2007-06-26T00:17:17 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2104:6e6a8e14de3b 1.7-storage-hwendel/MoinMoin/ (9 files in 3 dirs): first acl code
2007-06-26T00:28:02 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2105:f7a399a4ba4e 1.7-storage-hwendel/MoinMoin/ (Page.py storage/_tests/__init__.py): small fix to acl code
2007-06-26T00:41:46 <ThomasWaldmann> lanius: - self.__deleted = None you removed this accidentally?
2007-06-26T00:44:12 <xorAxAx> + for namespace in sorted(self.backends, reverse = True):
2007-06-26T00:44:15 <xorAxAx> PEP8 alarm :)
2007-06-26T00:45:07 <ThomasWaldmann> shit, xoraxax was faster :)
2007-06-26T00:46:23 <ThomasWaldmann> lanius: that real flag on current_revision is for what?
2007-06-26T00:48:49 <ThomasWaldmann> lanius: + if verb not in metadata:
2007-06-26T00:48:49 <ThomasWaldmann> + metadata[verb] = []
2007-06-26T00:48:49 <ThomasWaldmann> + metadata[verb].append(args.strip())
2007-06-26T00:48:52 <xorAxAx> lanius: http://hg.alexanderweb.de/moin-1.7-storage-hwendel/rev/59e4a2f5f28f is incomplete
2007-06-26T00:50:07 <ThomasWaldmann> metadata[verb] = metadata.setdefault(verb, []).append(args.strip())
2007-06-26T00:50:24 <xorAxAx> he fixed the reverse already
2007-06-26T00:51:15 <ThomasWaldmann> and verb.lower() must be done at an earlier place
2007-06-26T00:52:43 <ThomasWaldmann> lanius: + # special handling for list metadata like acl's --> http://members.aol.com/apostrophs/apostrophen.htm
2007-06-26T00:54:19 <ThomasWaldmann> and are you reversing acl order at that place?
2007-06-26T01:05:04 <ThomasWaldmann> forget that setdefault stuff, it doesnt mix well with append
2007-06-26T02:03:01 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2254:564dee5eb254 1.7/MoinMoin/ (request/__init__.py wikidicts.py events/wikidictsnotify.py): use event system to trigger wikidicts cache update
2007-06-26T02:05:58 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2255:b74a4230a123 1.7/MoinMoin/events/ (wikidictsrescan.py wikidictsnotify.py): renamed wikidictsnotify > wikidictsrescan
2007-06-26T08:07:59 <ThomasWaldmann> moin
2007-06-26T08:22:39 <ThomasWaldmann> http://www.selenic.com/pipermail/mercurial/2007-June/013496.html
2007-06-26T08:37:24 <lanius> ThomasWaldmann: self.__deleted wasn't removed, just a strange diff
2007-06-26T08:38:54 <lanius> xorAxAx: that was already fixed before
2007-06-26T08:39:02 <ThomasWaldmann> ah, maybe I've overlooked it
2007-06-26T08:45:34 <lanius> xorAxAx: where is it incomplete?
2007-06-26T08:45:47 <lanius> ThomasWaldmann: the set default stuff seems to work ok
2007-06-26T08:49:43 <lanius> ThomasWaldmann: the real=True argument is needed because of the Deleted stuff, in most cases you wan't to get the last existing revisions, only sometimes you wan't to get the revision which is in current which can be real +1 if the page is deleted
2007-06-26T08:49:56 <ThomasWaldmann> lanius: the result of l.append() is None, and a modified l.
2007-06-26T08:53:12 <lanius> ThomasWaldmann: that's ok, isn't it?
2007-06-26T08:53:56 <ThomasWaldmann> what?
2007-06-26T08:54:52 <lanius> ah, i just did the following
2007-06-26T08:54:53 <lanius> metadata.setdefault(verb, []).append(args.strip())
2007-06-26T08:55:03 <lanius> without the assignement
2007-06-26T08:55:05 <lanius> which seems to work
2007-06-26T08:57:01 <ThomasWaldmann> ah, ok, right
2007-06-26T08:57:49 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2106:db2c9ce6a3fa 1.7-storage-hwendel/MoinMoin/ (4 files in 3 dirs): small fixes all around
2007-06-26T08:58:56 <ThomasWaldmann> btw, use less apostrophes :)
2007-06-26T09:06:38 <ThomasWaldmann> those unquoteWikiname() you removed were nops?
2007-06-26T09:43:16 <lanius> nops?
2007-06-26T09:43:33 <lanius> yes they were since i moved that into the storage layer
2007-06-26T09:50:56 <xorAxAx> lanius: well, you didnt move, you only deleted code
2007-06-26T10:21:36 <lanius> ah, you mean the description
2007-06-26T10:21:42 <lanius> the move was in the commit before then, ok
2007-06-26T10:40:46 <grzywacz> @_@
2007-06-26T10:40:46 <moinBot> grzywacz: Error: "_@" is not a valid command.
2007-06-26T10:40:55 <grzywacz> Welcome to Poland, where Internet Just Stops Working. :E
2007-06-26T10:41:25 <xorAxAx> :-)
2007-06-26T10:41:43 <grzywacz> The whole day of breakage...
2007-06-26T10:44:10 <grzywacz> ThomasWaldmann: I know Vilnus is not far away, but I not in a position to spend this much money for transport and a place to stay.
2007-06-26T10:44:28 <grzywacz> ThomasWaldmann: maybe next year if the con ends up being oraganised somewhere near by ;)
2007-06-26T10:45:46 <xorAxAx> well, if nobody volounteers, it wont :)
2007-06-26T10:46:17 <xorAxAx> and i dont know anybody from poland who said that he wants to help organising it
2007-06-26T10:46:25 <xorAxAx> or team members of the orga team that are from poland
2007-06-26T10:46:46 <CIA-27> moin: Karol Nowak <grzywacz@sul.uni.lodz.pl> * 2266:cbad6fb20f2d 1.7-jabber-knowak/MoinMoin/user.py: Fix a typo.
