MoinMoin: MoinMoinChat/Logs/moin-dev/2006-05-27

2006-05-27T00:32:41  <ThomasWaldmann> lol, MSIE6 in place 8 of the worst tech products ever
2006-05-27T00:33:21  <xorAxAx> in place 8?
2006-05-27T00:35:30  <ThomasWaldmann> on
2006-05-27T00:35:34  <ThomasWaldmann> http://www.pcworld.com/reviews/article/0,aid,125772,pg,1,00.asp
2006-05-27T11:44:35  <Kepplar> ThomasWaldmann: I can't seem to get onto my branch
2006-05-27T11:54:29  <nwp> Kepplar: that makes you sound like a cat
2006-05-27T11:57:32  <Kepplar> heh
2006-05-27T12:02:56  <ThomasWaldmann> Kepplar: did you read the mercurial page in the mm wiki?
2006-05-27T12:06:00  <Kepplar> yes
2006-05-27T12:06:05  <Kepplar> you said use the plain url
2006-05-27T12:06:13  <Kepplar> yet you say the rep is called a
2006-05-27T12:06:26  <Kepplar> according to the page it says write the entire uri
2006-05-27T12:07:54  <ThomasWaldmann> how about just looking at hg.thinkmo.de with your browser? :)
2006-05-27T12:08:56  <Kepplar> ah.. its storage-a :)
2006-05-27T12:21:49  <fpletz> ThomasWaldmann: I have been looking through the source since yesterday and just thought about the parser-issue.. should lupy still be supported in 1.6 and further releases?
2006-05-27T12:22:09  <fpletz> the parser and lupy seem to be pretty strong tied together
2006-05-27T12:22:27  <xorAxAx> IMHO dropping it is an option if the alternative works well and the regex issue is solved
2006-05-27T12:22:30  <xorAxAx> fpletz: which parser?
2006-05-27T12:22:46  <xorAxAx> do you mean the query parser?
2006-05-27T12:22:48  <fpletz> in search.py, for the boolean searches
2006-05-27T12:22:49  <ThomasWaldmann> fpletz: no
2006-05-27T12:22:50  <fpletz> yeah
2006-05-27T12:22:58  <xorAxAx> yeah, they should they decoupled
2006-05-27T12:23:03  <xorAxAx> ThomasWaldmann: why not?
2006-05-27T12:23:17  <ThomasWaldmann> because it doesnt work correctly
2006-05-27T12:23:32  <xorAxAx> ah, ok
2006-05-27T12:23:42  <xorAxAx> i have remembered the question incorrectly, too early
2006-05-27T12:24:16  <xorAxAx> fpletz: the tight integration had a few problems, e.g. didnt moinmoin 1.3.5  run on python 2.2.2
2006-05-27T12:24:23  <xorAxAx> fpletz: so thats a bad example :)
2006-05-27T12:29:59  <fpletz> ok
2006-05-27T12:36:42  <Kepplar> ok
2006-05-27T12:36:46  <Kepplar> got mer working fully
2006-05-27T12:36:59  <Kepplar> is there a python eclipse plugin>
2006-05-27T12:37:30  <fpletz> yeah, pydev
2006-05-27T12:37:41  <fpletz> some student is even working on that for soc ;)
2006-05-27T12:37:42  <Kepplar> i usually do my development on vi, dont feel i want to this time :)
2006-05-27T12:38:02  <fpletz> I'll stick to vim
2006-05-27T12:38:45  <Kepplar> ThomasWaldmann: does the rep autogenerate epydoc on push?
2006-05-27T12:40:37  <fpletz> mmh, I got some errors this morning while trying to generate the api docs
2006-05-27T12:43:37  <Kepplar> btw guys, input on the specs for the storage engine interfaces would be useful as I will almost certainly miss things
2006-05-27T12:45:21  <ThomasWaldmann> Kepplar: no
2006-05-27T12:45:41  <ThomasWaldmann> Kepplar: soon
2006-05-27T12:47:11  <Kepplar> why would you want to pass a parser as an argument for canusecache on page?
