MoinMoin: MoinMoinChat/Logs/moin-dev/2007-07-13

2007-07-13T00:42:30  <neagulm> good night
2007-07-13T00:48:10  <dreimark> gn
2007-07-13T01:10:07  <grzywacz> :\
2007-07-13T01:10:34  <grzywacz> Maybe I should make an option in preferences, to switch between normal (text) notifications and data forms..?
2007-07-13T01:10:49  <grzywacz> That would make it a little bit cleaner to code. ;)
2007-07-13T01:10:59  <xorAxAx> why?
2007-07-13T01:11:08  <xorAxAx> ... is that related to cleanliness
2007-07-13T01:11:32  <grzywacz> Because the notification function would remain the way it is - it sends a message with subject and some elaborate body.
2007-07-13T01:11:45  <xorAxAx> why?
2007-07-13T01:11:53  <xorAxAx> would it change otherwise
2007-07-13T01:11:57  <grzywacz> The data forms approach requires the approach described on JabberSupport
2007-07-13T01:12:07  <grzywacz> With many args used to build a form, that is.
2007-07-13T01:12:15  <xorAxAx> and a switch in the prefs would help how?
2007-07-13T01:12:29  <grzywacz> One function would be called or the other.
2007-07-13T01:12:44  <xorAxAx> well
2007-07-13T01:12:45  <grzywacz> If there's no switch, I'd have to send both data for forms, and elaborate messages.
2007-07-13T01:12:54  <grzywacz> And then let the bot decide which to send.
2007-07-13T01:13:07  <xorAxAx> thats rather a design problem ... that shouldnt be solved by changing the UI
2007-07-13T01:13:10  <xorAxAx> good night
2007-07-13T01:13:13  <grzywacz> night
2007-07-13T01:13:29  <grzywacz> Of course it's a design problem.
2007-07-13T03:32:17  <zenhase> is it my connection or are there sometimes problems with POSTing edits and/or attachments to moinmo.in?
2007-07-13T03:33:45  <zenhase> hmm, it seems to work when i hit the upload button again while it's waiting for the first submit to finish (which it never does) o_O
2007-07-13T11:55:57  <zenhase> moin
2007-07-13T12:01:30  <ffesti> moin zenhase
2007-07-13T12:02:20  <zenhase> hello my mentor
2007-07-13T12:03:26  <xorAxAx> :)
2007-07-13T12:04:24  <zenhase> *cough* sorry xorAxAx :)
2007-07-13T12:05:20  <xorAxAx> hmm? :)
2007-07-13T12:05:23  <johill> xorAxAx: looks like you trained them well :P
2007-07-13T12:05:35  <xorAxAx> i was just grinning because you picked up the pattern :)
2007-07-13T12:06:10  <zenhase> xorAxAx: yes
2007-07-13T12:06:33  <zenhase> xorAxAx: i felt like blatantly copying the pattern between you and grzywacz ;)
2007-07-13T12:07:46  <grzywacz> :D
2007-07-13T12:07:51  <grzywacz> morning everyone
2007-07-13T12:07:53  <grzywacz> hello xorAxAx :)
2007-07-13T12:07:55  <xorAxAx> moin grzywacz
2007-07-13T12:07:56  <zenhase> hi there
2007-07-13T12:08:56  <zenhase> hmm, refreshing laugh of the morning: http://en.wikipedia.org/wiki/Wolfgang_Sch%C3%A4uble
2007-07-13T12:09:44  <zenhase> he is now officially the head of the 'Stasi'
2007-07-13T12:09:56  <zenhase> if wikipedia says so, it must be true!
2007-07-13T12:10:17  <grzywacz> xorAxAx, it would be cool if you got yourself svn psi so that I could show you the data forms and ask for comments, changes, etc
2007-07-13T12:10:28  <xorAxAx> ok :)
2007-07-13T12:10:30  <xorAxAx> svn url?
2007-07-13T12:11:03  <grzywacz> svn checkout http://svn.psi-im.org/psi/trunk psi
2007-07-13T12:17:20  <xorAxAx> grzywacz: build running :)
2007-07-13T12:19:16  <grzywacz> xorAxAx, cool, there's not much to see yet, but I can crash you just fine. :D
2007-07-13T12:19:25  <xorAxAx> :-p
2007-07-13T12:19:57  <grzywacz> I'm awfully good at breaking software. ;S
2007-07-13T12:27:19  <zenhase> i know some guy who is paid money to do exactly this
2007-07-13T12:29:08  <grzywacz> Yes, that happens.