2007-06-26T10:46:50 <grzywacz> huh, it works.
2007-06-26T10:46:50 <CIA-27> moin: Karol Nowak <grzywacz@sul.uni.lodz.pl> * 2267:e6ebe6853e51 1.7-jabber-knowak/MoinMoin/_tests/test_events.py: Unit test; check if page_change_message() raises an exception.
2007-06-26T10:46:51 <CIA-27> moin: Karol Nowak <grzywacz@sul.uni.lodz.pl> * 2268:495b09143c10 1.7-jabber-knowak/MoinMoin/_tests/test_events.py: Add a new unit test for _filter_subscriber_list()
2007-06-26T10:46:52 <grzywacz> ThomasWaldmann: thanks.
2007-06-26T10:47:24 <grzywacz> xorAxAx: I'm not talking about Poland. 500km in one direction or the other fits into my definition of "near by".
2007-06-26T10:47:27 <grzywacz> ;)
2007-06-26T10:47:31 <xorAxAx> hehe :)
2007-06-26T10:47:55 * xorAxAx is gone to university
2007-06-26T10:48:06 <grzywacz> ok
2007-06-26T12:36:57 <grzywacz> hg fails on windows ;S
2007-06-26T12:37:29 <grzywacz> diff/ci behave really bad. Maybe because of spaces in path, maybe because of some different obscure reasons.
2007-06-26T12:37:44 <ThomasWaldmann> 0.9.4?
2007-06-26T12:38:47 <grzywacz> 0.9.3
2007-06-26T12:39:33 <grzywacz> oh
2007-06-26T12:40:18 <ThomasWaldmann> improved Windows support <- from the ChangeLog
2007-06-26T12:41:00 <grzywacz> Yes, I see. But there's no binary. ;)
2007-06-26T12:41:11 <grzywacz> Maybe the one including fixes through 18th will be better.
2007-06-26T12:42:42 <grzywacz> Hm, yes. This actually works.
2007-06-26T12:47:57 <ThomasWaldmann> btw, did you run the test on win32?
2007-06-26T12:48:05 <ThomasWaldmann> testS
2007-06-26T12:48:14 <grzywacz> Unit tests? Yes. All pass.
2007-06-26T12:48:45 <ThomasWaldmann> ok, great
2007-06-26T12:49:06 <ThomasWaldmann> zenhase: how is it going?
2007-06-26T12:49:39 <grzywacz> Looks like I have to take a closer look at data forms now...
2007-06-26T12:50:54 <ThomasWaldmann> or the other events?
2007-06-26T12:51:40 <xorAxAx> grzywacz: yes, have you seen the issues described by thomas?
2007-06-26T12:52:09 <xorAxAx> grzywacz: currently a page rename event is missing and a deletion fires two events
2007-06-26T12:57:00 <grzywacz> xorAxAx: rename is already commited, not pushed
2007-06-26T12:57:07 <grzywacz> xorAxAx: I know about double firing.
2007-06-26T12:57:19 <xorAxAx> how often does rename fire? 5 times? :)
2007-06-26T12:57:29 <xorAxAx> how do you plan to mitigate this issue? :-)
2007-06-26T12:57:32 <grzywacz> Let me count.
2007-06-26T12:57:39 <grzywacz> rename+creation+delete
2007-06-26T12:57:40 <grzywacz> 3?
2007-06-26T12:57:42 <grzywacz> :P
2007-06-26T12:57:51 <xorAxAx> delete is 2 so its at least 4
2007-06-26T12:58:18 <grzywacz> The simplest workaround would be calling PageEditor's, um, save?, with argument disabling the page change notification.
2007-06-26T12:58:21 <grzywacz> And so on.
2007-06-26T12:58:31 <grzywacz> If the events were queued, they could be somehow merged.
2007-06-26T12:58:33 <grzywacz> But it's not the case.
2007-06-26T12:59:37 <xorAxAx> yes, you need to mute the events
2007-06-26T12:59:48 <grzywacz> Will be done.
2007-06-26T12:59:52 <xorAxAx> but only the right ones :)
2007-06-26T12:59:58 <grzywacz> Yes.
2007-06-26T13:00:02 <xorAxAx> nice
2007-06-26T13:00:17 <grzywacz> Hm.
2007-06-26T13:00:54 <grzywacz> Either changing PageEditor, or adding a "temporarely ignore these events" option to event handling code would do the trick.
2007-06-26T13:01:03 <grzywacz> The second option somehow seems cleaner.
2007-06-26T13:01:11 <xorAxAx> i dont think so
2007-06-26T13:01:11 <grzywacz> Or maybe not.
2007-06-26T13:01:20 <grzywacz> I'd have to know what to block, fugly.
2007-06-26T13:01:31 <xorAxAx> imagine some event handler edits a page which generates a new event
2007-06-26T13:01:36 <grzywacz> Yes, I know.
2007-06-26T13:01:39 <xorAxAx> it would be muted as well in the second case
2007-06-26T13:01:45 <grzywacz> I'll change the PageEditor slightly.
2007-06-26T13:01:46 <xorAxAx> so how about adding some kwargs?
2007-06-26T13:01:50 <xorAxAx> i think it makes sense here
2007-06-26T13:01:53 <grzywacz> That's what I want to do.
2007-06-26T13:02:10 <xorAxAx> maybe the storage refactoring will touch this as well and avoid the pageeditor usage at all
2007-06-26T13:02:17 <xorAxAx> so it might become cleaner again :)
2007-06-26T13:02:28 <grzywacz> We shall see. 8)
2007-06-26T13:03:26 <xorAxAx> hmm
2007-06-26T13:03:31 <grzywacz> There's one good thing about me being here. The monitor is much larger and eclipse's outliner fits into the screen. ;)
2007-06-26T13:03:31 <xorAxAx> a global flag isnt that bad maybe
2007-06-26T13:03:49 <xorAxAx> hah :)
2007-06-26T13:03:50 <grzywacz> This makes it so much easier to navigate through the code. ;)
2007-06-26T13:04:26 <xorAxAx> you just need to toggle the flag at the right time
2007-06-26T13:04:37 <grzywacz> xorAxAx: no, this won't work, because I'd have to know what events to disable exactly.
