2007-05-09T00:43:18 <ThomasWaldmann> johill: http://moinmoin.wikiwikiweb.de/MoinMoinBugs/BrowserPyIndent is this fixed? If yes, please close it / delete the page, if not, please supply more details.
2007-05-09T00:49:21 <xorAxAx> ThomasWaldmann: 00:41:11 < RememberPOL> What makes bazaar superior to mercurial?
2007-05-09T00:49:24 <xorAxAx> :-)
2007-05-09T01:30:35 <grzywacz> heh ;)
2007-05-09T12:26:20 <ThomasWaldmann> hehe, fi(Help) == Ohje :)
2007-05-09T12:26:52 <xorAxAx> :-)
2007-05-09T13:03:06 <shirish> Hi all, sorry to butt in but I'm looking for roadmap for 1.6 beta as to when it would be there either by some trac or some other system if its possible to know, the wiki doesn't make that info. easier to know.
2007-05-09T13:47:26 <ThomasWaldmann> shirish: http://moinmoin.wikiwikiweb.de/MoinMoinTodo/Release_1.6.0
2007-05-09T13:48:14 <shirish> ThomasWaldmann: thank you :)
2007-05-09T13:49:15 <shirish> ThomasWaldmann: its already released or about to be released? 1.6.0 ?
2007-05-09T13:49:32 <shirish> for wikipedia lists 1.5.7 as the most recent release
2007-05-09T13:49:47 <ThomasWaldmann> that's true
2007-05-09T13:50:24 <shirish> so when is the beta, RC cycle supposed to start?
2007-05-09T13:50:33 <ThomasWaldmann> don't hold your breath for 1.6 release
2007-05-09T13:51:10 <ThomasWaldmann> btw, if we don't want to discuss development topics, we should rather continue on #moin
2007-05-09T13:51:17 <shirish> sure
2007-05-09T15:31:26 <lanius> ThomasWaldmann: xorAxAx: what do we need to discuss next regarding the storage design?
2007-05-09T15:40:47 <xorAxAx> lanius: attribute-backup means that the UOW patterns needs some reference to check if the data has changed
2007-05-09T15:40:56 <xorAxAx> lanius: please describe your current planned storage pattern
2007-05-09T15:50:29 <lanius> xorAxAx: what exactly shell i describe?
2007-05-09T16:00:02 <xorAxAx> lanius: a sequence diagram for the save process
2007-05-09T16:00:16 <xorAxAx> did you realise that the UOW-pattern is bad?
2007-05-09T16:00:25 <xorAxAx> because of transactional issues
2007-05-09T16:01:53 <lanius> yes i did
2007-05-09T16:01:59 <lanius> i moved the save into the item
2007-05-09T16:02:22 <xorAxAx> ok, then how does the item know what of it to save?
2007-05-09T16:02:33 <xorAxAx> i.e. which revisions, data, metadata
2007-05-09T16:02:54 <lanius> the data class can have a changed flag
2007-05-09T16:03:00 <lanius> each revision as well
2007-05-09T16:03:14 <lanius> and metadata must keep track if the item was changed for each key/value pair
2007-05-09T16:03:52 <xorAxAx> so you have one data instance per revision? how would you manage security then?
2007-05-09T16:04:29 <lanius> on item level when accessing the revision
2007-05-09T16:04:55 <xorAxAx> please answer the first question
2007-05-09T16:05:02 <lanius> yes
2007-05-09T16:05:23 <xorAxAx> if there is just one instance, how should it differentiate if the caller may write to it?
2007-05-09T16:06:49 <lanius> then it needs to check the access rights on data level as well
2007-05-09T16:07:18 <xorAxAx> it would be much simpler if you handed an initialised instance that is opened in r or w
2007-05-09T16:07:34 <xorAxAx> (and the file wouldnt need to know anything about security then)
2007-05-09T16:12:27 <lanius> that means i would have a ReadableItem/WriteableItem ReadableData/WriteableData and ReadableMetadata/WriteableMetadata
2007-05-09T16:12:30 <lanius> ?
2007-05-09T16:22:52 <johill> ThomasWaldmann: I probably fixed that bug, deleted the page now
2007-05-09T16:27:33 <ThomasWaldmann> johill: ah, thanks
2007-05-09T16:28:50 <ThomasWaldmann> lanius: just a short comment on ACLs from me. 1.5 behaviour is that the most recent available acl is honoured.
