2008-07-23T00:14:09 <dennda> urgh
2008-07-23T00:20:41 <xorAxAx> good night
2008-07-23T00:21:52 <dennda> n8 xorAxAx
2008-07-23T00:22:20 <dennda> johill: ping
2008-07-23T00:24:54 <johill> dennda: what? :)
2008-07-23T00:24:58 <johill> just got back from a breakout
2008-07-23T00:26:11 <dennda> oh great, how late is it over there? :)
2008-07-23T00:26:41 <johill> 6 hours difference
2008-07-23T00:27:01 <dennda> johill: I am bad at calculating such things in my head ;)
2008-07-23T00:27:18 <dennda> Got a minute?
2008-07-23T00:27:22 <johill> no
2008-07-23T00:27:43 <dennda> Ok :)
2008-07-23T00:28:03 <dennda> Just trying to get rid of _write_file and get it to work
2008-07-23T00:28:14 <dennda> Wondered how to extract the revno from the webform as you suggested
2008-07-23T00:28:58 <dennda> But no problem, I can use a workaround for now
2008-07-23T00:37:50 <johill> well when you pull up the editor you store the item.get_reivsion(-1).revno into the form
2008-07-23T00:38:08 <johill> then when the user saves, you grab that from request.form and use it for create_revision(revno+1)
2008-07-23T00:38:19 <johill> that way, you automatically fail if somebody else saved in the meantime
2008-07-23T00:45:29 <dennda> Ah I see
2008-07-23T00:45:42 <dennda> So I need to put the revno there myself
2008-07-23T00:45:46 <dennda> need to figure how that works, though
2008-07-23T00:46:06 <johill> <input type="hidden" name="revno" value="7">
2008-07-23T00:46:45 <dennda> Well, yeah. I mean in the code
2008-07-23T00:47:00 <dennda> You don't want me to parse the HTML file, do you? :)
2008-07-23T00:47:53 <johill> on to get it out again?
2008-07-23T00:48:35 <johill> well to put it in you probably use MoinMoin.widget.html.INPUT?
2008-07-23T00:49:22 <PawelPacana> gn
2008-07-23T00:49:56 <dennda> I'll take a look at it
2008-07-23T00:50:13 <dennda> for now I'll just use a workaround to get it running
2008-07-23T00:51:33 <johill> and the browser parses the html file
2008-07-23T00:51:41 <johill> and when the user submits, the info shows up in request.form[]
2008-07-23T00:59:07 <johill> dinner now, then sleep
2008-07-23T00:59:21 <dennda> That early?
2008-07-23T00:59:34 <dennda> enjoy your meal
2008-07-23T01:17:50 *** irc.freenode.net sets mode: +o ChanServ
2008-07-23T03:29:32 * dennda emits a broadcast ping
2008-07-23T03:29:35 <dennda> anybody still alive?
2008-07-23T03:29:36 <dennda> :)
2008-07-23T03:30:15 <dennda> Looking for some hidden cache that moin creates. After editing a page that page is still shown even after I removed the underlay folder (and created it again) and my data/ folder
2008-07-23T03:36:34 <dennda> Ok it stored something to /tmp
2008-07-23T03:43:08 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station PERIOD com> default * 4349:d06a2f54dc88 1.8-storage-cdenter/MoinMoin/ (PageEditor.py storage/backends/memory.py): storage: PageEditor.py: Removing _write_file and starting to remove the concept of underlay to be a seperate thing. memory.py: Fixing a bug in list_revisions if there are none yet
2008-07-23T04:35:09 <dennda> hm ok I got a bit stuck fixing this obvious bug
2008-07-23T04:35:30 <dennda> Will try again tomorrow and request help otherwise :)
2008-07-23T04:35:49 <dennda> I hope I don't sleep too long
2008-07-23T05:25:10 <johill> dennda: you told it to store to /tmp ;)
2008-07-23T06:51:04 <CIA-53> Byeongweon [tasyblue@gmail.com] default * 3905:07d73a42f555 1.8-guieditor-mbyeongweon/wiki/htdocs/applets/moinFCKplugins/moinlink/fck_link.js: fix editing a link target does not work problem
2008-07-23T08:40:45 <CIA-53> Byeongweon [tasyblue@gmail.com] default * 3906:76dfe53612a0 1.8-guieditor-mbyeongweon/ (2 files in 2 dirs): fix manipulaion of interwiki on gui editor problem
2008-07-23T08:53:12 <ThomasWaldmann> byegonweon: if you fix a bug, please add the changeset url to the bug report
2008-07-23T08:53:35 <byegonweon> ThomasWaldmann: ok :)
2008-07-23T08:54:53 <byegonweon> ThomasWaldmann: such like this http://moinmo.in/MoinMoinBugs/GuiEditorInterwikiLinksMiss%22wiki%3A%22prefix#preview ?
2008-07-23T08:56:19 <ThomasWaldmann> yes
2008-07-23T08:56:45 <ThomasWaldmann> byegonweon: why do you do lstrip('/')? what does it mean?
2008-07-23T08:59:59 <byegonweon> ThomasWaldmann: when people insert '/' front of url remove that. to protect by people mistake.
2008-07-23T09:00:24 <byegonweon> ThomasWaldmann: I just find other reason.
2008-07-23T09:00:54 <byegonweon> ThomasWaldmann: that code help upper 2 line of code useless
2008-07-23T09:00:56 <byegonweon> if href == "/": # we used this as replacement for empty href
2008-07-23T09:00:56 <byegonweon> href = ""
2008-07-23T09:00:59 <byegonweon> this one
2008-07-23T09:02:25 <CIA-53> Byeongweon [tasyblue@gmail.com] default * 3907:0b1d6e252e1f 1.8-guieditor-mbyeongweon/MoinMoin/converter/text_html_text_moin_wiki.py: remove useless code. href.lstrip('/') did same work
2008-07-23T09:02:57 <ThomasWaldmann> the question is rather if it is always '/' that you need to strip, if if it could be also '/scriptname/'
2008-07-23T09:04:58 <ThomasWaldmann> also, i think you did not fix THAT bug on the wiki page - it was related to "wiki:" prefix that moin 1.8 does not use anymore anyway (but maybe some other interwiki bug)
2008-07-23T09:15:06 <byegonweon> ThomasWaldmann: oh. I misunderstood what ReimarBauer say about interwiki.