2006-05-27T12:50:13  <xorAxAx> if the outer parser doesnt support cahcing, then caching is disabled completly
2006-05-27T12:51:15  <Kepplar> im not that familar on parsers with moin
2006-05-27T13:16:42  <Kepplar> is there any real documentation for the project?
2006-05-27T13:16:55  <Kepplar> im stuggling to pull things apart
2006-05-27T13:17:00  <Kepplar> and whats the deal with the wikiutil?
2006-05-27T13:17:20  <fpletz> Kepplar: I read the development notes in the wiki for an overview
2006-05-27T13:17:41  <fpletz> http://moinmoin.wikiwikiweb.de/MoinDev
2006-05-27T13:17:46  <Kepplar> i have :)
2006-05-27T13:17:57  <fpletz> ok ;)
2006-05-27T13:18:31  <Kepplar> im from the generation UML everything under the sun :)
2006-05-27T13:18:35  <Kepplar> that^
2006-05-27T13:19:42  <fpletz> oh, hehe
2006-05-27T13:21:40  <Kepplar> I didnt realise the parsers were so tightly-coupled
2006-05-27T13:21:57  <Kepplar> with the existing storage enginre
2006-05-27T13:22:58  <Kepplar> i can't see any documentation on the wiki regarding architecture, summery of what a formatter, parser IS for example
2006-05-27T13:24:09  <ThomasWaldmann> a parser gets a bunch of src lines and transforms that into formatter calls for some output format
2006-05-27T13:24:43  <ThomasWaldmann> the formatter gets those calls and e.g. linebreak() emits a <br> for text_html formatter
2006-05-27T13:25:58  <Kepplar> src lines of what, from where?
2006-05-27T13:26:34  <Kepplar> oh the wiki mimetype
2006-05-27T13:27:16  <ThomasWaldmann> the lines from between {{{#!parser and }}} for example
2006-05-27T13:27:26  <ThomasWaldmann> or a whole page
2006-05-27T13:35:23  <ThomasWaldmann> how about having some docs/CHANGES.$NAME for each student?
2006-05-27T13:56:09  <ThomasWaldmann> i guess this means "yes" :)
2006-05-27T13:58:31  <ThomasWaldmann> xorAxAx: any pending main-1.6 changes?
2006-05-27T13:59:03  <fpletz> yes ;)
2006-05-27T14:12:48  <ThomasWaldmann> ok, I recreate the student repos now with the real stuff
2006-05-27T14:19:43  <ThomasWaldmann> done
2006-05-27T14:20:07  <ThomasWaldmann> fpletz: Kepplar: xorAxAx: please see docs/CHANGES.$yourname
2006-05-27T14:21:12  <ThomasWaldmann> and get a fresh clone of your hg.thinkmo.de repo
2006-05-27T14:24:59  <fpletz> ok
2006-05-27T14:31:40  <fpletz> abort: No space left on device
2006-05-27T14:31:42  <fpletz> not good ;)
2006-05-27T14:43:18  <ThomasWaldmann> use hg 0.9, it is more space-efficient :)
2006-05-27T14:44:04  <fpletz> well, it's 0.9.. but it's my fault anyway ;)
2006-05-27T14:51:17  <xorAxAx> ThomasWaldmann: yes, i want to write a doc page for mail import
2006-05-27T14:51:22  <xorAxAx> ThomasWaldmann: do you want to release 1.6.0
2006-05-27T14:51:24  <xorAxAx> ...?
2006-05-27T14:52:14  <ThomasWaldmann> not immediately at least
2006-05-27T14:52:27  <ThomasWaldmann> but the help page can be put on moinmaster wiki
2006-05-27T15:11:39  <Kepplar> added 3 changesets with 5 changes to 4 files (+1 heads)
2006-05-27T15:11:39  <Kepplar> (run 'hg heads' to see heads, 'hg merge' to merge)
2006-05-27T15:11:49  <Kepplar> what does this mean?
