MoinMoin: MoinMoinChat/Logs/moin-dev/2007-06-28

2007-06-28T00:33:41  <ThomasWaldmann> zenhase: avoid using max as identifier
2007-06-28T00:34:59  <zenhase> oh damnit
2007-06-28T00:35:09  <zenhase> it should be max_
2007-06-28T00:35:46  <zenhase> well, i try to think of namespace clash with python builtins, but i don't have all of them in my head at all times :>
2007-06-28T00:35:49  <ThomasWaldmann> or max_len
2007-06-28T00:36:10  <ThomasWaldmann> and it is "Sanitize"
2007-06-28T00:37:27  <ThomasWaldmann> why do you use _i18n?
2007-06-28T00:41:33  <zenhase> perhaps because i imported version a line above and needed to work around a nameclash there my brain continued this to a non-existant clash on i18n
2007-06-28T00:41:37  <ThomasWaldmann> eventlog (as it is) needs the request to determince the logfile location
2007-06-28T00:41:52  <zenhase> well, thats ok
2007-06-28T00:42:00  <zenhase> but add gets it passed in again :>
2007-06-28T00:42:18  <zenhase> and i wondered if there will be ever two distinct request-objects
2007-06-28T00:47:26  <ThomasWaldmann> no, once should be enough
2007-06-28T01:22:48  <CIA-27> moin: Marian Neagul marian@info.uvt.ro * 2299:ddadd61affcf 1.7-classify-mneagul/MoinMoin/ (_tests/test_classifier.py classifier/__init__.py): Unicode fixes
2007-06-28T01:35:03  <CIA-27> moin: Marian Neagul marian@info.uvt.ro * 2300:cca03ef5f1c5 1.7-classify-mneagul/MoinMoin/_tests/test_classifier.py: Fix stupid error
2007-06-28T02:09:44  <CIA-27> moin: Marian Neagul marian@info.uvt.ro * 2301:e54e051895fd 1.7-classify-mneagul/MoinMoin/ (_tests/test_classifier.py classifier/SpamBayes.py): Separate benchmarks and fix a bug in SBClassifier
2007-06-28T02:11:20  <CIA-27> moin: Marian Neagul marian@info.uvt.ro * 2302:a75165e5ab9f 1.7-classify-mneagul/MoinMoin/_tests/test_classifier.py: Small CodingStyle Fix
2007-06-28T02:13:17  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2303:d48a4d0a0c70 1.7-classify-mneagul/ (MoinMoin/server/STANDALONE.py moin.py): pycallgraph support
2007-06-28T02:13:18  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2304:07fdbda38be2 1.7-classify-mneagul/MoinMoin/ (Page.py config/__init__.py): move split_regex to config (so it is compiled only once)
2007-06-28T02:13:19  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2305:4197965b0820 1.7-classify-mneagul/MoinMoin/config/__init__.py: fix: use re.UNICODE for split_regex
2007-06-28T02:13:20  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2306:ece659592dd9 1.7-classify-mneagul/MoinMoin/ (_tests/test_wikidicts.py action/SyncPages.py wikidicts.py): wikidicts: refactoring, more tests
2007-06-28T02:13:21  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2307:2c8f25796120 1.7-classify-mneagul/ (4 files in 4 dirs): merged main
2007-06-28T02:13:26  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2308:564dee5eb254 1.7-classify-mneagul/MoinMoin/ (request/__init__.py wikidicts.py events/wikidictsnotify.py): use event system to trigger wikidicts cache update
2007-06-28T02:13:29  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2309:b74a4230a123 1.7-classify-mneagul/MoinMoin/events/ (wikidictsrescan.py wikidictsnotify.py): renamed wikidictsnotify > wikidictsrescan
2007-06-28T02:13:32  <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2310:691b6a650d94 1.7-classify-mneagul/MoinMoin/_tests/parser/test_text_moin_wiki.py:
