MoinMoin: MoinMoinChat/Logs/moin-dev/2008-03-27

2008-03-27T00:04:21  <dreimark> gn
2008-03-27T00:06:17  <ThomasWaldmann> gn dreimark
2008-03-27T00:38:56  <napi> xorAxAx; application is in. I need to finish writing the page to go in the /cv/ folder on my site but it should be up by tomorrow morning (got to finish bit of dissertation talking about Gtk vs Qt ... fun)
2008-03-27T00:40:13  <xorAxAx> napi: yeah, no hurry
2008-03-27T00:40:20  <xorAxAx> napi: why is gtk vs qt relevant? :)
2008-03-27T00:41:00  <napi> We have to explain *every* choice we made when doing our dissertation
2008-03-27T00:41:26  <napi> got a smegging a4 page on why I choce c/c++ over java lol
2008-03-27T00:41:34  <napi> much to the hate of my tutor.. stupid man
2008-03-27T00:41:47  <xorAxAx> :-)
2008-03-27T00:49:22  <dennda> napi: what did you apply for? :)
2008-03-27T00:49:42  <napi> web installation+admin
2008-03-27T00:50:17  <napi> done something similar before for a CMS I used to work on so got a pretty good idea in my head how to go about doing it
2008-03-27T00:50:33  <napi> (admittedly that cms was doing it all through database, so will be different for this editing the config files)
2008-03-27T00:51:01  <dennda> ok
2008-03-27T00:56:29  <napi> finishing this dissertation bit then going to write the 'cv' page then I'm watching Resident Evil Exctintion and going to bed.
2008-03-27T02:02:55  <napi> righty. resident evil time
2008-03-27T02:03:02  <napi> then i'm off to bed. laters all
2008-03-27T09:22:59  <mitsuhiko> deja vu
2008-03-27T09:23:04  <mitsuhiko> storage engine :)
2008-03-27T09:25:08  <ThomasWaldmann> moin
2008-03-27T09:25:59  <ThomasWaldmann> mitsuhiko: not quite
2008-03-27T09:26:08  <mitsuhiko> ThomasWaldmann: why?
2008-03-27T09:26:54  <ThomasWaldmann> last year was implementation of storage api, backend etc.
2008-03-27T09:27:26  <ThomasWaldmann> this year is adding some missing bits to it, getting it into production
2008-03-27T09:34:25  <dreimark> moin
2008-03-27T09:37:57  <xorAxAx> mitsuhiko: praise lanius who developed the initial backend stuff :)
2008-03-27T09:38:48  <lanius> praise?
2008-03-27T09:38:54  <lanius> for what exactly
2008-03-27T09:39:24  <xorAxAx> lanius: for the work :)
2008-03-27T09:40:08  <xorAxAx> he was greeting us this morning with:
2008-03-27T09:40:09  <xorAxAx> 09:22:59 < mitsuhiko> deja vu
2008-03-27T09:40:09  <xorAxAx> 09:23:04 < mitsuhiko> storage engine :)
2008-03-27T09:41:11  <lanius> :)
2008-03-27T09:45:55  <mitsuhiko> xorAxAx: wasn't the storage keppys project?
2008-03-27T09:46:48  <ThomasWaldmann> mitsuhiko: that was the year before and not too successful X)
2008-03-27T09:47:08  <mitsuhiko> heh. 3rd lap then? :D
2008-03-27T09:51:10  <ThomasWaldmann> heh, Twisted jumped from 2.4.0 to 8.0.1 :P
2008-03-27T09:55:42  <xorAxAx> mitsuhiko: no, 4th or 5th
2008-03-27T09:57:28  <mitsuhiko> ThomasWaldmann: it's by years now
2008-03-27T10:48:22  <dennda> good morning
2008-03-27T10:48:48  <dennda> sounds promising :)
2008-03-27T11:47:29  <zenhase> moin
2008-03-27T12:55:19  <dreimark> moin zenhase
2008-03-27T13:10:04  <zenhase> ah, someone is alive :D
2008-03-27T13:10:45  <zenhase> how's business? are there already several student applications?
2008-03-27T13:15:11  <napi> There's at least 1 :p
2008-03-27T13:21:23  <dreimark> zenhase: two currently
2008-03-27T13:21:34  <dreimark> zenhase: and will you add the next one?
2008-03-27T13:22:11  <zenhase> ahhh :)
2008-03-27T13:22:20  <zenhase> i am still undecided
2008-03-27T13:22:35  <zenhase> regarding the whole SoC, not moin in particular
2008-03-27T13:22:50  <dreimark> what would you prefer to do ?
2008-03-27T13:23:08  <zenhase> that's what i am trying to find out for myself right now
2008-03-27T13:23:35  <napi> zenhase, what you undecided about? whether to apply for GSoC at all or not?