2007-07-13T12:29:08  <grzywacz> Maybe I should look for a testing job in the future. ;)
2007-07-13T13:01:55  <xorAxAx> grzywacz: ping
2007-07-13T13:02:16  <xorAxAx> grzywacz: ah, the commands should not be case-sensitive ...
2007-07-13T13:02:32  <xorAxAx> and we have left the testing phase ... so you should consider usability now :)
2007-07-13T13:38:34  <grzywacz> arrr
2007-07-13T13:38:41  <grzywacz> Actually, that makes sense. :D
2007-07-13T13:53:09  <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2447:d442d4755b31 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: Use urlgrabber's callback for checksumming
2007-07-13T13:53:17  <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2448:8d422e535020 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: fix comment typo
2007-07-13T13:53:25  <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2449:65d560605d8e 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: removing old code
2007-07-13T13:53:28  <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2450:b0c513eedd55 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: Some comment/output changes
2007-07-13T13:53:34  <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2451:66ac934876e1 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: Discard my reporthook implementation and use urlgrabber's TextMeter, some other small fixes
2007-07-13T14:47:13  <ThomasWaldmann> moin
2007-07-13T14:47:45  <grzywacz> moin
2007-07-13T15:42:23  <ThomasWaldmann>  def handle(event):
2007-07-13T15:42:24  <ThomasWaldmann> +    """An event handler"""
2007-07-13T15:42:33  <ThomasWaldmann> %)
2007-07-13T16:02:08  <ThomasWaldmann> vpv:
2007-07-13T16:02:12  <ThomasWaldmann> +            print "Cheksum doesn't match, re-downloading"
2007-07-13T16:02:46  <ThomasWaldmann> (speling error)
2007-07-13T16:05:05  <ThomasWaldmann> +            os.chdir("../" + self.tmpdirname)
2007-07-13T16:05:20  <ThomasWaldmann> (use os.path.join)
2007-07-13T16:05:42  <ThomasWaldmann> +        if type == "sha":
2007-07-13T16:07:53  <vpv> so should I use "is" in the last case?
2007-07-13T16:08:16  <ThomasWaldmann> type is a builtin function, dont use it as variable name
2007-07-13T16:08:44  <ThomasWaldmann> maybe hash_type is a good name
2007-07-13T16:08:49  <xorAxAx> vpv: WTF? chdir is evvil
2007-07-13T16:08:49  <vpv> oh, that's right, I didn't even notice
2007-07-13T16:09:09  <xorAxAx> vpv: so have you started getting your stuff generally usable?
2007-07-13T16:09:28  <ThomasWaldmann> +        file = open(path)
2007-07-13T16:09:50  <ThomasWaldmann> don't use "file" as variable name for the same reasons
2007-07-13T16:10:24  <ThomasWaldmann> but do something like local_file = file(path)
2007-07-13T16:11:06  <ThomasWaldmann> file() is the preferred way, not open()
2007-07-13T16:15:03  <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2452:d34c3dbdbd8c 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: Small changes and some code reorganization that produces an ugly diff, you've been warned ;)
2007-07-13T16:15:04  <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2453:3e4b0f736abe 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: Basics of filelist.xml parsing, one typo fix
2007-07-13T16:15:09  <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2454:d4a64f2ba2b6 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: Some fixes requested by ThomasWaldmann
2007-07-13T16:15:22  <vpv> my newest changes should be there
2007-07-13T16:15:55  <vpv> xorAxAx: not yet, since the RPM repository handling is still not finished, I'll have to do some "overtime" this weekend to finish it
2007-07-13T16:16:47  <vpv> xorAxAx: do you have a better idea than using chdir in my case, I need to be able to change directories
2007-07-13T16:48:21  <xorAxAx> vpv: why do you need to change directories?
2007-07-13T16:50:46  <vpv> main reason are the file downloading tools that only seem to download the file to cwd, of course I could also move the files...
2007-07-13T16:52:05  <xorAxAx> "file downloading tools"? you dont really need external tools for that, do you?
2007-07-13T16:52:48  <vpv> no, but urllib2 and whatnot
2007-07-13T16:53:05  <xorAxAx> yeah, should be fine, nor?