2007-06-28T02:13:35  <CIA-27> moin: test_text_moin_wiki: moved to _tests/parser, renamed from test_parser_text_moin_wiki, added test for colorized
2007-06-28T02:13:38  <CIA-27> moin: parser python and nesting {{{ }}}
2007-06-28T02:13:40  <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2311:6327a199c666 1.7-classify-mneagul/MoinMoin/parser/text_moin_wiki.py:
2007-06-28T02:13:47  <CIA-27> moin: text_moin_wiki:fixed bug for Colorized parsers and nesting {{{ }}} (FeatureRequests/ParsersCanNotContain{{{}}})
2007-06-28T02:13:50  <CIA-27> moin: , should be added to 1.6 too
2007-06-28T02:13:52  <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2312:9c66c14f6b62 1.7-classify-mneagul/MoinMoin/ (2 files in 2 dirs): text_moin_wiki: completing nesting for colorized parsers and some tests
2007-06-28T02:13:55  <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2313:709eea9354c7 1.7-classify-mneagul/MoinMoin/parser/text_moin_wiki.py: text_moin_wiki: some small PEP8 changes, that version should go into 1.6 too
2007-06-28T02:14:00  <CIA-27> moin: Alexander Schremmer <alex AT alexanderweb DOT de> * 2314:c9a4fbbb2b2e 1.7-classify-mneagul/MoinMoin/ (2 files in 2 dirs): Merged main.
2007-06-28T02:14:03  <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2315:7e04a576e8c7 1.7-classify-mneagul/MoinMoin/_tests/parser/test_text_moin_wiki.py: test_text_moin_wiki: small PEP8 changes
2007-06-28T02:14:06  <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2316:ec8d3c66ad0c 1.7-classify-mneagul/MoinMoin/_tests/parser/test_text_moin_wiki.py: test_text_moin_wiki: typo fixed
2007-06-28T02:14:11  <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2317:c1682ea7ed8c 1.7-classify-mneagul/MoinMoin/ (2 files in 2 dirs): text_moin_wiki: fixed an other nesting {{{ }}} problem and added a new test
2007-06-28T02:14:14  <CIA-27> moin: Marian Neagul marian@info.uvt.ro * 2318:0481465602af 1.7-classify-mneagul/ (11 files in 10 dirs): Merge with 1.7
2007-06-28T03:49:53  <zenhase> *yawn*
2007-06-28T03:51:56  <zenhase> i did researching on how to refactor .send_page and what to do to get around all the output capturing stuff
2007-06-28T03:53:02  <zenhase> tomorrow i would like some chat about this and some ideas i got
2007-06-28T06:14:52  *** irc.freenode.net sets mode: +o ChanServ
2007-06-28T08:52:51  <ThomasWaldmann> moin
2007-06-28T13:09:22  <grzywa> back...
2007-06-28T14:52:44  <lanius> ThomasWaldmann: what is the "refresh" action good for
2007-06-28T14:53:35  <xorAxAx> lanius: to clean the page cache
2007-06-28T14:53:47  <xorAxAx> lanius: i.e. the bytecode that is cached from every page's content
2007-06-28T14:59:00  <lanius> why would you want to do that
2007-06-28T15:01:39  <xorAxAx> because some macros or parsers are buggy with regards to caching, also a page rendering pass that didnt find a macro will not try to search for it again until the cache is deleted (this is important in plugin installation use cases)
2007-06-28T15:02:07  <xorAxAx> hmm, maybe not buggy, but better "need to be triggered explicitly to regenerate contents"
2007-06-28T15:02:20  <xorAxAx> note that every plugin can describe how it should be handled cache-wise
2007-06-28T15:05:33  <lanius> hmm, wouldn't it be better to have the cache implicit?
2007-06-28T15:06:35  <xorAxAx> ?