2006-05-27T15:12:24  <xorAxAx> think of your repo as a DAG of revisions
2006-05-27T15:12:38  <xorAxAx> currently, you have one root but multiple leafs
2006-05-27T15:12:51  <xorAxAx> so you have to merge the leafs together
2006-05-27T15:13:05  <xorAxAx> if you want a colourful overview of the current state, checkout my branchview branch
2006-05-27T15:13:22  * Kepplar is lost
2006-05-27T15:13:22  <xorAxAx> i.e. fetch and install http://moin.pocoo.org:8080/hg-branchview-integration
2006-05-27T15:13:36  <xorAxAx> you know what a DAG is? :)
2006-05-27T15:13:41  <Kepplar> no
2006-05-27T15:13:48  <xorAxAx> directed acyclic graph
2006-05-27T15:14:12  <Kepplar> i fail to see how this has caused a new banch
2006-05-27T15:14:17  <xorAxAx> try that branchview thingie, helped me a lot to understand it
2006-05-27T15:14:17  <Kepplar> all i did was pull my repo
2006-05-27T15:14:35  <xorAxAx> did you have local commits?
2006-05-27T15:14:43  <Kepplar> eh?
2006-05-27T15:14:52  <xorAxAx> ah, never mind, i know the reason
2006-05-27T15:15:08  <xorAxAx> you pushed this minor test change to thomas repo
2006-05-27T15:15:16  <xorAxAx> and committed it locally as well
2006-05-27T15:15:17  <Kepplar> yep
2006-05-27T15:15:22  <Kepplar> i thougt i had to
2006-05-27T15:15:25  <xorAxAx> but thomas pruned that branch and recloned it
2006-05-27T15:15:35  <xorAxAx> so now he doesnt have that test change but you still have
2006-05-27T15:15:47  <Kepplar> so shouldnt the pull just get rid of it
2006-05-27T15:15:59  <xorAxAx> you can use pull -r revid to create a new local repo without that other branch
2006-05-27T15:16:09  <xorAxAx> use hg heads to see the revid of the correct branch
2006-05-27T15:16:20  <xorAxAx> no, why should pull delete revisions?
2006-05-27T15:16:25  <Kepplar> seriously all this has lost me
2006-05-27T15:16:39  <xorAxAx> i still recommend the branchview extension :)
2006-05-27T15:16:51  <xorAxAx> DVCS work differently than cvs/svn
2006-05-27T15:16:54  <Kepplar> maybe later, i dont like installing things which aren't apt get :P
2006-05-27T15:17:02  <xorAxAx> but once you understand it, its obvious :)
2006-05-27T15:17:15  <xorAxAx> you are using the deb package of moin? ugh %-)
2006-05-27T15:17:24  <Kepplar> i dont have moin really installed :P
2006-05-27T15:17:48  <xorAxAx> hehe
2006-05-27T15:18:14  <xorAxAx> you can try rm -rf and hg clone as well, but then you dont learn much :)
2006-05-27T15:18:19  <Kepplar> i really need a good IDE though, since its painful with the lack of documentation trying to see what calls what where
2006-05-27T15:19:24  <xorAxAx> yeah
2006-05-27T15:19:27  <xorAxAx> i can recommend komdo
2006-05-27T15:19:30  <xorAxAx> komodo
2006-05-27T15:20:46  <Kepplar> bah i wish there was a grand unified ide
2006-05-27T15:21:04  <xorAxAx> life is hard :)
2006-05-27T15:21:23  <Kepplar> ok
2006-05-27T15:21:31  <Kepplar> so i merged the changes and pushed it
2006-05-27T15:21:58  <Kepplar> what does parent mean in the changelog?
2006-05-27T15:22:15  <xorAxAx> which is sensible because your initial change actually made sense
2006-05-27T15:22:52  * xorAxAx sighs and points to http://moin.pocoo.org:8080/moin-1.6-main?cmd=branchview
2006-05-27T15:23:14  <xorAxAx> and the bottom, you see non-linear revisions. every changeset has at most 2 parent changesets
2006-05-27T15:23:42  <xorAxAx> i will mirror your branch in the next days as well, then you can checkout branchview on my server :)
2006-05-27T15:25:02  <Kepplar> cool
2006-05-27T15:25:21  <Kepplar> reason i cant be arsed to do much regarding that is this box is due for reformat soon
2006-05-27T15:28:35  <Kepplar> do we have build bot set up ?