2008-07-23T09:18:09 <byegonweon> ThomasWaldmann: could you tell me more detail about '/scriptname/'
2008-07-23T09:21:22 <ThomasWaldmann> wiki page urls can look like http://seattlewireless.net/moin.cgi/RecentChanges
2008-07-23T09:22:13 <ThomasWaldmann> (== not just beginning with / like it is in the standalone server)
2008-07-23T09:23:19 <ThomasWaldmann> at least REAL page urls. not sure what it exactly does for interwiki when feeding the gui editor / in the gui editor
2008-07-23T09:24:40 <byegonweon> mosquito mosquito !! it just like jungle :-(
2008-07-23T09:25:09 <TheSheep> badger badger badger
2008-07-23T09:28:19 <byegonweon> ThomasWaldmann: ok I understand what you say. and I found another negative effects of removing '/'
2008-07-23T09:28:57 <byegonweon> this is real jungle. I kill 2-3 mosquito every 5min :-(
2008-07-23T09:32:47 <ThomasWaldmann> byegonweon: there are electric fly flaps - very effective >:)
2008-07-23T09:35:02 <byegonweon> ThomasWaldmann: my clap evolve everyday. finally I have weapon clap.
2008-07-23T09:35:14 <dreimark> moin
2008-07-23T09:35:21 <byegonweon> moin
2008-07-23T09:35:29 <dreimark> TheSheep: mushroom
2008-07-23T09:37:54 <CIA-53> Byeongweon [tasyblue@gmail.com] default * 3908:c2e6b2b00277 1.8-guieditor-mbyeongweon/MoinMoin/converter/text_html_text_moin_wiki.py: back to previous code because of keep pagename
2008-07-23T09:44:00 <ThomasWaldmann> waldi: 4 days no commit
2008-07-23T09:57:12 <ThomasWaldmann> waldi: zenhase: byegonweon: dennda: and dreimark: TheSheep: (students not here yet): meeting today at 11:00 UTC (like every week)
2008-07-23T09:58:05 <byegonweon> ThomasWaldmann: roger
2008-07-23T09:58:56 <TheSheep> ThomasWaldmann: Pawel is traveling again at this time, he's not sure whether he'll have connectivity on the train, he'll be home at 15:00 CET...
2008-07-23T10:00:26 <ThomasWaldmann> timeline says August 11 is suggested pencils down date (to have a week improving tests, docs, etc), August 18 is the final deadline
2008-07-23T10:01:30 <ThomasWaldmann> that means students have about 2.5 - 3.5 weeks left and should make some plans about how to spend them on their project
2008-07-23T10:06:25 * ThomasWaldmann will also have an appointment at 12:00 UTC, so I can't be there at the end of the meeting.
2008-07-23T10:07:13 * xorAxAx is here
2008-07-23T10:08:47 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3997:6b187ac286ec 1.8-dom-bblank/MoinMoin/macro2/ (__init__.py _base.py _tests/test__base.py): Add new macro interface infrastructure
2008-07-23T10:08:48 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3998:9d2211f1c35b 1.8-dom-bblank/MoinMoin/converter2/macro.py: Macro converter - Support new macro interface
2008-07-23T10:08:49 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3999:97b159da8087 1.8-dom-bblank/MoinMoin/ (macro2/PageCount.py macro/PageCount.py): Migrate PageCount macro to new interface
2008-07-23T10:09:02 <ThomasWaldmann> moin waldi :)
2008-07-23T10:09:27 <waldi> yeah
2008-07-23T10:09:31 * ThomasWaldmann looks
2008-07-23T10:36:01 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4000:8df37b5156f3 1.8-dom-bblank/MoinMoin/converter2/ (_tests/test_html_out.py html_out.py): HTML output converter - Support list types
2008-07-23T10:36:02 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4001:7c7edd166d06 1.8-dom-bblank/MoinMoin/converter2/html_out.py: HTML output converter - Fix namespace in list support
2008-07-23T10:36:04 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4002:e7dc26cf18d0 1.8-dom-bblank/MoinMoin/converter2/ (_tests/test_creole_in.py creole_in.py): Creole input converter - Support list types
2008-07-23T10:50:45 <PawelPacana> moin
2008-07-23T10:56:23 <ThomasWaldmann> hi Pawel
2008-07-23T11:03:01 <TheSheep> PawelPacana: 68 failed <-- that's corrct count, or did I set up mercurial wrong?
2008-07-23T11:20:55 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4003:10e794345c14 1.8-dom-bblank/MoinMoin/ (converter2/macro.py macro2/_base.py): Macros - Add page name to interface
2008-07-23T11:20:56 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4004:646c7c36b605 1.8-dom-bblank/MoinMoin/ (macro/__init__.py macro2/GoTo.py): Migrate GoTo macro to new interface
2008-07-23T11:20:59 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4005:8d19011515ba 1.8-dom-bblank/MoinMoin/converter2/macro.py: Macro handling - Remove debugging code
2008-07-23T11:24:44 * dreimark is here
2008-07-23T11:25:15 <PawelPacana> TheSheep: wrong hg version
2008-07-23T11:25:32 <TheSheep> PawelPacana: that's hg clone from their repo...
2008-07-23T11:25:39 <PawelPacana> for now 4/105 should fail
2008-07-23T11:25:46 <TheSheep> or maybe it picked up a wrong one...
2008-07-23T11:26:08 <PawelPacana> TheSheep: maybe you experience problem i had few days before..
2008-07-23T11:26:23 <PawelPacana> try importing memctx from mercurial.context
2008-07-23T11:26:42 <TheSheep> PawelPacana: with 1.0.1 I get 101 failed
2008-07-23T11:27:31 <TheSheep> >>> from mercurial.context import memctx
2008-07-23T11:27:31 <TheSheep> >>> memctx
2008-07-23T11:27:31 <TheSheep> <class 'mercurial.context.memctx'>
2008-07-23T11:28:28 <TheSheep> ah, I didn't apply your patch
2008-07-23T11:33:47 <TheSheep> works with newer py.test :)
2008-07-23T11:38:36 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4006:35517c304e14 1.8-dom-bblank/MoinMoin/converter2/ (_tests/test_creole_in.py _wiki_macro.py macro.py): Macro handling - Rename macro-type to macro-context
2008-07-23T11:38:36 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4007:ee5bfba9a424 1.8-dom-bblank/MoinMoin/converter2/html_out.py: HTML output converter - Support attributes without namespace
2008-07-23T11:39:20 <dennda> johill: Yeah I realized
2008-07-23T11:40:54 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4360:c0ca7e4dc6b4 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Copyright up-to-date. Docstring clarification.