2007-06-28T15:06:46  <xorAxAx> please elaborate
2007-06-28T15:07:56  <lanius> implement it a way that the plugins must not bother with the cache
2007-06-28T15:08:21  <xorAxAx> thats not possible
2007-06-28T15:08:36  <xorAxAx> only plugins know if it is possible to cache their output
2007-06-28T15:08:41  <xorAxAx> and this works very well
2007-06-28T15:09:08  <xorAxAx> the only issue is that moin could invalidate it less often based on the current pseudo-events
2007-06-28T15:09:32  <xorAxAx> currently it is not cached at all if the plugin says that it subscribes to any pseudo-event
2007-06-28T15:09:50  <xorAxAx> moin could optimise this for common event realms like pages etc.
2007-06-28T15:10:04  <xorAxAx> have a look at various plugins, the name is "Dependencies" :)
2007-06-28T15:10:10  <lanius> k
2007-06-28T15:23:43  <lanius> http://somethingaboutorange.com/mrl/projects/nose/
2007-06-28T15:25:25  <xorAxAx> nose is a bad py.test remake  ::)
2007-06-28T15:25:36  <lanius> why is that
2007-06-28T15:25:58  <xorAxAx> because its less powerful than py.test and it was much later started than py.test ...
2007-06-28T15:26:52  <lanius> but it can do code coverage
2007-06-28T15:27:02  <xorAxAx> we can do so as well
2007-06-28T15:27:12  <xorAxAx> as i explained here a few days ago
2007-06-28T15:27:18  <lanius> ah, ok, have to read the logs then
2007-06-28T15:27:18  <xorAxAx> see our conftest file
2007-06-28T16:24:54  <zenhase_> re
2007-06-28T16:28:46  <vpv> weird, hg diff says "no newline" even when I put something like 5 newlines to the end of the file :P
2007-06-28T16:30:10  <zenhase> perhaps your editor strips them on save?
2007-06-28T16:30:46  <vpv> that might be possible but still weird, it's pydev and has worked until now...
2007-06-28T16:32:40  <vpv> works now, there were spaces in the end of the line before the last line
2007-06-28T16:36:50  <CIA-20> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2287:c1682ea7ed8c 1.7-maninfo-vpv/MoinMoin/ (2 files in 2 dirs): text_moin_wiki: fixed an other nesting {{{ }}} problem and added a new test
2007-06-28T16:36:53  <CIA-20> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2288:f7c762483617 1.7-maninfo-vpv/MoinMoin/ (2 files in 2 dirs): merge from main
2007-06-28T16:36:54  <CIA-20> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2289:f33ea579757d 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: The basic of Fedora CVS support, refactoring
2007-06-28T16:39:37  <CIA-20> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2290:2f3db1709b79 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: Add a comment explaining the incomplete state of CVS support
2007-06-28T16:51:25  <lanius> vpv: maybe there are some spaces in the new line
2007-06-28T16:57:07  <lanius> vpv: next time i read the complete log :D
2007-06-28T17:00:52  <vpv> :)
2007-06-28T17:11:57  <ser> vpv: where can read about cvs support?
2007-06-28T17:13:23  <vpv> What do you mean? I need CVS support for my man/info project and in the beginning it's going to be pretty Fedora specific...
