2007-08-01T00:14:11 <neagulm> good night
2007-08-01T00:47:04 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2583:dd005fd66306 1.7/MoinMoin/script/migration/ (1050800.py 1059999.py 1060000.py conv160.py): 1.5.x to 1.6 conversion (unfinished)
2007-08-01T00:54:36 *** ThomasWaldmann sets mode: -o ThomasWaldmann
2007-08-01T02:12:35 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2584:efac33f4b9e4 1.7/MoinMoin/script/migration/ (1050800.py 1059999.py _conv160.py conv160.py): 1.6 converter: rename module
2007-08-01T02:12:37 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2585:cb899abca9a8 1.7/MoinMoin/script/migration/ (1050800.py 1059999.py): 1.6 converter: add main function for testing, some fixes
2007-08-01T06:12:55 <dreimark> moin
2007-08-01T08:13:11 <ThomasWaldmann> moin
2007-08-01T08:22:23 * ThomasWaldmann reads about Mentor Summit 2007 plans
2007-08-01T09:05:22 <chacha_chaudhry> ThomasWaldmann: umask variable does not have its usual meaning in config.py file as in umask command. :-)
2007-08-01T09:13:42 * chacha_chaudhry finds it out in filesys ------------- umask AND with MODE rather then XOR. Simple :p
2007-08-01T09:37:59 * ThomasWaldmann wrote some soc roundup to the soc mentors list
2007-08-01T09:58:39 <ThomasWaldmann> chacha_chaudhry: non-development stuff please on #moin :)
2007-08-01T09:59:25 <chacha_chaudhry> ThomasWaldmann: aah, next will keep in mind :-)
2007-08-01T10:08:15 <xorAxAx> ThomasWaldmann: URL? :)
2007-08-01T10:09:27 <xorAxAx> ah, on the mailing list, ok
2007-08-01T10:25:02 <ThomasWaldmann> in short: Google Mentor Summit, 6. Oct 2007, up to 3 mentors/org, no guarantees for being invited
2007-08-01T10:27:18 <xorAxAx> hehe, i wonder whom they pick
2007-08-01T10:52:47 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2738:b8973bb24aa3 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: some updates handling improvements, needs even more testing
2007-08-01T10:57:46 <xorAxAx> grzywacz: ping
2007-08-01T11:03:32 <johill> ThomasWaldmann: somewhere on my files dir :)
2007-08-01T11:04:07 <johill> ThomasWaldmann: http://johannes.sipsolutions.net/files/linkconv.py.txt
2007-08-01T11:04:19 <johill> xorAxAx: I made a whole parser/cache/formatter test with dom
2007-08-01T11:04:48 <xorAxAx> cool, will have a look
2007-08-01T11:04:55 <xorAxAx> in the next days
2007-08-01T11:04:56 <ThomasWaldmann> johill: i guess i have that one already
2007-08-01T11:05:33 <ThomasWaldmann> i made the framework ready for markup conversion, I hope I can integrate it the next days.
