MoinMoin: MoinMoinChat/Logs/moin-dev/2009-06-14

2009-06-14T01:32:35  *** grzywacz has quit IRC
2009-06-14T01:41:23  *** amartani has joined #moin-dev
2009-06-14T02:23:50  *** amartani has left #moin-dev
2009-06-14T02:24:27  *** amartani has joined #moin-dev
2009-06-14T03:09:10  *** amartani has quit IRC
2009-06-14T04:11:22  <ThomasWaldmann> dennda: e.g. the HelpContents page is linked from navibar
2009-06-14T04:15:40  <ThomasWaldmann> dimazest: looks ok, except that MoinMoin/groups/backends/__init__.py:34 is superfluous
2009-06-14T09:39:45  <dreimark> moin
2009-06-14T09:39:57  * dreimark greets from Gehrden
2009-06-14T09:40:53  <dreimark> dimazest: I look at it after I am back in the evening
2009-06-14T10:16:27  <dreimark> bbl
2009-06-14T11:31:08  <ThomasWaldmann> re
2009-06-14T11:33:30  <xorAxAx> wow, i got a mail from my student but he still didnt appear here :-(
2009-06-14T12:13:52  <ThomasWaldmann> dimazest: why is retrieve_members in the Backend and not in Group?
2009-06-14T12:16:42  <dimazest> ThomasWaldmann: i tought it's better to keep it in backend
2009-06-14T12:17:03  <dimazest> should it be in a group?
2009-06-14T12:19:34  <ThomasWaldmann> not sure, therefore i was asking
2009-06-14T12:20:14  <ThomasWaldmann> as it deals with group members and not with namespace of group names, i thought it might be better in Group
2009-06-14T12:20:39  <dimazest> i think, it should be in backend, because then there is no need to inherit group
2009-06-14T12:20:48  <dimazest> except caching is needed
2009-06-14T12:21:42  <dimazest> but if caching code is moved int backend, and retrieve_members takes care of caching
2009-06-14T12:21:53  <dimazest> there is no need for Group calss for every backend
2009-06-14T12:22:27  <dimazest> there will be only one Group class for all backends (at least for wiki pages and config)
2009-06-14T12:23:23  * ThomasWaldmann looks again at current code
2009-06-14T12:24:28  <ThomasWaldmann> btw, has the public repo all changesets?
2009-06-14T12:25:01  <dimazest> comparing with ssh://moin/moin/1.9-groups-dmilajevs/
2009-06-14T12:25:02  <dimazest> searching for changes
2009-06-14T12:25:03  <dimazest> no changes found
2009-06-14T12:26:58  <ThomasWaldmann> ah, I thought you already did some more cleanup, but I see that was just a plan update for next week
2009-06-14T12:30:48  <ThomasWaldmann> btw, membergroups could be also an iterator
2009-06-14T12:32:30  <ThomasWaldmann> and maybe could be named either membersgroups or groups_of_member
2009-06-14T12:33:16  <ThomasWaldmann> or groups_with_member()
2009-06-14T12:33:28  <ThomasWaldmann> ask a native english speaker :)
2009-06-14T12:33:48  * dimazest needs to find one
2009-06-14T12:35:19  <ThomasWaldmann> there are some millions of them on this planet. :) maybe try on #moin and #python and just show the code or tell what it does.
2009-06-14T12:35:37  <dimazest> ok
2009-06-14T12:37:07  <ThomasWaldmann> btw, as BaseGroup provides _load_group and calls it from __init__, how about initializing self.members and self.member_groups to the empty set and using self.x in load_groups?
2009-06-14T12:38:04  <ThomasWaldmann> or self.members, self.member_groups = self._load_group()
2009-06-14T12:38:24  <dimazest> hg diff
2009-06-14T12:38:46  <ThomasWaldmann> that ... = None looks a bit superfluous
2009-06-14T12:39:46  <ThomasWaldmann> btw, for 4th week, please first finish groups stuff, then proceed to dicts
2009-06-14T12:41:35  <dimazest> ok, actually, this plan is just list of to do, order doen't matter
2009-06-14T12:44:04  <ThomasWaldmann>     33         return iter(self._groups.keys())   -> iterkeys?