2007-06-28T17:14:32  <ser> ah, ok - i hoped it will be usefull wider
2007-06-28T17:14:38  <vpv> But here's the description how I need to work with Fedora's systems: http://fedoraproject.org/wiki/SummerOfCode/2007/VillePekkaVainio/SourceRepository
2007-06-28T17:14:58  <xorAxAx> i am sure that it will be useful for everybody, at least if he assumes that the code will go into the main line :)
2007-06-28T17:15:07  <xorAxAx> but its perfectly ok to generalise code later on
2007-06-28T17:15:20  <vpv> that
2007-06-28T17:15:21  <xorAxAx> (if later on doesnt mean 10 years later :))
2007-06-28T17:15:23  <vpv> oops
2007-06-28T17:15:39  <vpv> that's the idea, I just want to get a proof of concept implementation now
2007-06-28T17:15:45  <xorAxAx> yeah, perfectly ok
2007-06-28T17:16:00  <ser> i was interested in because i need some way of connectivity between cvs and moin
2007-06-28T17:16:25  <xorAxAx> ser: i think it would be pretty unrelated, but in 1.7 you will be able to write a custom storage backend
2007-06-28T17:16:32  <xorAxAx> ser: e.g. one that displays text files from cvs
2007-06-28T17:16:40  <ser> ok, thanks
2007-06-28T17:16:44  <vpv> but what I'm doing with CVS is commands.getstatusoutput("cvs co ...") so it's not anything more fancy than that
2007-06-28T17:16:56  <xorAxAx> vpv: arent there cvs wrappers? :)
2007-06-28T17:17:03  <xorAxAx> so that you dont need to reinvent the parsing wheel
2007-06-28T17:17:28  <vpv> I haven't found any, or actually I found one but that was an initial release and seemed unmaintained after that
2007-06-28T17:17:39  <xorAxAx> hmm
2007-06-28T17:17:42  <xorAxAx> i just know svn wrappers
2007-06-28T17:41:07  <zenhase> now there
2007-06-28T17:42:24  <zenhase> yesterday i was trying to get a hold of where to start taking apart .send_page and move the different parts
2007-06-28T17:44:23  <zenhase> and i kinda ran into a blockage, because as it appears to me, there are so many interdependencies :/
2007-06-28T17:45:56  <dreimark> bbl
2007-06-28T17:46:46  <zenhase> you start at send_page, walk through things like parsers and end up somewhere at the formatters who write the actual output
2007-06-28T17:46:59  <xorAxAx> yes
2007-06-28T17:47:03  <xorAxAx> so? :)
2007-06-28T17:47:06  <zenhase> really quite intermingled stuff :/
2007-06-28T17:47:27  <xorAxAx> well
2007-06-28T17:47:32  <xorAxAx> thats unrelated to sendpage
2007-06-28T17:47:32  <zenhase> and i actually don't think that doing the output capturing thingy will work in the long term
2007-06-28T17:47:57  <xorAxAx> but some parts of the formatter stuff is weird, e.g. formatter.page
2007-06-28T17:48:34  <zenhase> i also read through some other projects code, especially trac
2007-06-28T17:49:04  <xorAxAx> well,  i dont see directly how parsers/formatters are an issue for the send_page stuff
2007-06-28T17:50:08  <zenhase> send_page calls them and they do their output directly themselves (the formatters that is)
2007-06-28T17:50:13  <xorAxAx> yes
2007-06-28T17:50:22  <xorAxAx> but thats unrelated to redesigning send page
2007-06-28T18:00:14  *** irc.freenode.net sets mode: +o ChanServ
2007-06-28T18:00:44  <xorAxAx> why not? how can it break?
2007-06-28T18:00:46  <xorAxAx> how do you want to fix this without redesigning the whole plugin interfaces?
2007-06-28T18:00:46  <zenhase> but actually it's kind of another dirty hack
2007-06-28T18:00:46  <xorAxAx> thereby rendering them useless
2007-06-28T18:00:47  <xorAxAx> well, it depends on your goals
2007-06-28T18:00:50  <xorAxAx> and there is no direct advantage to make all stuff return the formatter output instead of writing them to the request
2007-06-28T18:01:03  <xorAxAx> or can you think of some?
