MoinMoin: MoinMoinChat/Logs/moin-dev/2009-02-01

2009-02-01T09:08:44  <ronny> ThomasWaldmann: ever tooks a look at re.Scanner? its fscking awsome for tokenizers
2009-02-01T09:10:43  <ronny> TheSheep: this also goes to you
2009-02-01T11:59:00  <ThomasWaldmann> ronny: undocumented feature?
2009-02-01T12:04:14  <ronny> ThomasWaldmann: yeah
2009-02-01T12:04:51  <ronny> see http://bazaar.launchpad.net/~ronny-pfannschmidt/vellum/parser/annotate/head%3A/vellum/parser.py for a simple example use
2009-02-01T12:18:40  <ThomasWaldmann> ronny: nice :) but is that feature stable?
2009-02-01T12:20:00  <ronny> ThomasWaldmann: it seems they just didnt document it
2009-02-01T12:44:11  <TheSheep> Please try again
2009-02-01T12:44:13  <TheSheep> Sorry, there was a problem connecting to the Launchpad server.
2009-02-01T12:44:18  <TheSheep> O.o
2009-02-01T12:44:47  <TheSheep> ronny: I remember something like that when I was looking through re.py
2009-02-01T13:11:35  <ronny> TheSheep: its a pretty cool tool for tokenizers
2009-02-01T13:11:50  <ronny> TheSheep: unfortunately it cant limit the matching to subsets based on context
2009-02-01T13:13:28  <ronny> oh wait
2009-02-01T13:13:32  <ronny> oh GREAT
2009-02-01T13:13:48  <ronny> it stops on the first character it cant match
2009-02-01T13:14:06  <ronny> i think i have an idea on how to make nesting now
2009-02-01T13:46:04  <ronny> TheSheep: do you think this could be used to enhance the parser for moin?
2009-02-01T16:27:30  <ThomasWaldmann> ronny: you know we have some new parsers in the dom repo?
2009-02-01T16:27:55  <ThomasWaldmann> hi ernestas
2009-02-01T16:28:29  <ernestas> hi Thomas
2009-02-01T16:29:02  <ernestas> no ghopc this year?
2009-02-01T16:29:36  <ThomasWaldmann> long time not seen. no, not 2008/2009, but likely 2009/2010 from what I've heard.
2009-02-01T16:30:31  <ThomasWaldmann> but help is always welcome here :)
2009-02-01T16:33:35  <ronny> ThomasWaldmann: nope
2009-02-01T16:33:55  <ronny> ops, the "do" is actually missing"
2009-02-01T16:36:23  <ThomasWaldmann> ronny: waldi has rewritten some stuff to use a dom tree
2009-02-01T16:36:44  <ronny> hmm
2009-02-01T16:37:06  <ronny> why dom?
2009-02-01T16:37:07  <ThomasWaldmann> that's one of the things we also still need to merge
2009-02-01T16:38:06  <ThomasWaldmann> because you have trouble doing some stuff with the current approach
2009-02-01T16:38:27  <ThomasWaldmann> e.g. nested lists, includes, toc, etc.
2009-02-01T16:39:41  <ThomasWaldmann> ronny: if you like to write a parser, the dom repo still needs one for rst and docbook iirc
2009-02-01T16:41:05  <ronny> ThomasWaldmann: i wonder if it makes sense to grab docutils and wreck a bit havoc with it
2009-02-01T16:41:45  <ThomasWaldmann> i think we use docutils
2009-02-01T16:42:07  <ronny> ThomasWaldmann: i mean with the whole codebase of it
2009-02-01T16:42:15  <ronny> im in a refactoring mood
2009-02-01T16:42:41  <ronny> mostly related to my exam tomorow
2009-02-01T16:42:47  <ronny> interpreting law
2009-02-01T16:43:31  <ThomasWaldmann> haha
2009-02-01T16:45:07  <ronny> yeah
2009-02-01T16:47:47  * ThomasWaldmann knows that. Having a boring exam soon and having to learn for it somehow triggers all sorts of good coding ideas.
2009-02-01T16:53:40  <ronny> yeah
2009-02-01T16:53:54  <ronny> i rewrote vellums parser already and fixed tonns of things in pida
2009-02-01T16:54:12  <ronny> hmm
2009-02-01T16:55:02  <ronny> all coders should have shortyly impending boring exams for like 4 weeks - we'd get a new age of computing ^^
2009-02-01T16:55:42  <ThomasWaldmann> moin 2.1 in 2 months :))
2009-02-01T17:15:09  <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4266:38110c49d0a6 1.8/MoinMoin/search/Xapian.py: Xapian indexing: in update mode, do not try to re-index old revisions again
2009-02-01T17:26:08  <ernestas> If I want to update something (PO file, for example), what wiki master should I edit, 1.8 or 1.9?