2007-07-13T16:54:16  <vpv> should be, but why should chdir be avoided? right now with using it I'm avoiding creating any files directly under MoinMoin/script/
2007-07-13T16:54:59  <xorAxAx> is it in the manimpor tscript? do you ensure that you will change it back afterwards? then its fine maybe
2007-07-13T16:55:06  <xorAxAx> more annoying is the tool dependency
2007-07-13T16:58:41  <vpv> yeah, manimport. the way I have it now is that it creates a tempdir for itself every time it's run, then works in that dir and removes it when finished. it also creates a cachedir that should stay and be used when it is started next. it tries to avoid downloading big files and rather caching them if possible
2007-07-13T17:00:48  <vpv> Maybe I was a bit vague about the "tool" thing, it uses urlgrabber now which directly handles mirrors, redownloads, progress meters etc. It's pure python but not in stdlib. My idea is to use that only for the Fedora specific stuff, though.
2007-07-13T17:01:29  <xorAxAx> ah, well, if its python
2007-07-13T17:01:42  <xorAxAx> and i think it would be fine to have it generally as well
2007-07-13T17:02:03  <vpv> here's the url http://linux.duke.edu/projects/urlgrabber/
2007-07-13T17:02:12  <xorAxAx> but i rather think that you should be able to advise it to use a specific target dir
2007-07-13T17:03:01  <vpv> I agree, but even urllib2 doesn't support that, afaik
2007-07-13T18:13:37  <ThomasWaldmann> johill: hey, quite some refactoring in 1.7 recently :)
2007-07-13T18:15:03  <ThomasWaldmann> johill: got your mail, will look at the link stuff today evening
2007-07-13T19:22:52  <ThomasWaldmann> zenhase:  1. the ''Request'' is passed to every registered ''!RequestHandler''
2007-07-13T19:22:55  <ThomasWaldmann> why?
2007-07-13T19:23:49  <dreimark> bbl
2007-07-13T19:25:19  <grzywacz> Bang, another tree goes down...
2007-07-13T19:27:14  <ThomasWaldmann> you axed a dom tree?
2007-07-13T19:27:26  <grzywacz> Hehe, no. I printed yet another xep. ;)
2007-07-13T19:29:16  * ThomasWaldmann read some stuff about docbook xml printed on a dead tree (in the plane)
2007-07-13T19:30:01  <grzywacz> Oh, so you flew to Vilnus. 8)
2007-07-13T19:30:32  <ThomasWaldmann> yeah, 1400km is too much
2007-07-13T19:30:47  <grzywacz> Hm, true. Esp. through Poland. ;)
2007-07-13T19:31:31  <ThomasWaldmann> that turboprop machine was adventurous enough
2007-07-13T19:32:09  <grzywacz> Why so?
2007-07-13T19:33:06  <ThomasWaldmann> unusual :)
2007-07-13T19:33:36  <ThomasWaldmann> but well, maybe the IT breakdown they had in FFM was worse
2007-07-13T19:33:49  <grzywacz> IT? ;)
2007-07-13T19:34:02  <ThomasWaldmann> computer system
2007-07-13T19:34:09  <grzywacz> FFM? ;)
2007-07-13T19:34:18  <grzywacz> I've flown a "turboprop" plane once, it wasn't really weird. ;)
2007-07-13T19:34:20  <ThomasWaldmann> frankfurt (main)
2007-07-13T19:46:32  <ThomasWaldmann> it was a bit loud (had a seat in the rear, behind the engines)
2007-07-13T19:47:32  <grzywacz> Normal turbo jets are loud as well. ;S
2007-07-13T19:52:39  <ThomasWaldmann> zenhase:  1. the ''Page'' and desired format is passed to every registered ''!FormatHandler''
2007-07-13T19:52:42  <ThomasWaldmann> why?
2007-07-13T19:56:06  <xorAxAx> grzywacz: so, whats your plan for pyxmpp?
2007-07-13T19:57:01  <grzywacz> xorAxAx, I'm not sure, no contact yet, but I'll probably end up using libxml directly to get additional information from nodes passed in stanza objects.
2007-07-13T19:57:37  <xorAxAx> libxml?
2007-07-13T19:57:45  <grzywacz> pyxmpp is based on libxml2.
2007-07-13T19:57:49  <xorAxAx> grzywacz: you said that you want to implement data forms like described in the wiki?
2007-07-13T19:57:55  <xorAxAx> where can i find that info?
2007-07-13T19:57:59  <xorAxAx> ah
2007-07-13T19:58:13  <grzywacz> xorAxAx, I meant the proposed functions and lists of arguments.