2009-06-14T12:46:13  <ThomasWaldmann> config group -> _retrieve_members: code as it is is fine for unicode members
2009-06-14T12:47:35  <ThomasWaldmann> so people will have to write g = {u'FooGroup': set([u'BarMember', u'BazMember']), ...}
2009-06-14T12:48:06  <dimazest> yes
2009-06-14T12:49:07  <ThomasWaldmann> iirc we have some code in multiconfig that decodes stuff if it isn't unicode, let me take a look at it...
2009-06-14T12:50:34  <dimazest> http://paste.pocoo.org/show/123012/ changes
2009-06-14T12:51:07  <ThomasWaldmann> it just tries to decode the strings with 'utf-8', so people using ascii or utf-8 will get away with strings
2009-06-14T12:52:25  <ThomasWaldmann> dimazest: maybe use the assignment, you can delete some LOC then
2009-06-14T12:52:47  <ThomasWaldmann> (12:38) < ThomasWal> or self.members, self.member_groups = self._load_group()
2009-06-14T12:53:28  <dimazest> then there are problems with wiki backend, with caching
2009-06-14T12:53:47  <dimazest> probably, caching needs to be moved to the backend
2009-06-14T12:54:04  <dimazest> and then it will be possible to do so
2009-06-14T12:54:15  * ThomasWaldmann looks
2009-06-14T12:55:39  <ThomasWaldmann> what problem?
2009-06-14T12:57:05  <dimazest> ok, no problem, i just needed to return these values
2009-06-14T13:03:38  <dimazest> http://paste.pocoo.org/show/123013/
2009-06-14T13:08:30  <ThomasWaldmann> looks good, did you run the tests?
2009-06-14T13:08:39  <dimazest> sure
2009-06-14T13:10:49  * ThomasWaldmann likes nice changesets with more red than green
2009-06-14T13:18:41  <ThomasWaldmann> http://news.bbc.co.uk/2/hi/programmes/from_our_own_correspondent/8096988.stm
2009-06-14T14:39:45  *** amartani has joined #moin-dev
2009-06-14T14:51:02  *** grzywacz has joined #moin-dev
2009-06-14T16:57:22  <amartani> hi
2009-06-14T16:58:39  <dimazest> hi
2009-06-14T17:09:04  <ThomasWaldmann> dennda: did you already look at the acl error handling?
2009-06-14T17:26:09  <dennda> ThomasWaldmann: No
2009-06-14T17:42:05  <ThomasWaldmann> it should not blow away the complete ui :)
2009-06-14T17:43:34  <dennda> sure :-)
2009-06-14T17:43:44  <dennda> and the error displaying itself could be more...
2009-06-14T17:43:57  <dennda> beautiful :-)
2009-06-14T17:44:08  <dennda> currently it's more like "No, you may not. Now fuck off"
2009-06-14T17:44:38  <dennda> Need to figure out how to best show the forbidden-message
2009-06-14T17:44:56  <ThomasWaldmann> and more informative :) without disclosing too much.
2009-06-14T17:45:05  <dennda> yeah
2009-06-14T17:45:07  <dennda> i.e.
2009-06-14T17:45:10  <dennda> NOT as typo 3 did it
2009-06-14T17:45:25  <dennda> I almost fell off my chair when I saw their latest security fix
2009-06-14T17:45:40  <ThomasWaldmann> hm?
2009-06-14T17:45:43  <dennda> (the one that was exploited to deface wolfgang schaeubles page and that of the schalke 04
2009-06-14T17:45:54  <dennda> it was like
2009-06-14T17:46:43  <dennda> if provided_hash != stored_admin_hash: return "Sorry, your provided_hash did not match the stored admin hash LjahuIHAIUHiiHASIFuh"
2009-06-14T17:46:53  <ThomasWaldmann> lol
2009-06-14T17:47:16  <dennda> and the fix was to remove the %s equivalentfrom the string
2009-06-14T17:47:17  <ThomasWaldmann> ... please try again providing the correct hash you have now.