2007-05-09T16:29:28 <ThomasWaldmann> that means, if a page has 3 revisions and you look at rev 2, it will use acl of rev3 for that
2007-05-09T16:30:32 <ThomasWaldmann> for deleted pages, it will use the latest available acl. as the deleted rev has no data file and the acl in 1.5 is contained in the data file, it will use the latest non-deleted revision to get the acl
2007-05-09T16:30:59 <lanius> okey
2007-05-09T16:31:20 <lanius> it would be nice if there would be some method like getRights(user, acl) so that i don't need to rewrite that
2007-05-09T16:32:32 <ThomasWaldmann> well, as we want to move acl check to a lower level ...
2007-05-09T16:33:10 <lanius> yeah it will be called from there, but maybe some static utility method somewhere
2007-05-09T16:33:17 <ThomasWaldmann> but you see, you can't handle that on a revision level
2007-05-09T16:33:23 <lanius> yes
2007-05-09T16:34:06 <lanius> then we need a callback from the revision to the item
2007-05-09T16:34:29 <lanius> or in xorAxAx version we don't need it i think
2007-05-09T16:34:49 <johill> lanius: isn't databackend external api too?
2007-05-09T16:35:13 <lanius> Data is external api which uses databackend
2007-05-09T16:35:49 <johill> ah, Data isn't defined anywhere though, what's the point?
2007-05-09T16:35:59 <lanius> data implements databackend :)
2007-05-09T16:35:59 <johill> wouldn't it be sufficient to say that Item.data is a file-like object?
2007-05-09T16:36:16 <lanius> that's documented in the class diagram
2007-05-09T16:36:32 <lanius> but what are file-like objects? there are no interfaces in python...
2007-05-09T16:36:53 <johill> ah, ok
2007-05-09T16:37:15 <johill> yeah, true, there's no such thing as a file-interface syntactically, but semantically databackend is exactly that if you make it an abstract class
2007-05-09T16:37:59 <lanius> yes
2007-05-09T16:38:13 <johill> ok, gotcha
2007-05-09T16:38:15 <lanius> but i need a specification for it to implement it, since i can't implement again a file interface
2007-05-09T16:38:25 <lanius> s/again/against/
2007-05-09T16:38:50 <johill> can you rename Consistence -> Consistency? :) consistence isn't the right word ;)
2007-05-09T16:39:21 <johill> otoh, maybe it's pretty much synonymous? I just see consistency used everywhere in CS
2007-05-09T16:39:45 <lanius> :)
2007-05-09T16:40:45 <lanius> xorAxAx: that means i would have a ReadableItem/WriteableItem ReadableData/WriteableData and ReadableMetadata/WriteableMetadata
2007-05-09T16:40:50 <lanius> ?
2007-05-09T16:49:12 <lanius> johill: i think i can remove your last question then
2007-05-09T16:49:30 <lanius> ThomasWaldmann: you're ok with the indexes?
2007-05-09T16:49:48 <johill> lanius: yeah
2007-05-09T17:06:02 <dreimark> bbl
2007-05-09T17:13:45 <johill> lanius: btw, now your stuff is pretty much compatible with mine afaict
2007-05-09T17:15:06 <lanius> sounds good
2007-05-09T17:15:29 <johill> except you still stuff a lot of things into Item I guess
2007-05-09T17:15:34 <johill> I'll have to see what you plan to do there
2007-05-09T17:15:54 <johill> (like how you want to do acls, with a an item proxy or not etc)
2007-05-09T17:15:58 <lanius> yes
2007-05-09T17:16:16 <johill> ?
2007-05-09T17:16:26 <lanius> yes that not defined yet
2007-05-09T17:16:32 <johill> ok yeah
2007-05-09T17:16:33 <lanius> yes, that is not defined yet
2007-05-09T17:23:25 <ThomasWaldmann> lanius: i guess i wouldnt talk about indexes at all. each backend can do it's own tricks to speed up, so if it will be implemented in the backend implementation, it is not part of the api. each backend will require some special configuration keywords or a set of constants.