2008-03-27T13:23:44  <zenhase> napi: yeah
2008-03-27T13:23:57  <napi> well, here's the obvious question; why not? :)
2008-03-27T13:24:44  <zenhase> i am a bit anxious about not meeting expectations actually
2008-03-27T13:27:17  <dreimark> well we always stand up and try again if it does't go as we'd suspected
2008-03-27T13:27:48  <zenhase> let me think about it :)
2008-03-27T13:28:35  <zenhase> this evening i go climbing again with a friend who tried to motivate me for this years SoC
2008-03-27T13:29:39  <zenhase> i will have some more chat with him
2008-03-27T13:30:17  <napi> zenhase; thats a silly reason not to apply to be honest :)
2008-03-27T13:30:57  <napi> To be stupidly blunt it doesn't matter in the sense that the project loses nothing except a little amount of time which would be no different to them helping any random person who wants to help but gets no where in the end
2008-03-27T13:31:18  <napi> Also, it's very unlikely you'd make no progress over the project. Even if ultimately it doesn't get finished, any work done is work done. It can still be used
2008-03-27T13:31:46  <napi> And lastly, the mentors help you! As long as you set yourself a realistic goal, plan it well and stay motivated, there's no reason you can't finish it
2008-03-27T13:31:59  <napi> The mentors will tell you straight away if you're project is too much for GSoC
2008-03-27T13:32:14  <napi> If they accept, it's a pretty fair assumption that they believe it's do-able in the time alloted :)
2008-03-27T13:32:21  <napi> So grow a pair, man up, and do an application
2008-03-27T13:32:25  * napi glares
2008-03-27T13:33:11  <zenhase> napi: you know that i did a soc-project for moin last year? :)
2008-03-27T13:33:28  <napi> lol so what the hells the problem then?! :)
2008-03-27T13:33:31  <zenhase> which did not finish :)
2008-03-27T13:33:43  <napi> finish it this year then
2008-03-27T13:34:05  <napi> or set yourself a smaller task
2008-03-27T13:34:30  <zenhase> hmm
2008-03-27T13:35:25  <zenhase> i will clear my head with climbing this evening ... then afterwards i can make a plan about what i could do and if it's doable and such
2008-03-27T13:35:38  <zenhase> sounds like a plan?
2008-03-27T13:35:46  <napi> sounds like a good plan
2008-03-27T13:36:41  <zenhase> :)
2008-03-27T13:42:03  <dreimark> napi: the moin command has a parameter export package
2008-03-27T13:42:50  <dreimark> sometimes people like to setup a local wiki and to transfer this content to an other side
2008-03-27T13:43:20  <dreimark> they can use wikisync or packages
2008-03-27T13:43:25  <napi> aah
2008-03-27T13:43:27  <napi> I see what you mean
2008-03-27T13:44:51  <johill> dreimark: hm I can reproduce it
2008-03-27T13:45:02  <napi> Well then yeah, you could use that once you've setup the local wiki, export all the settings, then load those in to a new location I guess
2008-03-27T13:45:23  <dreimark> johill: I was wrong, (you talk about my statement yesterday?)
2008-03-27T13:45:33  <johill> dreimark: about you saying you cannot reproduce it
2008-03-27T13:45:55  <napi> first page of the installation could be offering choice of two options; start a new wiki, import settings for an existing wiki (allow upload for the settings file etc)
2008-03-27T13:46:42  <dreimark> johill: xorAxAx has explained me that I always have to get a edit conflict in this case
2008-03-27T13:46:53  <johill> dreimark: what did you expect?
2008-03-27T13:47:30  <dreimark> as it works now, because the saving is not done simultaniously, one is before the other
2008-03-27T13:47:37  <dreimark> but this was wrong
2008-03-27T13:48:06  <dreimark> It behaves a bit different, sometimes the second browser is reloaded with changed content
2008-03-27T13:48:06  <johill> ah ok yeah I see where the confusion comes in
2008-03-27T13:48:16  <johill> its a high-level conflict, not a low-level threading conflict
2008-03-27T13:48:21  <johill> though both can actually happen
2008-03-27T13:48:42  <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3752:c6e49c414651 1.7-storage-hwendel/BUGS: storage bug: rename vs. the rest
2008-03-27T13:48:45  <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3753:7830508b20cb 1.7-storage-hwendel/MoinMoin/config/multiconfig.py:
2008-03-27T13:48:45  <CIA-39> Don't index on DELETED meta key
2008-03-27T13:48:45  <CIA-39> Indexing on DELETED doesn't make sense because the index
2008-03-27T13:48:45  <CIA-39> cannot optimise the common search for NOT DELETED since
2008-03-27T13:48:47  <CIA-39> non-existence of the DELETED metadata key indicates that
2008-03-27T13:48:49  <CIA-39> the item is deleted.
2008-03-27T13:48:53  <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3754:7d4a1c53e26d 1.7-storage-hwendel/MoinMoin/search/term.py: introduce HasMetaDataValue search term
2008-03-27T13:48:56  <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3755:b59c83b55ea1 1.7-storage-hwendel/MoinMoin/user.py: use HasMetaDataValue to fix openid search
2008-03-27T13:48:58  <dreimark> do you have seen this changed content behaviour in the second browser too?
2008-03-27T13:48:59  <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3756:5f5ad4c1325e 1.7-storage-hwendel/MoinMoin/storage/backends/filesystem.py: fix index backend metadata search optimisation
2008-03-27T13:49:02  <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3757:a2739d056478 1.7-storage-hwendel/MoinMoin/search/term.py: search terms: implement proper boolean term objects
2008-03-27T13:49:07  <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3758:0432974922a4 1.7-storage-hwendel/MoinMoin/storage/backends/common.py: restore broken rename_item tests in "common" backend code so test suite doesn't go south
2008-03-27T13:49:11  <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3759:3f818d653043 1.7-storage-hwendel/MoinMoin/storage/_tests/test_backends_meta.py: test suite should not assume list_items is sorted
2008-03-27T13:49:16  <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3760:b571907c7382 1.7-storage-hwendel/docs/CHANGES.storage: add a few TODO items
2008-03-27T13:49:19  <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3761:ce77f9434b6f 1.7-storage-hwendel/docs/CHANGES.storage: more todo
2008-03-27T13:49:22  <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3762:203f6add9cea 1.7-storage-hwendel/MoinMoin/search/_tests/test_terms.py: use term.BOOL in term tests
2008-03-27T13:49:34  <dreimark> napi: :)
2008-03-27T13:49:51  <napi> dreimark :)
2008-03-27T13:51:43  <dreimark> johill: do you get this reload of content in the second browser sometimes too?