2009-06-14T17:47:25  <dennda> yeah
2009-06-14T17:47:48  <ThomasWaldmann> "don't drink and code" :)
2009-06-14T17:49:08  <dennda> schaeubles hash was "gewinner" or something similar
2009-06-14T17:49:26  <dennda> uncrackable
2009-06-14T17:49:31  <ThomasWaldmann> hehe
2009-06-14T17:50:29  <ThomasWaldmann> btw, we also still need a good plan for underlay
2009-06-14T17:50:39  <dennda> yeah well
2009-06-14T17:50:40  <ThomasWaldmann> or ... system/help pages
2009-06-14T17:50:52  <dennda> i guess you don't fancy writing a read only tar-backend :-)
2009-06-14T17:51:40  <ThomasWaldmann> that's maybe a bit slow for non-personal use
2009-06-14T17:52:12  <ThomasWaldmann> (you have to read the tar always from the beginning when seeking for some file, right?)
2009-06-14T17:52:17  <dennda> is there some install script that all users run when setting up their wiki?
2009-06-14T17:52:25  <dennda> because then we could pump it into the backend as I already do
2009-06-14T17:52:35  * dennda has no idea how tar works
2009-06-14T17:52:55  <CIA-19> Dmitrijs Milajevs <dimazest@gmail.com> default * 4788:415338464695 1.9-groups-dmilajevs/MoinMoin/groups/backends/ (__init__.py config_group.py wiki_group.py): Groups2009: Group._load_group() returns members and member_groups. Minor code refinements.
2009-06-14T17:52:56  <ThomasWaldmann> it's from the 70s, at that time with real tapes
2009-06-14T17:52:57  <CIA-19> Dmitrijs Milajevs <dimazest@gmail.com> default * 4789:3c0ce322472a 1.9-groups-dmilajevs/MoinMoin/groups/ (4 files in 4 dirs): Groups2009: BaseBackend.groups_with_member() method was added. Test.GroupManager.test_membergroups() removed.
2009-06-14T17:52:58  <CIA-19> Dmitrijs Milajevs <dimazest@gmail.com> default * 4790:89497de6165d 1.9-groups-dmilajevs/MoinMoin/ (10 files in 6 dirs): Groups2009: GroupManager became compose backend and was moved to backends directory. Unnecessary tests were removed.
2009-06-14T17:53:00  <CIA-19> Dmitrijs Milajevs <dimazest@gmail.com> default * 4791:a1018d829c3e 1.9-groups-dmilajevs/MoinMoin/ (15 files in 5 dirs): Groups2009: It is possible to import backends as 'from Moin.groups import WikiGroups, ConfigGroups, CompositeGroups'. Compose backend was renamed to Composite.
2009-06-14T17:54:10  <ThomasWaldmann> dimazest: the test are maybe not "unnecessary", they just should also move to the right place
2009-06-14T17:54:15  <dennda> either we clone it into the backend itself or we combine the read-only fs backend with some layering middleware
2009-06-14T17:54:41  <ThomasWaldmann> how do we handle updates/grades?
2009-06-14T17:54:50  <dennda> as we already do?
2009-06-14T17:55:02  <dennda> the user needs to download and unpack it, doesn't he?
2009-06-14T17:55:14  <dennda> (that's for the layering solution)
2009-06-14T17:55:46  <ThomasWaldmann> the question is more about the old pages, maybe even edited old pages
2009-06-14T17:56:32  <ThomasWaldmann> with the layering, it is no big problem, but we have no layering yet
2009-06-14T17:56:47  <ThomasWaldmann> and the other question is whether we want to keep it just for that
2009-06-14T17:58:52  <dennda> keep what?
2009-06-14T17:59:12  <ThomasWaldmann> to have layering just to deal with system pages upgrade
2009-06-14T17:59:37  <ThomasWaldmann> and the package system is also not quite up to what we need
2009-06-14T18:00:05  <ThomasWaldmann> (handling of metadata, maybe also of binary items?)
2009-06-14T18:00:37  <dennda> hm? binary items need metadata too
2009-06-14T18:01:01  <ThomasWaldmann> i didn't say the contrary :)
2009-06-14T18:01:42  <ThomasWaldmann> packagepages was coded to handle with pages and attachments, but not with separate metadata at the time it was written.
