MoinMoin: MoinMoinChat/Logs/moin-dev/2009-07-19

2009-07-19T00:00:36  <dreimark> hmm that self thingy should we get hidden
2009-07-19T00:14:49  *** Sepp has quit IRC
2009-07-19T00:15:53  *** Sepp has joined #moin-dev
2009-07-19T00:33:13  <CIA-36> Alexandre Martani <amartani AT gmail DOT com> default * 6154:71a224e31c14 2.0-storage-editor-amartani/MoinMoin/ (items/lock.py wikiutil.py): Real-time editor: Move lock functions to ItemLock. Call change_metadata on start of acquire method to avoid getting access denied, and avoid race condition.
2009-07-19T00:33:17  <CIA-36> Alexandre Martani <amartani AT gmail DOT com> default * 6155:da110421550a 2.0-storage-editor-amartani/MoinMoin/action/_tests/test_mobwrite.py: Real-time editor: faster test of large texts
2009-07-19T00:33:19  <CIA-36> Alexandre Martani <amartani AT gmail DOT com> default * 6156:0881863575ab 2.0-storage-editor-amartani/MoinMoin/ (_tests/test_caching.py caching.py): caching: accept str/unicode item name as arena
2009-07-19T00:33:21  <CIA-36> Alexandre Martani <amartani AT gmail DOT com> default * 6157:23fd5685f0a8 2.0-storage-editor-amartani/MoinMoin/action/mobwrite.py: Real-time editor: Don't create Item instance, just pass item name to CacheEntry
2009-07-19T00:33:24  <CIA-36> Alexandre Martani <amartani AT gmail DOT com> default * 6158:9ba803a61c51 2.0-storage-editor-amartani/MoinMoin/items/lock.py: Real-time editor: Make ItemLock use caching instead of Item metadata.
2009-07-19T00:33:25  <CIA-36> Alexandre Martani <amartani AT gmail DOT com> default * 6159:cc62b8d82ac3 2.0-storage-editor-amartani/MoinMoin/items/__init__.py: Real-time editor: Don't create storage.item for passing to ItemLock.
2009-07-19T01:01:25  <dreimark> gn
2009-07-19T01:03:04  <ThomasWaldmann> gn dreimark
2009-07-19T01:12:32  <ThomasWaldmann> http://fiber-space.de/wordpress/?p=1016
2009-07-19T01:41:02  *** Sepp has quit IRC
2009-07-19T01:53:04  *** amartani has quit IRC
2009-07-19T04:07:01  *** amartani has joined #moin-dev
2009-07-19T06:03:54  *** amartani has quit IRC
2009-07-19T06:07:18  *** amartani has joined #moin-dev
2009-07-19T07:25:09  *** amartani_ has joined #moin-dev
2009-07-19T07:27:06  *** amartani has quit IRC
2009-07-19T09:09:00  *** PawelPacana has quit IRC
2009-07-19T10:16:31  *** PawelPacana has joined #moin-dev
2009-07-19T11:36:39  <dimazest> moin
2009-07-19T11:37:57  <ThomasWaldmann> moin
2009-07-19T11:43:49  <dimazest> ThomasWaldmann: "A dict file does cache the pages which
2009-07-19T11:43:52  <dimazest> do fit to the page_group_regex variable"
2009-07-19T11:44:03  <dimazest> it is in the makecache.py
2009-07-19T11:44:19  <dimazest> what is the dict file?
2009-07-19T11:47:45  <ThomasWaldmann> i think the old wikidicts cache file is meant
2009-07-19T11:48:38  <ThomasWaldmann> just review what makecache does and fix the docstring
2009-07-19T11:49:43  <dimazest> is arena is the same as cache file?
2009-07-19T11:50:19  <ThomasWaldmann> no, it is the container
2009-07-19T11:52:11  <dimazest> i'm not sure do we need makecache for dicts/groups
2009-07-19T11:53:01  <ThomasWaldmann> we don't, but you have to fix the docstring :)
2009-07-19T11:55:39  <dreimark> we will need if if one has many groups or much dicts
2009-07-19T11:56:00  <dreimark> it is better if that cache could be recreated by cli
2009-07-19T11:56:47  <ThomasWaldmann> piece of cake
2009-07-19T11:57:23  <dreimark> moinbtw. how can i enable the wsgi debugger
2009-07-19T11:57:54  <ThomasWaldmann> pagelinks needs that, because all pages get rendered/parsed for pagelinks at once and that takes a while
2009-07-19T11:57:54  <dreimark> he found it
2009-07-19T11:58:32  <ThomasWaldmann> but the groups stuff is on-demand, you don't have that many groups and so noone will notice a delay
2009-07-19T12:00:10  <dimazest> ThomasWaldmann: i would remove mention of dicts from makecache
2009-07-19T12:01:03  <dimazest> http://paste.pocoo.org/show/129505/
2009-07-19T12:02:03  <dimazest> ThomasWaldmann: do we have dict file? I can't understand what is dict file
2009-07-19T12:03:39  <ThomasWaldmann> you killed it and replaced it by more sane stuff
2009-07-19T12:04:57  <ThomasWaldmann> paste -> ok
2009-07-19T12:12:28  * ThomasWaldmann needs some feedback on i18n.strings page sets. does someone have a better idea? see also the output of "make pagepacks" on LanguageSetup page
2009-07-19T12:14:16  <dreimark> is 'not_translated_system_pages' because of yet not translated?
2009-07-19T12:15:21  <dreimark> my fault should read
2009-07-19T12:17:54  <dreimark> s/not_translated_system_pages/supplies/
2009-07-19T12:23:26  <ThomasWaldmann> i rather mean the set count and the set contents
2009-07-19T12:25:21  <dreimark> looking at the list I was confused about not_translated_system_pages, translated_system_pages
2009-07-19T12:27:06  * ThomasWaldmann verified that 1.9 does not leak files
2009-07-19T12:27:13  *** PawelPacana has quit IRC
2009-07-19T12:31:37  <dreimark> HelpOnUserPreferences may be moved to essential_help_pages
2009-07-19T12:33:13  <dreimark> HelpOnLogin TOO
2009-07-19T12:35:37  <dreimark> HelpOnLanguages too
2009-07-19T12:36:03  <ThomasWaldmann> essential == it does not work / is not usable without it
2009-07-19T12:36:06  *** PawelPacana has joined #moin-dev
2009-07-19T12:36:44  <PawelPacana> moin
2009-07-19T12:37:04  <dreimark> in the past we have called HelpOnLanguages as Frontpage
2009-07-19T12:37:54  <dennda> It's LanguageSetup now
2009-07-19T12:38:13  <dreimark> we discuss 1.9
2009-07-19T12:38:40  * dennda shuts up
2009-07-19T12:38:42  <dreimark> ok found it it is in not_translated_system_pages
2009-07-19T12:38:46  <dreimark> sorry dennda
2009-07-19T12:38:55  <dennda> no need to be sorry, my bad :-)
2009-07-19T12:39:11  <dreimark> you were right I guess
2009-07-19T12:39:40  <dreimark> essential from the point of view of a user is different than from the technical side
2009-07-19T12:42:26  <ThomasWaldmann> essential does not mean comfortable or complete
2009-07-19T12:42:55  <ThomasWaldmann> it is the bare minimum. if you like to have more, you install more.
2009-07-19T12:44:11  <ThomasWaldmann> and i renamed it to LanguageSetup because that page won't be translated (unlike the help pages)
2009-07-19T12:44:29  <ThomasWaldmann> and it is the only page pre-installed in moin
2009-07-19T12:44:53  <dreimark> ok
2009-07-19T13:04:21  <dreimark> dimazest: I am not sure if we should rename self to dummy in multiconfig. "self" looks like strange.