2008-07-23T11:40:54 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4361:db338fca1a05 1.8-mercurialbackend-ppacana/MoinMoin/storage/_tests/test_backends_hg.py: (test_backends_hg) Typo fixed.
2008-07-23T11:42:14 <dennda> Jesus, my irssi layout freaked up again
2008-07-23T11:42:19 * dennda restarts
2008-07-23T11:47:30 <gizmach> moin
2008-07-23T11:47:49 <xorAxAx> moin melita
2008-07-23T11:48:01 <gizmach> hi alexander
2008-07-23T12:00:56 <gizmach> ThomasWaldmann: in LdapTestBase are defined setup_class and teardown_class and in test_groupldap I need to define also setup_class and teardown_class and a setup_class needs to use parms from LDAPTestBase setup_class. Have any ideas how to do it?
2008-07-23T12:11:51 <zenhase> moin
2008-07-23T12:12:01 <gizmach> zenhase: hi
2008-07-23T12:12:21 * zenhase grabs some lunch quickly
2008-07-23T12:12:35 <byeongweon> ThomasWaldmann: I should go somewhere now. I will be back soon bug about 10 minutes late this meeting I think.
2008-07-23T12:12:41 <zenhase> see you in 45minutes
2008-07-23T12:14:59 <ThomasWaldmann> byeongweon: ok, but don't miss it
2008-07-23T12:16:16 <dennda> I could need some help figuring out a bug in a few minutes
2008-07-23T12:16:41 <dennda> A bit harder than I thought due to my missing knowledge of parts of moin
2008-07-23T12:17:40 <dennda> For starters: If a Page is viewed in the browser, the Page-class is used, and the ultimate method that sends the data to the browser from there is the sendText method, right?
2008-07-23T12:22:32 <ThomasWaldmann> dennda: where do you have sendText?
2008-07-23T12:22:51 <dennda> send_page() I mean, sorry. Just got up
2008-07-23T12:23:13 <ThomasWaldmann> yeah, send_page does almost everything :|
2008-07-23T12:23:39 <dennda> ok, I think my problem is there, but I am not sure
2008-07-23T12:23:39 <ThomasWaldmann> see the docstring
2008-07-23T12:23:45 <dennda> Can you pull my latest changeset and run it?
2008-07-23T12:24:54 <dennda> You will notice that when saving a Page, the text is shown correctly. But when you click any other Page (like RecentChanges or something) the same text is shown, independant from the current page.
2008-07-23T12:25:00 * ThomasWaldmann clones
2008-07-23T12:25:48 <dennda> http://paste.pocoo.org/show/80136/ wikiconfig_local.py
2008-07-23T12:29:19 <ThomasWaldmann> (takes a bit, storage is under high load)
2008-07-23T12:32:14 <ThomasWaldmann> dennda: it complains about underlay
2008-07-23T12:33:38 * ThomasWaldmann untars underlay and wonders how it is used (it is not configured in wikiconfig)
2008-07-23T12:33:38 <dennda> ThomasWaldmann: Well, untar it
2008-07-23T12:34:13 <dennda> Hm, dunno, thought that worked automatically?
2008-07-23T12:34:27 <dennda> But that will go in my branch anyway...
2008-07-23T12:34:37 <xorAxAx> no, its only done for the tests
2008-07-23T12:34:55 <ThomasWaldmann> dennda: there is debug output not using the logging framework
2008-07-23T12:35:03 <ThomasWaldmann> self.arena_dir ...
2008-07-23T12:35:14 <dennda> Yes, I already removed that here
2008-07-23T12:35:32 <dennda> just ignore it
2008-07-23T12:35:36 <ThomasWaldmann> now I have UnicodeDecodeError
2008-07-23T12:35:52 <dennda> hmm, where?
2008-07-23T12:36:03 <dennda> Never faced that up to now
2008-07-23T12:37:35 <dennda> Did it load the front page when you started it?
2008-07-23T12:37:59 <ThomasWaldmann> wtf? it is showing me some file name definitely not in the memorybackend
2008-07-23T12:38:09 <xorAxAx> pyc files?
2008-07-23T12:38:19 <ThomasWaldmann> fresh clone
2008-07-23T12:38:23 <dennda> Why would they be cloned?
2008-07-23T12:38:32 <dennda> What's the traceback?
2008-07-23T12:43:16 * dennda quickly takes a shower before the meeting
2008-07-23T12:43:22 <ThomasWaldmann> dennda: it crashes due to an umlaut in a private document of me. I have no idea what it is accessing / how / why it is doing that.
2008-07-23T12:44:06 <ThomasWaldmann> (I didn't put that document into the workdir or anywhere else where your code should be accessing it)
2008-07-23T12:45:05 <ThomasWaldmann> (crash in AttachFile._get_files)
2008-07-23T12:45:58 <xorAxAx> its the phone-home feature
2008-07-23T12:46:13 <ThomasWaldmann> :P
2008-07-23T12:48:03 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4008:80ebc115292b 1.8-dom-bblank/MoinMoin/converter2/creole_in.py: Creole input converter - Partial support for parser sections
2008-07-23T12:48:19 <TheSheep> oh wow, I gotta see that
2008-07-23T12:48:26 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4009:92c1d7fd035f 1.8-dom-bblank/MoinMoin/converter2/creole_in.py: Creole input converter - Update TODO entries
2008-07-23T12:48:39 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4010:1caf57ecc13e 1.8-dom-bblank/MoinMoin/ (9 files in 4 dirs): Redifine converter interface to match the new macro interface
2008-07-23T12:48:51 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4011:2bf2fd65f214 1.8-dom-bblank/MoinMoin/converter2/creole_in.py: Creole input converter - Supply a string not a list to subsequent converter, fix parser name handling
2008-07-23T12:51:02 <ThomasWaldmann> dennda: obviously it is trying to process attachments from /tmp directory
2008-07-23T12:55:19 <ThomasWaldmann> dennda: 'StoredRevision' object has no attribute 'mtime'
2008-07-23T12:56:30 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4012:6a5b0fd63592 1.8-dom-bblank/MoinMoin/converter2/_tests/test__wiki_macro.py: Tests for pseudo macro handling - Add testcases for TableOfContents
2008-07-23T12:57:04 <zenhase> re
2008-07-23T12:58:22 <dreimark> strange one of my wikis gots lots of a user definition
2008-07-23T12:58:59 * gizmach here
2008-07-23T13:00:15 <xorAxAx> its 11:00 UTC
2008-07-23T13:00:25 <waldi> ding
2008-07-23T13:00:44 <ThomasWaldmann> ok. lets begin, topics are:
2008-07-23T13:00:44 <byeongweon> dong
2008-07-23T13:00:50 <ThomasWaldmann> student reporting about progress of that week (what's new, what works, what does not work, ...)