2009-02-01T17:30:47  <ThomasWaldmann> both :)
2009-02-01T17:31:22  <ThomasWaldmann> master18 will get into next 1.8 release(s), master19 into 1.9 later
2009-02-01T17:32:09  <ThomasWaldmann> ernestas: for the po file, you could try updating master18 first and then replace the master19 page with the same content
2009-02-01T17:32:46  <ThomasWaldmann> (likely there will be some mismatches, but that'll get sorted out when I run the update script for 1.9)
2009-02-01T17:39:17  <ThomasWaldmann> TheSheep: could you have a look at the trail css (again)? http://master18.moinmo.in/FrontPage looks strange (when logged in and a trail is displayed).
2009-02-01T17:40:23  <ThomasWaldmann> iirc you fixed it some time ago for the case when there is no logo, maybe that caused this effect?
2009-02-01T18:14:58  <dreimark> moin
2009-02-01T18:15:01  <dreimark> hi ernestas
2009-02-01T18:19:41  <ernestas> moin dreimark
2009-02-01T18:31:58  <dreimark> ernestas: do you have trained yourself a bit by python coding?
2009-02-01T18:45:10  <ernestas> sorry, have to go
2009-02-01T18:45:31  <ronny> ThomasWaldmann: can there be a history clean up for the 2.0 series
2009-02-01T18:45:43  <ronny> i really hate having around 70 mb repos
2009-02-01T18:49:23  <ThomasWaldmann> yeah, that's really a problem in the age of XXX(X) GB disks :P
2009-02-01T18:52:45  <ThomasWaldmann> we could do it, the question is rather what it'll break if we do
2009-02-01T19:00:46  <ronny> hmm, it would be less pressing if the stuff could be managed using nested repos + shallow clones later
2009-02-01T19:04:45  <ThomasWaldmann> why do you consider that as an improvement? at first, it looks like making things more complicated.
2009-02-01T19:09:48  <ronny> mainly inconviences in remote cloning and the massive underlay updates
2009-02-01T19:11:07  <dreimark> ronny: if you look at the suid problem I think it is related that in web.session FileSessionService.delete is not implemented but destroy_session
2009-02-01T19:11:54  <dreimark> ronny: which mercurial version do ypu use. I think 1.0.2 is much faster
2009-02-01T19:14:03  <ronny> dreimark: yeah, thats probably the main cause
2009-02-01T19:14:09  <ronny> dreimark: and im on hg crew
2009-02-01T19:14:24  <ronny> but hg cant fix __large__ stuff passing around
2009-02-01T19:14:27  <dreimark> ok, fine  I didn't know
2009-02-01T19:14:59  <dreimark> ronny: I think we will need another solution for underlay
2009-02-01T19:17:37  <dreimark> e.g. I think we can split underlay other than english pagepacks from the dev tree.
2009-02-01T19:18:02  <dreimark> and merge them at release time.
2009-02-01T19:19:17  <ronny> hmm
2009-02-01T19:19:40  <ronny> im not exactly a friend of underlay anyway - i prefer to start empty
2009-02-01T19:20:00  <dreimark> may be we can use a wiki with a hg storage repo later on for all the translations. so we can merge and don't have to edit twice
2009-02-01T19:20:55  <dreimark> ronny: it can't be completly empty at least Systempages aka RecentChanges must live there
2009-02-01T19:21:22  <dreimark> I think after the storage api is available we can have several solutions
2009-02-01T20:02:25  * ThomasWaldmann plays with hg convert
2009-02-01T20:03:22  <starshine> interesting point he raises, why should any one language have to be the first underlay?
2009-02-01T20:07:29  <ThomasWaldmann> because it doesn't work without :)
2009-02-01T20:07:48  <CIA-38> Reimar Bauer <rb.proj AT googlemail DOT com> default * 4521:64e3ae35a5eb 1.9/MoinMoin/web/serving.py: web.serving: Method should have 'self' as first argument
2009-02-01T20:07:49  <CIA-38> Reimar Bauer <rb.proj AT googlemail DOT com> default * 4522:a5ecf8f3303e 1.9/MoinMoin/web/serving.py: web.serving: typo in __call__ fixed
2009-02-01T20:08:45  <starshine> hm, what if someone wants no help then.  maybe they dont want it to look like a wiki.  It still needs underlay?