2007-06-26T13:04:44 <grzywacz> xorAxAx: this adds dependencies
2007-06-26T13:04:50 <xorAxAx> grzywacz: but where would other events come from?
2007-06-26T13:04:58 <xorAxAx> a page deletion fires a limited set of events
2007-06-26T13:05:06 <xorAxAx> just the delete event could fire others
2007-06-26T13:05:09 <grzywacz> xorAxAx: no idea, maybe there will be some other events related to security, logging, etc. in the future
2007-06-26T13:05:12 <xorAxAx> and that one will be run with events enabled
2007-06-26T13:05:32 <xorAxAx> ok, well
2007-06-26T13:05:51 * xorAxAx is gone to get some lunch
2007-06-26T13:06:00 <grzywacz> Enjoy your meal.
2007-06-26T13:12:26 <ThomasWaldmann> $4500 should be enough for some big screens :)
2007-06-26T13:13:26 <grzywacz> Pretty much so, yes. I'll probably buy one if I get it. ;-)
2007-06-26T13:22:10 <ThomasWaldmann> 2 and a dh gfx card :D
2007-06-26T13:23:08 <grzywacz> dh?
2007-06-26T13:23:33 <ThomasWaldmann> dual head
2007-06-26T13:24:27 <ThomasWaldmann> 2400x1600 for cheap
2007-06-26T13:37:26 <grzywacz> I have a dual head one. ;)
2007-06-26T13:37:39 <grzywacz> I'm not sure what's the max supported resolution, though.:)
2007-06-26T13:38:27 <ThomasWaldmann> even my old matrox g550 managed that (but only analog)
2007-06-26T13:39:50 <xorAxAx> even my porsche does 300 km/h on the autobahn :)
2007-06-26T13:40:04 <grzywacz> O_o;
2007-06-26T13:40:47 <xorAxAx> (i dont own a car)
2007-06-26T13:40:48 <ThomasWaldmann> xorAxAx: your 5cm big Porsche only gets that fast on a 1:100 autobahn
2007-06-26T13:42:20 <xorAxAx> i was just pointing out that a matrox card isnt bad
2007-06-26T13:42:24 * ThomasWaldmann .oO(and who wants a Porsche nowadays?)
2007-06-26T13:42:32 <grzywacz> I want a new bike. ;]
2007-06-26T13:42:59 <xorAxAx> but i own a stevens bike :)
2007-06-26T13:45:37 <grzywacz> I own a broken one. ;]
2007-06-26T14:06:54 <ThomasWaldmann> if you break it, you own the pieces :)
2007-06-26T14:13:06 <grzywacz> Or one broken piece. ;)
2007-06-26T14:29:35 <dreimark> and where are the unit tests ?
2007-06-26T14:36:20 <grzywacz> MoinMoin/_tests
2007-06-26T14:36:43 <grzywacz> O_o
2007-06-26T14:39:46 <xorAxAx> they dont test your broken bike, though
2007-06-26T14:51:48 <grzywacz> It's fubar, nothing to test.
2007-06-26T14:52:21 <zenhase> re
2007-06-26T14:52:33 <grzywacz> That's what heppens to bikes that survive fsckups from railway banks.
2007-06-26T14:52:34 <grzywacz> ;]
2007-06-26T14:52:57 <zenhase> ThomasWaldmann: i have been doing callgraphing yesterday ... pycallgraph is really a slow painful crawl :/
2007-06-26T14:53:52 <xorAxAx> grzywacz: railway banks?
2007-06-26T14:54:20 <grzywacz> xorAxAx: you know, like a slope on which railway tracks are built
2007-06-26T14:54:29 <grzywacz> I forward wheel slipped on a patch of sand. :P
2007-06-26T14:54:40 <zenhase> while i was waiting for graphs here and there and trying out different filters and trying to make the graphs i bit more manageable i wrote a small tracer myself o_O
2007-06-26T14:55:08 <zenhase> it serializes to pickles files and contains a stream of call/return-events
2007-06-26T14:55:18 <zenhase> s/files//
2007-06-26T14:55:31 <grzywacz> O_o'
2007-06-26T14:56:06 <zenhase> it's actually genshi for call-tracing, you can stream-filter the saved traces afterwards and serialize them to graphviz dot and i have also experimental udrawgraph support o_O
2007-06-26T14:56:17 <grzywacz> xorAxAx: and once, on a slope covered with clay, I learnt that one can actually *strafe* on a bike xD
2007-06-26T14:56:28 <grzywacz> (it was after rain and kind of slippery :P)
2007-06-26T14:58:25 <xorAxAx> grzywacz: hehe :)
2007-06-26T14:58:43 <xorAxAx> somehow i had less problems when i started using my "good" bike
2007-06-26T14:59:03 <grzywacz> My old bike is, well, old. 8+ years now or so.
2007-06-26T14:59:55 <zenhase> i will put it up somewhere, perhaps others will find it useful too
2007-06-26T15:01:03 <zenhase> it pulls down the actual tracing down from about 10minutes on my machines to 10secs and can therefore be put directly into running development wikicode for collecting traces
2007-06-26T15:01:22 <zenhase> and you have ALL calls traced and filter afterwards
2007-06-26T15:02:31 <zenhase> hope you don't mind that i decided to write supporting stuff yesterday
2007-06-26T15:04:32 <xorAxAx> zenhase: of course not
2007-06-26T15:04:44 * zenhase packages
2007-06-26T15:04:48 <xorAxAx> as long as you dont write a keppy's "rubik cube"
2007-06-26T15:04:53 <xorAxAx> (running gag! :))
2007-06-26T15:04:55 <ffesti> I was already wondering why it was so quiet
2007-06-26T15:05:27 <xorAxAx> zenhase: does it generate a dot file? or did say genshi because it generates dynamic clickable pages?