2007-08-01T11:05:47 <ThomasWaldmann> see my last commits
2007-08-01T11:06:05 * ThomasWaldmann posted some travel story from last year
2007-08-01T11:08:48 <johill> ThomasWaldmann: I mailed you an earlier copy, so better pull that one again
2007-08-01T11:09:04 <johill> ThomasWaldmann: also, I never had anything but this script, I made a simple find/xargs command line to convert all files
2007-08-01T11:13:54 <ThomasWaldmann> see my last commits for about what it all needs to do
2007-08-01T11:16:41 <grzywacz> xorAxAx, pong
2007-08-01T11:16:49 <xorAxAx> grzywacz: current status? :)
2007-08-01T11:17:34 <grzywacz> Hm, can we try ekiga in about 30 minutes? :)
2007-08-01T11:17:43 <grzywacz> I'll report there. ;D
2007-08-01T11:17:55 <xorAxAx> hmm, hehe
2007-08-01T11:17:59 <xorAxAx> ok
2007-08-01T11:18:19 <grzywacz> thanks :)
2007-08-01T11:20:13 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2586:a838bdc4e40a 1.7/MoinMoin/script/migration/_conv160.py: 1.6 converter: quicklinks/subscribed pages (unfinished)
2007-08-01T11:20:27 <ThomasWaldmann> brb
2007-08-01T11:30:11 <xorAxAx> grzywacz: i am currently trying to find out how to call a sipgate account without going via the phone network :)
2007-08-01T11:31:26 <grzywacz> xorAxAx, heh, ok. That can be useful. ;-)
2007-08-01T11:31:36 <grzywacz> (There's always skype, tsk tsk ;p)
2007-08-01T11:31:57 <xorAxAx> its not possible :) but i have another solution :)
2007-08-01T11:32:14 <xorAxAx> i just terminate my own sip account on my router
2007-08-01T11:36:41 <xorAxAx> sip:8006245@alexanderweb.homeip.net
2007-08-01T11:37:05 <xorAxAx> grzywacz: that should work :)
2007-08-01T11:37:14 <xorAxAx> (and connect you directly with my DECT phone :))
2007-08-01T11:37:40 <grzywacz> ok, I'll cal you in a few minutes when I'm done with my tea and breakfast, so as to make no strange noises during the conversation ;)
2007-08-01T11:37:54 <grzywacz> *call
2007-08-01T11:38:01 <xorAxAx> hehe, ok
2007-08-01T12:09:26 <grzywacz> woot, thanks
2007-08-01T12:09:38 <grzywacz> I must switch to thinking in English by tomorrow. ;)
2007-08-01T12:09:50 <grzywacz> I don't want to pause like during the interview. ;s
2007-08-01T12:10:19 <xorAxAx> well, you were quite good
2007-08-01T12:10:51 * xorAxAx knows some pypy-guy from poland who seemed to learn english while being active in the pypy project :)
2007-08-01T12:10:54 <grzywacz> not good enough, though
2007-08-01T12:11:06 <grzywacz> I used to do better in the past. ;D
2007-08-01T12:11:10 <xorAxAx> hehe
2007-08-01T12:26:00 <johill> xorAxAx: there's one major problem with how I structured it
2007-08-01T12:26:09 <xorAxAx> that is? :)
2007-08-01T12:26:14 <johill> that being that a macro can create anything but not everything is valid everywhere
2007-08-01T12:28:26 <xorAxAx> hmm, yes
2007-08-01T12:28:40 <xorAxAx> one could run a validation checker after every include
2007-08-01T12:29:05 <johill> http://git.sipsolutions.net/?p=dom.git;a=blob;f=PROBLEMS;hb=HEAD
2007-08-01T12:29:16 <johill> bit more detailed
2007-08-01T12:29:30 <xorAxAx> ok
2007-08-01T12:32:27 <johill> yes, that validator is one possibility
2007-08-01T12:32:35 <johill> I'm not sure it's good though because it's hard to fix it up
2007-08-01T12:32:47 <johill> and just knowing it's wrong doesn't do any good
2007-08-01T12:41:11 <xorAxAx> it does some good for macro developers :)
2007-08-01T12:41:20 <xorAxAx> but legacy is an issue, yes
2007-08-01T12:42:02 <johill> well not quite
2007-08-01T12:42:18 <xorAxAx> hmm?
2007-08-01T12:42:19 <johill> I think all uses cases are valid, i.e. having a macro that just expands to text or that expands to a whole section
2007-08-01T12:42:33 <xorAxAx> yes
2007-08-01T12:42:34 <johill> think TOC vs. date
2007-08-01T12:42:39 <xorAxAx> both are valid, yes
2007-08-01T12:42:42 <johill> so the parser already has to know where to stick it
2007-08-01T12:43:02 <xorAxAx> the parser sees the context: in p or not
2007-08-01T12:43:37 <johill> yes, but if say include is in p then it has to close the p and then add the include node
2007-08-01T13:01:52 <xorAxAx> neagulm: so why does d49c38fdfb37 have a mismatch between comment and diff? (This message has been postponed.)
2007-08-01T13:02:03 <neagulm> hello
2007-08-01T13:02:21 <xorAxAx> hi neagulm :)
2007-08-01T13:02:23 <neagulm> xorAxAx, It was a mistake, I forgot that it was added
2007-08-01T13:02:29 <xorAxAx> ok
2007-08-01T13:02:45 <neagulm> xorAxAx, it works ! :D
2007-08-01T13:03:21 <xorAxAx> the ART stuff?