2007-05-09T17:23:44 <ThomasWaldmann> s/it's/its/
2007-05-09T17:24:08 <johill> I concur
2007-05-09T17:24:31 <johill> and if it's backend-specific configuration then we don't need to talk about it at this level either
2007-05-09T17:24:35 <ThomasWaldmann> for the file backend I already suggested some dirty tricks :)
2007-05-09T17:25:45 <ThomasWaldmann> but we could talk about backend setup/configuration in general :) see my last q on that page.
2007-05-09T17:26:01 <johill> heh yeah
2007-05-09T17:26:44 <lanius> ok
2007-05-09T17:26:46 <johill> in my scheme it's just by setting an itemcollection ;)
2007-05-09T17:26:50 <johill> which can be a nested one
2007-05-09T17:27:06 <johill> that gives you the full power of python for config, but lanius wanted to do some namespaces tricks
2007-05-09T17:27:43 <ThomasWaldmann> johill: i guess that fstab like thing was a strange idea of me :)
2007-05-09T17:27:54 <johill> heh
2007-05-09T17:28:13 <johill> doesn't make much sense to me but hey ;)
2007-05-09T17:28:25 <johill> I'll just shut up and see how it turns out, I can always refactor the code later :P
2007-05-09T17:28:37 <lanius> it's not much different in my opinion
2007-05-09T17:28:42 <lanius> you set a backend class
2007-05-09T17:28:48 <lanius> and the backend class needs some configuration itself
2007-05-09T17:28:54 <ThomasWaldmann> the question is how we want to do such stuff like virtualpages
2007-05-09T17:29:07 <johill> well, yeah, but your dispatching backend thing has hugely complex configuration possibilities
2007-05-09T17:29:13 <lanius> what are virtualpages again
2007-05-09T17:29:25 <johill> while I just punt and if you want such complex configurations then you just have to nest itemcollections
2007-05-09T17:29:29 <ThomasWaldmann> lanius: some dirty patch of 1.5 :)
2007-05-09T17:29:43 <johill> lanius: basically a no-store store-backend that auto-generates pages
2007-05-09T17:29:57 <ThomasWaldmann> claiming some part of the namespace and not mapping it to page storage, but creating the page content somehow else
2007-05-09T17:30:07 <lanius> the configuration of dispatchingbackend could just look like such a stack
2007-05-09T17:30:10 <johill> imagine a backend that grabs the /calculate/ namespace and simply puts 4 into the page /calculate/2+2
2007-05-09T17:30:18 <ThomasWaldmann> e.g. from filesystem files, from database, ...
2007-05-09T17:30:18 <lanius> ThomasWaldmann: wouldn't that be just another backend?
2007-05-09T17:30:44 <johill> lanius: yeah, you're essentially unrolling that stack with lots of room for error
2007-05-09T17:30:49 <ThomasWaldmann> somehow, yes. it could be ro, though.
2007-05-09T17:30:53 <lanius> why errors?
2007-05-09T17:31:04 <johill> (error like if you say /path/ before /path/some/ then /path/some/ is ignored)
2007-05-09T17:31:06 <lanius> ThomasWaldmann: all save operations could just do nothing
2007-05-09T17:31:18 <johill> lanius: well, the UI shouldn't offer editing
2007-05-09T17:31:24 <lanius> johill: no error, just a definition
2007-05-09T17:31:54 <johill> lanius: sure, the configuration won't behave like you apparently want it to
2007-05-09T17:31:55 <lanius> i think the dispatchingbackend configuration would be a dict from namespace to backend
2007-05-09T17:32:06 <johill> what about nested namespaces then?
2007-05-09T17:32:16 <johill> get the longest match always?