2007-06-28T18:01:52  <xorAxAx> of course the best solution to have some sane stream or DOM interface inside
2007-06-28T18:02:06  <zenhase> seperation of responsibilities
2007-06-28T18:02:10  <zenhase> yes
2007-06-28T18:02:16  <xorAxAx> but thats a bit unrelated to the theming stuff itself
2007-06-28T18:02:23  <xorAxAx> and a big task itself
2007-06-28T18:02:26  <zenhase> that's right
2007-06-28T18:02:46  <xorAxAx> please elaborate about "seperation of responsibilities"
2007-06-28T18:03:16  <zenhase> that i don't think that a formatter is something that should output to the user
2007-06-28T18:03:50  <zenhase> parser/formatter the way i view it should mainly convert between different representations
2007-06-28T18:03:52  <xorAxAx> you mean the r.w(f.foo(...)) pattern?
2007-06-28T18:03:54  <zenhase> s/view/see
2007-06-28T18:04:56  <zenhase> it's like a chain of responsibility, where you put in stuff on the one end and get a result at the other
2007-06-28T18:05:30  <xorAxAx> yeah, but thats a very basic idea of refactoring and designing new interfaces, and i want to say orthogonal one to theming stuff
2007-06-28T18:05:31  <zenhase> and in this "conveyor-belt" like chain you can even do finegrained caching at every station
2007-06-28T18:05:45  <xorAxAx> its perfectly valid to think about this, it would shift the work from your soc focus, though
2007-06-28T18:05:54  <zenhase> i know
2007-06-28T18:06:21  <xorAxAx> (thats a nice task to tackle after you have finished (for some definition of finished) the theming stuff)
2007-06-28T18:06:23  <zenhase> but i would not like to base the result of my SOC on workarounds and hacks
2007-06-28T18:06:51  <xorAxAx> well, its not a workaround, its adapting the current basic interfaces
2007-06-28T18:07:12  <zenhase> actually i think that is an euphemism now .. ;P
2007-06-28T18:07:19  <xorAxAx> have you looked into genshi stream renderer if genshi likes the described callback pattern?
2007-06-28T18:07:47  <zenhase> it doesn't
2007-06-28T18:07:58  <zenhase> it expects markup-events at the end
2007-06-28T18:08:12  <xorAxAx> maybe you can subclass genshi classes to make it compliant to our ideas?
2007-06-28T18:08:21  <zenhase> if you call back from within the theme, you end up having the pageoutput first and then the themestuff
2007-06-28T18:08:23  <xorAxAx> well, you can serialise request.write stuff to markup events
2007-06-28T18:08:49  <zenhase> now thats basically output-capturing again
2007-06-28T18:09:03  <xorAxAx> no, seriously, IMHO you shouldnt wonder about this issue ... you cant currently fix it
2007-06-28T18:09:31  <xorAxAx> trying to achieve some perfect solution doesnt work if you go the first path of hacks to fix up first
2007-06-28T18:09:43  <xorAxAx> you need to work on units and their interfaces, and sometimes they are suboptimal
2007-06-28T18:10:32  <zenhase> well, but instead of adding to those suboptimal interfaces i think that it's worthwhile fixing them right away
2007-06-28T18:10:44  <zenhase> .send_page is a good example for this
2007-06-28T18:10:49  <xorAxAx> as i said, thats not possible without losing SOC focus
2007-06-28T18:10:56  <zenhase> i looked back to 2002
2007-06-28T18:11:00  <xorAxAx> send_page is more along the lines of your theming tasks :)
2007-06-28T18:11:08  <zenhase> i think .send_page at least doubled size
2007-06-28T18:11:13  <xorAxAx> yes, it was cleaner back then :)
2007-06-28T18:11:29  <zenhase> but it still should have been killed right back then
2007-06-28T18:11:32  <xorAxAx> i dont see how the described "workaround" is a problem in terms of complexity
2007-06-28T18:11:38  <zenhase> as the "TODO" item states
2007-06-28T18:11:45  <xorAxAx> have you had a look at the text_python formatter?
2007-06-28T18:11:49  <zenhase> yes
2007-06-28T18:11:54  <xorAxAx> it was written by your mentor and contains a pretty large hack
2007-06-28T18:12:01  <xorAxAx> and nobody extended it with his own code
2007-06-28T18:12:12  <xorAxAx> because the formatter API fits the use cases in terms of it
2007-06-28T18:12:25  <zenhase> hmm?