2009-06-14T18:02:02  <ThomasWaldmann> similar thing for xmlrpc stuff
2009-06-14T18:02:21  * ThomasWaldmann is a bit tempted to just dump everything to xml
2009-06-14T18:03:30  <ThomasWaldmann> wikisync also had a similar problem: was just written to deal with pages, no idea how it would deal with binary items
2009-06-14T18:05:10  <ThomasWaldmann> maybe we could summarize this all under "item serialization"
2009-06-14T18:09:06  <dennda> is it me or do we have a meeting since 9 minutes already?
2009-06-14T18:09:40  <dimazest> have no idea
2009-06-14T18:09:49  <amartani> I'm here..
2009-06-14T18:19:33  <dennda> could we do it now? I'll have to go in a bit
2009-06-14T18:22:48  <amartani> me too
2009-06-14T18:23:31  <dimazest> here is my status:
2009-06-14T18:23:32  <dimazest> LAST: i did WikiGroups, ConfigGroups, CompositeGroups backends.
2009-06-14T18:23:32  <dimazest> NEXT: i will update wiki page. Finish polishing existing backends and start to do sql backend.
2009-06-14T18:23:35  <dimazest> BLOCKERS: no
2009-06-14T18:24:38  <ThomasWaldmann> oops :)
2009-06-14T18:27:00  <ThomasWaldmann> dimazest: and dicts :)
2009-06-14T18:27:12  <ThomasWaldmann> (before you start sql)
2009-06-14T18:27:32  <dimazest> ThomasWaldmann: what concrete about dicts?
2009-06-14T18:27:59  <ThomasWaldmann> that they work in the same style as groups?
2009-06-14T18:29:01  <dimazest> ThomasWaldmann: do you mean bacends for dicts?
2009-06-14T18:29:15  <ThomasWaldmann> sure
2009-06-14T18:32:02  <dimazest> what backends are needed? one that takes definitions from config
2009-06-14T18:32:21  <dimazest> and one which takes dict definitions from wikipages
2009-06-14T18:32:26  <ThomasWaldmann> yes
2009-06-14T18:32:27  <dimazest> right?
2009-06-14T18:32:48  <ThomasWaldmann> it is almost the same thing as groups, but it is not only the keys, but also the values
2009-06-14T18:33:38  <dimazest> but how to implement, should i inherit from groups?
2009-06-14T18:34:46  <ThomasWaldmann> we'll have to see that, but not now
2009-06-14T18:35:25  <dimazest> ok
2009-06-14T18:35:29  <ThomasWaldmann> maybe implement in parallel, maybe with moving some common stuff into a new base
2009-06-14T18:35:59  <ThomasWaldmann> and you likely need a new package now (remember my "datastruct" package suggestion? :)
2009-06-14T18:36:04  <dimazest> ThomasWaldmann: i can start in parallel and then merge common code into a base
2009-06-14T18:36:29  <dimazest> ThomasWaldmann: so i cretae datastruct
2009-06-14T18:36:39  <dimazest> move groups into it
2009-06-14T18:36:45  <dimazest> and create dicts
2009-06-14T18:42:11  <ThomasWaldmann> my suggestion would be packages/{groups.py,dicts.py,backends/*.py}
2009-06-14T18:42:31  <ThomasWaldmann> and then we just import the important stuff to __init__.py
2009-06-14T18:48:07  <amartani> should I post my status?
2009-06-14T18:48:13  <dennda> (can I be next please?)
2009-06-14T18:48:26  <dimazest> ok
2009-06-14T18:48:37  <dimazest> sorry for delay, i had to reboot
2009-06-14T18:52:09  <ThomasWaldmann> ok, thanks dimazest, i guess we can talk about details later. thanks for your report.
2009-06-14T18:52:15  <ThomasWaldmann> amartani: you're next
2009-06-14T18:52:46  <amartani> LAST: Syncronization is working, mobwrite is running in-process. Listing users is working, but kind of hackish.
2009-06-14T18:52:50  <amartani> NEXT: Clean up modifications on mobwrite, try to get it upstream.
2009-06-14T18:52:53  <amartani> BLOQUERS: None.
2009-06-14T18:58:04  <ThomasWaldmann> amartani: about your work schedule, I had the impression that your daily work time on the project was a bit low.