2008-07-23T13:00:54 <ThomasWaldmann> student reporting about what blocks him (if anything)
2008-07-23T13:00:56 * dennda resurfaces
2008-07-23T13:00:57 <ThomasWaldmann> mentor and student defining next steps, what to do for unblocking
2008-07-23T13:01:04 <ThomasWaldmann> waldi: please begin
2008-07-23T13:01:10 <dennda> (Let's talk about that bug afterwards)
2008-07-23T13:01:23 <waldi> macro interface works, two examples implemented
2008-07-23T13:01:56 <waldi> converters uses a similar interface now, maybe they will be merged sometimes
2008-07-23T13:02:29 <ThomasWaldmann> you mean parser unified with macros?
2008-07-23T13:02:32 <waldi> yes
2008-07-23T13:03:45 <waldi> what does not work: parser arguments
2008-07-23T13:04:32 <ThomasWaldmann> (that stuff is a bit messy/underdevelopped anyway, would be nice if it was cleaned up)
2008-07-23T13:04:38 <waldi> yep
2008-07-23T13:05:22 <waldi> nothing blocking. so i think its finally time to do the moin wiki parser *urgs*
2008-07-23T13:05:33 <ThomasWaldmann> :)
2008-07-23T13:05:50 <ThomasWaldmann> can you show TOC/Include is working with the creole parser?
2008-07-23T13:06:03 <waldi> ups, include is missing, TOC works fine
2008-07-23T13:06:10 <waldi> okay, include ...
2008-07-23T13:06:48 <ThomasWaldmann> (and TOC with Include ...)
2008-07-23T13:06:53 <TheSheep> fun
2008-07-23T13:06:55 <waldi> the later is no problem
2008-07-23T13:07:02 <waldi> toc is implemented in the html converter
2008-07-23T13:07:19 <waldi> and include needs to be done even before macro handling
2008-07-23T13:08:21 <TheSheep> what if you include a page that has toc on it already? :0
2008-07-23T13:08:21 <ThomasWaldmann> maybe make sure that the creole stuff completely works, including the toc/include/footnotes stuff that was a long term problem for the old infrastructure
2008-07-23T13:08:39 <waldi> TheSheep: it will be shown more then once
2008-07-23T13:08:41 <xorAxAx> TheSheep: good question, that needs to work
2008-07-23T13:08:50 <xorAxAx> waldi: well, that will generate garbage output
2008-07-23T13:08:59 <xorAxAx> a TOC should only show the headings on its page
2008-07-23T13:09:01 <ThomasWaldmann> (and after that begin turning the wiki parser upside down)
2008-07-23T13:09:03 <xorAxAx> (or included pages)
2008-07-23T13:09:09 <xorAxAx> (and this feature is actually used :))
2008-07-23T13:09:20 <xorAxAx> in the moin wiki
2008-07-23T13:09:21 <waldi> xorAxAx: hmm, well, should be no problem
2008-07-23T13:09:35 <xorAxAx> waldi: you basically need information about where the boundaries are
2008-07-23T13:09:38 <ThomasWaldmann> ok, so plan until next meeting is?
2008-07-23T13:10:24 <waldi> include, think about table headers, toc per page boundary
2008-07-23T13:10:32 <ThomasWaldmann> zenhase / mitsuhiko - your are next after waldi has finished
2008-07-23T13:10:34 <waldi> begin moin parser
2008-07-23T13:10:36 <zenhase> ok
2008-07-23T13:10:50 <waldi> xorAxAx: they are available, i need them for macro handling anyway
2008-07-23T13:11:09 <waldi> xorAxAx: the macros wants to know the pagename and this needs to be changed according to the real page
2008-07-23T13:11:09 <xorAxAx> mitsuhiko: ping
2008-07-23T13:11:15 <xorAxAx> waldi: cool
2008-07-23T13:11:15 <ThomasWaldmann> ok, thanks waldi - great to see that nice new stuff working :)
2008-07-23T13:11:19 <dennda> He's on vacation as far as I know
2008-07-23T13:11:24 <zenhase> xorAxAx: this may not work
2008-07-23T13:11:31 <xorAxAx> dennda: hmm, he didnt say that to us
2008-07-23T13:11:31 <zenhase> yes
2008-07-23T13:11:36 <zenhase> i talked with him yesterday
2008-07-23T13:11:40 <xorAxAx> ok, zenhase go ahead
2008-07-23T13:11:53 <zenhase> what was done:
2008-07-23T13:13:07 <zenhase> i integrated the werkzeug-standalone-server into moins script-area
2008-07-23T13:13:14 <xorAxAx> (these text blocks can be prepared and pasted as well, btw)
2008-07-23T13:13:51 <zenhase> thomas pointed me to that there is still some issue with config options not being found
2008-07-23T13:14:13 <zenhase> i have patched that already, but it's not submitted yet iirc
2008-07-23T13:14:23 <zenhase> then i fixed XMLRPC
2008-07-23T13:14:46 <zenhase> except request_cli related code in script everything is working quite fine now
2008-07-23T13:15:05 <zenhase> performance seems same as ever, tho i didn't try out plain old CGI
2008-07-23T13:16:06 <zenhase> for what comes next: i will start the documentation for setting up WSGIzed moin
2008-07-23T13:16:21 <zenhase> erm wait, first request_cli, then documentation
2008-07-23T13:16:22 <ThomasWaldmann> is that new standalone wsgi server similar in speed as the old standalone?