2007-08-01T13:03:26 <neagulm> xorAxAx, yes
2007-08-01T13:03:28 <xorAxAx> how did you solve the labelling problem?
2007-08-01T13:03:36 <xorAxAx> cool
2007-08-01T13:04:03 <neagulm> xorAxAx, i didn't solve it but i could attach a label to the training pages
2007-08-01T13:04:43 <xorAxAx> hmm
2007-08-01T13:05:02 <xorAxAx> maybe its enough to list some pages that are classified very strongly into a specific catgory
2007-08-01T13:05:13 <xorAxAx> neagulm: or like that, yes
2007-08-01T13:05:24 <xorAxAx> s/catgory/cluster/
2007-08-01T13:05:38 <neagulm> i need to filter pages that only redirect to another page
2007-08-01T13:06:53 <neagulm> xorAxAx, i had some problems with this issue, after some playing around i discovered that the classifier had grouped them together
2007-08-01T13:07:13 <neagulm> xorAxAx, currently I'm testing with the underlay pages
2007-08-01T13:07:57 <xorAxAx> yeah, good idea
2007-08-01T13:08:20 <xorAxAx> do you feed it with wiki markup?
2007-08-01T13:08:26 <xorAxAx> i mean, raw page text
2007-08-01T13:08:40 * xorAxAx is gone
2007-08-01T13:09:36 <neagulm> xorAxAx, yes
2007-08-01T13:51:05 <dreimark> moin
2007-08-01T13:51:57 <dreimark> neagulm: How are you?
2007-08-01T13:52:49 <neagulm> hello dreimark
2007-08-01T13:53:10 <neagulm> dreimark, you finished your vacation ?
2007-08-01T13:53:37 <dreimark> No
2007-08-01T13:53:48 <neagulm> dreimark, I'm fine... trying to understand some inner workings of NN
2007-08-01T13:54:08 <dreimark> MM
2007-08-01T13:54:14 <neagulm> dreimark, :)
2007-08-01T13:54:27 <neagulm> NN = Neural Networks
2007-08-01T13:54:38 <dreimark> Ok,
2007-08-01T13:54:55 <dreimark> Oh, another shortcut
2007-08-01T13:55:37 <neagulm> dreimark, the ART classifier seems to be working... :D
2007-08-01T13:56:33 <neagulm> dreimark, you have joined #moin-dev from the hotel ?
2007-08-01T13:56:46 <dreimark> fine, I cant test from here, because to small device
2007-08-01T13:56:59 <dreimark> Yeah, from my tpc
2007-08-01T13:58:06 <dreimark> We will visit grand canyon in some hours
2007-08-01T13:58:57 <neagulm> nice
2007-08-01T13:59:28 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2739:40cc00b40ce3 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: fix parse_filelist() and parse_primary(), updates handling should finally work
2007-08-01T13:59:53 <dreimark> Yeah, you should visit the National Parks too
2007-08-01T14:00:01 <johill> vpv: ISTR there were two things I was supposed to commit to 1.7 from you, can you remind me what it was?
2007-08-01T14:00:37 <vpv> johill: I wish I remembered myself even, I'll search my repo a bit
2007-08-01T14:01:26 <johill> I think it was still about the xml rpc token stuff I thoroughly messed up
2007-08-01T14:01:54 <dreimark> MM shows 500
2007-08-01T14:03:29 <vpv> johill: ok, I found them, it was 9 days ago, says hg
2007-08-01T14:03:38 * johill looks
2007-08-01T14:03:46 <vpv> http://hg.moinmo.in/moin/1.7-maninfo-vpv/rev/849142a76ea1
2007-08-01T14:04:05 <vpv> that's one, I'm not sure if it's in main yet
2007-08-01T14:04:24 <dreimark> neagulm: do you have added docs for the ART?
2007-08-01T14:04:56 <johill> vpv: yeah that was it, applied
2007-08-01T14:04:57 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2587:64725664bd63 1.7/MoinMoin/xmlrpc/__init__.py: fix _cleanup_stale_tokens call
2007-08-01T14:05:36 <neagulm> dreimark, not yet, it is not fully implemented, I need to find some way to store the classifications. I need to be able to find in a fast way all pages that have been previously classified in a specific category...