2006-05-27T15:29:00  <xorAxAx> no
2006-05-27T15:30:03  <Kepplar> how do we test things then without installing it and sending that up to the repo?
2006-05-27T15:31:04  <xorAxAx> umm, you can run the unittests locally
2006-05-27T15:31:15  <Kepplar> I mean E2E
2006-05-27T15:32:38  <xorAxAx> E2E?
2006-05-27T15:32:44  <Kepplar> end to end
2006-05-27T15:32:45  * xorAxAx just knows J2EE :)
2006-05-27T15:33:05  <xorAxAx> there are no unittests that involve traversal of the request code, if you mean that :)
2006-05-27T15:33:39  <xorAxAx> besides that, it was typical that moin shipped 50-60 broken unittests to stress the quality of the code :)
2006-05-27T15:34:01  <xorAxAx> until tom moved them to the directory "broken" (?)
2006-05-27T15:34:39  <Kepplar> runtests.py wont run unless i install it
2006-05-27T15:35:39  <xorAxAx> ?
2006-05-27T15:35:49  <Kepplar> trst/runrests.py crashes
2006-05-27T15:35:51  <Kepplar> test*
2006-05-27T15:36:03  <xorAxAx> it might just show that your PYTHONPATH is set incorrectly
2006-05-27T15:36:15  <Kepplar> data_dir "testwiki/data" does not exists, or has incorrect ownership or
2006-05-27T15:36:15  <Kepplar> permissions.
2006-05-27T15:36:15  <Kepplar> Make sure the directory and the subdirectory pages are owned by the web
2006-05-27T15:36:15  <Kepplar> server and are readable, writable and executable by the web server user
2006-05-27T15:36:15  <Kepplar> and group.
2006-05-27T15:36:17  <Kepplar> It is recommended to use absolute paths and not relative paths. Check
2006-05-27T15:36:19  <Kepplar> also the spelling of the directory name.
2006-05-27T15:36:23  <Kepplar> aka it wants me to install it
2006-05-27T15:36:31  <xorAxAx> make test
2006-05-27T15:36:35  <xorAxAx> no, it doesnt
2006-05-27T15:36:44  <xorAxAx> thats just the normal "dir is wrong" error message
2006-05-27T15:36:55  <xorAxAx> make test correctly unpacks a test data dir
2006-05-27T15:37:53  <Kepplar> dude, seriously all this needs way better documentation
2006-05-27T15:38:07  <Kepplar> im also getting an error from make test
2006-05-27T15:40:32  <Kepplar> http://pastebin.com/741133
2006-05-27T15:41:52  <Kepplar> also how do you clean up the testwiki?
2006-05-27T15:43:59  <xorAxAx> its cleaned up on the next run :)
2006-05-27T15:44:07  <Kepplar> yea but its crashed :P
2006-05-27T15:44:10  <xorAxAx> of course it needs better docs, you can start today :)
2006-05-27T15:44:24  <xorAxAx> Kepplar: it just shows that a test is broken
2006-05-27T15:44:35  <xorAxAx> umm, i mean the code
2006-05-27T15:44:49  <xorAxAx> obviously the wiki parser is broken again
2006-05-27T15:46:05  <Kepplar> we'll need to decide on the interface specs soon (Everyone in general) because its going to have a massive effect on the wiki in general
2006-05-27T15:46:16  <xorAxAx> yep!
2006-05-27T15:46:18  <fpletz> yeah, my first real commit *g*
2006-05-27T15:46:29  <xorAxAx> note that you can commit to the main branch as well
2006-05-27T15:46:33  <xorAxAx> if you have found a bug etc.
2006-05-27T15:46:41  <xorAxAx> at least that would be my policy :)
2006-05-27T15:47:18  * xorAxAx creates mirrors
2006-05-27T15:50:41  <Kepplar> execute(self, request, parser, code)
2006-05-27T15:50:43  <Kepplar> Write page content by executing cache code
2006-05-27T15:50:51  <Kepplar> what does that actually do?
2006-05-27T15:50:54  <Kepplar> ThomasWaldmann: ?
2006-05-27T15:51:03  <xorAxAx> a little more context please :)
2006-05-27T15:51:10  <Kepplar> Page.py
2006-05-27T15:51:37  <xorAxAx> thats a page method?