2008-03-27T13:54:59  <johill> hm I haven't done the test many times
2008-03-27T13:55:04  <johill> I don't think so
2008-03-27T13:55:49  <johill> dennda: after fixing the bugs the way I propsed it'll actually mostly look like the class diagram I made ;)
2008-03-27T13:55:51  <dreimark> I ve seen it as i entered some more lines and probably  with the existing page
2008-03-27T14:04:42  <dennda> johill: great! (that means I don't have to create one) :)
2008-03-27T14:05:06  <johill> now, I expect lanius to not accept such changes
2008-03-27T14:05:10  <johill> and come up with more locking to fix it
2008-03-27T14:05:59  <dennda> well, if something is borked by design you should rather fix the design than build workarounds, shouldn't you?
2008-03-27T14:07:20  <johill> but that collides with the all-over-shadowing goal of having trivial backends
2008-03-27T14:10:28  <dennda> why is that? because you need to change the backends? (to raise appropriate errors, for instance)
2008-03-27T14:10:47  <johill> because the backends will have to do smarter things
2008-03-27T14:11:22  <johill> look at bug §6
2008-03-27T14:11:32  <johill> eh, bug #6, that I just added
2008-03-27T14:11:35  <dennda> oh you added a new one
2008-03-27T14:11:40  * dennda pulls
2008-03-27T14:12:52  <dennda> Looks like another issue derived from the lack of thread-safety, on first glance
2008-03-27T14:17:45  <johill> yeah
2008-03-27T14:17:57  <johill> but a bit worse because rename_item can't properly be locked against readers
2008-03-27T14:18:04  <johill> unless you want to take locks for all readers too...
2008-03-27T14:28:43  <dennda> What other options (those that fix those bugs rather than working around them) do you have?
2008-03-27T14:28:47  <dennda> Not that many, imho
2008-03-27T14:32:35  <johill> maybe yet-unthought ones
2008-03-27T14:33:00  <dennda> I just had a crazy idea that I discarded immediately
2008-03-27T14:33:59  <johill> heh, what_
2008-03-27T14:34:00  <johill> ?
2008-03-27T14:36:03  <dennda> make get_data() aware of changes that have happened to rev since it has been instantiated (utter bullshit, imho)
2008-03-27T14:36:52  <dennda> you could've used the observer pattern or something similar
2008-03-27T14:36:57  <johill> that can't really work across processes or you're thinking of something else
2008-03-27T14:37:53  <dennda> well, I said I discarded it immediately, didn't I? :)
2008-03-27T14:38:31  <johill> :)
2008-03-27T14:39:13  <dennda> still only two applications?
2008-03-27T14:44:18  <johill> haven't checked
2008-03-27T14:53:46  <ThomasWaldmann> currently 2
2008-03-27T14:54:07  <dennda> is that usual? only 4 days left, counting today
2008-03-27T14:54:10  <ThomasWaldmann> i expect most stuff arriving late, usually people work until almost last minute
2008-03-27T14:54:32  <ThomasWaldmann> (or at least last day :)
2008-03-27T14:55:27  <dennda> hopefully
2008-03-27T14:59:17  <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3763:93f710373743 1.7-storage-hwendel/BUGS: more explanation for bug 6
2008-03-27T15:00:05  <johill> maybe that helps clear up reservations about how these different semantics make backends impossible to implement by mere mortals ;)
2008-03-27T15:04:21  <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2612:c7641d8e269e 1.6/ (MoinMoin/parser/text_moin_wiki.py docs/CHANGES): wiki parser: removed protocol check for (image) transclusion
2008-03-27T15:11:41  <johill> lanius: can you comment on the bugs I point out in the file? maybe you have much better ideas for fixing them
2008-03-27T15:17:08  <johill> can you have docstrings for properties?
2008-03-27T15:19:20  <johill> I guess not. hrm
2008-03-27T15:36:29  <kikka> Good Morning.
2008-03-27T15:37:12  <johill> hm. maybe we should assign a unique ID metadata to each page for things like FeatureRequests/EmailNotifOnlyOnce
2008-03-27T15:39:46  <dreimark> this will give the possibility to rename pagelinks with renaming of pages too
2008-03-27T15:40:02  <johill> well
2008-03-27T15:40:10  <johill> only if you want to take the hit of storing IDs
2008-03-27T15:40:20  <johill> which then means you need a very efficient ID->name mapping
2008-03-27T15:43:02  <johill> if done, though, it should be the responsibility of the storage code so that sql storage for example can just use the table ID
2008-03-27T15:50:24  * johill leaves, will think a bit more about create_item later
2008-03-27T16:07:42  <xorAxAx> wohoo, now we get mails
2008-03-27T16:07:48  <xorAxAx> for every new application
2008-03-27T16:18:26  <dreimark> cool
2008-03-27T16:20:51  <dreimark> let's see when we get a new mail
2008-03-27T16:31:26  <godog> dennda: looking at my exam schedule I won't be able to attend GSoC this year, so the storage stuff is all yours if nobody shows up :)
2008-03-27T16:31:56  <godog> I'll idle^Whang here tho
2008-03-27T16:37:21  <dennda> xorAxAx: how was your exam?
2008-03-27T16:37:58  <xorAxAx> dennda: after 2.5 of 4 hours, i didnt have any more ideas to write down :)
2008-03-27T16:38:09  <xorAxAx> but i hope that the profs likes it nevertheless
2008-03-27T16:38:16  <dennda> xorAxAx: I don't know whether that is good or bad
2008-03-27T16:39:14  <xorAxAx> me neither
2008-03-27T16:39:26  <xorAxAx> i filled 9 pages
2008-03-27T16:40:18  <dennda> let's just hope the best
2008-03-27T16:40:33  <dennda> xorAxAx: so how many applications are there at the moment? sounds you got at least one new
2008-03-27T16:41:18  <xorAxAx> napi and you
2008-03-27T16:42:04  * dennda files bogus applications
2008-03-27T16:42:35  <xorAxAx> :)
2008-03-27T16:42:55  <napi> woo
2008-03-27T16:43:13  <dennda> that throws the question:
2008-03-27T16:43:31  <dennda> is the number of slots relative to the number of applications or the number of students?