2008-07-23T13:16:59 <zenhase> well and then it's basically testing testing, writing tests and improving documentation for developers
2008-07-23T13:17:09 <zenhase> ThomasWaldmann: yes
2008-07-23T13:17:24 <zenhase> ThomasWaldmann: only thing missing is SSL-support
2008-07-23T13:18:12 <ThomasWaldmann> ok, if it is (in real world usage), we maybe could drop the other one, but keep the same commandline api
2008-07-23T13:18:13 <zenhase> oh and there might be a 2.5 dependency problem with the standalone-server right now, which i need to fix
2008-07-23T13:18:35 <zenhase> because its based on an implementation from wsgiref
2008-07-23T13:18:41 <ThomasWaldmann> > MoinMoin.support
2008-07-23T13:18:59 <ThomasWaldmann> there is still lots of stuff in request/ and server/
2008-07-23T13:19:15 <zenhase> ok, then i'll put something there
2008-07-23T13:19:25 <zenhase> yes, i didn't delete it yet
2008-07-23T13:19:59 <zenhase> when i finished with request_cli i can safely delete the code in server and request packages
2008-07-23T13:20:04 <zenhase> it's totally unused then
2008-07-23T13:20:05 <ThomasWaldmann> maybe have a final look inside before you delete it, just to make sure :))
2008-07-23T13:20:15 <zenhase> of course
2008-07-23T13:20:37 <xorAxAx> zenhase: regarding cgi it would be interesting to know if the apache path_info issues are handled by some layer in between
2008-07-23T13:20:41 <ThomasWaldmann> gizmach: dreimark: you are next after zenhase has finished
2008-07-23T13:20:45 <dreimark> ok
2008-07-23T13:20:49 <gizmach> ok
2008-07-23T13:20:49 <xorAxAx> zenhase: cf. the path info code in requestbase
2008-07-23T13:21:07 <zenhase> xorAxAx: yeah, i have seen this code in the old request
2008-07-23T13:21:10 <xorAxAx> zenhase: these are moin-unrelated workarounds that make moin working on a wide range of systems
2008-07-23T13:22:13 <zenhase> xorAxAx: but if this is really something that needs fixing this is finally something that can be plugged in as middleware before any other moincode
2008-07-23T13:22:33 <zenhase> maybe you can later explain to me, what the exact problem with apache is there
2008-07-23T13:22:49 <ThomasWaldmann> ok, thanks zenhase (we talk about more details later, including proxy stuff)
2008-07-23T13:23:00 <zenhase> well, i think right now that's all i have to say
2008-07-23T13:23:00 <zenhase> ok
2008-07-23T13:23:27 <ThomasWaldmann> gizmach: dreimark: your stage :)
2008-07-23T13:23:32 <gizmach> thx
2008-07-23T13:23:54 <gizmach> now config group backend and wiki config backend works, tested with wiki and tests written
2008-07-23T13:23:54 <dreimark> gizmach: please report
2008-07-23T13:24:17 <gizmach> only I didn't erite test for multiple backends to use (will do that)
2008-07-23T13:24:42 <gizmach> I done some ldap implementation and test for it , but this is not complete
2008-07-23T13:25:24 <gizmach> also I fixed some errors dreimark and ThomasWaldmann told me yesterday but I need also to ask some things about the ldap groups
2008-07-23T13:25:30 <gizmach> on mailing lists
2008-07-23T13:25:38 <gizmach> and write more tests for acls
2008-07-23T13:26:22 <gizmach> well the most I done is that ldap code and finaly make sure that the other two backends work
2008-07-23T13:26:49 <gizmach> I need to simplify that code a bit and that is in plan when I make the ldap groups work
2008-07-23T13:26:59 <dreimark> gizmach: do you have thought about to simplify the code by using lambda ?
2008-07-23T13:27:11 <gizmach> dreimark: exactly that
2008-07-23T13:27:53 <dreimark> what is with documentatuion and snippets ?
2008-07-23T13:28:25 * ThomasWaldmann suggests to do snippets after stuff has found its final form
2008-07-23T13:28:39 <gizmach> also I started writing ldap conf snippet but I stoped because I think I will do that after ldap grups works because it was ThomasWaldmann idea and it's good to create ldap_conn somewhere else(before)
2008-07-23T13:29:11 <gizmach> and docmentation, I started some wroting on paper in bus need to revrite it and to do that more
2008-07-23T13:29:41 <gizmach> but the first thing is to make that group backendcode completly works
2008-07-23T13:29:58 <ThomasWaldmann> so goal until next meeting is?
2008-07-23T13:30:42 <gizmach> make the ldap group code works, test it, I also planed to add more complicated ldap entries for testing it
2008-07-23T13:31:08 <gizmach> and that to think about the createing of ldap_connection (where to do that) and than change he code upon it
2008-07-23T13:31:42 <gizmach> and try to make some sane and easy readable docmentation
2008-07-23T13:31:46 <ThomasWaldmann> i think you should be rather completely finished with group stuff until next meeting, or you will run out of time
2008-07-23T13:32:02 <dreimark> gizmach: http://code.google.com/soc/2008/moin/appinfo.html?csaid=E23681282C1A268A
2008-07-23T13:32:16 <gizmach> ThomasWaldmann: ok ll give my best
2008-07-23T13:32:17 <ThomasWaldmann> (there is still lists and dicts todo)
2008-07-23T13:32:43 <gizmach> :) ok
2008-07-23T13:33:04 <gizmach> I hope I can make that work completly
2008-07-23T13:33:04 <dreimark> gizmach: we need to discuss next meeting about 2. Extend moin's groups to database
2008-07-23T13:33:07 <ThomasWaldmann> and the caching stuff is not even touched/designed
2008-07-23T13:33:16 <gizmach> dreimark: ok
2008-07-23T13:33:47 <dreimark> gizmach: I mean we have to discuss if this can be done in time
2008-07-23T13:33:55 <gizmach> ThomasWaldmann: urgh yes I will try to hurry that
2008-07-23T13:33:59 <gizmach> dreimark: ok
2008-07-23T13:34:00 <ThomasWaldmann> ok, byeongweon and xorAxAx you are next after gizmach has finished
2008-07-23T13:34:09 <byeongweon> ok
2008-07-23T13:34:18 <dreimark> gizmach: any blockers?