2006-05-27T15:51:41  <Kepplar> yep
2006-05-27T15:51:45  <xorAxAx> in 1.6?
2006-05-27T15:51:53  <xorAxAx> looks weird
2006-05-27T15:52:03  <Kepplar> im reading the epydoc
2006-05-27T15:52:39  <Kepplar> ThomasWaldmann: can we arrnage a time to go over these relevent modules so I know whats depreciated, or why things exist
2006-05-27T15:52:48  <Kepplar>          get_current_from_pagedir(self, pagedir)
2006-05-27T15:52:49  <Kepplar> get the current revision number from an arbitrary pagedir.
2006-05-27T15:52:59  <Kepplar> for example I not surw why anyone would need that function
2006-05-27T15:58:40  <xorAxAx> umm, that sounds fairly broken for a storage interface :)
2006-05-27T16:02:15  <Kepplar> Kepplar: hence why I'll need ThomasWaldmann to go through with me whats there
2006-05-27T16:02:24  <Kepplar> I'm making modification to the interface specs now
2006-05-27T16:02:43  <xorAxAx> see http://moin.pocoo.org:8080/moin-1.6-storage-a?cmd=branchview for some colours, Kepplar  :)
2006-05-27T16:03:27  <xorAxAx> oh, your first/last merge wasnt clean. merges should be clean, i.e. only contain merging related changes
2006-05-27T16:03:41  <xorAxAx> i.e. you should have committed before hg merge
2006-05-27T16:05:36  <Kepplar> ah cool
2006-05-27T16:06:09  <Kepplar> ive lost you on that comment :)
2006-05-27T16:06:32  <xorAxAx> do you see your last commit in that graph?
2006-05-27T16:06:38  <xorAxAx> it says "marked user.py" ...
2006-05-27T16:07:21  <xorAxAx> now this changeset has 2 parents, i.e. you merged in that changeset both ("micro-") branches
2006-05-27T16:07:50  <xorAxAx> so you had actually two kind of changes in that "marked ..." commit: a merge and the addition of a line, which was unrelated to that merge
2006-05-27T16:07:55  <xorAxAx> Kepplar: does it make sense? :)
2006-05-27T16:08:53  <Kepplar> not really :P
2006-05-27T16:09:36  <xorAxAx> do you have a specific question?
2006-05-27T16:11:20  <xorAxAx> do you have the browser openned with the branchview page?
2006-05-27T16:13:04  <Kepplar> nah its ok, just nothing really makes sense
2006-05-27T16:13:09  <Kepplar> and i have a headache that doesnt help
2006-05-27T16:13:28  <Kepplar> how do you do colored text?
2006-05-27T16:14:57  <xorAxAx> on the wiki page?
2006-05-27T16:15:14  <xorAxAx> using the html formatter for example. or using table markup with CSS
2006-05-27T16:17:20  * Kepplar isnt able to find anything regarding formatters in help
2006-05-27T16:17:45  <xorAxAx> HelpOnActions and MoinDev talk about them
2006-05-27T16:19:07  <Kepplar> nothing on HelpOnActions
2006-05-27T16:19:28  <xorAxAx> "emit specially formatted views of a page. Parameter mimetype specifies the wanted mime-type."
2006-05-27T16:19:44  <xorAxAx> but it doesnt talk about formatters explicitly, yep
2006-05-27T16:39:31  <Kepplar> ok
2006-05-27T16:39:34  <Kepplar> gone through user.py now
2006-05-27T16:39:36  <Kepplar> hmm
2006-05-27T16:39:45  <Kepplar> i need to have a think about how to handle this then
2006-05-27T17:05:20  <ThomasWaldmann> re
2006-05-27T20:35:08  <fpletz> ThomasWaldmann: what about using xapian for the category stuff? this would basically just involve a new term prefix. I noticed this currently being pretty slow.
