2009-07-14T01:22:01 *** grzywacz has quit IRC
2009-07-14T02:41:24 *** starshine_away is now known as starshine
2009-07-14T03:53:04 *** starshine is now known as starshine_away
2009-07-14T04:42:57 *** starshine_away is now known as starshine
2009-07-14T07:43:22 *** starshine is now known as starshine_away
2009-07-14T08:45:32 <dreimark> moin
2009-07-14T08:46:01 <dimazest> moin
2009-07-14T08:46:36 <dimazest> dreimark: i'm still not shure, what to do with dicts
2009-07-14T08:47:37 <dimazest> because the way it works now (with formatter) won't work with mix of markups
2009-07-14T08:47:37 <dreimark> I know (me was to tired yesterday to get a good idea)
2009-07-14T08:47:43 <dimazest> so
2009-07-14T08:48:14 <dimazest> my proposal, for 1.9 use regexp (a least it will work as it worked)
2009-07-14T08:48:32 <dimazest> for 2.0 use new dom code
2009-07-14T08:48:52 <dreimark> waldi: pi
2009-07-14T08:48:56 <dreimark> ng
2009-07-14T08:49:11 <dimazest> may be dict shpuld return an object, whict then can be formatted to wiki/creole/whatever
2009-07-14T08:51:25 <dreimark> MOM
2009-07-14T09:00:53 <dreimark> can it return now something like {u'Two': {'wiki': u'2'}, u'One': {'wiki': u'1'}} ?
2009-07-14T09:02:05 <dreimark> or a tuple for {u'Two': {'wiki': u'2'}, u'One': {'wiki': u'1'}}
2009-07-14T09:02:08 <dreimark> shit
2009-07-14T09:02:29 <dreimark> {u'Two': ('wiki': u'2'), u'One': ('wiki': u'1')}
2009-07-14T09:02:39 <dreimark> s/:/,/
2009-07-14T09:02:48 * dreimark has slept too long
2009-07-14T09:04:09 <waldi> dreimark: 3.141527
2009-07-14T09:05:06 <dimazest> it can
2009-07-14T09:06:07 <dreimark> waldi: float exception 3.14159 if that is pi
2009-07-14T09:08:52 <dreimark> waldi: the dict refactoring will be later refactored in you domtree
2009-07-14T09:09:49 <ThomasWaldmann> re
2009-07-14T09:18:44 <waldi> dreimark: well, this is a simple tree walk
2009-07-14T09:28:26 <ThomasWaldmann> dimazest: if we can't get a well-defined semantics with the formatter approach, it is maybe easier / more compatible to keep the old approach
2009-07-14T09:30:03 <ThomasWaldmann> dimazest: please add a comment about the formatter-kind of solution and what problems have to get solved to do it
2009-07-14T09:30:04 <dimazest> that's my idea, other solutions look too hackish
2009-07-14T09:30:21 <dreimark> and forward a ToDo or XXX for the domtree refactoring
2009-07-14T09:30:51 <dimazest> ThomasWaldmann: are formatters needed for domtree?