2007-08-01T14:06:51 <vpv> johill: thanks, I don't think there was anything else, you seem not to have applied my aa428c7444e3 to main, where I just messed things up. If you have, then there's a fix for that...
2007-08-01T14:07:11 <neagulm> dreimark, it would be nice if I could use sqlite. But because sqlite is missing from standard python 2.3 and 2.4 I will need to find some other way in order to not introduce another dependency
2007-08-01T14:07:30 <johill> hm, aa428c7444e3 looks valid
2007-08-01T14:07:45 <johill> ah no, aa428c7444e3 is the fix
2007-08-01T14:08:44 <johill> do you need the user object stored?
2007-08-01T14:10:35 <vpv> I believe aa428c7444e3 is something I messed up
2007-08-01T14:11:15 <johill> somewhere you got the line above as self.u instead of u
2007-08-01T14:11:34 <vpv> because I'm fixing it for real in b41c924c4f4c
2007-08-01T14:11:35 <johill> so now in 1.7 it looks fine
2007-08-01T14:11:49 <vpv> self.u is bad, that was my mistake
2007-08-01T14:12:03 <johill> yeah, it looks like that in 1.7 now
2007-08-01T14:12:23 <dreimark> neagulm: oj
2007-08-01T14:12:29 <dreimark> Ok
2007-08-01T14:12:40 <johill> except in b41c924c4f4c you have _cleanup_stale_tokens again instead of self.
2007-08-01T14:13:56 <vpv> yes, for some reason I do, but 64725664bd63 was after b41c924c4f4c, where I fixed that
2007-08-01T14:14:15 <johill> ah ok
2007-08-01T14:14:35 <vpv> I'll take a look at main too, to be sure
2007-08-01T14:14:46 <johill> that one is a commit from main
2007-08-01T14:14:51 <johill> 64725664bd63 I mean
2007-08-01T14:16:05 <vpv> then I had http://hg.moinmo.in/moin/1.7-maninfo-vpv/rev/6b4fc04d700d which I wrote to fail if there happens to ne similar problems in getAuthToken, I'm not sure if that should go to main
2007-08-01T14:16:20 <vpv> s/ne/be/
2007-08-01T14:16:51 <johill> should I put that into 1.7 too then?
2007-08-01T14:16:53 <neagulm> dreimark, "o j" ?
2007-08-01T14:17:10 <vpv> johill: I'm not sure, if it looks ok to you
2007-08-01T14:18:07 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2588:5ae633bf2de5 1.7/MoinMoin/xmlrpc/_tests/test_xmlrpc.py: add test for xmlrpc getAuthToken
2007-08-01T14:18:24 <dreimark> neagulm: I try to
2007-08-01T14:18:35 <vpv> Ok, I just checked the xmlrpc stuff in main where all this confusion happened, it looks alright to me now :)
2007-08-01T14:18:38 <vpv> thanks johill
2007-08-01T14:19:14 <dreimark> neagulm: http://initd.org/tracker/pysqlite/wiki/pysqlite
2007-08-01T14:19:53 <neagulm> dreimark, yes but is not built-in in python stdlib
2007-08-01T14:20:14 <johill> vpv: thanks for checking :)
2007-08-01T14:21:57 <dreimark> neagulm: discuss with ThomasWaldmann to add it to support files if no other solution is possible
2007-08-01T14:22:45 <dreimark> I try to fall in sleep again, gn
2007-08-01T14:23:00 <neagulm> dreimark, good night
2007-08-01T14:23:10 <neagulm> dreimark, have fun!
2007-08-01T14:27:40 <grzywacz> http://googlesummerofcode.blogspot.com/2007/07/midterm-results-are-in.html
2007-08-01T14:32:53 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2746:e53250e1a138 1.7-maninfo-vpv/MoinMoin/ (5 files in 2 dirs): merge from main
2007-08-01T14:44:13 <ThomasWaldmann> neagulm: bsddb maybe?