2009-06-14T18:58:19  <ThomasWaldmann> if I am right, could you improve this?
2009-06-14T19:00:35  <amartani> ThomasWaldmann: I'll try my best.. this semester is being harder than I expected at university..
2009-06-14T19:01:35  <ThomasWaldmann> well, if you are too busy at workdays, maybe try to use the weekend
2009-06-14T19:01:49  <ThomasWaldmann> also, I didn't see you communicating much
2009-06-14T19:06:01  <amartani> sorry, I'll talk more here about my work..
2009-06-14T19:11:52  <ThomasWaldmann> amartani: the in-process running, how did you do it exactly?
2009-06-14T19:12:00  <ThomasWaldmann> waldi: are you with us?
2009-06-14T19:14:55  <amartani> I have moved the methods on mobwrite daemon that really does the work (from the ones that listen to connections, etc.), and did the setup it does when the daemon is starting inside the action..
2009-06-14T19:16:56  <amartani> as I am now using mobwrite code in an action, should I move it to support? (I had put it on contrib previously)
2009-06-14T19:25:48  <dennda> ok I will be gone for lunch now... sorry.
2009-06-14T19:25:53  <dennda> bbiab
2009-06-14T19:30:02  <ThomasWaldmann> ok, thanks amartani. dennda: you're next.
2009-06-14T19:33:29  <ThomasWaldmann> (btw, usually the idea is that not only more mentors are participating :)
2009-06-14T19:48:58  <amartani> going lunch..
2009-06-14T19:51:27  *** amartani has quit IRC
2009-06-14T19:54:48  <ThomasWaldmann> (btw, usually the idea is that not only students, but also the mentors are participating :)
2009-06-14T19:56:21  <ThomasWaldmann> xorAxAx: sunday evening seems to be a bad date. can you arrange a different one? plus tell your student that he also has to participate.
2009-06-14T20:03:01  <ThomasWaldmann> dennda: sorry, I overlooked your request (I was also busy a bit here locally), but we can talk after your lunch
2009-06-14T20:25:20  <dennda> ThomasWaldmann: np
2009-06-14T20:25:26  <dennda> ThomasWaldmann: now?
2009-06-14T20:28:25  <dennda> I'll just post my stuff
2009-06-14T20:28:50  <dennda> What was done: An ACL middleware was provided that makes sure permissions are checked before performing any security-relevant action. The current acl checks were removed.
2009-06-14T20:29:10  <dennda> Where it is necessary that permissions are checked for the UI (e.g. for displaying 'Immutable Page'), the old implementation was changed so as to use the new permission checking facility.
2009-06-14T20:29:21  <dennda> For development and betatesting a script was provided that copies the english underlay pages into a permanent filesystem backend.
2009-06-14T20:29:25  <dennda> NEXT:
2009-06-14T20:30:03  <dennda> ACL related functionality will be provided so as to allow beta tests.
2009-06-14T20:30:48  <dennda> The user should be able to use the wiki as he normally would. We need to take care all cases are handled. We still need to decide what happens with specific privileges.
2009-06-14T20:31:21  <dennda> We will then need some people for beta testing :-)
2009-06-14T20:37:39  <dennda> I applied to get my blog on planet.python.org because i figured that may be a good source for peer review
2009-06-14T20:38:56  <ThomasWaldmann> (20:30) <    dennda> ACL related functionality will be provided so as to allow beta tests.
2009-06-14T20:39:03  <ThomasWaldmann> that means exactly what?
2009-06-14T20:39:34  <dennda> uh bad wording
2009-06-14T20:39:39  <dennda> finalized
2009-06-14T20:39:44  <dennda> and no fancy new functionality
2009-06-14T20:39:52  <dennda> but e.g. the error handling
2009-06-14T20:40:40  <ThomasWaldmann> ok, what else?