2009-07-14T09:30:59 <waldi> it may even work with html then
2009-07-14T09:31:14 <ThomasWaldmann> it is all "converters" there, package converter2
2009-07-14T09:31:21 <dreimark> dimazest: waldi is doing this refactoring
2009-07-14T09:31:39 <ThomasWaldmann> dreimark: the todo can be added in dimazest's repo
2009-07-14T09:31:54 <ThomasWaldmann> it will get merged then anyway
2009-07-14T09:32:00 <dreimark> it was meant so, because of the merge
2009-07-14T09:32:45 <CIA-78> Dmitrijs Milajevs <dimazest@gmail.com> default * 4840:0988f3da0c95 1.9-groups-dmilajevs/MoinMoin/formatter/ (_tests/test_groups.py groups.py): Groups2009: formatter.groups for links collects links, not labels. Creole tets for a MoinMoin.formatter.groups
2009-07-14T09:32:56 <dreimark> typo
2009-07-14T09:34:50 <dimazest> my bad, dreimark could you review tests there
2009-07-14T09:38:52 <dreimark> will do
2009-07-14T09:43:41 <ThomasWaldmann> dimazest: del has no ()
2009-07-14T09:44:52 <ThomasWaldmann> and please check the creole tests
2009-07-14T09:44:56 <ThomasWaldmann> * first level
2009-07-14T09:44:57 <dreimark> dimazest: indenting wrong del(self.request.page)
2009-07-14T09:45:01 <ThomasWaldmann> ** second level
2009-07-14T09:45:13 <ThomasWaldmann> (there is no indenting with blanks in creole iirc)
2009-07-14T09:46:21 <ThomasWaldmann> (at least, it is not required for determine list level, not sure whether it is ignored if it is there)
2009-07-14T09:47:48 <dreimark> it is not ignored
2009-07-14T09:48:08 <dreimark> they become all first level lists and have then some * in the text
2009-07-14T09:48:56 <dimazest> ThomasWaldmann: http://paste.pocoo.org/show/128444/
2009-07-14T09:49:09 * dreimark wonders why assert len(members) == 2
2009-07-14T09:49:14 <dreimark> doesn't fail
2009-07-14T09:50:12 <dimazest> dreimark: which line?
2009-07-14T09:50:21 <dreimark> http://hg.moinmo.in/moin/1.9-groups-dmilajevs/rev/0988f3da0c95#l1.118
2009-07-14T09:51:21 <dimazest> there are two first level items
2009-07-14T10:28:52 * dreimark if there is a syntax error it should be handled different
2009-07-14T10:29:03 <dreimark> but it look like blanks are ignored
2009-07-14T10:49:37 <dreimark> window 14
2009-07-14T10:51:18 *** grzywacz has joined #moin-dev
2009-07-14T11:57:44 <dimazest> ThomasWaldmann: if i use regexp for dicts, should i remove dicts.py formatter?
2009-07-14T12:05:24 <ThomasWaldmann> well, if it doesn't work good enough, remove it
2009-07-14T12:06:06 <ThomasWaldmann> maybe add a link to the changeset that added it in a comment at the place where the other comments about this are (near the regex?)
2009-07-14T12:26:55 <PawelPacana> ThomasWaldmann: to which repo i have write access?
2009-07-14T12:28:56 <PawelPacana> from two: hg and newhg backends, hg will be killed by the end of the week and newhg renamed to hg
2009-07-14T12:37:09 <ThomasWaldmann> PawelPacana: did you try 2.0-storage?
2009-07-14T12:43:57 <PawelPacana> i get abort: no suitable response from remote hg!
2009-07-14T12:44:50 <CIA-78> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6064:d9e82caf4cda 2.0-storage/MoinMoin/storage/_tests/test_backends_newhg.py: removed bogus test - checked feature is in stable now
2009-07-14T12:44:52 <CIA-78> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6065:b076d216a0a3 2.0-storage/MoinMoin/storage/backends/newhg.py: (hg) refactored for mercurial 1.3 api compatibility
2009-07-14T12:45:28 * PawelPacana updated .hgrc
2009-07-14T12:47:21 <dreimark> PawelPacana: if there is yet no snippet for wikiconfig can you please add one
2009-07-14T12:51:18 <PawelPacana> some rename tests are failing, will fix it later
2009-07-14T12:51:54 <PawelPacana> dreimark: ok, where can i edit HelpOnConiguration anyway?
2009-07-14T12:52:05 <ThomasWaldmann> ah, looks like it is working now :)
2009-07-14T12:52:52 <dennda> dreimark: You mean for configuring the hg backend?