2007-07-13T19:58:31  <grzywacz> I have no visual design of forms, if you ask about it.
2007-07-13T19:58:40  <grzywacz> It's not possible to do any layouts anyway. ;)
2007-07-13T20:00:01  <grzywacz> Right now I'm reading about XHTML-IM, so that (maybe) pages could be easily presented to the xmpp client in a more readable format.
2007-07-13T20:00:13  <grzywacz> Wiki markup / full html don't meet the criteria of readability.
2007-07-13T20:00:15  <grzywacz> ;)
2007-07-13T20:00:39  <xorAxAx> hmm?
2007-07-13T20:00:56  <xorAxAx> if you edit a wikipage, you want to do it primarily using wiki markup
2007-07-13T20:01:00  <grzywacz> Yes.
2007-07-13T20:01:13  <grzywacz> But when you read it, you may want to get something readable.
2007-07-13T20:01:16  <xorAxAx> which proposed functions do you mean?
2007-07-13T20:01:19  <xorAxAx> yes
2007-07-13T20:01:38  <xorAxAx> i am only seeing commentpage and revertpage that may qualify here
2007-07-13T20:02:16  <grzywacz> Yes.
2007-07-13T20:02:35  <grzywacz> Maybe GetPage(HTML) as well, depending on what user actually wants to do.
2007-07-13T20:02:58  <xorAxAx> yes
2007-07-13T20:03:08  <xorAxAx> but you should think about stuff like page editing as well
2007-07-13T20:03:12  <xorAxAx> its not listed there
2007-07-13T20:04:14  <grzywacz> It is not, correct.
2007-07-13T20:04:27  <grzywacz> It's already implemented, actually, in a rather basic way.
2007-07-13T20:04:47  <xorAxAx> editing is implemented?
2007-07-13T20:04:57  <grzywacz> No, getting pages.
2007-07-13T20:05:09  <grzywacz> Erm, yeah.
2007-07-13T20:05:13  <xorAxAx> thats something completly different[tm] :)
2007-07-13T20:05:20  <xorAxAx> please do some cleanup on the page :)
2007-07-13T20:05:49  <grzywacz> Hm, ok.
2007-07-13T20:07:26  <grzywacz> xorAxAx, I'll decide on all the outstanding issues by tomorrow, ok? This will be the calling convention for data-forms enabled functions and UI issue of configuring data forms? I'm still a little bit confused about these. ;S
2007-07-13T20:07:36  <xorAxAx> ok
2007-07-13T20:08:04  <grzywacz> Ah, and once I know more about xhtml-im, I'll add my ideas about it to the JabberSupport page as well.
2007-07-13T20:08:22  <grzywacz> Splitting it would make it more readable, but it's too early to do it.
2007-07-13T20:08:59  <xorAxAx> hmm, its pretty long ... so maybe its not too early :)
2007-07-13T20:22:59  <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2455:9564be5b2ba3 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: parsing: now creates a dict for packages as key
2007-07-13T20:23:00  <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2456:fec2a870bcf6 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: parsing works now, fixed a bug, prints a nice list of packages in a repository and which man pages they have
2007-07-13T21:06:35  <CIA-27> moin: Marian Neagul marian@info.uvt.ro work * 2356:25e327b42998 1.7-classify-mneagul/ (6 files in 5 dirs): SpamBayes interface fixes
2007-07-13T21:10:17  <xorAxAx> why is work read?
2007-07-13T21:10:23  <xorAxAx> a branch name?
2007-07-13T21:47:38  <xorAxAx> neagulm:
2007-07-13T21:47:47  <xorAxAx> grzywacz: have you fixed the PEP8-unconformant local name?
2007-07-13T21:47:55  <neagulm> xorAxAx, ?
2007-07-13T21:48:05  <neagulm> xorAxAx, I don't understand
2007-07-13T21:48:07  <xorAxAx> neagulm: do you know what "work" means?
2007-07-13T21:48:11  <xorAxAx> in the line CIA printed
2007-07-13T21:49:22  <neagulm> xorAxAx, Yes, I know.