2009-06-14T20:41:24  <dennda> depends on what our target beta testing crowd is like
2009-06-14T20:41:42  <dennda> I'm not sure whether some proper ui for acl editing should be added
2009-06-14T20:41:52  <dennda> i.e. more than just that textarea that contains all the metadata
2009-06-14T20:42:53  <ThomasWaldmann> for now, I think the current UI is fine, because it shows you rather directly what you have
2009-06-14T20:43:16  <dennda> yeah well, we don't want morons as beta testers anyway
2009-06-14T20:43:50  <ThomasWaldmann> but the metadata keys could be sorted
2009-06-14T20:44:00  <dennda> I don't think there are any other show stoppers
2009-06-14T20:44:17  <dennda> well yeah, piece of cake...
2009-06-14T20:44:19  <ThomasWaldmann> starting the wiki should get easier
2009-06-14T20:44:28  <dennda> hmhm
2009-06-14T20:44:39  <dennda> we could do a check in the standalone wikiserver
2009-06-14T20:45:01  <dennda> and if there's no instance/.finnished_creation file or something, it starts the conversion
2009-06-14T20:46:41  <ThomasWaldmann> some minor style additions would be also nice (I am not sure *I* will try that, because anytime I touch CSS, it is mostly a waste of time :)
2009-06-14T20:47:18  * dennda has no idea of moinmoin theming
2009-06-14T20:47:37  <ThomasWaldmann> maybe some width=100% would be enough for the editors
2009-06-14T20:47:39  <dennda> i can try though
2009-06-14T20:47:53  <dennda> ah well yeah something like that...
2009-06-14T20:48:17  <fxhp> I've designed a new CSS template for moinmoin for my work.  Looks really nice.
2009-06-14T20:48:37  <dennda> we have a volunteer! :D
2009-06-14T20:48:50  <ThomasWaldmann> fxhp: :) I guess you are working on 1.8, right?
2009-06-14T20:48:57  <fxhp> Yes.
2009-06-14T20:49:01  <dennda> ThomasWaldmann: is that part jinja based already?
2009-06-14T20:49:12  <dennda> hm yeah i think it is
2009-06-14T20:49:16  <ThomasWaldmann> the 2.0 stuff is quite different and not production ready
2009-06-14T20:49:19  <dennda> but you mentioned some other part was not
2009-06-14T20:50:12  <ThomasWaldmann> dennda: there is a config switch between old theme code (more or less working for header/footer) and new jinja theme code (mostly not working / offering nothing yet)
2009-06-14T20:50:53  <ThomasWaldmann> basically we need someone writing a jinja2 template-based theme and redo/reorganize most css
2009-06-14T20:52:20  <ThomasWaldmann> dennda: what do you think about my xml serialization idea?
2009-06-14T20:54:10  <ThomasWaldmann> (for item packages [e.g. sys/help pages], for wiki xmlrpc, wiki sync, backup, restore, ...)
2009-06-14T20:56:35  <ThomasWaldmann> dennda: btw, we didn't get clear about how to handle syspages upgrades, regarding old and edit pages that are already there in storage
2009-06-14T20:57:00  <dennda> where did you propose said idea?
2009-06-14T20:57:00  <ThomasWaldmann> currently, there are 2 ways we could go:
2009-06-14T20:57:53  <ThomasWaldmann> a) have layering/underlay similar to 1.8, throw away all items in underlay storage, load all items with new content
2009-06-14T20:58:27  <ThomasWaldmann> todo: layering middleware, some script to remove / refill
2009-06-14T20:58:42  <ThomasWaldmann> (or adapt packagepages somehow)
2009-06-14T20:59:30  <ThomasWaldmann> b) no having layering, but just help/sys pages in normal storage. tag them somehow so we know what is a original sys/help page.
2009-06-14T20:59:54  <dennda> how do you handle changes, then? should change even happen for underlay pages?
2009-06-14T21:00:17  <ThomasWaldmann> todo: tagging. how to handle edited pages?
2009-06-14T21:00:41  <dennda> we can't sanely
2009-06-14T21:01:05  <dennda> so either you can't upgrade easily to new versions of the underlay or you can't change pages from the underlay
2009-06-14T21:01:10  <ThomasWaldmann> in 1.8 it is possible to edit underlay pages (that just transfers them to normal data layer)
2009-06-14T21:01:29  <dennda> ok
2009-06-14T21:01:35  <dennda> then we could at least preserve the 1.8 behaviour
2009-06-14T21:02:29  <ThomasWaldmann> the 1.8 behaviour was not without problems. e.g. if someone just fixed some typo in a system page locally, he overlayed every future underlay page.