2008-03-27T16:43:47  <dennda> (just kidding)
2008-03-27T16:45:33  <napi> hm
2008-03-27T16:45:44  <napi> xorAxAx, last chance to quiz me on my app on irc before I go to cardiff in a few hours
2008-03-27T16:45:48  <napi> :p
2008-03-27T16:46:07  <napi> getting my head screwed up on whether not many questions/comments is a good/bad/no thing
2008-03-27T16:46:15  <TheSheep> napi: why not just put the wikiconfig.py on a wiki page? ;)
2008-03-27T16:46:23  <xorAxAx> napi: its fine if you leave, we will ask you next week :)
2008-03-27T16:46:28  <napi> heh
2008-03-27T16:46:43  <xorAxAx> dennda: number of applications, roughly
2008-03-27T16:46:46  <napi> TheSheep; no time to actually write it. Am blasting out dissertation stuff at the moment cause I go away for 3 days this evening
2008-03-27T16:46:50  <dennda> xorAxAx: harr harr
2008-03-27T16:46:54  <dennda> sounds exploitable :)
2008-03-27T16:48:18  <TheSheep> dennda: no, humans decide
2008-03-27T16:54:30  <dennda> ah damn they are clever
2008-03-27T16:55:11  <xorAxAx> TheSheep: welll
2008-03-27T16:55:29  <xorAxAx> TheSheep: you could still construct applications that bump the slot limit
2008-03-27T16:55:34  <xorAxAx> they could even make sense
2008-03-27T17:18:22  <TheSheep> xorAxAx: too much work
2008-03-27T17:20:08  <zenhase> hmm
2008-03-27T17:20:25  <zenhase> and hello
2008-03-27T17:27:54  <dennda> zenhase: back from climbing?
2008-03-27T17:29:42  <zenhase> nope
2008-03-27T17:29:49  <zenhase> climbing is from 1900-2300
2008-03-27T17:30:21  <zenhase> i am back from town, where i bought my own climbing equipment
2008-03-27T17:30:34  <zenhase> started this about 3 weeks ago
2008-03-27T17:31:11  <zenhase> first time in years i really enjoy sports actually
2008-03-27T17:31:33  <dennda> oh, is it that much fun?
2008-03-27T17:31:49  <dennda> maybe I should give it a try
2008-03-27T17:31:50  <kikka> Do you climb with blackwing? :)
2008-03-27T17:31:52  <zenhase> imo yes
2008-03-27T17:31:55  <dennda> how much is the equipment?
2008-03-27T17:32:10  <zenhase> kikka: no, he is about 400km from here
2008-03-27T17:32:29  <zenhase> dennda: you don't need it right away, but i just invested about 130eu
2008-03-27T17:32:44  <zenhase> pair of climbing shoes and security belt
2008-03-27T17:33:05  <kikka> zenhase: oh.
2008-03-27T17:33:25  <dennda> uh expensive
2008-03-27T17:33:26  <godog> zenhase: I guess physical preparation is needed as well?
2008-03-27T17:33:47  <zenhase> hmm?
2008-03-27T17:33:49  <godog> basic preparation, that is
2008-03-27T17:34:02  <zenhase> you only need to learn basic security techniques
2008-03-27T17:34:16  <zenhase> which boils down to 2 knots and howto secure someone else
2008-03-27T17:34:39  <dennda> some muscles are needed as well, I guess
2008-03-27T17:34:39  <godog> I see, thanks
2008-03-27T17:34:40  <zenhase> learn it in 10min ... from there climbing IS your physical preparation :)
2008-03-27T17:34:44  <zenhase> dennda: nope
2008-03-27T17:34:50  <zenhase> dennda: it's more about technique
2008-03-27T17:35:06  <zenhase> also you don't take on the hard stuff from the beginning
2008-03-27T17:35:17  <zenhase> you can start climbing right away without any knowledge
2008-03-27T17:35:43  <zenhase> i mean, it's funny ... it's something our body already knows, nobody has to explain to you how to climb up somewhere
2008-03-27T17:35:50  <godog> intriguing
2008-03-27T17:36:04  <zenhase> and with time you learn how to use your body effectivly
2008-03-27T17:36:18  <kikka> Wow, 130eur aren't too much :) I invented 300eur+ for a wakeboarding course...
2008-03-27T17:36:57  <zenhase> well, you do not even need the equipment
2008-03-27T17:37:21  <zenhase> in indoor venues you can rent the equipment
2008-03-27T17:37:41  <zenhase> it's about 5eu for the day
2008-03-27T17:37:48  <zenhase> + entry fee
2008-03-27T17:38:02  <dennda> how much is the fee?
2008-03-27T17:38:09  <zenhase> depends on the venue
2008-03-27T17:38:26  <dennda> roughly?