2008-07-23T13:34:43 <gizmach> dreimark: well just small bugs and things I need to check on ldap groups and msiling lists
2008-07-23T13:35:22 <gizmach> well, I'm finished
2008-07-23T13:35:22 <dreimark> ok, thanks gizmach
2008-07-23T13:35:26 <gizmach> :)
2008-07-23T13:35:51 <xorAxAx> byeongweon: please start :)
2008-07-23T13:35:54 <byeongweon> last week I told you make automated testcase for moinmoin plugin
2008-07-23T13:36:19 <xorAxAx> yeah
2008-07-23T13:36:25 <byeongweon> I try several way. but I couldn't find better.
2008-07-23T13:36:39 <xorAxAx> what do you mean by "better"?
2008-07-23T13:36:40 <byeongweon> so. It still on the way.
2008-07-23T13:37:06 <byeongweon> selenium test is hard to use this case
2008-07-23T13:37:16 <xorAxAx> ah, why?
2008-07-23T13:37:26 <byeongweon> jsunit couldn't cover whole thing.
2008-07-23T13:37:54 <byeongweon> fckeditor is to complecated I think.
2008-07-23T13:38:11 <byeongweon> fckeditor dev team also try their project with selenium
2008-07-23T13:38:20 <xorAxAx> yeah, and the problem with selenium is?
2008-07-23T13:39:07 <byeongweon> hard to define testcase fit to moinmion plugin.
2008-07-23T13:39:29 <xorAxAx> ok, we can talk about it later
2008-07-23T13:39:34 <byeongweon> ok.
2008-07-23T13:39:36 <byeongweon> and
2008-07-23T13:39:48 <byeongweon> I clear undefined bugs
2008-07-23T13:40:00 <byeongweon> now I have 3 undefined bug.
2008-07-23T13:40:24 <byeongweon> they will clear today.
2008-07-23T13:40:49 <byeongweon> I also have 11 open bugs.
2008-07-23T13:40:56 <xorAxAx> regarding these bugs, i wonder how "couldnt reproduce", "fixed" and "seems like fixed" differ
2008-07-23T13:41:19 <xorAxAx> how did you reproduce them?
2008-07-23T13:41:28 <xorAxAx> using a moin-1.8 from the main repo?
2008-07-23T13:41:33 <byeongweon> yes.
2008-07-23T13:41:58 <byeongweon> the resson why separated them
2008-07-23T13:42:32 <xorAxAx> ok, as i wrote it would be nice if you could put the status updates on the particular bug pages
2008-07-23T13:42:32 <byeongweon> is I will try again "seems like fixed" and "couldn't reproduce"
2008-07-23T13:42:54 <byeongweon> xorAxAx: ok I will simplify that also
2008-07-23T13:43:18 <byeongweon> I fixed few bugs this week.
2008-07-23T13:43:27 <xorAxAx> yeah, saw that
2008-07-23T13:43:41 <byeongweon> during that works I found one bug is bigger than I thought
2008-07-23T13:43:53 <xorAxAx> what do you plan to do next?
2008-07-23T13:44:36 <byeongweon> next week. I will separate all bugs into fixed and open
2008-07-23T13:44:46 <byeongweon> and fixing bugs :)
2008-07-23T13:45:07 <byeongweon> I have question.
2008-07-23T13:45:07 <byeongweon> http://moinmo.in/MoinMoinBugs/GuiEditorCreolePluginMissingError
2008-07-23T13:45:14 <xorAxAx> yeah, we can handle that later
2008-07-23T13:45:30 <xorAxAx> as blockers you already listed the testing framework. anything else?
2008-07-23T13:45:34 <byeongweon> this one is hard to fix I think.
2008-07-23T13:45:54 <byeongweon> creole converter bug will be that one :)
2008-07-23T13:46:11 <byeongweon> maybe to fix this one we make html->creole converter.
2008-07-23T13:46:31 <ThomasWaldmann> byeongweon: that convertor is not in your task's scope
2008-07-23T13:46:50 <ThomasWaldmann> but you could try making a nicer error msg or avoiding that error somehow
2008-07-23T13:46:52 <byeongweon> ThomasWaldmann: hehe
2008-07-23T13:47:02 <byeongweon> ThomasWaldmann: ok thanks.
2008-07-23T13:47:11 <byeongweon> I think I'm finished
2008-07-23T13:47:39 <dennda> If I may? :)
2008-07-23T13:47:40 <xorAxAx> then its dennda turn
2008-07-23T13:47:43 <ThomasWaldmann> ok, dennda and johill now (and bye, I have to leave - will read backlog later)
2008-07-23T13:47:58 <dennda> I don't think johill is actually around, is he?
2008-07-23T13:48:11 <dennda> Anyway, let's get started
2008-07-23T13:48:49 <dennda> The last week I mainly spent getting to know how Page viewing / editing / saving works internally in order to adjust that code to use the new storage layer.
2008-07-23T13:48:58 <xorAxAx> dennda: its for everyone of us and the logs :)
2008-07-23T13:49:16 <dennda> That isn't quite easy, but I discussed the design and the internals with a few of you, and that really helped me understand it better.
2008-07-23T13:49:56 <dennda> One failed attempt to rewrite _write_file(), which is gone now. All methods calling this will be adjusted. This is part of a little refactoring.
2008-07-23T13:50:06 <dennda> Other things I did: Bug fixing, merging, adding tests...
2008-07-23T13:50:32 <xorAxAx> what do you plan to do?
2008-07-23T13:50:36 <dennda> Comments / questions / additions / thoughts so far? :)
2008-07-23T13:50:48 <dennda> Plan until next meeting: Get Page viewing / editing / saving working, throw out underlay if possible in that time.
2008-07-23T13:51:46 <xorAxAx> note that theoretically you dont need to change anything at all in the page rednering change :)
2008-07-23T13:51:51 <xorAxAx> s/change :/chain :/
2008-07-23T13:52:05 <dennda> theoretically
2008-07-23T13:52:08 <xorAxAx> esp. because there are 2 people working on that area currently
2008-07-23T13:52:32 <xorAxAx> well, if you have to touch stuff in that area, please consult your mentor. any blockers?
2008-07-23T13:52:35 <dennda> OK I will try to keep my changes there to a bare minimum
2008-07-23T13:52:51 <dennda> Blockers: None, except those nasty bugs and of course it is more difficult to understand and adjust that code, but I guess that doesn't count as blocker. :)
2008-07-23T13:53:41 <xorAxAx> no, thats reality. ok, nice. pawel is still traveling?