2007-08-01T14:45:22 <ThomasWaldmann> neagulm: if the stuff you want to store falls into category "item meta data", we hopefully will have some some moin-built-in storage for that soon
2007-08-01T14:45:40 <ThomasWaldmann> neagulm: see lanius' project
2007-08-01T14:46:06 <neagulm> ThomasWaldmann, exactly. I need a way to find fast all pages that are marked as belonging to a specific category
2007-08-01T14:46:57 <ThomasWaldmann> there is pagelinks cache and xapian
2007-08-01T14:47:26 <ThomasWaldmann> (otoh, maybe not everybody will have xapian activated)
2007-08-01T14:47:47 <neagulm> I was thinking about xapian for a while
2007-08-01T15:03:42 <neagulm> ThomasWaldmann, bsddb seems a good choice, i only need to store a list of categories (bsddb keys) and their members (bsddb values). The number of categories depends on the results from ART
2007-08-01T15:04:43 <neagulm> ThomasWaldmann, I need locking, fast access, ability to add or remove members from a category (this could be a problem with bsddb)
2007-08-01T15:04:57 * neagulm is going away for a few minutes
2007-08-01T15:23:39 <vpv> nice, I just got xmlrpc putpage to hang, strace says it's waiting on recv(5, :/
2007-08-01T15:32:22 <vpv> seems the whole moin was crashed, I had to kill -9 it
2007-08-01T15:39:43 <ThomasWaldmann> lanius: + items = self._items_all.keys()[:] is this copying really needed?
2007-08-01T15:41:02 <ThomasWaldmann> lanius: the request_cli stuff will have to get fixed in main
2007-08-01T15:46:41 <ThomasWaldmann> neagulm: please read PEP8 again. between member function "def"s is usually 1 empty line, between toplevel class defs or fn defs usually 2 empty lines.
2007-08-01T15:47:27 <ThomasWaldmann> (it can be that the pep8 checker does not warn you for every of theses cases because I disabled some stuff that gave false positives sometimes)
2007-08-01T15:49:29 <ThomasWaldmann> starpage action definitely lacks a better docstring (why is it called starpage? ART? Train for what? ...)
2007-08-01T15:50:11 <vpv> I'm seriously considering of porting my import script to use PageEditor locally, without xmlrpc. I have no idea what's going on, but in couple of my latest tests, it seems that the main moin.py and the script runner moin.py get into some sort of accept-recv deadloop
2007-08-01T15:51:00 <ThomasWaldmann> vpv: maybe test if it is multithreading related by choosing the SimpleServer instead of the ThreadPoolServer
2007-08-01T15:54:15 <vpv> I'll try that, first some coffee :)
2007-08-01T15:54:18 <lanius> ThomasWaldmann: yes it is needed
2007-08-01T15:54:26 <lanius> ThomasWaldmann: it was fixed but somehow not merged
2007-08-01T15:56:14 <ThomasWaldmann> lanius: for a normal dict .keys() hands out a copy, you hand out a reference to an internal list?
2007-08-01T15:57:42 <lanius> right
2007-08-01T15:57:50 <lanius> so i should hand out a copy
2007-08-01T15:58:31 <ThomasWaldmann> if it does no harm...
2007-08-01T15:58:50 <lanius> didn't think about that ;)
2007-08-01T15:58:52 <ThomasWaldmann> it would be safer to use I think
2007-08-01T16:03:27 <neagulm> ThomasWaldmann, I will fix now the PEP8 issue. The starpage action was introduced to allow testing the ART classifier: report a page to the classifier. It will be removed.
2007-08-01T16:04:43 <ThomasWaldmann> ah
2007-08-01T16:09:31 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2741:411124fac6e1 1.7-storage-hwendel/MoinMoin/ (4 files in 3 dirs): cleanups
2007-08-01T16:10:38 <lanius> ThomasWaldmann: i have the list_items method which takes a dictionary of filters
2007-08-01T16:10:53 <lanius> would you expect only the items returns which match all filters (and) or which match one of the filters (or)
2007-08-01T16:11:09 <ThomasWaldmann> and
2007-08-01T16:11:30 <lanius> ok
2007-08-01T16:11:32 <ThomasWaldmann> (you can simulate or by doing multiple calls and add the results)
2007-08-01T16:12:06 <lanius> is there a super special nice way in python to do and AND from two lists?