2007-05-09T17:32:22 <lanius> yes
2007-05-09T17:32:43 <johill> sounds ok I guess
2007-05-09T17:33:13 <johill> then again
2007-05-09T17:33:26 <johill> the whole backend thing still doesn't make sense to me
2007-05-09T17:33:43 <johill> you put it under 'internal' but now you're requiring to touch backend stuff from the configuration which is external
2007-05-09T17:34:15 <lanius> configuration is both internal and external
2007-05-09T17:34:50 <johill> yeah, but internal backend configuration should be associated with that backend class
2007-05-09T17:34:53 <johill> well
2007-05-09T17:34:58 <johill> I guess your config could be something like
2007-05-09T17:35:02 <johill> from ... import DispatchingBackend
2007-05-09T17:35:13 <johill> backend = DispatchingBackend(namespacedict)
2007-05-09T17:36:01 <lanius> yes
2007-05-09T17:36:12 <johill> while I had the impression you were proposing
2007-05-09T17:36:16 <johill> cfg.namespaces = namespacedict
2007-05-09T17:36:26 <johill> and then the dispatching backend is instantiated "somewhere"
2007-05-09T17:36:36 <lanius> well i don't care, both is possible
2007-05-09T17:36:41 <lanius> i don't see a real difference
2007-05-09T17:37:28 <johill> well, the difference to me is that in the latter case the dispatching backend is like the only choice and hardcoded internals, while in the former you can actually do much different things and if you don't need a dispatcher you can just go around it
2007-05-09T17:37:42 <lanius> who is you?
2007-05-09T17:38:12 <johill> anyone
2007-05-09T17:38:17 <johill> oh btw, a namespace dict doesn't cut it
2007-05-09T17:38:18 <lanius> a user or the wiki
2007-05-09T17:38:24 <johill> the admin
2007-05-09T17:38:35 <johill> unless each entry in the namespace dict is a list
2007-05-09T17:40:11 <johill> (then the default config would be something like
2007-05-09T17:40:26 <lanius> rigt a list
2007-05-09T17:40:37 <johill> DispatchingBackend({'/':[FileBackend(), FileBkacned(underlay=True)]})
2007-05-09T17:41:33 <johill> )
2007-05-09T17:41:39 <lanius> take a look at what i posted
2007-05-09T17:42:12 <johill> yeah
2007-05-09T17:42:59 <ThomasWaldmann> johill: your idea would be to just stack everything?
2007-05-09T17:43:01 <johill> looks like I imagined
2007-05-09T17:43:07 <johill> ThomasWaldmann: basically
2007-05-09T17:43:13 <johill> I was thinking to have one class for everything
2007-05-09T17:43:31 <johill> like collection = StackedCollection(FileCollection(), FileCollection(underlay=True))
2007-05-09T17:43:35 <johill> or
2007-05-09T17:43:42 <ThomasWaldmann> how do we get reasonable performance if we stack everything?
2007-05-09T17:43:57 <lanius> how do you do underlay if you stack?
2007-05-09T17:43:57 <ThomasWaldmann> for data/underlay it is of course ok (same as now)
2007-05-09T17:44:11 <lanius> ah you mean to check every backend, i understand
2007-05-09T17:44:12 <johill> do you think those few calls that go through matter?
2007-05-09T17:44:24 <ThomasWaldmann> but what if I want to mount some filesystem files as virtual pages under FS/*
2007-05-09T17:44:29 <johill> well the behaviour of stackedcollection would be to check one first
2007-05-09T17:44:33 <johill> ThomasWaldmann: that'd be
2007-05-09T17:44:46 <lanius> but if it does not exist it must check all
2007-05-09T17:45:19 <johill> collection = StackedCollection(NameSpaceCollection('/FS', FileSystemBackend()), FileCollection(), FileCollection(underlay=True))
2007-05-09T17:45:36 <ThomasWaldmann> in the fstab like way, we KNOW where something is located, while in the stacking approach, we have to go trough all layers to find it
2007-05-09T17:45:53 <johill> then make it the other way around
2007-05-09T17:46:14 <johill> anyway
2007-05-09T17:46:15 <lanius> seems complexer to achieve the same
2007-05-09T17:46:50 <ThomasWaldmann> although that approach is pretty and simple, I see some performance problems.
2007-05-09T17:47:28 <ThomasWaldmann> esp. if some of our powerusers start to play with many layers, performance will degrade 1/N for stuff on the lowest layer
2007-05-09T17:47:35 <johill> if you have those you can define a name-space type collection too
2007-05-09T17:48:12 <johill> like NameSpaceCollection({'/':StackedCollection(...), '/FS': FSCollection()})
2007-05-09T17:49:09 <ThomasWaldmann> johill: would that be the same as lanius stuff just better separated into layering and composing?