2007-06-26T15:10:41 <grzywacz> rubik cube? ;>
2007-06-26T15:11:03 <grzywacz> I see there are more legends going around. Oo;
2007-06-26T15:13:54 <xorAxAx> grzywacz: see and marvel - http://hg.thinkmo.de/moin/1.6-storage-a/file/d0f2f482ae3f/MoinMoin/util/rubik.py
2007-06-26T15:16:38 <xorAxAx> 90% docstrings, a gpl contradictment and weird code make up this wonderful peace of software
2007-06-26T15:16:43 <grzywacz> ...
2007-06-26T15:17:01 <xorAxAx> umm, piece
2007-06-26T15:17:33 <grzywacz> My head is already spinning. ;-)
2007-06-26T15:21:04 <zenhase> xorAxAx: it generates a stream of events like explained in the python docs for settrace/setprofile
2007-06-26T15:22:05 <zenhase> they currently look like (<EVENT>, <fullyqualifiedfuncname>, (<file>, <line>))
2007-06-26T15:22:55 <xorAxAx> zenhase: how do you eval them? you still need to generate a dot file to make it usable, no?
2007-06-26T15:27:04 <zenhase> xorAxAx: seperate steps ... somewhere in my code i have something like t = Trace(); t.start_trace(); <... stuff ...>; result = t.parse(); streampickle(result, outputfile)
2007-06-26T15:27:52 <zenhase> or better streampickle(result | decee, outfile) ... decee kills all C_CALL/C_RETURN events from the trace
2007-06-26T15:27:54 <xorAxAx> zenhase: i know, i am just saying that it is of no direct use, yet :)
2007-06-26T15:28:29 <xorAxAx> and i am curious what your plan is
2007-06-26T15:29:12 <zenhase> and then on a cmdline i do ./dotfile.py -o graph.png -I 'MoinMoin.*' -I 'genshi.*' -X '*.__*__' <tracefile> for example
2007-06-26T15:30:16 <zenhase> it basically assembles a temporary dotfile with filtered trace and outputs it
2007-06-26T15:30:21 <xorAxAx> but dotfile.py doesnt exist yet, right?
2007-06-26T15:30:25 <zenhase> it does
2007-06-26T15:30:59 <xorAxAx> coool
2007-06-26T15:31:07 <xorAxAx> that wasnt clear to me
2007-06-26T15:31:49 <zenhase> http://eigensex.org/~flowtech/simpletrace-0.01beta.tar.gz
2007-06-26T15:31:53 <zenhase> there it is
2007-06-26T15:33:25 <zenhase> feel free to report bugs, but it seems to work right now
2007-06-26T15:34:01 <zenhase> oh fuck
2007-06-26T15:34:09 <zenhase> s/fuck/f*ck/
2007-06-26T15:34:50 <zenhase> bug from the start
2007-06-26T15:36:02 <xorAxAx> can you upload the resulting png somewhere? :)
2007-06-26T15:37:24 <zenhase> http://eigensex.org/~flowtech/testout.png
2007-06-26T15:37:50 <zenhase> this is from a trace with empty moin-cache
2007-06-26T15:40:40 <zenhase> a i forgot escaping " in function-names o_O
2007-06-26T15:45:05 <grzywacz> will it kill my firefox? ;]
2007-06-26T15:45:44 <zenhase> as with pycallgraph
2007-06-26T15:46:26 <grzywacz> I got a black picture. sad
2007-06-26T15:46:41 <dreimark> moin neagulm
2007-06-26T15:50:58 <zenhase> i made a quick fix for the function names in dotfile
2007-06-26T15:52:04 <zenhase> pycallgraph did not have this problem, since it threw away some calls right away (those calls in genshi-templates would not show there for example)
2007-06-26T15:52:50 <zenhase> and those (and also lambdas) can have names with " and % in them o_O
2007-06-26T15:53:39 <zenhase> well, uploaded again (without version number change)
2007-06-26T15:54:08 <zenhase> i will go look through my graphs and traces now and start working on .send_page again
2007-06-26T16:09:54 <neagulm> hello dreimark
2007-06-26T16:10:33 <neagulm> dreimark, I'm at school and i can't find a microphone anywhere ... :(
2007-06-26T16:11:33 * grzywacz playing with pyxmpp
2007-06-26T16:15:05 <CIA-27> moin: Karol Nowak <grzywacz@sul.uni.lodz.pl> * 2269:9a71ea0c941b 1.7-jabber-knowak/MoinMoin/ (4 files in 2 dirs): Add support for PageRenamed event. Not tested yet.
2007-06-26T16:15:06 <CIA-27> moin: Karol Nowak <grzywacz@sul.uni.lodz.pl> * 2270:38f4f0311a98 1.7-jabber-knowak/MoinMoin/ (PageEditor.py action/revert.py): Attempt to remove situations when unnecessary PageChangedEvent is sent.
2007-06-26T16:20:06 <dreimark> neagulm: seems something is broken for the packets of skype
2007-06-26T16:24:43 * grzywacz ponders what to break next;)
2007-06-26T17:01:40 <grzywacz> Ok, afk. Cooking dinner and stuff.
2007-06-26T17:43:49 <dreimark> moin starshine
2007-06-26T17:44:37 <dreimark> we have talked some time ago about some nice icons for the spambayes classifications
2007-06-26T17:44:56 <dreimark> did you have some in preparation ?
2007-06-26T17:50:12 <zenhase> what would a antispam-icon look like?