2008-03-27T17:38:31  <zenhase> right now i am going to a venue that belongs to the deutscher alpin verband
2008-03-27T17:38:47  <zenhase> and pay 3eu for the evening
2008-03-27T17:39:07  <zenhase> commercial ones usually charge about 5-6eu for students
2008-03-27T17:39:11  <dennda> that's ok
2008-03-27T17:39:22  <zenhase> yeah, you usally stay there for some hours
2008-03-27T17:39:31  <zenhase> so it's really not that expensive
2008-03-27T17:39:58  <dennda> maybe I'll try that with my girlfriend
2008-03-27T17:40:13  <zenhase> hrhr
2008-03-27T17:40:33  <godog> I though you are supposed to climb walls, not people
2008-03-27T17:40:35  * godog runs
2008-03-27T17:40:48  <godog> thought even
2008-03-27T17:41:31  <dennda> :/
2008-03-27T17:42:33  <zenhase> what's damn strange: most of the people i met during the last few weeks there climbing ... they are computer science students or at least math
2008-03-27T17:42:46  <zenhase> seems like in karlsruhe it's some kind of geek sport
2008-03-27T17:46:33  <dreimark> hmm the old Hits macro reading event-log at once is much much faster as the current one
2008-03-27T17:47:34  <dennda> xorAxAx: willing to review my application now? I changed quite some parts
2008-03-27T17:52:18  <dreimark> bbl
2008-03-27T18:07:42  <napi> righty. leaving for my train in 20 minutes- any questions from mentors about my app before I go? :)
2008-03-27T18:07:55  <napi> (or from any one else in fact... doesn't just have to be mentors)
2008-03-27T18:08:30  <gizmach> hi all
2008-03-27T18:08:40  <napi> hey giz
2008-03-27T18:09:23  <gizmach> napi: :)
2008-03-27T18:21:41  <gizmach> napi: Do u know what to include in my application for gsoc in The Abstract of an accepted student application is displayed on the public GSoC web site. and for Detailed Description . I know it's probbably a stupid question but I'm very confused
2008-03-27T18:22:59  <napi> umm
2008-03-27T18:23:23  <napi> abstract just do an overview of the project, explaining (mostly in layman's terms) what you're going to do and why you're going to do it
2008-03-27T18:23:26  <gizmach> napi: :) it's ok I found something on faq for gsoc students applicaions
2008-03-27T18:23:27  <napi> don't put any personal info in there
2008-03-27T18:23:29  <napi> kk
2008-03-27T18:23:35  <gizmach> :))
2008-03-27T18:23:40  <gizmach> napi: :) thx
2008-03-27T18:23:44  <napi> nps
2008-03-27T18:30:57  <napi> Right i'm off. have a good weekend all. see you sunday
2008-03-27T19:14:07  <grzywacz> moin
2008-03-27T19:37:29  <gizmach> grzywacz: moin
2008-03-27T19:37:41  <grzywacz> gizmach, hey
2008-03-27T19:59:09  <gizmach> I'm still interested in DOM project at gsoc is here someone I could ask some questions about that project
2008-03-27T20:00:32  <dennda> gizmach: I guess you can just ask your questions and they will be answered as soon as someone knowing the answer reads it :)
2008-03-27T20:05:52  <gizmach> dennda: well, I'm interested if esign and implement a DOM tree based interface to the output formatter means that you would like to have something like derivation tree ?
2008-03-27T20:08:02  <ThomasWaldmann> gizmach: -v
2008-03-27T20:09:25  <gizmach> ThomasWaldmann: ?
2008-03-27T20:10:27  <damjan> -v I guess means be more verbose
2008-03-27T20:14:32  <gizmach> ThomasWaldmann: well I was considering an idea of make recursive derivation tree - from a left recursion or something like that get a DOM tree based interface... Maybe I didn't understand the project idea.
2008-03-27T20:14:38  <gizmach> damjan: thx :)
2008-03-27T20:14:52  <gizmach> ThomasWaldmann: If this is a bit clearer now
2008-03-27T20:16:09  <ThomasWaldmann> the idea is to make the wiki parser build a treelike representation of the wiki page
2008-03-27T20:16:30  <ThomasWaldmann> then do misc post processing
2008-03-27T20:16:37  <ThomasWaldmann> finally generate output from the tree
2008-03-27T20:17:45  <ThomasWaldmann> (the tricky part is to do that in a very generalized way, so that macros, includes, non-text-type objects work as they should)
2008-03-27T20:17:52  <gizmach> ThomasWaldmann: than you could probabbly from the grammar defined in the parser construct a left or some new overriden type of parser to generate that type
2008-03-27T20:18:08  <gizmach> ThomasWaldmann:  yes I was thinking about that
2008-03-27T20:18:52  <ThomasWaldmann> there is no grammar
2008-03-27T20:19:00  <ThomasWaldmann> it's a bunch of regexes
2008-03-27T20:19:22  <gizmach> ThomasWaldmann: but the regexes define how it is parsing
2008-03-27T20:20:08  <gizmach> I lost the link on currend parser , could you please send it to me, I can't find it right now
2008-03-27T20:20:35  <ThomasWaldmann> hg.moinmo.in
2008-03-27T20:20:43  <gizmach> thanks
2008-03-27T20:20:57  <waldi> well, regexps are equivalent to grammars
2008-03-27T20:21:39  <gizmach> waldi: I thaught so
2008-03-27T20:22:30  <waldi> but if you want to define a regexp for email addresses you'll find the limitatiins
2008-03-27T20:25:17  <ThomasWaldmann> gizmach: look at the text_moin_wiki parser, the text_creole parser, Include macro, other macros, TOC, and the discussion on the wiki
2008-03-27T20:28:19  <gizmach> ThomasWaldmann: ok, but I really think some kind of recursion parser should really work, the only question is will it be too slow, and how to handle with that
2008-03-27T20:29:12  <ThomasWaldmann> gizmach: wiki pages are not programming code
2008-03-27T20:29:46  <ThomasWaldmann> thus, a parser needs to be much more forgiving than e.g. a python language parser
2008-03-27T20:30:18  <ThomasWaldmann> and for the slowness: we need caching of results, then speed of parsing does not matter
2008-03-27T20:31:42  <gizmach> ThomasWaldmann: ok... i figure tht
2008-03-27T20:33:21  <ThomasWaldmann> the current parser/formatter combo uses a nice hack for this: the text_python formatter output python code that generates output for a wiki page when executed
2008-03-27T20:33:41  <ThomasWaldmann> that code is then compiled to bytecode and stored on disk
2008-03-27T20:34:10  <ThomasWaldmann> if another request for that page arrives, it loads the bytecode and executes it
2008-03-27T20:34:47  <gizmach> ThomasWaldmann: then after compiling to bytecode you don't have any work with parser if I understand it? Or am I wrong?