2007-06-28T18:12:39  <xorAxAx> i mean, not every that is dirty will be abused and made even worse
2007-06-28T18:12:50  <xorAxAx> and currently, you need to adapt to the other working pattern of the rendering chain
2007-06-28T18:13:18  <xorAxAx> no way around it and redesigning is too large for your current focus (of course you can do it later and then rip out the workaround, at least partly)
2007-06-28T18:13:29  <xorAxAx> and dont forget compatiblity :)
2007-06-28T18:14:06  <zenhase> backward compatibility will not work for the complete lifecycle of a project
2007-06-28T18:14:09  <zenhase> :>
2007-06-28T18:14:24  <xorAxAx> sure, win xp can run win 1 programs :)
2007-06-28T18:14:30  <xorAxAx> well, it needs to balanced of course
2007-06-28T18:15:32  <zenhase> well, win is also a perfect example where backward compat accumulated so much cruft it's sometimes a real pita to develop there
2007-06-28T18:16:02  <xorAxAx> well, surely depends, they have different platforms
2007-06-28T18:16:15  <xorAxAx> i think MFC is a pita for example
2007-06-28T18:16:17  <zenhase> remember thomas talking about this software which costs 150$ on osx and 350$ on win? i really think it has to do with developer-pain at least in some parts
2007-06-28T18:16:36  <zenhase> (and develtools for win costing a lot of cash)
2007-06-28T18:16:51  <xorAxAx> or the mac port is already amortized
2007-06-28T18:17:08  <xorAxAx> ... more fully
2007-06-28T18:17:36  <zenhase> in terms of userbase i doubt that ;D
2007-06-28T18:17:49  <zenhase> brb, toilet
2007-06-28T18:20:36  <zenhase> re
2007-06-28T18:25:19  <zenhase> well back to the main subject then
2007-06-28T18:28:07  <zenhase> last item was: compatibility
2007-06-28T18:28:50  <zenhase> if i do kill .send_page ... this will already break several things in terms of compat, doesn't it?
2007-06-28T20:35:24  <ThomasWaldmann> re
2007-06-28T20:39:10  <ThomasWaldmann> zenhase: quite some stuff uses it for displaying some msg
2007-06-28T20:40:28  <ThomasWaldmann> otoh, that msg=... stuff should be replaced anyway
2007-06-28T20:41:42  <zenhase> i saw something nice in trac
2007-06-28T20:41:54  <zenhase> would be cool for templates too
2007-06-28T20:42:03  <zenhase> they have a Chrome object
2007-06-28T20:42:37  <zenhase> it stands for everything actions or the like want to add to the final output, mostly stuff in <head>
2007-06-28T20:44:07  <zenhase> then, when rendering with the genshi themes, the theme has access to this chrome-object and can add links, alternates, metas from the chrome to the head
2007-06-28T20:44:38  <zenhase> in moin i would do this in _base.html so theme authors don't get really bothered
2007-06-28T20:45:30  <zenhase> this would also clean up the hardcoded html-stuff that's in the configs
2007-06-28T20:46:49  <ThomasWaldmann> but only if we switch completely to genshi
2007-06-28T20:46:54  <zenhase> you would put lists of css or js files in the config to have them added to the chrome by default and _base.html puts them into the output automatically
2007-06-28T20:46:58  <zenhase> nope
2007-06-28T20:47:07  <zenhase> this could also be hacked into ThemeBase
2007-06-28T20:47:17  <zenhase> so it works with old themes too
2007-06-28T20:47:58  <zenhase> just reach the chrome around by attaching it to request for example
2007-06-28T20:48:12  <zenhase> this is how trac does it too btw iirc
2007-06-28T20:49:26  <ThomasWaldmann> well, if it works, feel free :)
2007-06-28T20:49:47  <zenhase> it would have methods like add_stylesheet(filename, media='screen'), add_meta(**kwargs) and the like
2007-06-28T20:50:19  * ThomasWaldmann must mentally recovery from 4hrs windows, including vista %)
2007-06-28T20:50:29  <zenhase> and in genshi i would not have to deal with those ${Markup(config.user_head)} stuff anymore :>
2007-06-28T20:51:14  <starshine> o_O /me rushes ThomasWaldmann a lemonade and some penguin mints
2007-06-28T20:51:17  <zenhase> and systemmessages could be put there too imo
2007-06-28T20:52:18  <zenhase> starshine: these caffeinated mints?