2007-05-09T17:49:29 <johill> yeah
2007-05-09T17:49:59 <ThomasWaldmann> ok, then it matches my first idea of having it seperate :)
2007-05-09T17:50:20 <ThomasWaldmann> (same for the ACL stuff, btw :)
2007-05-09T17:50:30 <lanius> what means seperate?
2007-05-09T17:50:32 <johill> it's just less stuff to do for each piece
2007-05-09T17:50:51 <ThomasWaldmann> lanius: not doing layering and composing in one class
2007-05-09T17:50:53 <johill> you have everything from the dict lookup to the list/stack traversal in one dispatcher, my approarch separates those
2007-05-09T17:51:55 <lanius> what is the user of underlay=true in your example
2007-05-09T17:51:56 <ThomasWaldmann> my feeling is that would be simpler and better to debug / configure
2007-05-09T17:51:59 <lanius> s/user/use/
2007-05-09T17:52:24 <johill> lanius: it's to tell the filecollection that it should do the underlay dir
2007-05-09T17:52:36 <lanius> what is the difference to a not-underlay-dir
2007-05-09T17:52:37 <johill> of course, it's a dummy config parameter
2007-05-09T17:52:41 <johill> none
2007-05-09T17:52:44 <johill> it should probably be
2007-05-09T17:53:05 <johill> FileCollection(path='...'), FileCollection(path='...underlay...')
2007-05-09T17:53:11 <lanius> yeah
2007-05-09T17:53:19 <johill> just a dummy parameter to see how it works together
2007-05-09T17:53:28 <lanius> this meens that i split the DispatchingBackend in a StackedBackend and a NamespaceBackend
2007-05-09T17:53:44 <johill> yeah, basically
2007-05-09T17:55:26 <lanius> what is the python rule for variable names
2007-05-09T17:55:28 <ThomasWaldmann> lanius: maybe rather user "layer"
2007-05-09T17:55:38 <lanius> instead of stack?
2007-05-09T17:55:44 <ThomasWaldmann> stack in CS usually doesnt include transparency
2007-05-09T17:55:50 <ThomasWaldmann> yes
2007-05-09T17:57:38 <lanius> this_is_a_variable=
2007-05-09T17:57:39 <lanius> ?
2007-05-09T17:57:58 <johill> read pep8
2007-05-09T17:58:06 <johill> http://www.python.org/dev/peps/pep-0008/
2007-05-09T17:59:06 * ThomasWaldmann really appreciates we have that time for planning. last year it proceeded much quicker in SOC.
2007-05-09T18:00:57 <lanius> take a look at my configuration example
2007-05-09T18:01:23 <ThomasWaldmann> lanius: btw, you could log in :)
2007-05-09T18:01:30 <johill> looks good to me
2007-05-09T18:01:57 <lanius> ThomasWaldmann: too lazy :-)
2007-05-09T18:02:05 <johill> use openid ;) oh wait :P
2007-05-09T18:02:23 <johill> I'm off, later
2007-05-09T18:02:28 <lanius> ^^
2007-05-09T18:03:56 <ThomasWaldmann> lanius: it maybe could be even FileBackend(path=..., ...) (not putting all into a dict, but let python do that via **kw)
2007-05-09T18:05:19 <grzywacz> moin
2007-05-09T18:05:50 <lanius> i wanted to make it more readable instead of putting all into one line
2007-05-09T18:05:52 <ThomasWaldmann> and you could use "from MoinMoin.storage.backends import layered, FILE, ...
2007-05-09T18:06:05 <ThomasWaldmann> not "file", though :)
2007-05-09T18:06:18 <lanius> :)
2007-05-09T18:06:43 <ThomasWaldmann> the point is not repeating the class of things
2007-05-09T18:07:02 <ThomasWaldmann> e.g. we use from MoinMoin.macro import BR (not BR_macro)
2007-05-09T18:07:26 <lanius> but repeating it makes it more clear
2007-05-09T18:08:01 <ThomasWaldmann> in the end, it will be 3..6 lines in your config
2007-05-09T18:08:29 <lanius> if we do it with **kw it will be one line of configuration and some lines of import
2007-05-09T18:08:47 <ThomasWaldmann> and if you don't used handmade dicts, you can nest all the constructors directly
2007-05-09T18:09:00 <lanius> yes, one line, not very readable
2007-05-09T18:09:32 <ThomasWaldmann> no, you can split it into one statement, but multiple lines:
2007-05-09T18:09:42 <ThomasWaldmann> be = A(
2007-05-09T18:09:45 <lanius> yes
2007-05-09T18:09:50 <ThomasWaldmann> x = B(...),
2007-05-09T18:09:52 <lanius> but still not very readable
2007-05-09T18:09:56 <lanius> i will add a comment
2007-05-09T18:09:56 <ThomasWaldmann> y = C(...)