2008-03-27T20:35:34  <ThomasWaldmann> not in many cases at least.
2008-03-27T20:35:58  <ThomasWaldmann> of course some stuff needs to stay dynamic.
2008-03-27T20:37:03  <gizmach> ThomasWaldmann: of course , OK now I understand why speed is not so important
2008-03-27T20:38:14  <ThomasWaldmann> (if caching and dependencies work ok :)
2008-03-27T20:42:36  <gizmach> ThomasWaldmann: also I don't understand this "Keep stack of all open tags as state (don't use the DOM tree directly for this)" on http://moinmo.in/WikiDomFormatter , could you please explain it better
2008-03-27T20:49:57  <gizmach> ThomasWaldmann: it would probably be a continuing of this
2008-03-27T21:01:32  <ThomasWaldmann> gizmach: that was mostly written by FlorianFesti
2008-03-27T21:02:00  <gizmach> ThomasWaldmann: I saw his name
2008-03-27T21:03:44  <ThomasWaldmann> gizmach: i think it is because of forgiveness
2008-03-27T21:06:25  <gizmach> ThomasWaldmann: what?
2008-03-27T21:06:33  * gizmach has blond moments all day
2008-03-27T21:07:01  <dennda> umm
2008-03-27T21:07:20  <ThomasWaldmann> gizmach: imagine that content:
2008-03-27T21:07:24  <ThomasWaldmann> '''foo
2008-03-27T21:07:40  <ThomasWaldmann> so what would the output be like?
2008-03-27T21:07:51  * waldi .o0( parse error ... )
2008-03-27T21:08:04  <waldi> <strong>foo</strong>
2008-03-27T21:10:22  * dennda just got to know that google won't shoot you if you already start coding within the "Community Bonding Period", which is quite great
2008-03-27T21:16:54  <gizmach> accordingly to http://moinmo.in/MoinDev/WikiDom?action=show&redirect=MoinMoinDOM diagram you don't need to touch your parser, after macros processors just to make postprocessing. I'm not sure how difficult will be to overwrite parser.
2008-03-27T21:17:15  <TheSheep> '''foo ''bar '''baz'' foo'''
2008-03-27T21:21:09  <TheSheep> gizmach: the dom tree cannot represent "crossed" tags like these
2008-03-27T21:21:47  <TheSheep> gizmach: if you keep track of them as state during parsing, you can generate a DOM tree that approximates them
2008-03-27T21:22:39  <TheSheep> gizmach: so instead of <b>foo <i>bar</b> baz</i> you get <b>foo <i>bar</i></b><i> baz</i>
2008-03-27T21:24:37  <gizmach> TheSheep: ok, than you make some osrt of utomata that you can't close some outer open tag if you don't close inner tag , with every new open tag you get into new state if automata
2008-03-27T21:25:26  <gizmach> and if you want to close an outer tag u got error probbably
2008-03-27T21:25:36  <gizmach> (just trying to represent a problem to myself)
2008-03-27T21:25:37  <TheSheep> gizmach: you just have variables like in_em, in_strong, etc. or something equivalent
2008-03-27T21:25:42  * waldi wrote a stack-parser which just pops elements in such a case
2008-03-27T21:26:04  <TheSheep> waldi: then the 'bar' wouldn't be italic
2008-03-27T21:26:23  <TheSheep> waldi: I do the same in the creole parser, I think
2008-03-27T21:27:42  <TheSheep> s/'bar'/'baz'
2008-03-27T21:27:45  <TheSheep> sorry
2008-03-27T21:29:42  <gizmach> TheSheep: I meand the stack parser but I didn't know to say it good
2008-03-27T21:29:51  <gizmach> ups
2008-03-27T21:34:50  <gizmach> TheSheep: also u have something like out_em
2008-03-27T21:34:53  <gizmach> ?
2008-03-27T21:38:21  <kikka> Night
2008-03-27T21:38:49  <gizmach> so when you get in_foo you probably instance a child node and after the out_foo you go to parent node back
2008-03-27T22:01:23  <xorAxAx> re
2008-03-27T22:04:50  <dennda> hey xorAxAx, wb
2008-03-27T22:05:06  <xorAxAx> i painted the room of my brother together with my father :-)
2008-03-27T22:05:35  <dennda> fun
2008-03-27T22:20:48  <TheSheep> gizmach: sorry, I'm back :)
2008-03-27T22:20:56  <TheSheep> gizmach: no, in_foo would be a bool variable
2008-03-27T22:21:16  <TheSheep> gizmach: when you encounter an opening tag, you set it to true
2008-03-27T22:22:21  <TheSheep> gizmach: and when closing -- to false
2008-03-27T22:22:42  <TheSheep> gizmach: this way when you close a tag, you can check if any of the contained tags need reopening
2008-03-27T22:23:39  <TheSheep> gizmach: using multiple variables for keeping state is much more convenient for the programmer -- although probably not for the mathematician
2008-03-27T22:25:03  <gizmach> TheSheep: oh nowI understand
2008-03-27T22:26:08  <TheSheep> gizmach: and since you can't nest ''' or '', you don't need a stack in there
2008-03-27T22:26:25  <gizmach> TheSheep: yes now it's ok
2008-03-27T22:26:27  <gizmach> :)
2008-03-27T22:26:28  <gizmach> thx
2008-03-27T22:26:30  <xorAxAx> dont forget that the parser is not the primary goal :-)
2008-03-27T22:26:41  <gizmach> xorAxAx: ?