2009-07-14T12:52:59 <dreimark> PawelPacana: snippets should go into wiki/config/more_samples
2009-07-14T12:53:01 <dreimark> dennda: yes
2009-07-14T12:53:05 <dennda> ok
2009-07-14T12:53:10 <ThomasWaldmann> PawelPacana: there is no 2.0 master wiki, but you can use CHANGES, sample snippets or some wiki page
2009-07-14T12:55:22 <PawelPacana> ok
2009-07-14T12:56:49 <CIA-78> Dmitrijs Milajevs <dimazest@gmail.com> default * 4841:6241e92270b6 1.9-groups-dmilajevs/MoinMoin/ (2 files in 2 dirs): Groups2009: Indention removed in creole test for the MoinMoin.formatter.groups. Right del usage in the wiki_groups backend.
2009-07-14T12:56:51 <CIA-78> Dmitrijs Milajevs <dimazest@gmail.com> default * 4842:866fe615893f 1.9-groups-dmilajevs/MoinMoin/ (3 files in 3 dirs): Groups2009: The wiki_dicts backend uses regexp to retrieve dict definition from a page. In 2.0 this should be changed. dicts.py formatter was removed.
2009-07-14T14:08:23 <dreimark> bbl
2009-07-14T14:20:55 <ThomasWaldmann> dimazest: indentation
2009-07-14T14:21:42 <ThomasWaldmann> (although that other word exists, too, not sure about the difference)
2009-07-14T14:49:12 <dimazest> ThomasWaldmann: they are synonyms
2009-07-14T15:10:34 <dreimark> and guido's word reed the py t-short
2009-07-14T15:10:40 <dreimark> s/ee/ea/
2009-07-14T15:16:08 *** aldebrn has joined #moin-dev
2009-07-14T17:46:45 *** birkenfeld has joined #moin-dev
2009-07-14T17:46:47 <birkenfeld> TheSheep: ping
2009-07-14T17:55:44 <dennda> what do you want, intruder?
2009-07-14T18:09:05 <dreimark> he
2009-07-14T18:50:31 <dimazest> does ldap support recursive groups?
2009-07-14T18:52:07 <ThomasWaldmann> moin birkenfeld :)
2009-07-14T19:25:37 *** amartani has joined #moin-dev
2009-07-14T19:32:14 *** aldebrn has left #moin-dev
2009-07-14T19:47:33 <amartani> hi
2009-07-14T19:47:51 <dreimark> moin amartani
2009-07-14T20:09:32 <dreimark> bbl
2009-07-14T20:24:16 <dimazest> dreimark: ThomasWaldmann http://paste.pocoo.org/show/128589/ this is draft of lazy backend
2009-07-14T20:25:07 <dimazest> a little bit tricky, as first implementation
2009-07-14T20:42:45 *** devilsadvocate has quit IRC
2009-07-14T20:42:57 *** devilsadvocate has joined #moin-dev
2009-07-14T20:47:41 <ThomasWaldmann> dimazest: that stuff looks strange and not very lazy
2009-07-14T20:55:10 <dimazest> ThomasWaldmann: i need to think about it a little bit more
2009-07-14T20:59:55 <ThomasWaldmann> maybe kill that _retrieve_members in that class
2009-07-14T21:15:27 <amartani> dennda: is MoinMoin.caching going to be changed to use storage or something like this?
2009-07-14T21:17:36 <ThomasWaldmann> maybe not now, but sooner or later, we'll have to change it
2009-07-14T21:17:46 <ThomasWaldmann> why do you ask?
2009-07-14T21:18:15 <ThomasWaldmann> btw, dennda: iirc we still need tell() on item revs
2009-07-14T21:20:41 <amartani> I was thinking about moinmoin not writing anything to the filesystem, only to the selected backend, and caching now stores things on files.
2009-07-14T21:21:40 <amartani> also, some methods of caching are still expecting Page objects.
2009-07-14T21:22:08 <ThomasWaldmann> for now, just use the caching api. for later, e.g. when someone writes a GAE backend, we need to make caching more flexible
2009-07-14T21:27:56 <ThomasWaldmann> amartani: what exactly do you mean?