2009-07-19T13:12:22  <dreimark> http://google.com/codesearch?hl=de&lr=&q=lang:python+"lambda+self:"&sbtn=Suche
2009-07-19T13:12:28  <dimazest> i do not know is it better or not
2009-07-19T13:12:29  <dreimark> hmm self wins
2009-07-19T13:13:06  <dreimark> 7.000 times used versa 77 dummy
2009-07-19T13:13:45  <dimazest> i think it should be self, because first method argument must be named self, isn't it?
2009-07-19T13:14:12  <dreimark> yes
2009-07-19T13:14:22  <dimazest> and there we define method (ok, using lambda, but still)
2009-07-19T13:22:59  <ThomasWaldmann> self > cfg
2009-07-19T13:24:36  <dimazest> ThomasWaldmann: i'm not sure that it is good idea to rename self to cfg
2009-07-19T13:24:48  <dreimark> only in multiconfig.py
2009-07-19T13:27:14  <dreimark> cfgwas used twice before in ggs codesaerch
2009-07-19T13:27:54  <ThomasWaldmann> dimazest: well, it is easier understandable
2009-07-19T13:28:36  <dimazest> ok
2009-07-19T13:31:48  <CIA-36> Dmitrijs Milajevs <dimazest@gmail.com> default * 4860:0437b25ee607 1.9-groups-dmilajevs/MoinMoin/script/maint/ (cleancache.py makecache.py): Groups2009: Docstring update for script.maint.cleancache and script.maint.makecache.
2009-07-19T13:31:51  <CIA-36> Dmitrijs Milajevs <dimazest@gmail.com> default * 4861:e3773a70a520 1.9-groups-dmilajevs/MoinMoin/formatter/groups.py: Groups2009: formatter.groups refactoring. Docstring improvement, some attributes of the formatter became internal.
2009-07-19T13:31:53  <CIA-36> Dmitrijs Milajevs <dimazest@gmail.com> default * 4862:2f087d85e97d 1.9-groups-dmilajevs/docs/CHANGES.dmilajevs: Groups2009: CHANGES update.
2009-07-19T13:31:55  <CIA-36> Dmitrijs Milajevs <dimazest@gmail.com> default * 4863:c9eedd899bb7 1.9-groups-dmilajevs/MoinMoin/ (57 files in 5 dirs): merged moin/1.9
2009-07-19T13:33:59  *** PawelPac1na has joined #moin-dev
2009-07-19T13:33:59  *** PawelPacana has quit IRC
2009-07-19T13:35:30  <dimazest> ThomasWaldmann: http://paste.pocoo.org/show/129508/ like this?
2009-07-19T13:51:24  *** PawelPac1na has quit IRC
2009-07-19T13:54:57  *** PawelPacana has joined #moin-dev
2009-07-19T13:55:54  <ThomasWaldmann> yes
2009-07-19T13:56:31  <ThomasWaldmann> also fix it in the docstring
2009-07-19T13:57:39  <ThomasWaldmann> and don't you need to encapsulate the default value so it shows up in HelpOnConfiguration correctly?
2009-07-19T13:58:10  *** grzywacz has quit IRC
2009-07-19T14:01:41  *** PawelPac1na has joined #moin-dev
2009-07-19T14:02:51  *** grzywacz has joined #moin-dev
2009-07-19T14:04:09  <ThomasWaldmann> btw, the reference to HelpOnDictionaries is not very helpful there as it has no info about configuration.
2009-07-19T14:04:44  <ThomasWaldmann> (and it should not have, help pages are primarily for using features, not for configuring)
2009-07-19T14:10:53  <ThomasWaldmann> dimazest: you didn't update your diary since a week
2009-07-19T14:11:33  <ThomasWaldmann> dreimark: as a mentor, you should keep an eye on that
2009-07-19T14:14:57  <dreimark> ThomasWaldmann: I've requested it already
2009-07-19T14:15:44  <ThomasWaldmann> also, the page contents need updating in general
2009-07-19T14:16:04  <dreimark> 19:37 < dimazest> dreimark: i still need to update diary, i remember about it
2009-07-19T14:16:21  <dreimark> but that was on 17th
2009-07-19T14:16:29  <ThomasWaldmann> I think the snippets should all live in wiki/config/...
2009-07-19T14:16:55  <ThomasWaldmann> the api docs could be moved to the appropriate module docstrings
2009-07-19T14:17:37  <ThomasWaldmann> current plan needs update
2009-07-19T14:18:15  <ThomasWaldmann> plan for next week needs to be made
2009-07-19T14:18:43  <dreimark> nice feature that one could now define CompositeGroups(request, WikiGroups(request), ConfigGroups(request, self.first_backend_groups)
2009-07-19T14:19:06  <dreimark> so you can setup some basic groups and one could overwrite them in the wiki
2009-07-19T14:19:56  <ThomasWaldmann> i still don't think you should define same groups at 2 places, that is just confusing
2009-07-19T14:20:48  <ThomasWaldmann> dreimark: about sqlagroups (still mentioned on the project page) - we decided to drop that, right?
2009-07-19T14:22:37  <ThomasWaldmann> dennda: still waiting for you plan, btw.
2009-07-19T14:22:40  <ThomasWaldmann> +r
2009-07-19T14:24:35  <ThomasWaldmann> dimazest: """Review HelpOnGroups on the master19.moinmo.in""" there is no such page.
2009-07-19T14:28:29  <ThomasWaldmann> dimazest: http://master19.moinmo.in/HelpOnAccessControlLists#Groups this is currently the only docs for groups
2009-07-19T14:30:59  <ThomasWaldmann> maybe the part about how to define groups should be split off to a HelpOnGroups page, but HelpOnAccessControlLists should still contain infos about group usage in ACLs plus a link to HelpOnGroups then.
2009-07-19T14:31:17  <dennda> I'll have some time on thursday evening and friday at least, I think
2009-07-19T14:32:15  <dreimark> ThomasWaldmann: yes sqla is dropped
2009-07-19T14:32:29  * dreimark edits the page
2009-07-19T14:49:20  <dreimark> ThomasWaldmann: new repo for xapian refactoring?
2009-07-19T15:12:23  <dreimark> dimazest: please add your means and ideas to http://moinmo.in/Xapian2009
2009-07-19T15:23:08  *** PawelPac1na has quit IRC
2009-07-19T15:25:12  *** PawelPac1na has joined #moin-dev
2009-07-19T15:39:40  <ThomasWaldmann> http://master19.moinmo.in/HelpOnGroups please review / improve
2009-07-19T15:40:34  <ThomasWaldmann> i am still a bit undecided about xapian. so lets first see some plan. :)
2009-07-19T15:40:56  *** PawelPac1na has left #moin-dev
2009-07-19T15:52:01  <CIA-36> Dmitrijs Milajevs <dimazest@gmail.com> default * 4864:eabd8725c940 1.9-groups-dmilajevs/MoinMoin/config/multiconfig.py: Groups2009: self changed to cfg in groups and dicts initializing functions.
2009-07-19T15:52:04  <CIA-36> Dmitrijs Milajevs <dimazest@gmail.com> default * 4865:f486af162cdf 1.9-groups-dmilajevs/MoinMoin/datastruct/backends/ (4 files): Groups2009: Doctstrings to the datastruct module were added. Copyright notice cleanup.
2009-07-19T15:55:44  *** PawelPacana has quit IRC
2009-07-19T15:55:47  *** PawelPacana has joined #moin-dev
2009-07-19T15:59:54  <dreimark> Due to that first match algorithm, you must sort your ACLs: first single usernames, then special groups, then more general groups, then Known and at last All.
2009-07-19T16:00:15  <dreimark> me has already replaced should by must
2009-07-19T16:01:06  <dreimark> the only special group left over in that list description is Trusted.
2009-07-19T16:01:55  <ThomasWaldmann> ?