2007-06-26T17:54:56 <ThomasWaldmann> http://www.golem.de/0505/38197-spam.jpg
2007-06-26T17:56:23 <ThomasWaldmann> http://galerie.sthnet.com/data/media/33/spam.jpg
2007-06-26T17:59:26 <zenhase> hrhr, try to get that in 16x16px? :>
2007-06-26T18:03:38 <starshine> did I hear pixels ? :D
2007-06-26T18:03:54 <dreimark> zenhase: read about MoinMoinChat/Logs/moin-dev/2007-05-16
2007-06-26T18:04:19 <dreimark> zenhase: wrong unit ;)
2007-06-26T18:05:09 <starshine> slice of mail open with "spam can" sticking out, prohibit sign over it.
2007-06-26T18:13:25 <dreimark> hehe
2007-06-26T18:15:01 <zenhase> picture of a nigerian prince with prohibit sign, but this could be interpreted in too many political uncorrect ways
2007-06-26T18:15:25 <zenhase> a blue pill perhaps?
2007-06-26T18:20:45 <starshine> torch and pitchfork ;P
2007-06-26T18:46:19 <dreimark> bbl
2007-06-26T19:14:32 <xorAxAx> grzywacz: how are you using the description class field on the event classes?
2007-06-26T19:16:40 <xorAxAx> grzywacz: IMHO the kwarg page_edit should be renamed to send_notification
2007-06-26T19:50:11 <grzywacz> xorAxAx: the description gets displayed in the list of subscribable events in user preferences
2007-06-26T19:50:25 <grzywacz> xorAxAx: regarding the name - ok, maybe "notify" to make it shorter.
2007-06-26T19:50:28 <xorAxAx> thats bad
2007-06-26T19:50:34 <grzywacz> Why so?
2007-06-26T19:50:45 <xorAxAx> then you need to wrap it with a pseudo _() call in the module and call gettext later
2007-06-26T19:51:10 <grzywacz> Hm, right. That's my i18n todo item, I guess. ;)
2007-06-26T19:51:13 <xorAxAx> (i.e. you need to define a _ = lambda x: x)
2007-06-26T19:51:21 <grzywacz> sure
2007-06-26T19:51:41 <xorAxAx> well, i dont see the point of deferring it :)
2007-06-26T19:52:35 <xorAxAx> yeah, notify is ok
2007-06-26T19:52:50 <grzywacz> Ok, will do. Afk for now again. ;-)
2007-06-26T20:48:52 <ThomasWaldmann> re
2007-06-26T21:14:54 <neagulm> hello dreimark
2007-06-26T21:15:10 <dreimark> hi neagulm
2007-06-26T21:18:17 <dreimark> ThomasWaldmann: ping
2007-06-26T21:21:00 <dreimark> ThomasWaldmann: lanius xorAxAx ping
2007-06-26T21:21:30 <xorAxAx> pong
2007-06-26T21:23:07 <dreimark> xorAxAx: neagulm How are the plans for revisions of attachments and using filters to have extracted text from attachments
2007-06-26T21:23:19 <xorAxAx> hmm?
2007-06-26T21:24:00 <dreimark> xapian and classification needs extracted text from attachments
2007-06-26T21:25:30 <xorAxAx> yes, so?
2007-06-26T21:25:40 <dreimark> and if attachments are currently deleted it gives trouble if some of them were used for classification of spam
2007-06-26T21:25:48 <dreimark> or ham
2007-06-26T21:26:11 <xorAxAx> well, whats the question? :)
2007-06-26T21:26:21 <dreimark> so the question is is it already planned to have some kind of cache for each attachment
2007-06-26T21:26:30 <dreimark> for each revision
2007-06-26T21:26:46 <xorAxAx> cache ...
2007-06-26T21:26:51 <xorAxAx> why is that necessary?
2007-06-26T21:27:37 <dreimark> the classification could be replayed which doesn't work if something is missing
2007-06-26T21:28:01 <xorAxAx> well, its invalid to reclassify an unexistant attachment
2007-06-26T21:28:05 <xorAxAx> why should that be wanted?
2007-06-26T21:29:32 <dreimark> neagulm: ?
2007-06-26T21:31:12 <neagulm> xorAxAx, dreimark IMHO an attachment should be deleted physically when there are no links to it, when there is no reference to it. Similar to to the concept of hard-links from operating systems
2007-06-26T21:32:18 <xorAxAx> neagulm: why?
2007-06-26T21:32:20 <neagulm> xorAxAx, dreimark each page revision should link to the attachment, deleting an attachment in a specific revision should not affect other revisions
2007-06-26T21:32:29 <xorAxAx> neagulm: i have many use cases that would conflict with such behaviur
2007-06-26T21:32:34 <xorAxAx> please present your current problems
2007-06-26T21:32:40 <xorAxAx> we will find a sane solution
2007-06-26T21:33:06 <xorAxAx> neagulm: what do you mean by "other revisions"?
2007-06-26T21:33:13 <xorAxAx> does the corpus contain all revisions?
2007-06-26T21:33:51 <neagulm> xorAxAx, In the corpus I save only the page name and specific revision classified by the users
2007-06-26T21:34:35 <neagulm> xorAxAx, The problem is that when something happens to the classifier database we will not get the same behavior from the classifier because some of the attachments used are not present any more
2007-06-26T21:34:59 <xorAxAx> well, we cant avoid that
2007-06-26T21:35:10 <neagulm> xorAxAx, dreimark :)
2007-06-26T21:35:13 <xorAxAx> and changing external semantics to fix in internal problem sounds a bit weird
2007-06-26T21:35:41 <xorAxAx> if you want to keep reproducible results, you must trigger on attachment deletes
2007-06-26T21:35:52 <xorAxAx> how do you handle attachments that are linked from other pages?