2007-07-13T21:49:46  <neagulm> xorAxAx, I didn't do to much in the last days, I had school problems
2007-07-13T21:49:47  <xorAxAx> so tell me, the curious guy
2007-07-13T21:49:57  <xorAxAx> oh, sorry you misunderstood me
2007-07-13T21:50:07  <xorAxAx> have a look at the last line by the CIA bot
2007-07-13T21:50:22  <xorAxAx> it says "work" near your name
2007-07-13T21:50:32  <xorAxAx> and i am wondering what it means
2007-07-13T21:52:48  <neagulm> ah, Yes. :)) It's a branch name, I forgot to switch branches. OTOH i'm feeling guilty for not working enough :)
2007-07-13T21:53:58  <xorAxAx> hehe :)
2007-07-13T21:54:12  <xorAxAx> will you add general classification capabilities?
2007-07-13T21:54:15  <xorAxAx> i.e. general categories
2007-07-13T21:55:05  <neagulm> xorAxAx, I want general categories but SB and Reverend can't do that much... I will have to integrate Orange or something else....
2007-07-13T21:55:16  <neagulm> xorAxAx, Currently just spam and language classification
2007-07-13T21:55:20  <grzywacz> xorAxAx, yes
2007-07-13T21:55:24  <grzywacz> xorAxAx, fixed, no pushed
2007-07-13T21:55:26  <xorAxAx> ok
2007-07-13T21:55:38  <xorAxAx> grzywacz: would be nice if you could ack that :)
2007-07-13T21:55:47  <xorAxAx> neagulm: it would be pretty useful
2007-07-13T21:55:57  <xorAxAx> neagulm: IMHO more useful than spam classification
2007-07-13T21:56:35  <neagulm> xorAxAx, have you read my initial proposal?
2007-07-13T21:56:43  <xorAxAx> yes
2007-07-13T21:56:50  <xorAxAx> but i cant remember it, it was garbage collected
2007-07-13T21:56:53  <xorAxAx> why?
2007-07-13T21:57:54  <neagulm> I proposed general text classification...
2007-07-13T21:58:07  <neagulm> xorAxAx, brb ~5 minutes..
2007-07-13T21:58:49  <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2457:38a8dd44ac59 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: Importing from a repository works to a certain point, crashes with problems in ismanfile()
2007-07-13T21:59:08  <xorAxAx> yes
2007-07-13T21:59:52  <grzywacz> xorAxAx, ack what? O_o
2007-07-13T22:00:10  <xorAxAx> grzywacz: things that i spot and comment about
2007-07-13T22:00:18  <xorAxAx> so i know that they arent lost
2007-07-13T22:00:56  <grzywacz> xorAxAx, oh, sure. Sorry.
2007-07-13T22:01:13  <vpv> xorAxAx: the last commit is a substantial improvement, now I have to fix some bugs and it should be ready for a somewhat public test
2007-07-13T22:01:49  <vpv> otoh I've worked for 11.5 hours today :P
2007-07-13T22:07:02  <zenhase> ThomasWaldmann: you mean why not instantiate the handler directly?
2007-07-13T22:09:50  <zenhase> the idea is to not impose a particular mechanism how an appropiate handler is determined in the architecture
2007-07-13T22:10:51  <zenhase> the overhead for the handlers checking if they can meet the request is comparably low
2007-07-13T22:13:35  <zenhase> especially if opposed to the loss in flexibility you get from a key value style handler lookup (like expecting code for a specific mimetype to be in a properly named module with properly named class/functions)
2007-07-13T22:14:54  <zenhase> i give an example:
2007-07-13T22:16:04  <zenhase> just pretend there is something like a parser/formatter for images to represent them in a nice way
2007-07-13T22:16:20  <zenhase> in html that is ... like exif information
2007-07-13T22:16:36  <zenhase> all markupped nicely
2007-07-13T22:17:08  <zenhase> someone has written this one with a generic image library
2007-07-13T22:17:30  <zenhase> and so his handler can actually match almost all image/* mimetypes
2007-07-13T22:18:33  <zenhase> but lets say because it is so generic someone implements another handler for image/png only, but with even more metadata support
2007-07-13T22:19:09  <zenhase> for png this handler claims to be more appropiate
2007-07-13T22:20:51  <zenhase> in trac for example the mimeview-package has an api for those formatters where every handler responds with a number between 0 and 9 to indicate how good he is at converting this particular request
2007-07-13T22:21:30  <zenhase> 0 meaning "can't do", 9 meaning "i am THE perfect handler"