2009-07-19T16:02:07  <dreimark> I wonder currently if it matters to do
2009-07-19T16:02:57  <dreimark> #acl SomeGroup:read,write SomeUser:admin,write,read Trusted:read
2009-07-19T16:05:48  <ThomasWaldmann> that might do something you do not want
2009-07-19T16:06:49  <dennda> Especially if SomeUser is in SomeGroup, afair
2009-07-19T16:07:03  <ThomasWaldmann> yes
2009-07-19T16:08:26  <ThomasWaldmann> about xapian:
2009-07-19T16:09:07  <ThomasWaldmann> xappy, flax, working in xapian in mm2.0
2009-07-19T16:09:51  <ThomasWaldmann> (xappy = getting rid of dead xapwrap)
2009-07-19T16:10:34  <ThomasWaldmann> that are things influencing what we should do
2009-07-19T16:14:13  <dreimark> it makes sense to refactor the current xapian code in 2.0 because it will be done anyway
2009-07-19T16:14:27  <dreimark> and work in 1.9 will get lost
2009-07-19T16:18:17  <dreimark> .oO the acl line seems to work
2009-07-19T16:18:40  <ThomasWaldmann> all acl lines should work :)
2009-07-19T16:18:59  <dreimark> the one from the example which doesn't fit the description
2009-07-19T16:19:13  <ThomasWaldmann> ?
2009-07-19T16:19:50  <dreimark> 16:02 < dreimark> #acl SomeGr...
2009-07-19T16:20:30  <ThomasWaldmann> that line does exactly what the specification tells it does, but likely not what the author wanted
2009-07-19T16:20:41  <dreimark> the description tells #acl User... Trusted... SomeGroup...
2009-07-19T16:21:33  <dennda> Well you have to be aware that it's first match
2009-07-19T16:23:05  <ThomasWaldmann> dreimark: what description?
2009-07-19T16:24:03  <dreimark> http://master19.moinmo.in/HelpOnAccessControlLists#Processing_logic_on_a_single_page
2009-07-19T16:24:10  <dreimark> the info line
2009-07-19T16:26:14  <ThomasWaldmann> i don't see any problem
2009-07-19T16:26:18  *** PawelPacana has quit IRC
2009-07-19T16:26:28  *** PawelPacana has joined #moin-dev
2009-07-19T16:27:07  <dreimark> dennda: ThomasWaldmann must instead of should is ok
2009-07-19T16:27:53  <ThomasWaldmann> well, it is not required that acl lines make sense
2009-07-19T16:28:02  <dreimark> I got this now too
2009-07-19T16:28:05  <ThomasWaldmann> even nonsense will be acceptable to the system
2009-07-19T16:28:37  <ThomasWaldmann> but if you define "must" as "you must do it that way so it makes sense", then it is ok
2009-07-19T16:29:13  <dreimark> yes
2009-07-19T16:29:49  <ThomasWaldmann> and i likely intentionally left out Trusted, because this is a complicated issue in itself.
2009-07-19T16:31:04  <dreimark> usually I don't use Trusted and Known
2009-07-19T16:42:03  <dimazest> dreimark: ThomasWaldmann i need to go now
2009-07-19T16:42:38  <dimazest> and i'm not sure that i will be here at 18:00
2009-07-19T16:43:03  <dimazest> here are my 3 status lines
2009-07-19T16:43:05  <dimazest> LAST: Lazy backend was implemented. Configuration update (got rid of *_manager_init). Configuration snippets for groups and dicts. Various docstring/code improvements.
2009-07-19T16:43:08  <dimazest> NEXT: Minor Groups/dicts code changes, if needed. Groups/dicts documentation improvements.  Look for xapian related code. Start xapian project wiki page.
2009-07-19T16:43:11  <dimazest> BLOCKERS: no
2009-07-19T16:43:58  <dreimark> dimazest: please read the backlog and we need to discuss with ThomasWaldmann later where xapian related stuff best would be done
2009-07-19T16:44:06  <dreimark> also please fix the missing diary
2009-07-19T16:44:33  <dreimark> on some days it is enough to write EP2009
2009-07-19T16:46:57  <dimazest> dreimark: i've updated diary for some days, based on commits
2009-07-19T16:47:30  <dreimark> ok
2009-07-19T16:47:56  <dreimark> I've started a http://moinmo.in/Xapian2009
2009-07-19T16:48:05  <dimazest> Wednesday and Thursday i was thinking on lazy backend and not commiting, so on Friday there was bunch of commits
2009-07-19T16:48:48  * ThomasWaldmann found a rev.read()
2009-07-19T16:49:03  <dimazest> ok see you, bbl
2009-07-19T16:49:09  * ThomasWaldmann imagines putting a dvd.iso in the wiki...
2009-07-19T16:49:10  <dreimark> just ok vu
2009-07-19T16:49:19  <dreimark> cu
2009-07-19T16:50:17  <ThomasWaldmann> btw, that rather sounds like 1 day than 1 week
2009-07-19T16:53:05  <dreimark> dimazest: superuser seems to be broken in your repo
2009-07-19T16:53:42  <dreimark> it seems to work to change to a user but any new request switches back
2009-07-19T16:54:36  <dreimark> (me tried a new repo)
2009-07-19T17:01:07  <ThomasWaldmann> dreimark: does it work in 1.9?
2009-07-19T17:05:37  <dreimark> yes
2009-07-19T17:09:50  *** amartani_ has quit IRC
2009-07-19T17:18:28  <CIA-36> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6080:9f31d8005397 2.0-storage/MoinMoin/storage/backends/newhg.py: (hg) module docstring updated
2009-07-19T17:18:30  <CIA-36> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6081:1de0678c44c0 2.0-storage/MoinMoin/storage/backends/newhg.py: (hg) proper encoding setting
2009-07-19T17:18:32  <CIA-36> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6082:0a1cf7a857bf 2.0-storage/MoinMoin/storage/ (3 files in 3 dirs): (hg) killed deprecated version of hg backend
2009-07-19T17:18:34  <CIA-36> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6083:f04ff609b02e 2.0-storage/MoinMoin/storage/ (4 files in 2 dirs): (hg) renamed newhg to hg
2009-07-19T17:25:54  <ThomasWaldmann> dimazest: http://master19.moinmo.in/HelpOnDictionaries http://master19.moinmo.in/HelpOnGroups
2009-07-19T17:26:10  <ThomasWaldmann> PawelPacana: thanks for cleaning up :)
2009-07-19T17:27:25  * ThomasWaldmann pulls and runs tests
2009-07-19T17:29:19  <ThomasWaldmann> dennda: devilsadvocate: dimazest: dreimark: waldi: xorAxAx: (and amartani who seems to be away): meeting in 30 mins
2009-07-19T17:29:46  * dennda has not that much to say
2009-07-19T17:30:05  <ThomasWaldmann> then you are rather late with planning
2009-07-19T17:30:18  <dennda> ?
2009-07-19T17:30:54  <ThomasWaldmann> (for you NEXT: :)
2009-07-19T17:31:40  <dennda> Well let me rephrase: Except telling you that I'll have some time on thursday evening / friday I don't have that much to say :-)
2009-07-19T17:32:30  <ThomasWaldmann> you should have a plan about what to do on thursday/friday
2009-07-19T17:35:53  <ThomasWaldmann> your diary is not updated since a week
2009-07-19T17:40:26  <PawelPacana> ThomasWaldmann: more to come soon, summed up checklist of incoming changes is here: http://moinmo.in/PawelPacana/MercurialBackend/5
2009-07-19T17:42:04  <ThomasWaldmann> dennda: (18:08) <    dennda> Next: Finnish ACL privs finally; Improve docs; Working System Pages
2009-07-19T17:42:28  <dennda> That still holds true...