2009-07-14T21:29:50 *** devilsadvocate has quit IRC
2009-07-14T21:29:50 *** amartani has quit IRC
2009-07-14T21:29:50 *** ThomasWaldmann has quit IRC
2009-07-14T21:29:54 *** ThomasWaldmann has joined #moin-dev
2009-07-14T21:30:07 *** amartani has joined #moin-dev
2009-07-14T21:30:18 *** devilsadvocate has joined #moin-dev
2009-07-14T21:30:52 <ThomasWaldmann> amartani: what exactly do you mean?
2009-07-14T21:31:15 <amartani> ThomasWaldmann: about expecting Page objects?
2009-07-14T21:31:55 <ThomasWaldmann> yes
2009-07-14T21:35:51 <amartani> ThomasWaldmann: shouldn't it be converted to work only with Items?
2009-07-14T21:40:59 <ThomasWaldmann> yes, or at least something similar to items
2009-07-14T21:41:31 <ThomasWaldmann> e.g. item data is revisioned, you don't want that for cache
2009-07-14T21:43:36 <amartani> ThomasWaldmann: right now, I think it is not possible to create a CacheEntry using an Item object as arena.
2009-07-14T21:44:55 <ThomasWaldmann> what happens?
2009-07-14T21:53:35 <amartani> ThomasWaldmann: caching.py:30 no Item.page_name
2009-07-14T21:57:39 <ThomasWaldmann> looks like an easy fix
2009-07-14T22:05:13 * ThomasWaldmann fixes
2009-07-14T22:09:17 <CIA-78> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6066:efb8fcd22967 2.0-storage/MoinMoin/caching.py: caching: fix AttributeError when called with Item instance for arena param
2009-07-14T22:09:19 <CIA-78> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6067:18f4214b488c 2.0-storage/MoinMoin/storage/ (_tests/test_backends_newhg.py backends/newhg.py): merge main
2009-07-14T22:09:35 <ThomasWaldmann> amartani: ^^
2009-07-14T22:10:24 <TheSheep> birkenfeld: pong
2009-07-14T22:10:30 <amartani> ThomasWaldmann: thanks. I hadn't look into it before, didn't know it was simple.
2009-07-14T22:11:40 <ThomasWaldmann> my laptop runs out of power soon (and I left the power supply at work...)
2009-07-14T22:13:29 <ThomasWaldmann> E from mercurial.error import RepoError
2009-07-14T22:13:29 <ThomasWaldmann> > ImportError: No module named error
2009-07-14T22:13:34 <ThomasWaldmann> PawelPacana: ^^
2009-07-14T22:14:05 <ThomasWaldmann> hg 1.1.2 here
2009-07-14T22:14:30 <PawelPacana> it is correct for 1.3
2009-07-14T22:15:20 <PawelPacana> i can wrap it in try: block
2009-07-14T22:16:00 <PawelPacana> but question is: which is the first version of hg which allows empty commits in a row
2009-07-14T22:16:39 <ThomasWaldmann> no idea
2009-07-14T22:16:43 <PawelPacana> as this feature is essenttial for backend
2009-07-14T22:16:55 <ThomasWaldmann> #mercurial?
2009-07-14T22:17:40 <PawelPacana> well, good point :)
2009-07-14T22:19:52 <PawelPacana> shouldnt we target latest stable, anyway?
2009-07-14T22:23:13 <amartani> ThomasWaldmann: shouldn't it be from MoinMoin.items import Item?
2009-07-14T22:23:56 <ThomasWaldmann> hmm, maybe :)
2009-07-14T22:24:39 <ThomasWaldmann> but then likely another place that already did that check also needs that change
2009-07-14T22:26:28 <PawelPacana> ?