2009-06-14T21:02:44  <ThomasWaldmann> we needed that EditedSystemPages action/macro to handle that.
2009-06-14T21:03:25  <ThomasWaldmann> if we would just tag all sys/help pages and create new revisions for them when editing or when updating them, we maybe get saner behaviour.
2009-06-14T21:05:38  <ThomasWaldmann> (and we don't need layering, at least not for that)
2009-06-14T21:08:15  <ThomasWaldmann> let me play it through for b)
2009-06-14T21:09:10  <ThomasWaldmann> assume we mark all moin-dev provided system/help pages with CategorySystem
2009-06-14T21:09:42  <ThomasWaldmann> with 2.0 we provide RecentChanges rev 0
2009-06-14T21:09:59  <ThomasWaldmann> same for HelpOnFooBar rev 0
2009-06-14T21:10:34  <ThomasWaldmann> then someone finds a typo or something else he doesn't like and edits RecentChanges, creates modified rev 1
2009-06-14T21:10:59  <ThomasWaldmann> he also does not like HelpOnFooBar and deletes it (== empty rev 1)
2009-06-14T21:11:30  <ThomasWaldmann> now we release 2.0.1 a week later
2009-06-14T21:13:23  <ThomasWaldmann> we'll provide some sys/help page bundles with it with (maybe) new versions of the items
2009-06-14T21:14:19  <ThomasWaldmann> it installs the new RecentChanges as rev 2 (optionally first compare whether it has different sha1)
2009-06-14T21:14:41  <ThomasWaldmann> same for HelpOnFooBar rev 2
2009-06-14T21:15:24  <ThomasWaldmann> if wiki admin likes old versions better, he could just revert to rev 1
2009-06-14T21:17:44  <ThomasWaldmann> we could also provide a script to remove ALL sys/help pages to get back to a defined state when someone accidentally has installed page packs he does not really need
2009-06-14T21:19:13  <ThomasWaldmann> (that can be done by either deleting or nuking the items)
2009-06-14T21:19:43  <ThomasWaldmann> btw, I think there was still an open question about this: do we support nuking items on backend api level?
2009-06-14T21:21:21  <ThomasWaldmann> dennda: still there? :)
2009-06-14T21:26:26  <ThomasWaldmann> btw, we could even use wikisync master wiki -> local wiki for that, giving the sync process some page list from i18n.strings (or a translation of it)
2009-06-14T21:27:20  <ThomasWaldmann> xorAxAx: ^^ what does the wikisync dev think about this? :)
2009-06-14T21:34:02  <dreimark> moin
2009-06-14T21:34:24  <dreimark> currently aou could purge underlay or replace it by a new one3
2009-06-14T21:34:33  <ThomasWaldmann> hi dreimark
2009-06-14T21:35:06  * dreimark is watching hitchiker too
2009-06-14T21:35:11  <dreimark> h
2009-06-14T21:35:58  * dreimark was caught in a traffic jam
2009-06-14T21:36:08  <dreimark> (more than once)
2009-06-14T21:36:24  * ThomasWaldmann switches to Pro7
2009-06-14T21:36:31  <dreimark> :)
2009-06-14T21:36:41  <dennda> ThomasWaldmann: sure ;)
2009-06-14T21:38:23  <ThomasWaldmann> dennda: sure what?