2008-03-27T22:26:54  <xorAxAx> gizmach: there is no tree api or formatter support yet
2008-03-27T22:26:58  <xorAxAx> and that needs to be done first
2008-03-27T22:27:07  <TheSheep> *nod* *nod*
2008-03-27T22:27:23  <TheSheep> gizmach: this is actually the hard part
2008-03-27T22:28:39  <TheSheep> gizmach: you need classes for representing the tree, utility functions for generating and manipulating it, and then you need to make moin actually use all that
2008-03-27T22:28:58  <dennda> xorAxAx: whenever it fits into your schedule: would be great to have some feedback on the application (if you don't mind)
2008-03-27T22:29:37  <xorAxAx> dennda: yes, but not today :)
2008-03-27T22:29:42  <gizmach> xorAxAx: well, for the tree I had some ideas but as I see, maybe it's to difficult
2008-03-27T22:29:44  <xorAxAx> dennda: do you need a comment to fix your applciation?
2008-03-27T22:29:48  <xorAxAx> (technically)
2008-03-27T22:30:01  <dennda> xorAxAx: no problem
2008-03-27T22:30:02  <xorAxAx> gizmach: well, generating an AST is not anything hard ...
2008-03-27T22:30:05  <TheSheep> gizmach: I was thinking about splitting the formatter into two parts -- one that would have the current formatter API, so that all the moin code can use it, and it would generate the tree, and second part that would output html from that tree
2008-03-27T22:30:12  <xorAxAx> gizmach: deciding how the nodes should look like is :)
2008-03-27T22:30:31  <gizmach> xorAxAx: oh thats better, now I understand the problem and it's good
2008-03-27T22:30:37  <xorAxAx> :-)
2008-03-27T22:30:41  <dennda> xorAxAx: well I don't know whether it needs to be fixed or not. that's why I ask :)
2008-03-27T22:30:42  <TheSheep> gizmach: then you can slowly and gradually make all the moin code not use that "adapter" formatter, but generate the tree (or parts of it) diectly
2008-03-27T22:30:47  <gizmach> now I know wht to write in my application, at least one part
2008-03-27T22:30:50  <xorAxAx> TheSheep: johill and i discussed about that already and johill implemented some demo
2008-03-27T22:31:07  <TheSheep> xorAxAx: can I see it?
2008-03-27T22:31:09  <xorAxAx> TheSheep: back then i claimed that you can still run oldclass stuff if you e.g. use pickle
2008-03-27T22:31:19  <xorAxAx> TheSheep: the code? no idea where he put it
2008-03-27T22:31:32  <xorAxAx> johill: do you still have it somewhere?
2008-03-27T22:31:34  <xorAxAx> gizmach: :-)
2008-03-27T22:31:47  <xorAxAx> gizmach: so one problem here is this for example:
2008-03-27T22:32:00  <xorAxAx> gizmach: moinmoin supports including pages. included pages might appear in the toc
2008-03-27T22:32:15  <gizmach> TheSheep: I done some geeration tree for one of my subjects at college but in c# , it's a part of compiler so I had some ideas
2008-03-27T22:32:21  <xorAxAx> gizmach: if you have a tree where a node means "insert a TOC here" - how would you need which topics this toc should print?
2008-03-27T22:32:25  <gizmach> just need to see if there will work
2008-03-27T22:32:36  <xorAxAx> gizmach: i.e. you would want to include first, but then the include might have a TOC again etc.
2008-03-27T22:33:15  <TheSheep> one thing with dom that worries me -- you cannot output anything until your dom tree is finished. or rather, you could, but implementing it would be pretty tricky
2008-03-27T22:33:30  <xorAxAx> TheSheep: so? :)
2008-03-27T22:33:37  <xorAxAx> TheSheep: its cached :)
2008-03-27T22:33:54  <gizmach> xorAxAx: thats the point I get it
2008-03-27T22:34:11  <TheSheep> xorAxAx: this smells like JIT compling...
2008-03-27T22:34:22  <TheSheep> compiling
2008-03-27T22:34:45  <xorAxAx> TheSheep: why?
2008-03-27T22:35:03  <xorAxAx> well, yes, this smells like partial evaluation :-)
2008-03-27T22:35:07  <xorAxAx> which is something else :)
2008-03-27T22:35:11  <gizmach> when I wrote an applicationcould I send it to one of you just to rewiew it?