2009-07-19T17:45:09  <ThomasWaldmann> well, maybe because you did not work on it enough?
2009-07-19T17:46:05  <dennda> It's not like I was on vacation
2009-07-19T17:46:28  <ThomasWaldmann> i know, I am rather talking about time managment
2009-07-19T17:47:25  <ThomasWaldmann> as I see you have quite some exams in those 2 weeks. but that should not mean that SOC works almost comes to a halt for 2 weeks.
2009-07-19T17:49:49  <dennda> Well that's why I've been trying to stick around here. Don't want you to think that I've gone away for no good reason. It's just not easy to make sure everyone's happy (especially if my GPA is concerned)
2009-07-19T17:50:11  <dennda> Can we just say that I will work twice as hard in the time remaining? :-)
2009-07-19T17:50:38  <ThomasWaldmann> when do lectures stop?
2009-07-19T17:50:47  <dennda> next wednesday
2009-07-19T17:50:51  *** PawelPacana has quit IRC
2009-07-19T17:50:54  *** PawelPacana has joined #moin-dev
2009-07-19T17:51:02  <dennda> is my last exam and then it's all over
2009-07-19T17:51:24  <dennda> err
2009-07-19T17:51:25  <dennda> no
2009-07-19T17:51:26  <dennda> thursday
2009-07-19T17:51:28  <dennda> 30th
2009-07-19T17:52:14  <dennda> lectures end on monday
2009-07-19T17:52:50  <ThomasWaldmann> when looking at your home page, it tells you will be rather busy until the end of the month
2009-07-19T17:53:26  <ThomasWaldmann> and after that there is just 2 weeks SOC left - that is just not enough if you more or less vanish for 2 weeks now.
2009-07-19T17:55:17  <ThomasWaldmann> thus, I think you should plan your days wisely and the plan should include some time for SOC
2009-07-19T17:56:13  <dennda> If we could come up with a very detailed list of what things need to be done in what order, that would help to allow me to work on it whenever there's time in between.
2009-07-19T17:56:28  <dennda> (More detailed than what you asked for, which is basically acl privs, system pages)
2009-07-19T17:57:29  <CIA-36> Bastian Blank <bblank@thinkmo.de> default * 6615:b597a6704271 2.0-storage-dom-bblank/MoinMoin/converter2/ (_tests/test_moinwiki_in.py moinwiki_in.py): Moin Wiki input converter - Support numeric entities
2009-07-19T17:57:44  <ThomasWaldmann> hmm, didn't we already talk about how to handle sys pages?
2009-07-19T17:58:14  <ThomasWaldmann> didn't I already tell how to improve interbackend rename?
2009-07-19T17:58:35  <dennda> Yes, I should make a list out of it
2009-07-19T18:00:09  <ThomasWaldmann> ok, meeting time :)
2009-07-19T18:00:12  <xorAxAx> good evening, welcome to the meetin
2009-07-19T18:00:13  <dreimark> it's time
2009-07-19T18:00:15  <xorAxAx> first newsitem for today: we have one student less. my student, devilsadvocate, unfortunately didnt pass the midterm evals
2009-07-19T18:00:17  <ThomasWaldmann> :)
2009-07-19T18:00:38  <xorAxAx> the remaining students may paste now
2009-07-19T18:00:53  <ThomasWaldmann> (that :) was about the parallel postings of "we have meeting time")
2009-07-19T18:01:02  <ThomasWaldmann> dennda: please begin
2009-07-19T18:01:59  <dennda> LAST: As already discussed I'm in the hot phase of this semester. The most important exams are coming up and I have hence been mostly busy with learning/doing excercises. I've not done major coding.
2009-07-19T18:02:27  <dennda> NEXT: I'll provide a detailed plan in the next hour of times when I'll be available for coding and list items that I'll do whenever there's time.
2009-07-19T18:02:33  <dennda> BLOCKERS: None
2009-07-19T18:03:00  <ThomasWaldmann> ok, thanks. maybe cleanup the project page while doing so.
2009-07-19T18:03:15  <dennda> .oO(didn't I do that lately? will check)
2009-07-19T18:03:39  <ThomasWaldmann> waldi: do you know where your student is?
2009-07-19T18:03:47  <dreimark> dimazest told already at 16:43
2009-07-19T18:04:12  <waldi> ThomasWaldmann: no
2009-07-19T18:04:20  <xorAxAx> bad communication
2009-07-19T18:04:36  <ThomasWaldmann> ok, as he is not here, waldi can you tell a bit about the project status?
2009-07-19T18:05:11  <xorAxAx> dennda: please paste your lines all at once next time
2009-07-19T18:05:30  <waldi> he tried to do some work on the caching and acl checks
2009-07-19T18:06:11  <ThomasWaldmann> and what was the outcome? :)
2009-07-19T18:07:18  <waldi> i can only comment on the commits, had no time to check them yet
2009-07-19T18:07:29  *** amartani has joined #moin-dev
2009-07-19T18:07:31  <waldi> according to him, it works
2009-07-19T18:07:35  <waldi> ah, there he is
2009-07-19T18:07:46  <ThomasWaldmann> please check it in practice
2009-07-19T18:07:51  <ThomasWaldmann> hi amartani :)
2009-07-19T18:07:52  <amartani> sorry
2009-07-19T18:07:57  <amartani> hi
2009-07-19T18:08:12  <ThomasWaldmann> you just came in the right moment, post your status please
2009-07-19T18:10:09  <ThomasWaldmann> amartani: ^^
2009-07-19T18:10:28  <xorAxAx> he is probably preparing it
2009-07-19T18:10:37  <amartani> yes, sorry.
2009-07-19T18:10:41  <xorAxAx> we should give failure points for late preparation
2009-07-19T18:10:47  <ThomasWaldmann> :P
2009-07-19T18:11:30  * ThomasWaldmann .oO( xorAxAx is experienced with failure points now )
2009-07-19T18:11:44  <xorAxAx> yeah, i overflowed my internal counter
2009-07-19T18:11:50  <xorAxAx> while countin
2009-07-19T18:12:58  <dreimark> -points per minute
2009-07-19T18:13:18  <ThomasWaldmann> xorAxAx: if you have time, maybe have a look at the xapian stuff also (dimazest's near-future todo)
2009-07-19T18:13:49  <xorAxAx> ok
2009-07-19T18:13:51  <ThomasWaldmann> xorAxAx: esp. if you have followed what exactly johill wanted to do with the search
2009-07-19T18:14:14  <amartani> LAST: Rewrite locking system to work with item and using caching.
2009-07-19T18:14:16  <xorAxAx> no, i didnt
2009-07-19T18:15:02  <amartani> NEXT: Make some things like "save and continue" work outside mobwrite so other parts can use it also. Also, there is some strange bugs I would like to look deeper.
2009-07-19T18:15:10  <amartani> BLOQUERS: None.
2009-07-19T18:16:27  <ThomasWaldmann> any questions?
2009-07-19T18:16:49  <ThomasWaldmann> amartani: did you do some practical test with 3 (or so) concurrent editors?
2009-07-19T18:16:54  <dreimark> strange bugs aren't blockers?
2009-07-19T18:17:31  <ThomasWaldmann> blockers means that you can't continue working because you do not know what to do
2009-07-19T18:17:57  <amartani> I have commented about it here, sometimes I get a werkzeug error.
2009-07-19T18:18:07  <ThomasWaldmann> (or that you'ld have to wait for sth not in your scope)
2009-07-19T18:18:31  <ThomasWaldmann> amartani: did you also ask mitsuhiko?
2009-07-19T18:18:56  <amartani> ThomasWaldmann: Yes, it usually works well.
2009-07-19T18:19:09  <amartani> ThomasWaldmann: no, who is mitsuhiko?