2009-06-14T21:43:04  <ThomasWaldmann> (we can also continue after HHGTTG)
2009-06-14T21:44:20  <dennda> sure i'm still there
2009-06-14T21:45:27  <dennda> ThomasWaldmann: actually i gotta leave for today soon, i'll give feedback tomorrow if there's no other urgent thing
2009-06-14T21:49:34  <dennda> ideally we wouldn't need to support nuking
2009-06-14T21:49:42  <dennda> but i think practicality beats us here
2009-06-14T21:51:00  <dennda> not as an UI option for the normal user of course :-)
2009-06-14T21:52:13  <ThomasWaldmann> ok, so if we can't nuke via API, we can't "throw away" old items (except by external means, which doesn't generally work then)
2009-06-14T21:54:55  <ThomasWaldmann> so b) looks easier if we don't want to force users to use fs* for sys pages
2009-06-14T21:57:15  <ThomasWaldmann> and some day, when we have revision purge and purge policies, we can flatten the stuff again
2009-06-14T21:57:31  <dennda> it would be "nuke revision" btw
2009-06-14T21:57:33  <dennda> not "nuke item"
2009-06-14T21:58:12  <ThomasWaldmann> it's both :)
2009-06-14T21:58:41  <ThomasWaldmann> if we kill an item from sys/help items, it could be purge item
2009-06-14T22:34:34  <dreimark> zipfile unpack has the same meta data missing problem
2009-06-14T22:37:16  <ThomasWaldmann> yes, with there difference that there is no metadata by definition then. for packages, there should be metadata (and initially we have them).
2009-06-14T22:37:41  <ThomasWaldmann> dreimark: we could try writing a xml backend
2009-06-14T22:38:19  <ThomasWaldmann> doing a backup / creating a package would then be just to clone some items to that backend
2009-06-14T22:38:54  *** franklin_ has joined #moin-dev
2009-06-14T22:39:12  <dreimark> hmm I have used packages to creaqte offline new pages/attachments
2009-06-14T22:39:17  <ThomasWaldmann> same for restore / installing a package
2009-06-14T22:40:39  <xorAxAx> ThomasWaldmann: yes, syncing should work for that
2009-06-14T22:40:45  <dreimark> for those I don't have metadata in the wiki.
2009-06-14T22:40:56  <xorAxAx> sorry for missing the meeting, i was asleep
2009-06-14T22:41:09  <xorAxAx> who didnt attent besides my student?
2009-06-14T22:41:48  <ThomasWaldmann> well, some mentors missed the date, too
2009-06-14T22:42:40  <ThomasWaldmann> and I also was fixing the neighbor's computer, but had my laptop online and with me :)
2009-06-14T22:44:50  <xorAxAx> reimar said he cant come
2009-06-14T22:46:41  <dreimark> xorAxAx: I was going back from gehrden near by paderborn at that time
2009-06-14T22:47:29  <xorAxAx> you mean gehrden near hannover? wikipedia doesnt know another gehrden :)
2009-06-14T22:47:49  <fxhp> I'm not positive what jinja is
2009-06-14T22:48:08  <dreimark> xorAxAx: http://www.schloss-gehrden.de/
2009-06-14T22:48:28  <dennda> fxhp: jinja.pocoo.org
2009-06-14T22:48:31  <dreimark> http://www.schloss-gehrden.de/lage_anfahrt.html
2009-06-14T22:48:32  <dennda> a templating engine
2009-06-14T22:48:49  <xorAxAx> ah, what did you do there, dreimark?
2009-06-14T22:50:11  <dreimark> wedding party of some friends (we slept there)
2009-06-14T22:50:52  <xorAxAx> ah
2009-06-14T22:51:05  <dreimark> (I never thought before on getting a room in a "schloss")
2009-06-14T22:51:15  <xorAxAx> you could have visited me :)
2009-06-14T22:52:05  *** franklin has quit IRC
2009-06-14T22:52:05  <dreimark> I know, if we have staid longer it would have been possible. We stayed there with all other relatives
2009-06-14T22:53:31  <xorAxAx> ah
2009-06-14T22:57:25  <dreimark> in the past it was a "Jugend Herberge" http://www.ahgz.de/regional-und-lokal/Herberge-wird-Schloss,200012157526.html
2009-06-14T23:00:17  <dreimark> ThomasWaldmann: waldi should be a bit more visible too
2009-06-14T23:02:20  <ThomasWaldmann> dreimark: he is here, you can speak directly to him :)
2009-06-14T23:03:09  <dreimark> waldi: do you talk in queries or mail to amartani?
2009-06-14T23:11:50  <dreimark> I do see only a few times waldi / amartani together on moin-dev.
2009-06-14T23:18:30  * ThomasWaldmann tries to hack a xml backend
2009-06-14T23:23:48  <dreimark> good night
2009-06-14T23:28:31  <ThomasWaldmann> gn dreimark