2006-05-27T20:37:16  <ThomasWaldmann> keep that idea for later
2006-05-27T20:37:44  <ThomasWaldmann> (maybe on a wiki page :)
2006-05-27T20:38:16  <fpletz> yup ;)
2006-05-27T20:41:16  <ThomasWaldmann> (that stuff might be influenced by metadata stuff we do in another branch)
2006-05-27T20:41:38  <xorAxAx> fpletz: yes, right, it should work like that :)
2006-05-27T20:41:53  <xorAxAx> and meta-data needs its own terms as well
2006-05-27T21:52:58  <mvirkkil> xorAxAx: So, what did you mean with "general DOM framework that mitigates such issues by e.g. fixing up the calls generated by the parsers."?
2006-05-27T22:05:53  <xorAxAx> mvirkkil: it means that this dom idea that you have already implemented can be pushed onto its own layer, in the first step residing in the docbook formatter, which can then be reused by other parsers
2006-05-27T22:06:41  <xorAxAx> i.e. more concretly, you could write (if you like to) an abstract formatter class that handles problems that arise because e.g. a para wasnt opened for the subclass
2006-05-27T22:19:17  <xorAxAx> mvirkkil: does that make sense?
2006-05-27T22:22:07  <Kepplar> Evening
2006-05-27T22:22:19  <Kepplar> ThomasWaldmann: reponded on the interface page
2006-05-27T22:23:10  <ThomasWaldmann> moin Kepplar
2006-05-27T22:23:13  <Kepplar> heya
2006-05-27T22:28:40  <ThomasWaldmann> Kepplar: as we don't have any section code, nor any concrete ideas or plans for it, maybe just ignore that stuff
2006-05-27T22:29:35  <Kepplar> ThomasWaldmann: I suppose so, if it does change, then making an index argument option won't break code later that uses the interface
2006-05-27T22:29:38  <Kepplar> ok :)
2006-05-27T22:29:54  <xorAxAx> IMHO section editing should be independent from the storage code
2006-05-27T22:30:06  <Kepplar> xorAxAx: any reason?
2006-05-27T22:30:50  <xorAxAx> Kepplar: because you would need to transform this multi segment view into every other layer above the storage code
2006-05-27T22:31:04  <xorAxAx> i.e. every other chunk of code has to know that there are multiple segments
2006-05-27T22:31:26  <xorAxAx> but section editing just matters when it comes to page editing
2006-05-27T22:31:38  <Kepplar> not true
2006-05-27T22:31:43  <xorAxAx> no?
2006-05-27T22:31:47  <Kepplar> because giving an index would be optional parameter
2006-05-27T22:31:58  <Kepplar> leaving it blank effectively treats the page as a normal page
2006-05-27T22:32:13  <xorAxAx> just think about this - how do you want to segment a page?
2006-05-27T22:32:27  <xorAxAx> and how do you ensure that the segment layout is still valid for concurrent editors?
2006-05-27T22:32:42  <xorAxAx> you cannot simply write n+1 segments if a user added a paragraph
2006-05-27T22:32:52  <xorAxAx> because that might break the editing for the other concurrent editor
2006-05-27T22:33:02  <xorAxAx> IMHO this is trouble some at this level, at least for this use case
2006-05-27T22:48:57  <Kepplar> hmm ok
2006-05-27T23:01:53  <mvirkkil> xorAxAx: Yeah, it makes sense (sorry about the lag, network trouble)
2006-05-27T23:02:44  <mvirkkil> xorAxAx: Basically extending FormatterBase in to something like FormatterDom
2006-05-27T23:02:56  <xorAxAx> yeah
2006-05-27T23:03:01  <mvirkkil> xorAxAx: And then implementing docbook formatter on top of that.
2006-05-27T23:03:07  <xorAxAx> yeah
2006-05-27T23:03:27  <xorAxAx> and this dom formatter might want to enforce some properties of the DOM
2006-05-27T23:03:35  <xorAxAx> i.e. when to have para and when not etc.
2006-05-27T23:03:41  <mvirkkil> xorAxAx: Yeah.. In case someone else uses som sort of dom, but right now docbook module is the only one=
2006-05-27T23:03:45  <mvirkkil> ?
2006-05-27T23:04:03  <xorAxAx> hmm, other formatters do as well
2006-05-27T23:04:26  <mvirkkil> xorAxAx: Ok. So all non-dom workarounds would be in one place.. It really does make sense.