2008-07-23T13:53:47 <xorAxAx> then TheSheep can talk a bit perhaps
2008-07-23T13:54:16 <TheSheep> yup
2008-07-23T13:55:03 <TheSheep> Pawel has finished the first iteration of the backend and now is working on some advanced stuff that was made possible with the recent changes in the development version of hg
2008-07-23T13:55:37 <TheSheep> he's gotten rid of the working dir -- it's not neede anymore, all commits are from meory
2008-07-23T13:55:41 <TheSheep> memory
2008-07-23T13:56:10 <xorAxAx> cool
2008-07-23T13:56:18 <TheSheep> he also found a way to store the revision metainformation efficiently in the revision additional data
2008-07-23T13:56:27 <xorAxAx> how is metadata handeled currently?
2008-07-23T13:56:31 <TheSheep> not in comments, but in a special dict designed for various helper data
2008-07-23T13:56:47 <TheSheep> it's an 'extra' dict saved with the revision
2008-07-23T13:57:00 <TheSheep> hg uses it internally to store branch names and tags
2008-07-23T13:57:10 <TheSheep> but api allows extending it
2008-07-23T13:57:17 <xorAxAx> is there fast lookup possible in it?
2008-07-23T13:57:17 <dreimark> how does it do large item commits (e.g. 1GB files)
2008-07-23T13:57:52 <dreimark> where are the limits?
2008-07-23T13:58:32 <TheSheep> dreimark: it uses mercurial's write methods directly
2008-07-23T13:58:51 <TheSheep> dreimark: the limits would be those of mercurial we didn't check them
2008-07-23T13:59:48 <xorAxAx> TheSheep: how are multiple heads or conflicts handled? is it possible to merge two branches on the command line with hg merge?
2008-07-23T13:59:53 <TheSheep> dreimark: oh, sorry, my fault, the current version that uses memory commits uses StringIO for data
2008-07-23T13:59:59 <TheSheep> dreimark: so it's limited
2008-07-23T14:00:17 <xorAxAx> well, thats not a conceptual limitation
2008-07-23T14:00:22 <TheSheep> xorAxAx: merging is done automatically but fails if there are conflicts
2008-07-23T14:00:50 <xorAxAx> so, regarding the meeting, i think we are done now, and continue with individual talks. after talking to TheSheep i want to discuss things with zenhase and byeongweon
2008-07-23T14:01:05 <byeongweon> xorAxAx: ok
2008-07-23T14:01:06 <xorAxAx> TheSheep: what kind of conflicts?
2008-07-23T14:01:21 <TheSheep> xorAxAx: modifications to the same line for example
2008-07-23T14:01:27 * dreimark has to leave for 1 ~ 2 h
2008-07-23T14:01:48 <xorAxAx> but the edit conflict detection is currently done in the editing code (and makes perfect sense there)
2008-07-23T14:02:01 <TheSheep> xorAxAx: sure
2008-07-23T14:02:25 <xorAxAx> so i still havent understood how you connect the "linear history, conflict resolving" and the "DAG, multiple possible heads" layers
2008-07-23T14:03:52 <xorAxAx> what if somebody pulls a remote branch
2008-07-23T14:03:58 <xorAxAx> does that work?
2008-07-23T14:04:00 <TheSheep> xorAxAx: we heven't gotten into interface yet but I guess we should start soon
2008-07-23T14:04:03 <xorAxAx> how would hg merge handle the dict?
2008-07-23T14:04:14 <xorAxAx> interface? thats all the storage interface, no?
2008-07-23T14:04:39 <TheSheep> xorAxAx: no, what you're saing about conflict resoultion at higher layers is about higher layers
2008-07-23T14:05:26 <TheSheep> xorAxAx: we need to override that and the info action somehow
2008-07-23T14:05:45 <TheSheep> xorAxAx: also decide which head to display
2008-07-23T14:05:55 <xorAxAx> TheSheep: so you plan to break the whole storage abstraction by introducing DAG history?
2008-07-23T14:06:17 <xorAxAx> thats harder than it sounds, esp. because of the breakage
2008-07-23T14:06:27 <TheSheep> xorAxAx: no, normal info action will work by showing a flattened view
2008-07-23T14:06:42 <xorAxAx> ah, you want to offer a dag-like view through different channels
2008-07-23T14:06:54 <TheSheep> xorAxAx: yes, an additional api
2008-07-23T14:07:00 <xorAxAx> so people may not use "hg merge" at all on moin repos?
2008-07-23T14:07:10 <xorAxAx> because hg would corrupt the metadata?
2008-07-23T14:07:20 <xorAxAx> (i guess that needs to be tested)
2008-07-23T14:07:48 <TheSheep> xorAxAx: no, you can also do commits by hand, and they don't contain metadata either
2008-07-23T14:07:58 <TheSheep> xorAxAx: hg merge with commit is such a manual commit
2008-07-23T14:08:35 <xorAxAx> TheSheep: and the backend handles that? cool
2008-07-23T14:08:47 <xorAxAx> sounds like a neat plan in the end
2008-07-23T14:08:55 <TheSheep> xorAxAx: this is not yet implemented but we wanted to fall back to the metadata of one of the parents, with some exceptions like author, data and comment that can be taken from normal commit data
2008-07-23T14:09:03 <xorAxAx> nowadays hg even has an official js-driven dag renderer
2008-07-23T14:09:04 <TheSheep> xorAxAx: it's not implemented yet
2008-07-23T14:09:08 <xorAxAx> that you could use for the info action
2008-07-23T14:09:12 <TheSheep> xorAxAx: yes
2008-07-23T14:09:23 <xorAxAx> a lot to do for 3 weeks ;-)
2008-07-23T14:09:49 <TheSheep> yes
2008-07-23T14:09:49 <xorAxAx> ok. zenhase, i checked the old code regarding the httpd compatiblity
2008-07-23T14:10:17 <xorAxAx> zenhase: apache 1.3 had an issue where i would suggest to ignore it (because there arent many old apaches left)
2008-07-23T14:10:23 <xorAxAx> zenhase: the issue in apache 2 was fixed
2008-07-23T14:10:53 <xorAxAx> zenhase: only IIS seems to remain - but i suggest to wait until some iis user comes up with a middleware that fixes that
2008-07-23T14:11:16 <xorAxAx> zenhase: generally speaking, path_info may be "truncated" because the system may do path translations according to the rfc :-(
2008-07-23T14:11:27 <xorAxAx> zenhase: what is the wsgi backend using to dispatch requests?
