Backend configuration: you will have to mention the Backend() in wikiconfig anyway, right?
2009-06-11T13:43:11  <ThomasWaldmann> if yes, I don't think there is a need for cfg.config_groups, the wiki admin can use an arbitrary name and just give it to Backend(my_config_groups).
2009-06-11T13:44:17  <ThomasWaldmann> please let your files end with one \n
2009-06-11T13:44:59  <ThomasWaldmann> used.  To find group members, it parses theses pages and extracts the
2009-06-11T13:45:02  <ThomasWaldmann> typo
2009-06-11T13:45:22  <dimazest> ok, i'll kill cfg.config_groups
2009-06-11T13:47:16  * ThomasWaldmann has 1 failing test, btw.
2009-06-11T13:48:04  <dimazest> i know, about this test
2009-06-11T13:54:42  <CIA-19> Dmitrijs Milajevs <dimazest@gmail.com> default * 4782:167db6117bb4 1.9-groups-dmilajevs/MoinMoin/groups/ (9 files in 4 dirs): Groups2009: pep8 fixes and typo elimination. BaseGroup._iter optimization. Test of group names mapping was removed.
2009-06-11T13:58:39  <ThomasWaldmann> dimazest: scripts.account.homepage still uses wikidicts groups
2009-06-11T13:59:38  <dimazest> i'll check it
2009-06-11T14:04:39  <ThomasWaldmann> ^^ looks good
2009-06-11T14:09:30  <ThomasWaldmann> wiki_group.py:81 -> return iter(...)
2009-06-11T14:14:00  <ThomasWaldmann> (then you'll get a different error :)
2009-06-11T14:14:44  <CIA-19> Dmitrijs Milajevs <dimazest@gmail.com> default * 4783:3cea38408fe6 1.9-groups-dmilajevs/MoinMoin/ (6 files in 4 dirs): Groups2009: GroupManager.__init__ takes arbitrary number of backends.
2009-06-11T14:14:45  <CIA-19> Dmitrijs Milajevs <dimazest@gmail.com> default * 4784:6a109416e2b2 1.9-groups-dmilajevs/MoinMoin/ (3 files in 3 dirs): Groups2009: cfg.config_groups was removed.
2009-06-11T14:20:27  <ThomasWaldmann> nice to see it converge to something simple, elegant and usable :)
2009-06-11T14:21:07  <dimazest> ;) for the last time i'm only killing, removing and so on
2009-06-11T14:23:01  <ThomasWaldmann> i have another thing to remove, in some cases
2009-06-11T14:23:50  <ThomasWaldmann> can you check that update/load_cache stuff? is it still used?
2009-06-11T14:24:12  <dimazest> it's next candidate for removal
2009-06-11T14:25:08  <dimazest> but wikidictsrescan uses it
2009-06-11T14:26:52  <ThomasWaldmann> i somehow feel that we don't need that any more (for groups), right?
2009-06-11T14:28:42  <ThomasWaldmann> and after that, please check whether GroupManager is in fact rather a special "Compose" backend (a backend than does not direct storage, but composes other backends to a new one)
2009-06-11T14:28:43  <dimazest> yes, because group object is created in Backend.__getitem__
2009-06-11T14:29:07  *** devilsadvocate has quit IRC
2009-06-11T14:29:24  <ThomasWaldmann> if you move membergroups method to Backend base class, you could just derive the ComposeBackend from that
2009-06-11T14:29:42  <dimazest> it is. i had feeling, that GroupManager is actually a Backend
2009-06-11T14:30:03  <dimazest> then life becomes even easier
2009-06-11T14:30:08  <ThomasWaldmann> and then, we do not need the ComposeBackend any more, except if the wiki admin really wants to compose backends
2009-06-11T14:30:18  <dimazest> yes
2009-06-11T14:30:40  <ThomasWaldmann> (could be also called middleware instead of backend)
2009-06-11T14:31:09  <ThomasWaldmann> btw, in storage repo, we'll have a similar thing for wiki items
2009-06-11T14:31:50  <dimazest> so should i use middleware instead of backend?
2009-06-11T14:32:30  <ThomasWaldmann> doesn't matter much
2009-06-11T14:33:46  <dimazest> ThomasWaldmann: if in other part of moin, same thing called middleware it's better to use middleware, not backend
2009-06-11T14:34:20  <dimazest> just to be "konsequent"
2009-06-11T14:34:41  <ThomasWaldmann> in storage, it is still all under backends/
2009-06-11T14:35:14  <ThomasWaldmann> btw, groups/__init_.py will be rather empty then (except docstring)
2009-06-11T14:35:33  <dimazest> true
2009-06-11T14:37:57  <ThomasWaldmann> we maybe can put some imports there for the stuff usually used
2009-06-11T14:38:31  <ThomasWaldmann> so one can just do "from MoinMoin.groups import WikiGroups, ConfigGroups"
2009-06-11T14:40:47  <ThomasWaldmann> brb
2009-06-11T15:52:10  <CIA-19> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5932:a949da574801 2.0-storage/MoinMoin/storage/ (__init__.py error.py): storage api docstrings: fixes typos, improved formatting, added notes
2009-06-11T15:52:48  <ThomasWaldmann> dennda: please read ^^ (you have at least to look at the XXX stuff)
2009-06-11T16:28:39  <ThomasWaldmann> dennda: please review storage api's use of "long"
2009-06-11T16:43:11  <dennda> ThomasWaldmann: as in <type 'long'>?
2009-06-11T17:10:39  <ThomasWaldmann> yes
2009-06-11T17:17:40  <dennda> ThomasWaldmann: what's wrong with that?
2009-06-11T17:18:01  <dennda> (Other than it was removed in py3k)
2009-06-11T17:21:05  <dennda> ...and the fact that there's no obvious usecase
2009-06-11T17:26:14  <dennda> !"$!$
2009-06-11T17:26:27  <dennda> need to get a new router... that bloody thing dies every 10-20 mins now
2009-06-11T17:26:39  <dennda> and then I have to unplug it
2009-06-11T17:30:39  <ThomasWaldmann> dennda: i had the same recently 2.6 kernel/drivers vs. wrt54gs
2009-06-11T17:31:59  <dennda> ThomasWaldmann: It's the router itself... it's a d-link di-524
2009-06-11T17:32:11  <dennda> in other words: crappy piece of !"$"
2009-06-11T17:33:01  <dennda> I've got a pretty good heuristic to decide whether it's the router itself or not
2009-06-11T17:33:25  <dennda> if roommate opens the door and rushes to the router at the same time, looking furious, it's the router.
2009-06-11T17:33:51  <ThomasWaldmann> hehe
2009-06-11T18:15:54  <ThomasWaldmann> yay, fileserver backend starts working reasonable...
2009-06-11T18:27:30  <CIA-19> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5933:ef8f64d3a4ef 2.0-storage/MoinMoin/storage/backends/fileserver.py: storage: fileserver backend: better mimetype metadata
2009-06-11T18:33:20  <CIA-19> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5934:e4e702592ed5 2.0-storage/MoinMoin/items/__init__.py: items: fix crash when sending items with no hash in metadata
2009-06-11T19:23:57  <CIA-19> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5935:998a6f11d8ae 2.0-storage/MoinMoin/storage/backends/fileserver.py: storage: fileserver backend: improve rendering
2009-06-11T20:12:35  <ThomasWaldmann> btw, the disks in the new moinmo.in server survived a badblocks with 0 reallocated sectors
2009-06-11T22:08:00  *** amartani has joined #moin-dev