2007-05-09T18:09:58 <ThomasWaldmann> )
2007-05-09T18:10:36 <ThomasWaldmann> and you still can use handmade dicts with B(**mydict)
2007-05-09T18:10:43 <lanius> can you?
2007-05-09T18:11:00 <ThomasWaldmann> the python interpreter is your friend :)
2007-05-09T18:11:04 <lanius> yeah :-)
2007-05-09T18:11:32 <lanius> ok, have to go, but we mad another step :-)
2007-05-09T18:11:36 <lanius> cu
2007-05-09T18:14:39 <ThomasWaldmann> cu :)
2007-05-09T18:22:12 <xorAxAx> rehi
2007-05-09T18:23:58 <grzywacz> rehi xorAxAx ;)
2007-05-09T18:24:23 <xorAxAx> hi student
2007-05-09T18:24:25 <xorAxAx> :-)
2007-05-09T18:25:08 <grzywacz> Should I call you "My Mentor"? :>
2007-05-09T18:26:16 * xorAxAx digs in wikipedia
2007-05-09T18:26:19 <xorAxAx> how about "His Most Reverend Excellency"?
2007-05-09T18:26:49 <xorAxAx> :-)
2007-05-09T18:26:56 <grzywacz> That's too long. How about "HMRE"?
2007-05-09T18:26:57 <grzywacz> ;-)
2007-05-09T18:27:15 <xorAxAx> hehe
2007-05-09T18:28:30 <grzywacz> Quite close to "BDFL". ;>
2007-05-09T18:29:18 <vpv> now that you're talking about students and mentors, I will be doing MoinMoin related stuff in a Finnish Summercode event during the summer. Take a look at http://www.coss.fi/web/coss/developers/summercode/2007 and "man/info editor
2007-05-09T18:29:30 <vpv> " for a short description
2007-05-09T18:29:46 <xorAxAx> ah, hi vpv
2007-05-09T18:30:00 <xorAxAx> you submitted an application, right?
2007-05-09T18:30:43 <xorAxAx> "A simple Wiki-based diff tool to see what was added/subtracted across versions would be implemented." -- hmm, moin already does that :)
2007-05-09T18:30:49 <vpv> I did, for GSoC, but for this project I'm working technically for Fedora
2007-05-09T18:31:23 <vpv> and I have a mentor from Fedora, but I'm still interested in working as closely with you as possible
2007-05-09T18:31:45 <xorAxAx> maybe you have too much time and may want to merge the branch that a gsoc-student from fedora left over last year
2007-05-09T18:31:54 <xorAxAx> his plan were the docbook enhancements :)
2007-05-09T18:32:03 <xorAxAx> he is from finnland as well
2007-05-09T18:32:24 <xorAxAx> hmm, that project looks much smaller than gsoc :-)
2007-05-09T18:33:44 <vpv> we'll see how it goes, though the edited pages from moinmoin to man/info formats and back upstream may take some time and planning
2007-05-09T18:33:51 <xorAxAx> vpv: do you plan to use moin 1.6 or 1.7 as a base? :)
2007-05-09T18:34:06 <xorAxAx> ah, you are even interested in editing them
2007-05-09T18:34:18 <xorAxAx> (we only allow our students to use 1.7 :))
2007-05-09T18:34:57 <vpv> I'd really have to talk to the Fedora people about that, but if you recommend 1.7 to me, I'll recommend it to them ;)
2007-05-09T18:35:25 <xorAxAx> do you know if they are using the new docbook stuff?
2007-05-09T18:35:51 <xorAxAx> because its not publicly merged yet, so they would need somebody merging the stuff internally ...