2007-08-01T16:12:36 * ThomasWaldmann once tried to find stuff in a online shop that implemented only "or" search. It sucked. X)
2007-08-01T16:13:08 <ThomasWaldmann> this is better for sets
2007-08-01T16:13:25 <lanius> sets?
2007-08-01T16:13:34 <lanius> ah nice
2007-08-01T16:13:40 <ThomasWaldmann> yeah, Python has sets :)
2007-08-01T16:13:52 <lanius> never knew :D
2007-08-01T16:13:57 <ThomasWaldmann> and the usual set operations
2007-08-01T16:14:27 <ThomasWaldmann> if you use it, you have to do some import magic like seen when grepping for Set in the src
2007-08-01T16:15:43 <vpv> I think I may have fixed my problem by setting the variable that contains the multicall object to "None" after each loop iteration
2007-08-01T16:16:42 <lanius> ThomasWaldmann: why do i need this magic?
2007-08-01T16:17:08 <ThomasWaldmann> because py 2.3 didn't have it built-in
2007-08-01T16:17:53 <lanius> so if i tell you another passage which is not py 2.3 compatible i don't have to do it
2007-08-01T16:18:48 <ThomasWaldmann> then we rather have to fix that other thing
2007-08-01T16:19:35 <lanius> do we really wan't to rely on 2.3?
2007-08-01T16:20:20 <ThomasWaldmann> quite some people still use it
2007-08-01T16:20:48 <ThomasWaldmann> what's the problem?
2007-08-01T16:21:01 <lanius> i hate doing silly stuff for compatibility reasons :D
2007-08-01T16:21:56 <ThomasWaldmann> 2.3 is not THAT much different
2007-08-01T16:22:14 <ThomasWaldmann> (we won't support py 1.5 :)
2007-08-01T16:23:24 <lanius> what about generator expressions?
2007-08-01T16:25:09 <lanius> reversed() and sorted()
2007-08-01T16:26:04 <lanius> and the simpler string substitution with %(
2007-08-01T16:26:09 <lanius> i don't think we are 2.3 compatible anymore ;)
2007-08-01T16:26:21 <grzywacz> We are. I use 2.3 for testing.
2007-08-01T16:26:34 <lanius> strange that i find sorted used all over the place
2007-08-01T16:27:05 <grzywacz> orly?
2007-08-01T16:27:06 <lanius> ok, i find the interesting hack to support it ;)
2007-08-01T16:27:23 <lanius> but i'd like to use generator expressions
2007-08-01T16:27:50 <grzywacz> Sorry. :)
2007-08-01T16:28:10 <grzywacz> And no, sorted is not used. At least my eclipse doesn't find it anywhere in Moin's sources.
2007-08-01T16:28:24 <lanius> mine does 4 times
2007-08-01T16:28:37 <lanius> two times in support/__ini__.py to support it
2007-08-01T16:28:54 <lanius> and once in each of AdvancedSearch and my code :D
2007-08-01T16:29:04 <grzywacz> I see no sorted in support ;-p
2007-08-01T16:29:21 <lanius> i see
2007-08-01T16:39:05 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2742:9403f1aa09e0 1.7-storage-hwendel/MoinMoin/ (13 files in 11 dirs): make set stuff python 2.3 compatible and move it to support
2007-08-01T16:42:47 * neagulm is going offline...
2007-08-01T16:42:56 <neagulm> seeya
2007-08-01T16:50:32 <ThomasWaldmann> lanius: it doesn't work that way (moving set stuff to support)
2007-08-01T16:51:55 <ThomasWaldmann> (or rather: I guess it will ONLY work for 2.3 then :)
2007-08-01T16:52:43 <ThomasWaldmann> so please revert and just do it like the others for now
2007-08-01T16:59:53 <lanius> ThomasWaldmann: why should it not work?