2007-06-28T20:52:41  * ThomasWaldmann feels already better :))
2007-06-28T20:53:54  <starshine> zenhase: yeah those are the ones. 3 of them has same caffeine as a soda
2007-06-28T20:54:09  <ThomasWaldmann> it really sucks do debug errors like "error, can't init load path" (of course it doesn't tell you which program triggered that error, or which path it is talking about)
2007-06-28T20:54:17  <starshine> a friend of mine didn't know that when he was given a tin of them for the Penguin Bowl game :)
2007-06-28T20:54:37  <starshine> it's ok you're back in good ol python now
2007-06-28T20:55:26  <zenhase> starshine: hmm .. i think i stick to club-mate when it comes caffeine
2007-06-28T20:56:27  * ThomasWaldmann prefers other molecules
2007-06-28T20:56:41  <zenhase> at the last c3 there was a guy who had to be shipped to hospital because he overdosed on raw caffeine pills
2007-06-28T20:58:09  <starshine> I figured the lemon and mint would be good :D
2007-06-28T20:58:29  <starshine> .o( what sort of moin icon for caffeine?
2007-06-28T21:01:47  <zenhase> a plain coffee bean?
2007-06-28T21:02:00  <zenhase> this will attract many java guys :)
2007-06-28T21:03:28  <ThomasWaldmann> we need a maintainer for twikidraw :D
2007-06-28T21:08:29  <xorAxAx> well, it wasnt easy to stop puking when i looked at the code how they handle the form upload
2007-06-28T21:08:46  <xorAxAx> maybe we should have a look if they optimised that in newer versions
2007-06-28T22:26:54  <xorAxAx> lanius: http://exogen.case.edu/revisions.png
2007-06-28T22:26:55  <xorAxAx> ThomasWaldmann: http://exogen.case.edu/revisions.png
2007-06-28T22:28:20  <xorAxAx> http://blog.case.edu/bmb12/2007/06/pagoda_revisions
2007-06-28T22:31:34  <xorAxAx> it looks interesting but i havent figured out how e.g. "node" works
2007-06-28T22:51:04  <dreimark> moin
2007-06-28T22:58:29  <dreimark> hi neagulm
2007-06-28T22:58:42  <neagulm> hello dreimark
2007-06-28T22:59:42  <dreimark> did you have read the logs from yesterday night ?
2007-06-28T23:01:16  <neagulm> dreimark, is it about "test_classifier.py  the last test fails" ?
2007-06-28T23:01:21  <dreimark> neagulm: starshine was requesting what kind of icons for spam / ham should be used
2007-06-28T23:01:42  <dreimark> neagulm: that test does now work, here
2007-06-28T23:01:56  <neagulm> dreimark, something like "garbage bin"
2007-06-28T23:02:17  <neagulm> dreimark, please consult the resulting TCR values for the two classifiers
2007-06-28T23:02:42  <neagulm> dreimark, SB have obtains a better score then Reverend... :)
2007-06-28T23:07:44  <dreimark> neagulm: this 9.0 x  where does it come from ?
2007-06-28T23:07:52  <dreimark> why is it set to 9.0
2007-06-28T23:09:32  <neagulm> dreimark, a magic number :). It represents the cost of penality of missclasifying a ham page as spam. It's presented in the article I have sent you.