2007-07-13T22:23:27  <zenhase> or the other thing with requesthandlers: currently it's basically the actions
2007-07-13T22:23:35  <ThomasWaldmann> zenhase: i still dont see how it is useful
2007-07-13T22:23:44  <ThomasWaldmann> we know the action, e.g. show
2007-07-13T22:23:52  <xorAxAx> zenhase: i am not sure if that should be called request handler :)
2007-07-13T22:23:59  <ThomasWaldmann> and we just need 1 handler doing this
2007-07-13T22:24:14  <zenhase> xorAxAx: well, it's basically what request.run dispatches to
2007-07-13T22:24:18  <ThomasWaldmann> so why should that handler not be named "show" to be directly findable
2007-07-13T22:24:49  <zenhase> ThomasWaldmann: i was getting to the requesthandlers now ... it's another reason there, not the same as for formatters
2007-07-13T22:25:25  <zenhase> my problem with the actions is, that most of them act on pages, then others again do not
2007-07-13T22:25:44  <zenhase> then again others are basically not implemented as real actions right now
2007-07-13T22:25:54  <zenhase> (editors, userform ... all code somewhere else)
2007-07-13T22:26:05  <zenhase> xmlrpc does not really fit the action scheme too imo
2007-07-13T22:26:25  <zenhase> as it's more an interface to the wiki as a whole
2007-07-13T22:26:40  <ThomasWaldmann> give a single example (except from the mimetype hierarchy) where this is useful
2007-07-13T22:27:52  <ThomasWaldmann> btw, we already handle the mimetype hierarchy by some simple fallback in the module name
2007-07-13T22:30:51  <zenhase> in the case of the actions: does anybody else feel a bit uneasy about "there are pages and actions on them, but some of those actions are actually something totally unrelated to the page"
2007-07-13T22:31:16  <ThomasWaldmann> please one thing after the other
2007-07-13T22:31:24  <zenhase> ok
2007-07-13T22:31:37  <ThomasWaldmann> we are still at the question why this querying stuff is usefull
2007-07-13T22:32:50  <ThomasWaldmann> aka 3.
2007-07-13T22:35:31  <zenhase> well, it's putting the responsibility for the selection of an appropiate handler for the request to the requests themselves. you free yourself from a very specific mechanism
2007-07-13T22:35:42  <zenhase> s/requests/request handlers
2007-07-13T22:36:03  <ThomasWaldmann> and?
2007-07-13T22:36:04  <zenhase> that is what i try to illustrate with the example above
2007-07-13T22:37:03  <zenhase> well, you request a page, so it's only request-handlers acting on page that would be appropiate in the first place
2007-07-13T22:37:54  <zenhase> so in a scheme where you decide from the start which handler to use, you have to pick a parameter to decide which one actually to use
2007-07-13T22:38:11  <zenhase> this is what the 'action'-GET-param is right now
2007-07-13T22:39:04  <ThomasWaldmann> we always have this action parameter. if it is not in the url, we just default to show.
2007-07-13T22:39:04  <zenhase> but then you introduce something new, that does not fit the action-on-a-page scheme
2007-07-13T22:39:13  <zenhase> so you end up either hacking it into this scheme
2007-07-13T22:39:44  <zenhase> or you could have something like the scheme i proposed, where you can decide according to other parameters
2007-07-13T22:40:07  <zenhase> but you do not want to add all those parameters to the dispatching mechanism for every idea you get
2007-07-13T22:40:27  <ThomasWaldmann> ok, give 1 example
2007-07-13T22:40:39  <zenhase> ok
2007-07-13T22:41:04  <zenhase> let's say someone gets the idea, to write a JSON-interface to the whole wiki
2007-07-13T22:41:40  <zenhase> a scrap that
2007-07-13T22:41:53  <zenhase> i just stick to the xml-rpc example, because it's already there
2007-07-13T22:42:47  <zenhase> xmlrpc as i have seen does not act on pages, but provides calls to do all kinds of things with the wiki via xmlrpc
2007-07-13T22:43:29  <zenhase> it's not a real single action
2007-07-13T22:43:40  <ThomasWaldmann> and?
2007-07-13T22:44:11  <zenhase> but a whole multitude of functions that all act differently on the wiki
2007-07-13T22:44:15  <ThomasWaldmann> you just need a url for it, the rest is handled internally in it
2007-07-13T22:44:56  <zenhase> you can use any url right now, right?