2009-07-19T18:19:31  <ThomasWaldmann> == Armin Ronacher aka "Mr. Werkzeug"
2009-07-19T18:19:32  <dreimark> he is the author of werkzeug
2009-07-19T18:19:52  <ThomasWaldmann> usually seen on #pocoo
2009-07-19T18:20:38  <amartani> ThomasWaldmann: thanks, I'll check with him.
2009-07-19T18:20:49  <ThomasWaldmann> amartani: btw, how is communication with mobwrite author?
2009-07-19T18:22:07  <dennda> (are we in the FFA discussion phase of the meeting?)
2009-07-19T18:22:30  <amartani> he is not answering on mailing lists, there was even some posts from bespin devs about their changes to mobwrite.
2009-07-19T18:23:03  <ThomasWaldmann> yes, the official status part is done. feel free to ask anything. :)
2009-07-19T18:23:37  <ThomasWaldmann> hmm, pity, would've been great to see some more collaboration
2009-07-19T18:23:52  <dennda> ThomasWaldmann: Ok, so just to make sure: The syspages thing is not about providing a fully tested solution that persists even after the sun explodes, but more about providing a framework and adding information to syspages that is necessary to handle them sanely later?
2009-07-19T18:24:54  <amartani> he probably is having to work on something else. he even said more than a month ago about features comming "next week"
2009-07-19T18:25:00  <ThomasWaldmann> the code is currently still having quite some references to "underlay" (which is pointless because there is no underlay)
2009-07-19T18:25:15  <dennda> ok so all of that needs to be ripped out as well?
2009-07-19T18:25:29  <ThomasWaldmann> and we need some means to install system pages and handle them for upgrades
2009-07-19T18:25:47  <dennda> yes that'd be my next question
2009-07-19T18:25:54  <dennda> how would the UI for that look like?
2009-07-19T18:25:58  <dennda> CLI script?
2009-07-19T18:26:32  <ThomasWaldmann> there needs to be some core code, optionally some script interface, some web ui
2009-07-19T18:26:51  <ThomasWaldmann> see install packages etc.
2009-07-19T18:26:53  <dennda> Where would that core code fit in?
2009-07-19T18:27:59  <ThomasWaldmann> it is called by script and by action
2009-07-19T18:28:28  <ThomasWaldmann> amartani: ok, let's see
2009-07-19T18:29:03  <dennda> MoinMoin.storage.upgrade_syspages("path/to/new/syspages/package.xml")?
2009-07-19T18:29:31  <ThomasWaldmann> yes
2009-07-19T18:30:01  <ThomasWaldmann> and before that, we could even fix the tests maybe (so they have sys pages when needed)
2009-07-19T18:30:36  <dennda> tbh, I don't know how that works/should be fixed
2009-07-19T18:31:59  <ThomasWaldmann> problem is that backend is empty
2009-07-19T18:32:19  <ThomasWaldmann> some tests are ok with that, some others expect ~4 pages in there
2009-07-19T18:33:33  <dennda> didn't you say you were already working on that?
2009-07-19T18:33:48  <ThomasWaldmann> yes, I was, but it didn't work as expected
2009-07-19T18:33:54  <xorAxAx> why xml and not package pages?
2009-07-19T18:34:17  <ThomasWaldmann> xorAxAx: the new page packages will be xml
2009-07-19T18:34:33  <ThomasWaldmann> because the old format was rather limited
2009-07-19T18:34:52  <ThomasWaldmann> ehrm "item packages"
2009-07-19T18:34:55  <dennda> http://moinmo.in/4ct10n/info/Storage2009?action=diff&rev2=62&rev1=61
2009-07-19T18:36:29  <ThomasWaldmann> ok
2009-07-19T18:37:03  <ThomasWaldmann> dennda: idea for syspage unserialization: maybe it is easier to use a temp memorybackend as unserialize target first
2009-07-19T18:37:17  <ThomasWaldmann> (== use unmodified unserialize code)
2009-07-19T18:37:32  <dennda> why?
2009-07-19T18:37:39  <ThomasWaldmann> and then copy revisions from temp to real
2009-07-19T18:38:08  <ThomasWaldmann> because then you don't have to modify the unserializer
2009-07-19T18:38:22  <dennda> didn't look at it
2009-07-19T18:38:27  <dennda> did you hardcode membackend?
2009-07-19T18:38:33  <ThomasWaldmann> no
2009-07-19T18:38:54  <ThomasWaldmann> but the unserializer currently is only happy with empty backend
2009-07-19T18:39:01  <dennda> ah
2009-07-19T18:39:05  <dennda> ok
2009-07-19T18:39:19  <ThomasWaldmann> but maybe that just should be fixed :)
2009-07-19T18:39:34  <dennda> we can safely defer that
2009-07-19T18:40:42  <ThomasWaldmann> yes, first some "kill underlay" cleanup, add some metadata k/v, adjust all code to use that instead of underlay detection
2009-07-19T18:41:15  <ThomasWaldmann> (or even i18n.strings lists)
2009-07-19T18:43:49  <ThomasWaldmann> dennda: btw, you can test tell() with PIL (see Image item ops)
2009-07-19T18:43:50  *** PawelPacana has quit IRC
2009-07-19T18:44:27  <dennda> you mean "test" as in "testcases"?
2009-07-19T18:44:39  <ThomasWaldmann> no
2009-07-19T18:44:51  <ThomasWaldmann> test as in "it works in real life"
2009-07-19T18:45:11  <dennda> I know what tell does
2009-07-19T18:45:15  <dennda> it just tells where you are, no?
2009-07-19T18:45:22  <ThomasWaldmann> yes
2009-07-19T18:45:55  *** PawelPacana has joined #moin-dev
2009-07-19T18:48:38  *** PawelPacana has quit IRC
2009-07-19T18:52:55  *** amartani_ has joined #moin-dev
2009-07-19T19:06:04  <CIA-36> Bastian Blank <bblank@thinkmo.de> default * 6616:83b08105d234 2.0-storage-dom-bblank/MoinMoin/converter2/_tests/test_moinwiki_in.py: Tests for Moin Wiki input converter - Cleanup
2009-07-19T19:06:05  <CIA-36> Bastian Blank <bblank@thinkmo.de> default * 6617:e667c9b52bbb 2.0-storage-dom-bblank/MoinMoin/converter2/_tests/test_moinwiki_in.py: Tests for Moin Wiki input converter - Use unicode strings
2009-07-19T19:06:08  <CIA-36> Bastian Blank <bblank@thinkmo.de> default * 6618:100cb4ef2bf0 2.0-storage-dom-bblank/MoinMoin/converter2/ (_tests/test_moinwiki_in.py moinwiki_in.py): Moin Wiki input converter - Support symbolic entities
2009-07-19T19:06:10  <CIA-36> Bastian Blank <bblank@thinkmo.de> default * 6619:995852c16842 2.0-storage-dom-bblank/MoinMoin/converter2/moinwiki_in.py: Moin Wiki input converter - Cleanup regular expressions
2009-07-19T19:06:58  *** amartani has quit IRC
2009-07-19T19:41:04  <CIA-36> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6084:2b5e651d72c1 2.0-storage/MoinMoin/storage/serialization.py: serialization: provide unserialize function, improve docstrings
2009-07-19T19:51:46  *** PawelPacana has joined #moin-dev
2009-07-19T19:55:00  <ThomasWaldmann> hmm, i have conflicting ideas for unserialize
2009-07-19T19:55:59  <ThomasWaldmann> do we ever want unserialize to write to less-than-current+1 revision?
2009-07-19T19:56:25  <dreimark> you mean to the past revision ?