2007-06-26T21:36:17 <xorAxAx> thats all a bit delicate
2007-06-26T21:36:23 <neagulm> xorAxAx, in no special way
2007-06-26T21:36:27 <xorAxAx> esp. because there is no reverse mapping available
2007-06-26T21:36:38 <xorAxAx> neagulm: well, then that would be the first design bug, no? :)
2007-06-26T21:36:48 <xorAxAx> (i have already talked about this issue once)
2007-06-26T21:39:09 <xorAxAx> or you drop this reproducibility constraint
2007-06-26T21:39:28 <xorAxAx> i have no idea if that is feasible because i am not informed about most of your stuff
2007-06-26T21:39:52 <neagulm> xorAxAx, dreimark hmm
2007-06-26T21:40:34 <neagulm> xorAxAx, dreimark For me it would be ok to drop this constraint... But I don't know if users would like it :)
2007-06-26T21:40:56 <xorAxAx> what would be the issues?
2007-06-26T21:42:53 <dreimark> * I begin to hate mercurial 0.9.1
2007-06-26T21:43:08 <xorAxAx> hmm?
2007-06-26T21:43:54 <neagulm> the users would not be able to reproduce the same db. This might be required when upgrading SB (when the upgrade affects the db schema), when the db gets corrupt (the SB docs mention that this could occur), etc
2007-06-26T21:43:59 <neagulm> brb ~5 min
2007-06-26T21:45:58 <xorAxAx> yeah, the DB would be considered stale
2007-06-26T21:46:00 <dreimark> xorAxAx: neagulm it is important to reproduce the same db or nearly the same
2007-06-26T21:47:18 <dreimark> one of the solution we thought today about was to keep for each revision of an attachment the extracted data somewhere
2007-06-26T21:48:31 <dreimark> that would be useful for xapian search too for searching in revisions
2007-06-26T21:48:35 <vpv> ThomasWaldmann or others, you pointed me at a multilang feature, any more info on that?
2007-06-26T21:49:24 <vpv> it seems to be something the Fedora localization project could also use
2007-06-26T21:50:51 <ThomasWaldmann> vpv: you mean sistersites?
2007-06-26T21:51:20 <neagulm> dreimark, xorAxAx it is important if we want to get the same behavior after retraining. If we consider that this is not really important then we can skip this requirement.
2007-06-26T21:52:09 <vpv> ThomasWaldmann: you gave me link which had a multilang feature of some sort, I can't remember more since I was busy with other stuff at the time. Maybe it is sistersites then ;)
2007-06-26T21:53:24 <neagulm> dreimark, xorAxAx obviously this needs user testing to see what is the proper behavior...
2007-06-26T21:53:31 <xorAxAx> neagulm: well, the problem is that the db is stale
2007-06-26T21:53:32 <ThomasWaldmann> vpv: http://en.test.wikiwikiweb.de/TestWiki
2007-06-26T21:53:35 <xorAxAx> neagulm: and you need to update it
2007-06-26T21:54:48 <ThomasWaldmann> vpv: this is simply a wiki farm, each wiki has the language abbreviation as interwiki name. sistersites gives the navigation to sister wikis having a page of the same(!) page name.
2007-06-26T21:56:00 <vpv> right, I found something from moinmo.in on this subject
2007-06-26T21:56:36 <ThomasWaldmann> if your page names are really the same (like man pages for commands, the commands are the same for every language), it will work.
2007-06-26T21:57:14 <ThomasWaldmann> if the pagenames are not the same (like translation arbitrary words as pagenames), it won't work like this.
2007-06-26T21:57:16 <dreimark> neagulm: the question is is it more important to have revisions of normal pages or how much important are attachments
2007-06-26T21:58:01 <dreimark> neagulm: xorAxAx if someone wants to regocnize attachment spams he needs a clear definition what is ok and what not
2007-06-26T21:58:26 <dreimark> so he will have to train again and I gues he dislikes that
2007-06-26T21:58:36 <xorAxAx> ?
2007-06-26T21:58:55 <xorAxAx> well, neagulm wants to keep a stale db
2007-06-26T21:59:06 <xorAxAx> if he thinks that the users dislike that, then he shouldnt do that
2007-06-26T21:59:13 <xorAxAx> but update his db (or better, corpus)
2007-06-26T21:59:15 <dreimark> yesterday we have had attachment spams on MM. later on the sb project would be able to reject them
2007-06-26T21:59:34 <dreimark> but if we don't have a db we have to train it sometimes again
2007-06-26T21:59:36 <vpv> ThomasWaldmann: yes, I'm starting to get it now, it won't work for Fedora wiki translations probably, but they have other ideas for that. But isn't a wikifarm a bit too much for the man/info thing too? I'd like to have it all in a single wiki instance. Or am I totally mistaken here.
2007-06-26T21:59:39 <neagulm> dreimark, depends on the way we do tokenization (if we send features from attachments or just signal their presence to the classifier)
2007-06-26T21:59:57 <xorAxAx> vpv: there is no other concrete solution - and that is the wikipedia model
2007-06-26T22:00:04 <xorAxAx> vpv: feel free to think about it
2007-06-26T22:00:33 <ThomasWaldmann> vpv: running a wikifarm with moin is not a big issue
2007-06-26T22:00:45 <neagulm> dreimark, xorAxAx Another possibility is to use a separate db for attachments, handle them separately from normal pages. This way the problem would disappear by magic
2007-06-26T22:01:09 <xorAxAx> neagulm: sounds much saner
2007-06-26T22:01:14 <dreimark> neagulm: xorAxAx :)
2007-06-26T22:01:29 <xorAxAx> neagulm: nevertheless you need to be notified when attachments are deleted
2007-06-26T22:01:45 <xorAxAx> but maybe you can ignore that until the storage stuff is done
2007-06-26T22:01:58 <xorAxAx> in that case, you can handle it using the usual events
2007-06-26T22:02:35 <vpv> ThomasWaldmann, xorAxAx: I do agree with the wikipedia model, but for my use, I originally had just namespaces in mind. So that you would have something like en/1/apropos and de/1/apropos.
2007-06-26T22:02:37 <neagulm> dreimark, xorAxAx Just warn the user that an attachment is spam and the problem is solved. We could even use the data got from the attachment classifier and feed it to the page classifier.