2007-07-13T22:45:14  <zenhase> anything pointing somewhere into the wiki
2007-07-13T22:45:16  <ThomasWaldmann> it just needs action=xmlrpc(2) at the end
2007-07-13T22:47:59  <zenhase> i feel my arguments being not strong enough
2007-07-13T22:48:37  <zenhase> till now it was something like a ... erm .. 'bauchgefuehl'?
2007-07-13T22:49:36  <ThomasWaldmann> well, I guess you have to show how it is useful with at least one concrete thing - if that is not possible, it maybe isn't useful, but just overengineered
2007-07-13T22:49:46  <zenhase> can i think about it again? i think i need to think of a better and more immanent usecase
2007-07-13T22:50:39  <ThomasWaldmann> i can currently think of only one, trivial application:
2007-07-13T22:50:40  <zenhase> i don't want to point to trac again in this matter, because i want to prove that it's worth for moin :>
2007-07-13T22:51:28  <ThomasWaldmann> realising the default action to be "show" and having some handler for "unknown action" if nobody claims the requested action
2007-07-13T22:51:55  <ThomasWaldmann> but this is a bit too trivial to do without that stuff
2007-07-13T22:53:36  <zenhase> that's right .. as said i will think about it again. i am still convinced of my idea, i just need something more than intuition to convince you too ... or realize my mistake
2007-07-13T22:53:55  <ThomasWaldmann> ok
2007-07-13T22:54:12  <ThomasWaldmann> same question is there for FormatHandler
2007-07-13T22:55:25  <ThomasWaldmann> (answer should include something aside from the trivial handling of some x/* mimetype by a generic x handler)
2007-07-13T22:55:59  <ThomasWaldmann> (or handling */* mimetype by a generic "any item" handler)
2007-07-13T22:57:54  <ThomasWaldmann> and btw, i guess you dont really want to touch parser/formatter stuff in the 2nd half of your time, but just create some workaround catching its output
2007-07-13T22:58:32  <xorAxAx> well, IMHO thats orthogonal and compatibility is crucial (which doesnt mean that there could be a new system)
2007-07-13T22:58:57  <xorAxAx> i.e. capturing should be in place as long as the new system isnt in place
2007-07-13T23:00:32  <ThomasWaldmann> the point is that the time will not be enough to create a new system
2007-07-13T23:00:46  <xorAxAx> well
2007-07-13T23:00:50  <xorAxAx> depends on the work!
2007-07-13T23:00:52  <xorAxAx> :-)
2007-07-13T23:01:02  <ThomasWaldmann> looks at first half :)
2007-07-13T23:01:06  <ThomasWaldmann> -s
2007-07-13T23:01:40  <xorAxAx> dark clouds approach
2007-07-13T23:03:50  <ThomasWaldmann> and handling everything except that "new parser/formatter system" will be enough work to keep zenhase busy
2007-07-13T23:05:32  <xorAxAx> compare the storage refactoring - you need 3 british industrial grade coders or one BA student from student :)
2007-07-13T23:05:41  <xorAxAx> s/m student/m germany/
2007-07-13T23:05:49  <ThomasWaldmann> hehe
2007-07-13T23:11:48  <zenhase> hmm
2007-07-13T23:34:02  <dreimark> moin
2007-07-13T23:35:53  <dreimark> hi neagulm
2007-07-13T23:36:32  <neagulm> hello dreimark
2007-07-13T23:39:52  <johill> ThomasWaldmann: :)
2007-07-13T23:40:37  <dreimark> neagulm: did you started some part of the documentation ?
2007-07-13T23:41:44  <neagulm> dreimark, nope, currently i'm doing changes to the way we detect a spamy page, it's not uniform
2007-07-13T23:42:05  <ThomasWaldmann> http://icanhascheezburger.com/
2007-07-13T23:42:08  <neagulm> dreimark, I will write the docs after fixing some issues
2007-07-13T23:42:46  <dreimark> neagulm: ok
2007-07-13T23:43:50  <grzywacz> xorAxAx, ok, feeding Moin-generated html directly to xmpp clients won't work.
2007-07-13T23:43:50  <neagulm> dreimark, I will write some documentation about using the SpamBayes XML-RPC interface tomorrow... It will require some testing...
2007-07-13T23:44:17  <grzywacz> xorAxAx, the set of allowed tags and attributes is heavily reduced and xhtml compliance is required.