2009-02-01T20:09:02  <dreimark> starshine: the SystemPages must live in underlay not any help page
2009-02-01T20:09:23  <dreimark> currently we don't have a separation between sys pages and others
2009-02-01T20:09:43  <ThomasWaldmann> there might be a time in the far distant future when moin will be able to run without system pages, but currently we rather have the problem to get 1.9 into releaseable state
2009-02-01T20:14:06  <ThomasWaldmann> dreimark: web.frontend: syntactically pass statement added < is that due to some pylint warning? a class works pretty well just with a docstring.
2009-02-01T20:15:10  <dreimark> yes it was
2009-02-01T20:15:23  <dreimark> ok, that was not necessary
2009-02-01T20:19:32  <ronny> ThomasWaldmann: whats the plan for replacing the system tables?
2009-02-01T20:20:06  <ronny> special views like in mediawiki?
2009-02-01T20:21:23  <dreimark> system tables?
2009-02-01T20:24:31  <dreimark> web.utils.check_forbidded is broken too
2009-02-01T20:24:45  <dreimark> s/ed/en/
2009-02-01T20:26:54  <ThomasWaldmann> ronny: we have lots of code to fix and to merge, thus it is currently a bad time for bigger new features
2009-02-01T20:27:25  <starshine> I was merely cheaking on the nature of the requirement, any redesign this is definitely not the time yet.
2009-02-01T20:27:50  <starshine> gah, I can't make a straight sentence this morning.  I think I need coffee. :)
2009-02-01T20:28:13  <ThomasWaldmann> :)
2009-02-01T20:32:34  <CIA-38> Reimar Bauer <rb.proj AT googlemail DOT com> default * 4523:12ccfa96b1e4 1.9/MoinMoin/web/utils.py: web.utils: typo in parameter kick fixed
2009-02-01T20:32:35  <CIA-38> Reimar Bauer <rb.proj AT googlemail DOT com> default * 4524:3e660579be91 1.9/MoinMoin/web/utils.py: web.utils: missing import of caching added
2009-02-01T20:34:32  <ronny> ThomasWaldmann: after 2.0?
2009-02-01T20:44:59  <ThomasWaldmann> you can save lots of work if you do such stuff after the dom repo is merged
2009-02-01T20:54:39  <ThomasWaldmann> ronny: hg convert with some excludes is getting the repo down to 32MB. But it seems to have problems with the tags.
2009-02-01T21:01:40  <ThomasWaldmann> .hgtags@d8d0e2b317ee, line 33: tag '1.8.0' refers to unknown node
2009-02-01T21:01:40  <ThomasWaldmann> .hgtags@d8d0e2b317ee, line 34: tag '1.8.1' refers to unknown node
2009-02-01T21:01:53  <ronny> what hg version?
2009-02-01T21:02:00  <ThomasWaldmann> otoh, the .hgtags file later looks valid
2009-02-01T21:02:15  <ThomasWaldmann> 1.1
2009-02-01T21:02:20  <ronny> tags are kinda tricky
2009-02-01T21:02:29  <ronny> since they life in the history they tag
2009-02-01T21:56:38  <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4525:b29b90a71625 1.9/MoinMoin/web/ (flup_frontend.py frontend.py):
2009-02-01T21:56:38  <CIA-38> revert (wrong) frontend related changesets (see below for details)
2009-02-01T21:56:38  <CIA-38> Changesets reverted:
2009-02-01T21:56:38  <CIA-38> 7f17aad0ab8987f4a04f65f9a281303190d541c4 web.frontend: fixed FrontEnd.run for CGI
2009-02-01T21:56:38  <CIA-38> 74e850a410ecf4e92d3f37fdf7ca369d7c9d94bb web.flup_frontend: fixed missing definition of server_types.
2009-02-01T21:56:42  <CIA-38> c6fa175c7b62c51de10877f2875ebf8e47cfd9d5 web.frontend: fixed undefined run_server
2009-02-01T22:01:01  * ThomasWaldmann would've liked to use hg backout, but creating 3 backout changesets + 3 merge changesets for this is just too much.
2009-02-01T22:42:41  <CIA-38> Reimar Bauer <rb.proj AT googlemail DOT com> default * 4526:1d06e13ce1ee 1.9/MoinMoin/web/frontend.py: web.frontend: the flup CGIRequest uses sys.exit(0) to terminate so we should not run into an exception by that.
2009-02-01T22:45:10  <dreimark> MoinMoin/formatter/_tests/test_formatter.py[156] FsF.
2009-02-01T23:01:35  <CIA-38> Ronny Pfannschmidt <Ronny.Pfannschmidt@gmx.de> default * 4527:b29674295262 1.9/MoinMoin/auth/_tests/test_auth.py: fixed normal auth tests
2009-02-01T23:01:36  <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4528:699fed126481 1.9/MoinMoin/auth/_tests/test_auth.py: merged http://ronny.uberhost.de/hg/moin-hacks/