2007-06-26T22:02:40 <neagulm> xorAxAx, yes
2007-06-26T22:04:14 <xorAxAx> vpv: thats fine
2007-06-26T22:09:55 <dreimark> neagulm: sounds good
2007-06-26T22:14:08 <dreimark> ThomasWaldmann: xorAxAx push creates new remote branches! again if i like to push
2007-06-26T22:14:44 <dreimark> I have added _tests/parser/
2007-06-26T22:15:05 <xorAxAx> dreimark: then you didnt merge
2007-06-26T22:15:09 <xorAxAx> dreimark: try `hg heads`
2007-06-26T22:15:32 <dreimark> two heads
2007-06-26T22:16:59 <dreimark> hg update updates 1 file but hg push does not work
2007-06-26T22:19:16 <ThomasWaldmann> hg merge
2007-06-26T22:19:28 <dreimark> there is nothing to merge
2007-06-26T22:19:46 <ThomasWaldmann> (and commit afterwards)
2007-06-26T22:19:57 <dreimark> nothing changed
2007-06-26T22:20:28 <dreimark> abort: push creates new remote branches!, (did you forget to merge? use push -f to force)
2007-06-26T22:20:53 <xorAxAx> oh well
2007-06-26T22:20:59 <xorAxAx> merge first before attempting to push
2007-06-26T22:21:05 <xorAxAx> its explained in the wiki
2007-06-26T22:21:15 <dreimark> yeah but there was nothing to merge
2007-06-26T22:21:27 <xorAxAx> yeah, then you didnt read the docs :)
2007-06-26T22:21:39 <xorAxAx> you can only merge if your workdir is in a revision that is a merge point
2007-06-26T22:21:41 <dreimark> xorAxAx: it is the third time now and i read the docs
2007-06-26T22:21:45 <xorAxAx> a potential merge point
2007-06-26T22:21:46 <ThomasWaldmann> did you pull?
2007-06-26T22:21:50 <xorAxAx> dreimark: even the mercurial wiki?
2007-06-26T22:22:40 <dreimark> yeah on both and I did a new clone too and i recognized that the directory wa not changed if I pulled my old version
2007-06-26T22:23:35 <xorAxAx> well, only merging helps here
2007-06-26T22:23:37 <xorAxAx> nothing else :)
2007-06-26T22:24:29 <dreimark> and if it doesnt work ?
2007-06-26T22:24:52 <xorAxAx> 22:21:39 < xorAxAx> you can only merge if your workdir is in a revision that is a merge point
2007-06-26T22:24:57 <xorAxAx> then you didnt obey that rule :)
2007-06-26T22:25:37 <xorAxAx> you need to hg up to the other head first
2007-06-26T22:25:42 <xorAxAx> i.e. hg up REV
2007-06-26T22:27:45 <xorAxAx> then you should be able to merge
2007-06-26T22:28:00 <xorAxAx> (before doing these things, make sure that you dont have any changes in the workdir
2007-06-26T22:29:04 <dreimark> unfortunately this doesnt work
2007-06-26T22:31:21 <xorAxAx> what does it say?
2007-06-26T22:32:01 <dreimark> 0 files updated, 0 files merged, 0 files removed, 0 files unresolved
2007-06-26T22:32:20 <dreimark> and hg merge tells abort: there is nothing to merge, just use 'hg update' or look at 'hg heads'
2007-06-26T22:32:38 <dreimark> hg heads tells two heads
2007-06-26T22:33:06 <dreimark> and update
2007-06-26T22:33:11 <dreimark> 0 files updated, 0 files merged, 0 files removed, 0 files unresolved
2007-06-26T22:36:33 <dreimark> 0.9.2 has this feature added merge now follows renames and copies, and 0.9.3 * fix a merge copy/rename corner case so I guess my version is broken
2007-06-26T22:37:05 <dreimark> so next weekend the system here gots a large update
2007-06-26T22:39:33 <xorAxAx> dreimark: no, your version is not broken
2007-06-26T22:39:43 <xorAxAx> dreimark: please use this pastebin: paste.pocoo.org
2007-06-26T22:39:49 <xorAxAx> please paste the output of hg heads
2007-06-26T22:42:16 <dreimark> http://paste.pocoo.org/show/1632
2007-06-26T22:42:40 <xorAxAx> now you have to enter:
2007-06-26T22:42:46 <xorAxAx> hg up b74a4230a123
2007-06-26T22:43:06 <xorAxAx> then hg merge
2007-06-26T22:43:25 <dreimark> abort: there is nothing to merge, just use 'hg update' or look at 'hg heads'
2007-06-26T22:43:39 <xorAxAx> hg merge 709eea9354c7
2007-06-26T22:43:53 <dreimark> abort: there is nothing to merge, just use 'hg update' or look at 'hg heads'
2007-06-26T22:44:17 <xorAxAx> does it work, dreimark?
2007-06-26T22:44:32 <dreimark> no
2007-06-26T22:44:36 <dreimark> (did you forget to merge? use push -f to force)
2007-06-26T22:44:57 <dreimark> no changes in heads
2007-06-26T22:45:42 <xorAxAx> you should not push ...
2007-06-26T22:46:46 <xorAxAx> dreimark: well, create a bundle and i can try the merge
2007-06-26T22:46:59 <dreimark> yeah I don't, good idea
2007-06-26T22:47:22 <xorAxAx> hg bundle foofile remoterepoofTom
2007-06-26T22:52:47 <dreimark> xorAxAx: http://moinmoin.wikiwikiweb.de/ReimarBauer?action=AttachFile&do=get&target=reimars_changes_to_the_parser
2007-06-26T22:53:39 <xorAxAx> is that 1.6?
2007-06-26T22:53:43 <dreimark> 1.7
2007-06-26T22:54:07 <xorAxAx> didnt you want to fix it in 1.6?