2007-05-09T18:36:08 <vpv> I'm not sure if they are
2007-05-09T18:51:26 <ThomasWaldmann> vpv: hey, cool :)
2007-05-09T18:52:09 <ThomasWaldmann> (that was some of the ideas for linuxwiki.org, but never had time to do it :)
2007-05-09T18:54:45 <ThomasWaldmann> vpv: we can create a branch for your work, of course. did you look at mercurial yet?
2007-05-09T18:54:49 <vpv> ThomasWaldmann: hi! I'm very interested about the project, if/when I get it done, it'll be cool and useful for many other projects besides Fedora too.
2007-05-09T18:55:38 <ThomasWaldmann> whether 1.6 or 1.7 is better for you depends a bit on how intrusive your work would be
2007-05-09T18:56:16 <ThomasWaldmann> if it is rather separate (plugin-like), you could use 1.6. If it turns stuff upside down, rather 1.7.
2007-05-09T18:56:27 <vpv> ThomasWaldmann: I have basic knowledge of mercurial, but as the project will start in 28th of May, I haven't gotten a lot into it yet
2007-05-09T18:57:39 <vpv> But the projects were announced on a couple of Finnish IT websites and magazines today, so the ball if starting to roll, so to say :)
2007-05-09T18:58:57 <vpv> So far I think it could be made like a plug-in, but on the other hand I wouldn't like it to "bit rot" with 1.6 if you are releasing 1.7 soon-ish
2007-05-09T19:03:23 <vpv> though it seems Fedora is still using 1.5 and about to update to 1.6 soon, so I really have to talk about this with my mentor and other Fedora people
2007-05-09T19:04:02 <xorAxAx> no, 1.7 wont be released soonish
2007-05-09T19:35:19 <ThomasWaldmann> bbl
2007-05-09T20:24:37 <mpathy> vpv: who is mentoring you (ThomasWaldmann using mpathy's machine :)
2007-05-09T20:31:10 <vpv> mpathy: Karsten Wade from Fedora
2007-05-09T20:31:43 <vpv> I would be glad if I had a mentor from MoinMoin too, I have been in contact with Thomas before about my project
2007-05-09T20:43:51 <mpathy> moment please, now mpathy = mpathy :) ThomasWaldmann is back soon with his own machine
2007-05-09T20:47:41 <mpathy> he boots now ;)
2007-05-09T20:49:21 <vpv> ThomasWaldmann: oh sorry, I didn't notice you were mpathy there :) but yes, if I could get a mentor or atleast a "semi-mentor" from you too, that'd be great
2007-05-09T20:55:55 <ThomasWaldmann> re :)
2007-05-09T20:57:36 <ThomasWaldmann> ah, karsten wade, known guy :)
2007-05-09T20:59:25 <xorAxAx> a lover of documentation and integration
2007-05-09T21:38:03 <ThomasWaldmann> vpv: before beginning the project, you must swear to do great documentation :)
2007-05-09T21:41:18 <vpv> I'll do my best :) My intention is to work as closely with MoinMoin as possible, so that the deliverables would be useful to you and all Linux distros using MoinMoin. That's also something that was talked about in the COSS event final where they chose my project.
2007-05-09T21:41:45 <xorAxAx> vpv: integration is the most important goal :)
2007-05-09T21:43:54 <vpv> yeah, I think having my code in a branch of your mercurial would be helpful with that. And it would maybe help other MoinMoin people to test my code...
2007-05-09T21:47:30 <ThomasWaldmann> vpv: btw, we would also prefer having the docs in the moin wiki rather than having to search them somewhere in the fedora wiki afterwards X)
2007-05-09T21:48:11 <xorAxAx> of course you can choose to keep them only in the fedora wiki
2007-05-09T21:48:19 <xorAxAx> but integration will be very hard or impossible then
2007-05-09T21:49:22 <vpv> ThomasWaldmann: I think that could be done, I could then put links to moin wiki in the fedora wiki.
2007-05-09T21:50:16 <xorAxAx> vpv: also, fedora will automatically get the docs pages with the next update MoinMoin: MoinMoinChat/Logs/moin-dev/2007-05-09 (last edited 2007-10-29 19:16:27 by localhost)