2009-07-19T19:56:36  <dreimark> an old one
2009-07-19T19:56:52  <ThomasWaldmann> yes (esp if it is not there [any more])
2009-07-19T19:58:16  <dreimark> I think we don't want. this seems to give some problems similiar to a wrong systemtime on a linux box which goes into the past
2009-07-19T19:58:23  <ThomasWaldmann> e.g. if we have an item with 1000 revs, we could serialize 0..900, store the xml elsewhere and destroy those revs
2009-07-19T19:59:02  <dreimark> for elsewhere it is ok, but not for overwriting the master
2009-07-19T19:59:31  <dreimark> serialize 0..900 into a package item and destroying afterwards is similiar
2009-07-19T19:59:32  <ThomasWaldmann> in that case, a valid usecase would be to restore the xml to the wiki
2009-07-19T19:59:55  <ThomasWaldmann> (as past revisions, not as new revisions)
2009-07-19T20:00:20  <dreimark> that won't give past editlog eventlog entries
2009-07-19T20:01:05  <ThomasWaldmann> editlog is in revision metadata, so it would
2009-07-19T20:01:10  <dreimark> ok
2009-07-19T20:01:16  <ThomasWaldmann> (and I don't care much about eventlog)
2009-07-19T20:02:01  <dreimark> the first revision in that case after destroying is 901 or?
2009-07-19T20:02:08  <ThomasWaldmann> hmm, i guess i need 2 modes: keep revno as in xml, try to write to exactly that revision
2009-07-19T20:02:21  <ThomasWaldmann> and just write to next free rev
2009-07-19T20:02:32  <dreimark> because if you link to revisions you do have the same problem as for rename page if that is changed to 1
2009-07-19T20:03:00  <ThomasWaldmann> i don't think we would renumber
2009-07-19T20:03:04  <dennda> that allows messing with the past!
2009-07-19T20:03:20  <dennda> (and as we all know from countles tv series and movies, that almost always goes wrong) :-)
2009-07-19T20:03:39  <ThomasWaldmann> destroy is also messing with the past :P
2009-07-19T20:03:54  <dennda> well yeah but at least then it's obvious that someone messed
2009-07-19T20:04:00  <dreimark> well that is quite similiar to a routing of 900 revisions by a different backend then
2009-07-19T20:04:07  <dreimark> and merging into an other one
2009-07-19T20:04:27  <dennda> you can't route revisions
2009-07-19T20:04:30  <dennda> you route items
2009-07-19T20:04:39  <dreimark> mercurial does
2009-07-19T20:04:50  <dreimark> or we do with merge
2009-07-19T20:05:25  <dreimark> yes we can't do it by the current code but you could do it manually or?
2009-07-19T20:05:42  <dreimark> you could decide at some revision to use a different backend or?
2009-07-19T20:05:51  <dennda> do what? manipulate old revision's contents/metadata?
2009-07-19T20:06:18  <dreimark> cycle space
2009-07-19T20:06:41  <dennda> you lost me
2009-07-19T20:07:19  <dreimark> I meant manually I can do make a copy of the last revision of an item into a new backend
2009-07-19T20:07:52  <dreimark> while the other still exists in the other backend
2009-07-19T20:08:36  <dennda> for what reason?
2009-07-19T20:08:50  <dennda> you'd need a new item in the new backend
2009-07-19T20:13:11  <dreimark> if you don't want to destroy the items you also could create an item.package of the oldest revision (compressed data) and have that in place instead of 900 revisions
2009-07-19T20:14:09  <ThomasWaldmann> hmm, we also have a problem with updating existing items: item metadata is not revisioned
2009-07-19T20:14:10  <dreimark> ThomasWaldmann: I get in mind a different kind of package
2009-07-19T20:14:47  <ThomasWaldmann> so, if we restore old revs, we likely do not want to overwrite item metadata.
2009-07-19T20:14:52  <dreimark> can we have a package which is not necessarily unpacked to access its content?
2009-07-19T20:15:00  <ThomasWaldmann> but do we want that if we write new revs?
2009-07-19T20:15:18  <ThomasWaldmann> what item metadata do we have, btw?
2009-07-19T20:16:01  * ThomasWaldmann .oO(can I take a shower without getting wet?)
2009-07-19T20:16:08  <dreimark> hehe
2009-07-19T20:17:06  <dreimark> may be may wish is more like move to a different backend
2009-07-19T20:17:24  <dennda> that's possible
2009-07-19T20:17:29  <dennda> (easily)
2009-07-19T20:17:37  <dreimark> and move back?
2009-07-19T20:17:42  <dennda> sure
2009-07-19T20:17:53  <dennda> clone, use new backend
2009-07-19T20:18:02  <waldi> ThomasWaldmann: you want a sonic shower?
2009-07-19T20:18:06  <dreimark> and not of all revisions?
2009-07-19T20:18:28  <dennda> dreimark: if you skip them / destroy them after cloning...
2009-07-19T20:18:56  <dennda> ThomasWaldmann: I think we need to be careful not to get featuritis before we have something that we can actually release :-)
2009-07-19T20:19:05  <ThomasWaldmann> http://pastebin.com/f1953b9ef
2009-07-19T20:19:32  <ThomasWaldmann> dennda: that are quite basic issues
2009-07-19T20:20:33  <dennda> i see your point
2009-07-19T20:20:49  <dreimark> hmm, was sure ewe have also revisions of meta data, don't we?
2009-07-19T20:21:01  <dennda> for basic syspages renewal i don't see why we'd need to mess with the items metadata, though
2009-07-19T20:21:06  <dennda> no
2009-07-19T20:21:11  <dennda> revisions are for revisioning
2009-07-19T20:21:21  <dennda> metadata is for unrevisioned extra information
2009-07-19T20:21:26  <dennda> items have metadata AND revisions
2009-07-19T20:21:31  <dennda> revisions have metadata AND content
2009-07-19T20:21:33  <dennda> it's that simple
2009-07-19T20:22:01  <ThomasWaldmann> dennda: do we have any item metadata (aside from user profiles)?
2009-07-19T20:22:14  <dennda> hm I'd need to check myself
2009-07-19T20:22:16  <dreimark> e.g. we have the mimetype in meta or?
2009-07-19T20:22:28  <ThomasWaldmann> rev meta
2009-07-19T20:22:32  <dennda> I'm not sure what is in rev metadata and what is in item metadata
2009-07-19T20:22:42  <ThomasWaldmann> each rev can have a different mimetype
2009-07-19T20:23:13  <dreimark> ok and that is stored in the items meta data
2009-07-19T20:23:25  *** amartani_ is now known as amartani
2009-07-19T20:23:27  <ThomasWaldmann> no
2009-07-19T20:23:38  <dennda> why would it?
2009-07-19T20:23:39  <ThomasWaldmann> in the rev metadata
2009-07-19T20:23:48  <dennda> the item really is just a container
2009-07-19T20:23:58  <dennda> revisions are our first class citizens, so to speak
2009-07-19T20:24:41  <dennda> of course you can use an item without revisions for storage of data that should not be revisioned, such as user data
2009-07-19T20:25:06  <dennda> and store that data in the item's metadata
2009-07-19T20:29:53  <ThomasWaldmann> btw, still quite many XXX in storage
2009-07-19T20:31:33  <dennda> we should only use XXX or TODO
2009-07-19T20:31:36  <dennda> that makes grepping easier
2009-07-19T20:34:06  <ThomasWaldmann> i don't think they are there  because it is hard to grep them :P
2009-07-19T20:42:28  <PawelPacana> ;)
2009-07-19T20:42:52  <dennda> don't smile, you left a few there yourself :-)
2009-07-19T20:43:15  <PawelPacana> couldn't resist
2009-07-19T20:44:42  <PawelPacana> fs and fs17 are leaders in XXX
2009-07-19T20:52:47  <ThomasWaldmann> btw, i find myself rather often to use list_revisions -> [-1] -> add one to create a new revno
2009-07-19T20:53:17  <ThomasWaldmann> shouldn't that be part of storage api?