2007-08-01T17:23:15 <ThomasWaldmann> because set is in builtins, not in support (for > 2.3)
2007-08-01T17:23:36 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2743:fed41f92d150 1.7-storage-hwendel/MoinMoin/ (storage/backends.py support/__init__.py): small fix to set stuff
2007-08-01T17:23:37 <lanius> now it works
2007-08-01T17:24:09 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2747:466e5e97002c 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: big package cache file handling updates
2007-08-01T17:26:05 <lanius> ThomasWaldmann: but i get other unit test failures with python 2.3
2007-08-01T18:40:35 <ThomasWaldmann> lanius: weird X) We could also implant it into builtins module, see config.py for similar hacks
2007-08-01T18:41:31 <lanius> config.py?
2007-08-01T18:52:13 <ThomasWaldmann> ehrm, config/__init__.py
2007-08-01T18:53:59 <lanius> ThomasWaldmann: i don't see real magic there?
2007-08-01T18:54:16 * ThomasWaldmann looks
2007-08-01T18:54:47 <ThomasWaldmann> lanius: the logging_fix
2007-08-01T18:55:14 <ThomasWaldmann> (it is done there because config is imported very early before too much happens)
2007-08-01T18:55:29 <lanius> ah there, and how can we alter the builtins?
2007-08-01T18:58:31 <lanius> ah, nice
2007-08-01T18:58:36 <lanius> yeah, i would prefer that indeed
2007-08-01T19:12:41 <lanius> sorry for not being able to execute "hg commit --exclude" the correct way
2007-08-01T19:12:49 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2744:bf710305d105 1.7-storage-hwendel/MoinMoin/ (conftest.py support/__init__.py): more 2.3 compatibility
2007-08-01T19:12:52 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2745:c116cc4d2a14 1.7-storage-hwendel/MoinMoin/ (18 files in 12 dirs): make 2.3 compatibility work like magic
2007-08-01T19:12:54 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2746:74587ba21614 1.7-storage-hwendel/MoinMoin/support/python23.py: make 2.3 compatibility work like magic
2007-08-01T19:13:19 <lanius> the storage stuff should not be in that commit
2007-08-01T19:15:46 <ThomasWaldmann> (the 2.3 stuff should not even be done in your repo :)
2007-08-01T19:17:12 * ThomasWaldmann leaves to LUG meeting, maybe bbl
2007-08-01T20:08:46 <xorAxAx> alter the builtins?
2007-08-01T20:08:47 <xorAxAx> UGH
2007-08-01T20:09:01 <lanius> why not
2007-08-01T20:09:15 <xorAxAx> because moin might run with n other apps in the same process
2007-08-01T20:09:40 <xorAxAx> so please dont patch builtins
2007-08-01T20:09:48 <xorAxAx> just use import
2007-08-01T20:09:57 <lanius> should be no problem
2007-08-01T20:10:08 <xorAxAx> what?
2007-08-01T20:10:26 <lanius> moin with n other apps in the same process and patching builtins
2007-08-01T20:10:34 <lanius> but how can it run in the same process with something else at all
2007-08-01T20:11:58 <lanius> ah, process != thread, ok, but if somebody is doing that, he is doing something really silly, isn't he?
2007-08-01T20:12:22 <xorAxAx> lanius: not at all, no
2007-08-01T20:12:30 <lanius> so why should he do that
2007-08-01T20:12:30 <xorAxAx> lanius: mod_py works like that, twisted as well etc. pp.
2007-08-01T20:12:37 <xorAxAx> to save memory?
2007-08-01T20:12:46 <lanius> so still no problem
2007-08-01T20:12:50 <xorAxAx> thats a perfectly fine setup
2007-08-01T20:12:58 <xorAxAx> whats your exact plan?
2007-08-01T20:13:04 <xorAxAx> putting Set as set into built-ins?
2007-08-01T20:13:10 <lanius> right
2007-08-01T20:13:10 <xorAxAx> that will break a lot of applications
2007-08-01T20:13:19 <lanius> why should it
2007-08-01T20:13:21 <xorAxAx> because Set has an incompatible API compared to set
2007-08-01T20:13:30 <lanius> it is only done if set is not available
2007-08-01T20:13:58 <lanius> and when set is not available other applications should not have used it at all
2007-08-01T20:14:21 <xorAxAx> thats wrong
2007-08-01T20:14:26 <lanius> why
2007-08-01T20:14:43 <xorAxAx> if its available, the application expects it to behave like "set" and not like "Set"
2007-08-01T20:14:57 <xorAxAx> so if you make it available, it sees it as being available
2007-08-01T20:15:01 <xorAxAx> so it will fail
2007-08-01T20:15:36 <lanius> but the application does not expect to be set available if running python 2.3
2007-08-01T20:16:17 <xorAxAx> ???