2006-05-27T23:04:48  <xorAxAx> yep
2006-05-27T23:04:54  <mvirkkil> xorAxAx: What would the api look like?
2006-05-27T23:05:01  <xorAxAx> no idea :)
2006-05-27T23:05:22  <mvirkkil> xorAxAx: :D
2006-05-27T23:08:44  <mvirkkil> xorAxAx: One weird solution would be to make some sort of generic xml tree out of the page. Then one could either use plain xslt or python to generate whatever you want.
2006-05-27T23:09:30  <mvirkkil> xorAxAx: Or then just make some sort of python tree of the thing, and pass that to the formatters.
2006-05-27T23:10:55  <mvirkkil> xorAxAx: Thoughts?
2006-05-27T23:11:26  <xorAxAx> mvirkkil: that idea is not new :)
2006-05-27T23:12:01  <mvirkkil> xorAxAx: Not surprised.
2006-05-27T23:12:05  <xorAxAx> thats a solution, there are some hints about it on the wiki, but not a cooked plan
2006-05-27T23:14:52  <mvirkkil> xorAxAx: Well, basically it would be quite close to what the docbook thing does now. It's just that I'm not sure that implementing a formatter on top of it would be much easier. Sure it would remove some hacks, but one would have to walk the xml tree in turn.
2006-05-27T23:15:08  <xorAxAx> yeah
2006-05-27T23:15:17  <xorAxAx> but you can always have the old interface back
2006-05-27T23:15:41  <mvirkkil> xorAxAx: But I would kinda like using xslt to do the transformation. After all that's what it was created for (xstl that is).
2006-05-27T23:15:49  <xorAxAx> by requesting calls to that adhere to the old interface
2006-05-27T23:16:15  <xorAxAx> python doesnt ship an xslt transformer
2006-05-27T23:17:00  <mvirkkil> xorAxAx: ok...
2006-05-27T23:17:17  <xorAxAx> i mean, there is 4suite and moin can already do xslt transformation
2006-05-27T23:17:22  <xorAxAx> (for 3 years)
2006-05-27T23:17:33  <xorAxAx> but its a little bit too heavy weight IMHO
2006-05-27T23:17:54  <mvirkkil> xorAxAx: What other formatters use dom?
2006-05-27T23:18:30  <xorAxAx> just have a look
2006-05-27T23:18:36  <xorAxAx> at least the dom_xml does
2006-05-27T23:18:42  <xorAxAx> (no idea how thomas calls it now)
2006-05-27T23:19:23  <mvirkkil> yah, I figured that one did :P
2006-05-27T23:21:39  <mvirkkil> xorAxAx: I think one first step that might be good, is to make some auxiliary class, that contains constants used for the hacks...
2006-05-27T23:21:49  <xorAxAx> yeah
2006-05-27T23:22:06  <mvirkkil> stuff like "need to break these if we gat a heading".
2006-05-27T23:22:22  <mvirkkil> or something.
2006-05-27T23:23:05  <mvirkkil> But I'll continue thinking about this tomorrow, I need to do some administrative stuff before I go to sleep, and it's already past midnight.
2006-05-27T23:23:55  <xorAxAx> yeah, ok
2006-05-27T23:40:16  <ThomasWaldmann> main.wikiwikiweb.de == auto-updating test wiki for moin/1.6 main branch
2006-05-27T23:45:53  <ThomasWaldmann> xorAxAx: if you make your next push to your *aschremmer repo, check if sync.wikiwikiweb.de runs on updated code afterwards
2006-05-27T23:46:21  <xorAxAx> ThomasWaldmann: cant you test such things? i am sure that i will complain :)
2006-05-27T23:46:47  <ThomasWaldmann> it is tested with main, but who knows :)
2006-05-27T23:50:26  <ThomasWaldmann> fpletz: dito
2006-05-27T23:50:29  <ThomasWaldmann> Kepplar: dito
2006-05-27T23:50:49  <ThomasWaldmann> using xapian.wikiwikiweb.de and storage.wikiwikiweb.de
2006-05-27T23:53:04  <fpletz> thanks, nice work