2009-07-19T20:55:28  <dennda> item.last_revno?
2009-07-19T20:56:53  <ThomasWaldmann>     def _create_revision(self, item, revno):
2009-07-19T20:56:53  <ThomasWaldmann>         """
2009-07-19T20:56:53  <ThomasWaldmann>         Takes an item object and creates a new revision. Note that you need to pass
2009-07-19T20:56:56  <ThomasWaldmann>         a revision number for concurrency reasons. The revno passed must be
2009-07-19T20:56:58  <ThomasWaldmann>         greater than the revision number of the item's most recent revision.
2009-07-19T20:57:24  <ThomasWaldmann> i don't quite get that. if 2 concurrent callers both pass 42, one will fail.
2009-07-19T20:58:15  <dennda> well, that's the point
2009-07-19T20:58:23  <ThomasWaldmann> hmm, i guess it is to trigger a merge.
2009-07-19T20:58:49  <ThomasWaldmann> the one who fails has likely based his edit on 41 also.
2009-07-19T20:59:32  <dennda> if you just have rev = item.create_revision(<noargshere>) you'd go on happily and try to commit but ooooh somebody was faster (and you have already put your data into that revision)
2009-07-19T21:00:40  <ThomasWaldmann> well, it could number the revs automatically, but that doesn't trigger a merge then.
2009-07-19T21:01:30  <dennda> yes, that's the next point
2009-07-19T21:02:42  <dennda> I could draw a petri net for this
2009-07-19T21:02:55  * dennda shoots himself
2009-07-19T21:05:13  *** amartani has quit IRC
2009-07-19T21:08:03  <ThomasWaldmann> dennda: yes, item.last_revno or .current_revno
2009-07-19T21:08:31  <ThomasWaldmann> that at least would avoid x differently broken implementations at misc. other places
2009-07-19T21:10:06  <ThomasWaldmann> as we assume revnos to be 0...n, it could just return -1 if there is no rev at all
2009-07-19T21:10:28  <dennda> hm
2009-07-19T21:10:42  <ThomasWaldmann> if you add 1 then to get next, it is correct
2009-07-19T21:10:46  <dennda> if we used next_rev instead we wouldn't have that problem :-)
2009-07-19T21:10:54  <ThomasWaldmann> or that
2009-07-19T21:11:14  <dreimark> saner
2009-07-19T21:12:11  <dennda> well that can be added easily to the generic level as no backend needs to implement that
2009-07-19T21:14:11  <ThomasWaldmann> at least at the beginning :)
2009-07-19T21:14:55  <ThomasWaldmann> (because it will have to either call _list_revisions to get a complete list of them , take the last revno there and add 1,
2009-07-19T21:15:09  <PawelPacana> something bad is happening, when i configure 2.0 to use hgbackend and create new page, fill it with data, save revision - it doesnt redirect me to that created revision but shows initial 'Create new item?' page again, yet I cannot create this item again (because it exists)
2009-07-19T21:15:19  <ThomasWaldmann> or it has to call _get_item(..., -1) and lookup revno there and add one)
2009-07-19T21:15:29  <PawelPacana> i verified and created items exist
2009-07-19T21:15:52  <PawelPacana> still hg doesnt pass only destroy_* tests
2009-07-19T21:16:13  <ThomasWaldmann> then it thinks it does not exist or it is deleted
2009-07-19T21:16:24  <PawelPacana> and on FSBackend these scenario proceeds as expected
2009-07-19T21:17:25  <ThomasWaldmann> does it work with old hg now, shall I try?
2009-07-19T21:18:10  <PawelPacana> old hg?
2009-07-19T21:18:25  <dennda> ThomasWaldmann: I'd just go with the former in case somebody gets lazy loading wrong
2009-07-19T21:18:54  <PawelPacana> didn't check
2009-07-19T21:19:18  <PawelPacana> curious, as backend tests pass
2009-07-19T21:23:10  <ThomasWaldmann>     data_backend = MercurialBackend('hgback')
2009-07-19T21:23:15  <ThomasWaldmann> is that enough?
2009-07-19T21:23:32  <ThomasWaldmann> assuming that there is a directory ./hgback/
2009-07-19T21:23:32  <PawelPacana> http://moinmo.in/PawelPacana/MercurialBackend/Configuration
2009-07-19T21:23:49  <PawelPacana> yes, this is enough
2009-07-19T21:27:25  <ThomasWaldmann> Starting underlay conversion...
2009-07-19T21:27:25  <ThomasWaldmann> Conversion of underlay failed. Please remove the instance folder and retry.
2009-07-19T21:28:36  <ThomasWaldmann> hmm, it seems to access irrelevant crap...
2009-07-19T21:28:52  <ThomasWaldmann> after i removed the (unrelated) instance/ folder, it worked X)
2009-07-19T21:29:35  <ThomasWaldmann> wtf, it recreate it as a FSBackend...
2009-07-19T21:29:59  <ThomasWaldmann> dennda's TODO grows :)
2009-07-19T21:30:25  <PawelPacana> well, that's create_if_missing in wikiserver.py
2009-07-19T21:31:52  <ThomasWaldmann> PawelPacana: strange behaviour confirmed
2009-07-19T21:33:53  <PawelPacana> where is that check if item exist made?
2009-07-19T21:38:18  <ThomasWaldmann> see MoinMoin/items/__init__.py
2009-07-19T21:39:09  <ThomasWaldmann> it's a bit tricky
2009-07-19T21:39:29  <ThomasWaldmann> begins in line 53
2009-07-19T21:41:55  <ThomasWaldmann> PawelPacana: is there a problem with rev metadata storage?
2009-07-19T21:42:06  <dennda> (i know)
2009-07-19T21:43:04  <PawelPacana> why should be? and if it was, why tests pass?
2009-07-19T21:43:40  <ThomasWaldmann> where should i find revision metadata on disk?
2009-07-19T21:43:49  <ThomasWaldmann> (i see the rev data)
2009-07-19T21:44:39  <PawelPacana> revision metadata is changeset dicts
2009-07-19T21:44:53  <PawelPacana> you'll have to dig through mercurial api
2009-07-19T21:44:53  <ThomasWaldmann> can i view it?
2009-07-19T21:45:13  <PawelPacana> although it is not hard
2009-07-19T21:45:17  <PawelPacana> just a sec..
2009-07-19T21:46:31  <ThomasWaldmann> (because the mimetype rev metadata is essential for that mechanism)
2009-07-19T21:49:57  <PawelPacana> ThomasWaldmann: http://paste.pocoo.org/show/129578/
2009-07-19T21:50:35  <PawelPacana> every key prefixed _moin_ is meta
2009-07-19T21:50:48  <PawelPacana> blah, prefixed _meta_
2009-07-19T21:52:12  <ThomasWaldmann> that outputs nothing
2009-07-19T21:52:31  <PawelPacana> ?
2009-07-19T21:53:00  <ThomasWaldmann> when i execute the code from the pastebin, it outputs nothing
2009-07-19T21:53:36  <PawelPacana> well, it outputs a lot on my repo
2009-07-19T21:53:48  <PawelPacana> did you change data dir?