2007-06-28T23:11:26  <neagulm> dreimark, in the spam classification domain (specificaly in benchmarking) there are used the following values: 1, 9 and 999. The higher the value the more are penalized false positives
2007-06-28T23:12:18  <dreimark> ahh ok. Please could you add a comment for that in the file too, I guess others will ask that too. Not now may be ins some weeks or years ;)
2007-06-28T23:13:50  <neagulm> currently I have some problems with locking... After the last merge's I get a strange exception...
2007-06-28T23:13:57  <neagulm> dreimark, i'm in debugging mode
2007-06-28T23:14:06  <dreimark> starshine_away: (starshine) neagulm please exchange your minds about which icons should be added to the theme, zenhase keep care that it fits into your code
2007-06-28T23:15:41  <dreimark> neagulm: mostly it were changes from me and  ThomasWaldmann wikidict refactoring
2007-06-28T23:18:59  <CIA-20> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2107:b80aab797cbd 1.7-storage-hwendel/MoinMoin/ (Page.py storage/fs_moin16.py): fix editing pages
2007-06-28T23:22:06  <lanius> puh, this one was hard, took some time to debug that
2007-06-28T23:22:34  <dreimark> neagulm: we should try to have on monday or tuesday a chat with skip?
2007-06-28T23:22:56  <dreimark> tommorrow evening I will not have much time
2007-06-28T23:23:05  <neagulm> Monday is fine.
2007-06-28T23:24:11  <dreimark> ok, i send him an mail, I suppose to have the same time as always
2007-06-28T23:24:30  <neagulm> dreimark, :)
2007-06-28T23:24:35  <dreimark> lanius: what was it about?
2007-06-28T23:25:50  <neagulm> dreimark, until then I will work for the interface for SpamBayes Server (even if this is not the right solution).
2007-06-28T23:26:44  <dreimark> neagulm: the current system is quite finished from the logic or?
2007-06-28T23:26:58  <dreimark> without using the server?
2007-06-28T23:27:27  <neagulm> dreimark, in general yes, but there are some unfinished things with the tokenizer. It needs more work...
2007-06-28T23:28:29  <lanius> dreimark: going through the save page stuff so that it works with my storage stuff
2007-06-28T23:29:16  <lanius> and the missing page thing
2007-06-28T23:31:05  <dreimark> lanius: :)
2007-06-28T23:33:38  <dreimark> neagulm: please feel free to add some ducumentation to http://moinmoin.wikiwikiweb.de/MoinMoinTodo/Release_1.7 too
2007-06-28T23:34:04  <xorAxAx> lanius: hah, the missing page thing :)
2007-06-28T23:34:07  <xorAxAx> very funny thing
2007-06-28T23:34:21  <xorAxAx> i remember how i added some code that even shows something if the missingpage is missing
2007-06-28T23:34:28  <lanius> few lines but big effect
2007-06-28T23:35:13  <dreimark> neagulm: and the SpamBayes server part currently is only usefull for the text part and not attachments. I think the last one should discussed with skip before working on that
2007-06-28T23:35:44  <dreimark> or whats your opinion
2007-06-28T23:35:47  <dreimark> ?
2007-06-28T23:39:37  <neagulm> dreimark, I agree. The attachment part will need more changes in other parts of MM. I'm not really sure how to disclose the attachment classification to the user
2007-06-28T23:43:46  <dreimark> Probably it needs to be shown in AttachList, using icons but may be other places in the theme would be nice too
2007-06-28T23:55:53  <starshine> dreimark: neagulm: saw your note about icons.
2007-06-28T23:58:37  <dreimark> starshine: neagulm it would be fine if you could prepare some nice icons we could add to the modern theme (classic, rightsidebar)
2007-06-28T23:59:14  <starshine> for which purposes and what sizes would you like?