2007-07-13T23:45:07  <neagulm> dreimark, SB's core_server.py keeps all training "messages" for later review and this could be a small problem for the enduser. After each classification SB make's a new copy of the classified "message", in our case page
2007-07-13T23:45:12  <xorAxAx> grzywacz: well :)
2007-07-13T23:45:38  <zenhase> ThomasWaldmann: oh noes!
2007-07-13T23:45:47  <xorAxAx> grzywacz: currently, parsers do some sanity compensation and the html formatter does less
2007-07-13T23:45:49  <zenhase> ThomasWaldmann: i still have a lol-trauma from gpn6
2007-07-13T23:45:52  <neagulm> dreimark, with >100 classifications of the same page it would mean that SB has in his archive >100 copies of the same page.
2007-07-13T23:46:24  <xorAxAx> grzywacz: one could extend the html formatter to output sane (x)html only in all cases, some bits for there are already in there
2007-07-13T23:46:37  <xorAxAx> grzywacz: and reducing the tag set is easy as well given that you can simply subclass it
2007-07-13T23:47:11  <grzywacz> xorAxAx, but it doesn't work so atm, so I'd rather not do it. ;)
2007-07-13T23:47:42  <xorAxAx> hmm?
2007-07-13T23:48:04  <xorAxAx> of course thats off the spot compared to your focus
2007-07-13T23:48:14  <xorAxAx> but i would consider it doable, but too large
2007-07-13T23:48:18  <dreimark> neagulm: sounds bad, and is not necessary page?rev=xx if it has  a connection to the wiki or isn't it?
2007-07-13T23:48:19  <xorAxAx> grzywacz: so whats your plan? :)
2007-07-13T23:48:23  <grzywacz> Sure it is doable. :D
2007-07-13T23:48:42  <grzywacz> xorAxAx, no xhtml-im for now.
2007-07-13T23:48:49  <grzywacz> xorAxAx, but I'd like to discuss the data forms issue.
2007-07-13T23:49:00  <grzywacz> Like, how to use it most efficiently.
2007-07-13T23:49:08  <grzywacz> Some input would be welcomed. :)
2007-07-13T23:49:15  <xorAxAx> with me or them? :)
2007-07-13T23:49:21  <grzywacz> Like... I have some ideas, but maybe you'd like them used for something?
2007-07-13T23:49:25  <grzywacz> 8)
2007-07-13T23:49:36  <johill> ohh I have many ideas ;)
2007-07-13T23:49:48  <xorAxAx> me too, lets sell them on ebay
2007-07-13T23:49:55  <johill> like integrating it with the openid server so it asks you on jabber
2007-07-13T23:50:39  <johill> but you already had a bunch too, with revert and whatever
2007-07-13T23:50:51  <johill> besides, my jabber client probbaly doesn't support it anyway ;)
2007-07-13T23:51:12  <neagulm> dreimark, i think I don't understand completely what you mean
2007-07-13T23:51:12  <grzywacz> The latest psi and tkabber do.
2007-07-13T23:51:16  <xorAxAx> johill: you have to implement them :)
2007-07-13T23:51:42  <johill> yeah like I don't have anything else to do ;)
2007-07-13T23:52:04  <johill> anyway, gotta get some sleep, night all
2007-07-13T23:52:16  <xorAxAx> johill: its a pity, i can only choose 6 out of 9 courses in the next semester
2007-07-13T23:52:22  <grzywacz> night johill
2007-07-13T23:52:24  <xorAxAx> thats like school
2007-07-13T23:52:37  <grzywacz> xorAxAx, why there's a limit?
2007-07-13T23:52:44  <xorAxAx> there are only 9 courses
2007-07-13T23:52:50  <xorAxAx> and i need to do 6 to use my time
2007-07-13T23:52:58  <xorAxAx> and stay in the mandated time frame
2007-07-13T23:53:45  <xorAxAx> gn johill
2007-07-13T23:54:11  <dreimark> neagulm: if it has to save the page it needs to know the first revisions which was used for the categorization. if it knows that revision number it does not need to save it because it could request it from the wiki too.
2007-07-13T23:55:11  <dreimark> may be this is slower as the current concept
2007-07-13T23:56:03  <neagulm> dreimark, yes... but I don't think the SB team will be willing to implement this type of behavior. But of course I could update my patch and propose it to Skip
2007-07-13T23:57:52  <dreimark> neagulm: does he has responded to your mail already?
2007-07-13T23:58:30  <neagulm> dreimark, not yet
2007-07-13T23:59:15  <dreimark> neagulm: do ask him about that after he has responded