2007-06-26T22:55:02 <dreimark> yeah but tests for 1.7 and i can later copy it to 1.6, ThomasWaldmann wanted to be sure I don't break something else
2007-06-26T22:55:27 <xorAxAx> ok
2007-06-26T22:56:13 <xorAxAx> abort: unknown parent 0131f0aabb18!
2007-06-26T22:56:22 <xorAxAx> how did you call bundle?
2007-06-26T22:56:40 <dreimark> hg bundle reimars_changes_to_the_parser http://hg.moinmo.in/moin/1.7
2007-06-26T22:57:04 <xorAxAx> hmm, my mercurial mirrors are broken, ok
2007-06-26T22:57:39 <xorAxAx> dreimark: i just had to do merge
2007-06-26T22:57:46 <xorAxAx> dreimark: maybe you didnt commit your merge?
2007-06-26T22:57:51 <xorAxAx> anyway, i will do so now
2007-06-26T22:58:08 <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2256:691b6a650d94 1.7/MoinMoin/_tests/parser/test_text_moin_wiki.py:
2007-06-26T22:58:08 <CIA-27> moin: test_text_moin_wiki: moved to _tests/parser, renamed from test_parser_text_moin_wiki, added test for colorized
2007-06-26T22:58:08 <CIA-27> moin: parser python and nesting {{{ }}}
2007-06-26T22:58:10 <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2257:6327a199c666 1.7/MoinMoin/parser/text_moin_wiki.py:
2007-06-26T22:58:12 <CIA-27> moin: text_moin_wiki:fixed bug for Colorized parsers and nesting {{{ }}} (FeatureRequests/ParsersCanNotContain{{{}}})
2007-06-26T22:58:15 <CIA-27> moin: , should be added to 1.6 too
2007-06-26T22:58:17 <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2258:9c66c14f6b62 1.7/MoinMoin/ (2 files in 2 dirs): text_moin_wiki: completing nesting for colorized parsers and some tests
2007-06-26T22:58:18 <dreimark> fine I got, nothing changed
2007-06-26T22:58:20 <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2259:709eea9354c7 1.7/MoinMoin/parser/text_moin_wiki.py: text_moin_wiki: some small PEP8 changes, that version should go into 1.6 too
2007-06-26T22:58:23 <CIA-27> moin: Alexander Schremmer <alex AT alexanderweb DOT de> * 2260:c9a4fbbb2b2e 1.7/MoinMoin/ (2 files in 2 dirs): Merged main.
2007-06-26T22:58:33 <xorAxAx> dreimark: nothing changed when doing commit?
2007-06-26T22:58:38 <xorAxAx> ok, that sounds weird
2007-06-26T22:59:13 <dreimark> nothing
2007-06-26T22:59:43 <xorAxAx> but your hg is old
2007-06-26T22:59:46 <xorAxAx> maybe its a bu
2007-06-26T22:59:47 <xorAxAx> g
2007-06-26T23:00:24 <dreimark> i guess its a bug too and lets see after an update if it occures again
2007-06-26T23:03:50 <ThomasWaldmann> http://www.heise.de/open/artikel/91563 (german)
2007-06-26T23:05:31 <ThomasWaldmann> 0.9.4 changelog looks nice anyway
2007-06-26T23:07:09 <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2261:7e04a576e8c7 1.7/MoinMoin/_tests/parser/test_text_moin_wiki.py: test_text_moin_wiki: small PEP8 changes
2007-06-26T23:08:41 <xorAxAx> why is there a dysfuntional comments link in the edit bar in 1.6?
2007-06-26T23:10:54 <ThomasWaldmann> where?
2007-06-26T23:11:08 <dreimark> xorAxAx: http://test.wikiwikiweb.de/CommentsAsCollapseableDivs you mean that link to toggle comments?
2007-06-26T23:12:18 <xorAxAx> yes, why isnt it called "toggle comments"?
2007-06-26T23:12:23 <xorAxAx> like this, its pretty confusing
2007-06-26T23:12:36 <ThomasWaldmann> screen space
2007-06-26T23:12:36 <xorAxAx> because nothing happens when clicking it
2007-06-26T23:12:42 <xorAxAx> very bad reason
2007-06-26T23:13:06 <xorAxAx> to reduce usability just to same 2 em
2007-06-26T23:13:11 <xorAxAx> s/same/safe/
2007-06-26T23:14:01 <ThomasWaldmann> it also jumps less
2007-06-26T23:14:36 <xorAxAx> well, toggle would stay toggle
2007-06-26T23:15:31 <ThomasWaldmann> yeah, but it doesnt help much
2007-06-26T23:15:44 <ThomasWaldmann> on/off would help more, but jumps
2007-06-26T23:15:54 <xorAxAx> i dont think so
2007-06-26T23:16:03 <dreimark> show and hide
2007-06-26T23:16:13 <xorAxAx> and you can avoid jumps by forcing ot have a fixed width
2007-06-26T23:16:27 <xorAxAx> that is as large as the longest words
2007-06-26T23:16:27 <dreimark> show comment / hide comment toggled
2007-06-26T23:16:48 <ThomasWaldmann> in german "Kommentare anzeigen"
2007-06-26T23:17:28 <ThomasWaldmann> rather long
2007-06-26T23:18:03 <xorAxAx> oh well
2007-06-26T23:20:59 <ThomasWaldmann> fixed width, how would that work?
2007-06-26T23:21:12 <ThomasWaldmann> (with i18n and unknown client fonts?)
2007-06-26T23:21:36 <xorAxAx> i could imgine a few solutions
2007-06-26T23:22:23 <xorAxAx> hmm
2007-06-26T23:22:30 <xorAxAx> you could render both texts and capture the length
2007-06-26T23:22:37 <xorAxAx> and then resize the span to the length
2007-06-26T23:42:04 <dreimark> gn all
2007-06-26T23:42:13 <ThomasWaldmann> gn dreimark MoinMoin: MoinMoinChat/Logs/moin-dev/2007-06-26 (last edited 2021-02-16 00:17:01 by Bind)