2009-07-14T22:27:20 <ThomasWaldmann> PawelPacana: well, if it would work with the stuff in ubuntu or debian, that wouldn't be too bad :)
2009-07-14T22:28:31 <TheSheep> PawelPacana: it broke for me in 1.3
2009-07-14T22:28:52 <TheSheep> PawelPacana: it worked fine with previous versions, so I assume they removed the empty_ok flag in 1.3
2009-07-14T22:29:29 <PawelPacana> TheSheep: but i experienced it before 1.3
2009-07-14T22:29:49 <TheSheep> I mean the commit with empty_ok broke in 1.3
2009-07-14T22:30:22 <PawelPacana> and I did some empty commits in a row about month ago
2009-07-14T22:30:25 <TheSheep> I suppose you can find the line in the hg repo and nnotate...
2009-07-14T22:30:28 <TheSheep> annotate
2009-07-14T22:30:32 <TheSheep> aka blame :)
2009-07-14T22:30:43 <PawelPacana> but installed 1.3 yesterday
2009-07-14T22:31:46 <PawelPacana> ThomasWaldmann: Jaunty has 1.1.2, that is 3 releases behind
2009-07-14T22:32:04 <PawelPacana> i guess it is the same on debian
2009-07-14T22:33:15 <PawelPacana> TheSheep: what also broke in 1.3?
2009-07-14T22:33:34 * ThomasWaldmann raise BatteryEmptyError
2009-07-14T22:33:47 <TheSheep> PawelPacana: ui.ui no longer takes arguments, you have to set the settings after creating it
2009-07-14T22:34:07 * TheSheep send some juice to ThomasWaldmann
2009-07-14T22:35:13 <amartani> PawelPacana: Debian stable has 1.0.1
2009-07-14T22:35:17 <TheSheep> http://selenic.com/repo/hg/annotate/25e9c71b89de/mercurial/localrepo.py#776
2009-07-14T22:35:28 <TheSheep> this is the revision where they dropped the files argument
2009-07-14T22:37:12 <PawelPacana> amartani: woah, thats old - empty commits were no-go that time
2009-07-14T22:39:09 <dreimark> re
2009-07-14T22:42:00 <TheSheep> http://selenic.com/repo/hg/annotate/223000a687b0/mercurial/localrepo.py#771 <-- this is the commit where they removed empty_ok
2009-07-14T22:42:25 <TheSheep> 14 May this year
2009-07-14T22:43:18 <TheSheep> the release before 1.3 was 1.2.1 in March
2009-07-14T22:43:49 <TheSheep> so 1.3 is the first release with empty_ok removed
2009-07-14T22:44:08 <TheSheep> oh, there were some changes in the locking mechanism too
2009-07-14T22:44:55 <PawelPacana> let me check empirically
2009-07-14T22:54:04 <dreimark> hmm why consumes one of the lasts test_search tests so much time
2009-07-14T23:01:39 <amartani> what is request.clock?
2009-07-14T23:07:38 <dreimark> amartani: Helper class for code profiling we do not use time.clock() as this does not work across threads This is not thread-safe when it comes to multiple starts for one timer. It is possible to recursively call the start and stop methods, you should just ensure that you call them often enough :)
2009-07-14T23:08:34 <amartani> dreimark: thanks.
2009-07-14T23:08:52 <dreimark> MM.util.clock.py
2009-07-14T23:23:19 <PawelPacana> TheSheep: eat that http://paste.pocoo.org/show/128622/ on 1.1.2
2009-07-14T23:24:34 <PawelPacana> TheSheep: what was empty_ok argument for?
2009-07-14T23:25:41 <PawelPacana> for empty commit message?
2009-07-14T23:28:21 <TheSheep> not sure
2009-07-14T23:28:32 <TheSheep> I know it made Hatta throw up
2009-07-14T23:37:24 <PawelPacana> http://paste.pocoo.org/show/128628/ <-- better, passes on 1.2.1, and 1.1.2
2009-07-14T23:46:18 <PawelPacana> uhm,that test is bogus, should check file revisions MoinMoin: MoinMoinChat/Logs/moin-dev/2009-07-14 (last edited 2009-07-13 23:30:01 by IrcLogImporter)