2007-08-01T20:16:21 <xorAxAx> of course it does
2007-08-01T20:16:30 <xorAxAx> if its available
2007-08-01T20:16:38 <xorAxAx> lanius: in python, you dont do version checks
2007-08-01T20:16:42 <xorAxAx> you just check if names are defined
2007-08-01T20:18:27 <xorAxAx> so if you name it like set, it must work like se
2007-08-01T20:18:28 <xorAxAx> t
2007-08-01T20:20:27 <lanius> well this issue may appear in 0.00001% of all cases
2007-08-01T20:22:48 <xorAxAx> ARGH
2007-08-01T20:30:57 <grzywacz> lol
2007-08-01T20:36:32 <xorAxAx> lanius: so i would be very pleased if you could fix your stuff
2007-08-01T20:36:52 <xorAxAx> also sorted should fail fast if there are unknown kwargs
2007-08-01T21:23:24 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2747:beecab2bd338 1.7-storage-hwendel/MoinMoin/storage/ (7 files in 2 dirs): fix up python magic
2007-08-01T21:28:24 <xorAxAx> lanius: why the __implements__?
2007-08-01T21:29:03 <xorAxAx> lanius: and what should TestRevision.test do now? seems a bit pointless
2007-08-01T21:29:18 <lanius> to see that it is an interface, it does nothing, but some editor know it
2007-08-01T21:29:23 <lanius> it does a testForSuccess
2007-08-01T21:29:41 <xorAxAx> no, here it does nothing besides 4 getattr ops
2007-08-01T21:29:46 <xorAxAx> def test(self):
2007-08-01T21:29:46 <xorAxAx> - assert isinstance(self.revision.data, DataBackend)
2007-08-01T21:29:46 <xorAxAx> - assert isinstance(self.revision.metadata, MetadataBackend)
2007-08-01T21:29:46 <xorAxAx> + self.revision.data
2007-08-01T21:29:47 <xorAxAx> + self.revision.metadata
2007-08-01T21:29:55 <xorAxAx> lanius: "some editor"?
2007-08-01T21:29:56 <lanius> yes, getattr == testForSuccess
2007-08-01T21:30:01 <lanius> xorAxAx: eclipse e.g.
2007-08-01T21:30:17 <xorAxAx> so you added it to assist for code completion in pydev?
2007-08-01T21:30:37 <lanius> well, python does not have interfaces
2007-08-01T21:30:43 <lanius> it's just for documentation purpose
2007-08-01T21:30:48 <xorAxAx> python has protocols
2007-08-01T21:31:01 <lanius> protocols?
2007-08-01T21:31:03 <xorAxAx> and zope.interfaces :)
2007-08-01T21:31:12 <xorAxAx> yes, protocols is a concept ...
2007-08-01T21:31:19 * xorAxAx searches a link
2007-08-01T21:31:25 <lanius> i tried to make it look like zope.interfaces
2007-08-01T21:32:26 <xorAxAx> types and classes implement various methods
2007-08-01T21:32:38 <xorAxAx> and to see which protocol is implemented, you check for the attribute or simply try it
2007-08-01T21:48:20 <lanius> have to check it tomorrow
2007-08-01T21:49:12 <xorAxAx> hmm, what?
2007-08-01T21:49:30 <xorAxAx> that was just a general explanation of the term
2007-08-01T21:49:46 <xorAxAx> nothing that can be specifically implemented :)
2007-08-01T21:57:03 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2748:895d24dd1742 1.7-storage-hwendel/MoinMoin/storage/external.py: do more python magic
2007-08-01T22:22:14 <xorAxAx> lanius: + def raiseException(self, size=None):
2007-08-01T22:22:19 <xorAxAx> lanius: PEP8 violation
2007-08-01T22:22:58 <lanius> indeed
MoinMoin: MoinMoinChat/Logs/moin-dev/2007-08-01 (last edited 2007-10-29 19:20:42 by localhost)