2008-03-27T22:35:53  <xorAxAx> gizmach: yes, you can also paste the url here if you dont fear any evil students copying it (we would notice that :-))
2008-03-27T22:36:00  <gizmach> :d
2008-03-27T22:36:18  <gizmach> well it's not the problem only I need to write it :D I have a concept now
2008-03-27T22:36:23  <gizmach> basic concept
2008-03-27T22:37:30  <gizmach> do I need to give you all explanation of what will I do or not, cause I really don't know everything, I only have Ideas where to start and time to study that and try to implement if i got the project
2008-03-27T22:38:14  <gizmach> then the first thing will be designing and implement tree class
2008-03-27T22:38:55  <xorAxAx> no, you dont need to give us a complete solution
2008-03-27T22:39:47  <gizmach> xorAxAx: ok,.. cause I really got only an idea and even don't know if tht will work properly
2008-03-27T22:40:26  <gizmach> if that doesn't work I'll try something else cause I plan to spend a lot on time on it (I need some hobby to stay away from my parents at summer)
2008-03-27T22:41:48  <dreimark> gizmach: just write down a rough plan how it could be solved. we can then discuss afterwards what is missing or is unclear
2008-03-27T22:42:16  <gizmach> dreimark: then I go and write it:D
2008-03-27T22:47:44  <dreimark> good night
2008-03-27T22:47:52  <gizmach> dreimark: night
2008-03-27T22:48:12  <gizmach> xorAxAx: and TheSheep thx for helping :)
2008-03-27T23:01:23  <gizmach> xorAxAx: sorry but I didn't understand your including pages example
2008-03-27T23:01:49  <gizmach> was that the example when will I need to rewrite an old parser totree generation
2008-03-27T23:02:07  <xorAxAx> gizmach: no, imagine that you have an ast that contains the parsed wiki page
2008-03-27T23:02:12  <gizmach> ok
2008-03-27T23:02:31  <xorAxAx> and now the main parts of the necessary task is to write stuff that operates on this AST
2008-03-27T23:03:00  <xorAxAx> and one question is how to evalute nodes that mean "insert TOC here" or e.g. "insert another wiki page here"
2008-03-27T23:03:34  <gizmach> ok I figure it now
2008-03-27T23:09:10  <gizmach> ok summary my tasks will be to: (1) - design and implement tree formatter api , that include (a) decide what the nodes should look like , (b) design operations on AST, (c) write a firts part of formatter for generating a tree, (d) write a seccond part for output tree in html , then (2) make all moin code to generate tree parts - rewrite a parser , (3) write tests for all stages , I just need to write my ideas for every stage
2008-03-27T23:09:52  <xorAxAx> gizmach: i dont understand C
2008-03-27T23:10:02  <xorAxAx> the idea is to let everything generate trees
2008-03-27T23:10:09  <xorAxAx> and make the formatter API legacy
2008-03-27T23:12:17  <gizmach> well yes but TheSheep said maybe it would be ok to split it in two parts, so yes everythong will generate a tree and output it to html, but maybe the first task is OK lets generate a tree
2008-03-27T23:12:38  <xorAxAx> welllll :-)
2008-03-27T23:12:45  <xorAxAx> not sure what he wanted to split
2008-03-27T23:13:07  <xorAxAx> TheSheep: what did you want to split?
2008-03-27T23:13:12  <gizmach> maybe in two diffrent classes
2008-03-27T23:13:16  <xorAxAx> gizmach: the new goal is to have an api thats passing around trees
2008-03-27T23:13:24  <xorAxAx> gizmach: but then you have legacy code that you want to use
2008-03-27T23:13:44  <xorAxAx> gizmach: so you need a trick to collect formatters calls and generate a tree
2008-03-27T23:13:57  <xorAxAx> the hard part here is that often, the calls to the formatter api are not wellformed
2008-03-27T23:14:13  <gizmach> gizmach: yes I saw the problem of legacy code
2008-03-27T23:14:20  <xorAxAx> then, you might want some tree 2 formatter code that lets you allow old formatters
2008-03-27T23:15:18  <gizmach> ok.........
2008-03-27T23:15:31  <gizmach> (thinking about what you said)
2008-03-27T23:22:25  <gizmach> probbably it disconected me
2008-03-27T23:22:58  <xorAxAx> maybe you can pose a question to clear your mind :-)
2008-03-27T23:23:51  <xorAxAx> do you know how the current "formatter API" looks like? :)
2008-03-27T23:23:58  <xorAxAx> that might clear up some things
2008-03-27T23:24:46  <gizmach> xorAxAx: no unfortunatley I don't :(
2008-03-27T23:26:38  <xorAxAx> then looking at it might help - i guess looking at moin's code is most useful for that
2008-03-27T23:28:15  <gizmach> :) well that's the creole if I'm not wrong
2008-03-27T23:31:41  <xorAxAx> no, creole is a parser
2008-03-27T23:33:50  <gizmach> pardon me
2008-03-27T23:33:52  <gizmach> ups
2008-03-27T23:34:29  <gizmach> I found it
2008-03-27T23:34:32  <gizmach> my fault :D
2008-03-27T23:34:45  <xorAxAx> but waldi already applied for this task :)
2008-03-27T23:35:01  <zenhase> re
2008-03-27T23:35:09  <xorAxAx> at 13:24 pdf
2008-03-27T23:35:11  <xorAxAx> pdt
2008-03-27T23:35:39  <gizmach> xorAxAx: well, I should think about that earlier :)
2008-03-27T23:36:28  <xorAxAx> hmm
2008-03-27T23:36:42  <gizmach> xorAxAx: what now :D
2008-03-27T23:38:01  <xorAxAx> so it might make sense to choose a different task - maybe on the parser side?
2008-03-27T23:38:23  <gizmach> xorAxAx: I don't get it
2008-03-27T23:38:39  <xorAxAx> hmm? :)
2008-03-27T23:39:08  <gizmach> that part of sentence "on the parser side" i didn't get it
2008-03-27T23:39:31  <xorAxAx> parser side vs. backend side (formatter etc.
2008-03-27T23:39:49  <zenhase> hmm, parser/formatter discussions
2008-03-27T23:39:53  * zenhase gets interested
2008-03-27T23:40:31  <gizmach> xorAxAx: but do you still think about the dom or something else
2008-03-27T23:40:47  <xorAxAx> gizmach: yes, but i have to see how waldi's proposal compares to this task
2008-03-27T23:41:06  <gizmach> oh, OK then
2008-03-27T23:41:08  <xorAxAx> hmm, i think i am too exhausted to think about this today, i will review the applications tomorrow and we can continue to talk about it :-)
2008-03-27T23:41:50  <gizmach> xorAxAx: I'm exausted to , I'm just going to write something about myself in the appliction and leave the conversation for tomorrow
2008-03-27T23:42:18  <gizmach> would u like my e-mail addres to let me know if I'm not on my PC
2008-03-27T23:42:45  <xorAxAx> gizmach: hmm, wont you be in irc tomorrow?
2008-03-27T23:46:15  <gizmach> xorAxAx: yes but at the evening
2008-03-27T23:46:51  <gizmach> xorAxAx: I'm going to see some of my relatives after classes tomorrow
2008-03-27T23:47:50  <xorAxAx> evening is fine
2008-03-27T23:48:22  <gizmach> xorAxAx: great :)