2009-07-19T21:54:01  <ThomasWaldmann> list(data._repo.changelog)
2009-07-19T21:54:01  <ThomasWaldmann> []
2009-07-19T21:54:36  <ThomasWaldmann> ah, wait
2009-07-19T21:55:10  <ThomasWaldmann> had forgotten /data when editing the path
2009-07-19T21:56:19  <ThomasWaldmann> {'_meta_edit_log_action': "S'SAVE'\np1\n.", '_meta_mimetype': 'Vtext/moin-wiki\np1\n.', '_meta_edit_log_hostname': 'Vlocalhost\np1\n.', '_meta_edit_log_extra': "S''\n.", '_rev': '0', '_meta_sha1': "S'207e7ae7c7c61d77ad79be51537523123ecc2b75'\np1\n.", '_name': 'test', '_meta_edit_log_userid': "S''\n.", '_meta_edit_log_addr': "S'127.0.0.1'\np1\n.", 'branch': 'default', '_id': '098f6bcd4621d373cade4e832627b4f6', '_meta_edit_log_comment':
2009-07-19T21:56:30  <PawelPacana> yeah, values are pickled
2009-07-19T21:57:09  <PawelPacana> because this dict is later processed by  "\0".join()
2009-07-19T21:57:16  <PawelPacana> by hg
2009-07-19T21:57:41  <PawelPacana> and AFAIK meta could not necesarily be str
2009-07-19T21:58:14  <ThomasWaldmann> i guess i need to add some debug code to Item
2009-07-19T22:05:24  <ThomasWaldmann> 2009-07-19 22:04:40,847 DEBUG MoinMoin.items:79 Item u'test3', got mimetype 'application/x-unknown' from revision meta
2009-07-19T22:05:57  <CIA-36> Bastian Blank <bblank@thinkmo.de> default * 6620:41746d410431 2.0-storage-dom-bblank/MoinMoin/converter2/moinwiki_in.py: Moin Wiki input converter - Lookup sub converters via new type, give arguments to sub converters
2009-07-19T22:05:58  <CIA-36> Bastian Blank <bblank@thinkmo.de> default * 6621:ea610876396d 2.0-storage-dom-bblank/MoinMoin/converter2/html_out.py: HTML output converter - Move definitions
2009-07-19T22:05:59  <CIA-36> Bastian Blank <bblank@thinkmo.de> default * 6622:4f568d7cc975 2.0-storage-dom-bblank/MoinMoin/converter2/html_out.py: HTML output converter - Move attribute conversion into own classes
2009-07-19T22:06:01  <CIA-36> Bastian Blank <bblank@thinkmo.de> default * 6623:cb808f885662 2.0-storage-dom-bblank/MoinMoin/converter2/html_out.py: HTML output converter - Rename attributes class
2009-07-19T22:06:08  <ThomasWaldmann>         mimetype = rev.get(MIMETYPE) or 'application/x-unknown' # XXX why do we need ... or ..?
2009-07-19T22:06:11  <ThomasWaldmann>         logging.debug("Item %r, got mimetype %r from revision meta" % (name, mimetype))
2009-07-19T22:06:26  <ThomasWaldmann> -> rev.get(MIMETYPE) gives falsy result
2009-07-19T22:06:48  <ThomasWaldmann> PawelPacana: ^^
2009-07-19T22:08:18  <ThomasWaldmann> PawelPacana: get_revision(-1) works for your stuff?
2009-07-19T22:09:15  <PawelPacana> implemented
2009-07-19T22:10:43  <PawelPacana> and passed tests where it is used
2009-07-19T22:13:43  <waldi> boah, I don't want to debug the old code to understand how it works
2009-07-19T22:15:09  <ThomasWaldmann> PawelPacana: 2009-07-19 22:14:40,246 DEBUG MoinMoin.items:80 Item u'test4', rev meta dict: {'dit_log_userid': '', 'dit_log_extra': '', 'sha1': 'e950c1517ee0d7e20454d22c306c4c501a7cf11c', 'dit_log_addr': '127.0.0.1', 'dit_log_action': 'SAVE', 'imetype': u'text/moin-wiki', 'dit_log_hostname': u'localhost', 'dit_log_comment': ''}
2009-07-19T22:15:18  <ThomasWaldmann> that's the meta data :P
2009-07-19T22:15:38  <PawelPacana> imetype
2009-07-19T22:15:52  <ThomasWaldmann> no idea why that passes the tests X)
2009-07-19T22:16:03  <PawelPacana> X]
2009-07-19T22:17:00  <PawelPacana> ok, i'll try to write failing test and fix it
2009-07-19T22:17:06  <ThomasWaldmann> obviously the tests are rather fault tolerant
2009-07-19T22:17:46  <ThomasWaldmann> waldi: what code?
2009-07-19T22:18:36  <waldi> ThomasWaldmann: the id-generation
2009-07-19T22:19:04  <waldi> it is split over 5 files and uses a special stack
2009-07-19T22:21:25  <waldi> and code of this form "if _id_debug and id:" makes me believe, that it was not easy to write either
2009-07-19T22:23:30  <CIA-36> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6085:23a686c909a5 2.0-storage/wiki/config/more_samples/mercurial_backend_wikiconfig_snippet: (hg) configuration snippet for hg backend
2009-07-19T22:25:37  <ThomasWaldmann> waldi: yeah, that's a bit weird
2009-07-19T22:31:24  <waldi> wikiutil.anchor_name_from_text
2009-07-19T22:31:30  <waldi> juhu
2009-07-19T22:33:34  <ThomasWaldmann> ? :)
2009-07-19T22:47:36  <CIA-36> Bastian Blank <bblank@thinkmo.de> default * 6624:0684dc57c6e0 2.0-storage-dom-bblank/MoinMoin/converter2/html_out.py: HTML output converter - Generate speaking ids
2009-07-19T22:47:57  <dreimark> dimazest: isSuperUser check seems sometimes not working
2009-07-19T22:57:26  <dimazest> dreimark: isSuperUser connect to groups?
2009-07-19T23:01:37  <dreimark> not really. I investigated if you switch in desktopedition to another user then you should stay as this user until you logout this user
2009-07-19T23:01:48  <dreimark> but you jump back with the next request
2009-07-19T23:01:56  <dreimark> it looks like the session file is correct
2009-07-19T23:02:34  <dreimark> (you can test this by having two users in your wiki and calling settings switch user)
2009-07-19T23:03:09  <ThomasWaldmann> dreimark: you checked that it works correctly in 1.9?
2009-07-19T23:03:20  <dreimark> I did it works there for me
2009-07-19T23:05:30  * dreimark fetches a new 1.9 clone 
2009-07-19T23:06:07  <dimazest> dreimark: the same for me
2009-07-19T23:06:18  <dreimark> do you have tried 1.9 ?
2009-07-19T23:07:03  <dimazest> no
2009-07-19T23:14:41  <dimazest> dreimark: the same with 1.9
2009-07-19T23:15:09  <dreimark> hmm, wonders why I can't repeat it (still fetching)
2009-07-19T23:23:27  <dreimark> I think I have an idea what goes wrong. in auth.setup_setuid it checks userobj.isSuperUser(). this depends on userobj.valid. It seems not to get a vlidation to be a True user
2009-07-19T23:38:59  <CIA-36> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6086:d91e8109f9c1 2.0-storage/MoinMoin/storage/ (_tests/test_backends_hg.py backends/hg.py): (hg) fixed stupid mistake when stripping revisions metadata key names
2009-07-19T23:42:45  <PawelPacana> tests were good, they just didn't check revision metadata with key starting with one of _, m, e, t, a
2009-07-19T23:43:06  <PawelPacana> which was problematic on improper use of lstrip
2009-07-19T23:43:09  <PawelPacana> anyway, fixed
2009-07-19T23:43:09  <dennda> PawelPacana: 6 -> len("_meta_") please
2009-07-19T23:43:25  <dennda> where that string is preferably some constant
2009-07-19T23:43:55  *** grzywacz has quit IRC
2009-07-19T23:49:18  <dennda> well, not too important
2009-07-19T23:49:27  <dennda> it should be clear there where that magic 6 came from
2009-07-19T23:50:18  <CIA-36> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6087:29daf78805c2 2.0-storage/MoinMoin/storage/backends/hg.py: (hg) refined metadata prefixing
2009-07-19T23:50:23  <PawelPacana> well, you're right
2009-07-19T23:52:55  <dennda> thanks :-)