2008-07-23T14:11:42 <xorAxAx> path_info?
2008-07-23T14:12:51 <xorAxAx> (also there had been these path_info issues in webservers where there was no script_name defined in the config and the server doesnt know what to derive as PI)
2008-07-23T14:12:54 <TheSheep> wsgi has standarized path_info and script_path
2008-07-23T14:13:05 <TheSheep> script_name sorry
2008-07-23T14:13:28 <xorAxAx> "The remainder of the request URL's "path", designating the virtual "location" of the request's target within the application." - ok, that disallows path mangling
2008-07-23T14:13:52 <xorAxAx> so interface wise all issues are in the cgi wsgi adapter
2008-07-23T14:13:57 <zenhase> argl, my wireless is a bit unstable right now
2008-07-23T14:14:12 <zenhase> yes
2008-07-23T14:14:57 <xorAxAx> ok, so lets ignore it for now
2008-07-23T14:15:23 <zenhase> dispatch is done as always, get the action -> decide if it's xmlrpc or not -> get the page from path_info -> handle action
2008-07-23T14:15:38 <zenhase> i think it's ok to do so
2008-07-23T14:15:39 <xorAxAx> byeongweon: regarding, http://moinmo.in/MoinMoinBugs/GuiEditorCreolePluginMissingError -- thats a bug in the edit action. it has to handle different default parsers gracefully and disallow the gui editor
2008-07-23T14:15:53 <xorAxAx> byeongweon: you can add that check if you like (shouldnt be too hard)
2008-07-23T14:16:11 <TheSheep> zenhase: I'm sorry I didn't follow your project, there will be no url redesign after all?
2008-07-23T14:16:45 <byeongweon> xorAxAx: yes. I consider it
2008-07-23T14:16:55 <zenhase> TheSheep: there isn't anything wrong with the urls, is there?
2008-07-23T14:18:29 <TheSheep> zenhase: tehy are not very restful are they?
2008-07-23T14:18:32 <xorAxAx> byeongweon: so regarding the unittesting framework selenium - where is the exact issue regarding complexity - cant you figure out how to write the test at all or doesnt it work?
2008-07-23T14:18:43 <byeongweon> xorAxAx: is it should check before display gui editor link?
2008-07-23T14:18:59 <zenhase> TheSheep: but what i have done is to replace this stupid ' "%s/%s' % (script_name, quoteWikiNameUrl(....)) stuff
2008-07-23T14:19:04 <xorAxAx> byeongweon: yes, there should be some logic that decides whether the gui link should be shown
2008-07-23T14:19:13 <byeongweon> xorAxAx: ok
2008-07-23T14:19:19 <TheSheep> zenhase: with Href()?
2008-07-23T14:19:24 <zenhase> TheSheep: the urls are very restful for most of moin, every url references a page e.g. resource
2008-07-23T14:19:32 <byeongweon> xorAxAx: to test selenium.
2008-07-23T14:19:43 <zenhase> TheSheep: but the lower workings of moin isn't restful ;)
2008-07-23T14:19:53 <zenhase> TheSheep: yes, with href
2008-07-23T14:20:03 <zenhase> every request-object has an href and abs_href attribute
2008-07-23T14:20:06 <TheSheep> zenhase: shouldn't things like page info, attachemnts, ect. be sub-resources?
2008-07-23T14:20:16 <zenhase> href is initialized with script_name and abs_href with the full script-url including host
2008-07-23T14:20:28 <byeongweon> xorAxAx: what we test about moinmoin plugin is
2008-07-23T14:20:35 <xorAxAx> byeongweon: hmm, the selenium site has been down for a few days. fckeditor is using selenium?
2008-07-23T14:20:54 <zenhase> TheSheep: well, this will interfer with subpages
2008-07-23T14:21:09 <byeongweon> xorAxAx: they try to use selenium 5 month ago
2008-07-23T14:21:13 <TheSheep> zenhase: that's why a careful design is needed :)
2008-07-23T14:21:20 <xorAxAx> byeongweon: ah. what are they using now?
2008-07-23T14:21:24 <gizmach> gtg (programming)
2008-07-23T14:21:36 <byeongweon> xorAxAx: but that test code doesn't contain fckeditor main branch.
2008-07-23T14:21:48 <TheSheep> zenhase: for example I work around it by %-encoding slashes in page names in urls
2008-07-23T14:22:13 <byeongweon> xorAxAx: they banch that codes and
2008-07-23T14:22:22 <byeongweon> xorAxAx: leave it along about 5 month.
2008-07-23T14:22:37 <byeongweon> http://dev.fckeditor.net/browser/FCKtest
2008-07-23T14:22:38 <zenhase> TheSheep: this is somehow ugly imo :o
2008-07-23T14:22:46 <TheSheep> zenhase: and by having the action name first, and page last so only the first slash when no action specified is provblematic
2008-07-23T14:22:47 <zenhase> having different slashes in the url
2008-07-23T14:23:23 <TheSheep> zenhase: it makes sense from the semantic point of view
2008-07-23T14:23:26 <byeongweon> it's hard to test such a thing like fckeditor I think.
2008-07-23T14:23:26 <zenhase> hmm
2008-07-23T14:23:56 <TheSheep> byeongweon: on the EuroPython there was that guy showing a unittest framework for javascript...
2008-07-23T14:24:07 <TheSheep> byeongweon: xson or so
2008-07-23T14:24:27 <byeongweon> TheSheep: what framework?
2008-07-23T14:24:39 <xorAxAx> byeongweon: yeah, seems orphaned
2008-07-23T14:24:44 <zenhase> TheSheep: for a restful interface having action-names being part of the URL (meaning part of the slashy part) doesn't really make sense, does it?
2008-07-23T14:24:57 <xorAxAx> byeongweon: ok, then we concentrate on bug fixing for next week
2008-07-23T14:25:15 <TheSheep> zenhase: it does, what different 'actions' (or rather views) show are in fact different resources
2008-07-23T14:25:35 <TheSheep> zenhase: the term 'action' is not restful itself :0
2008-07-23T14:25:41 * zenhase still finds the name FCKeditor to sound quite offensive ;)
2008-07-23T14:25:58 <xorAxAx> i think there are a lot of unittest frameworks
2008-07-23T14:26:00 <xorAxAx> for js
2008-07-23T14:26:01 <byeongweon> xorAxAx: ok. I'm ready for killing them all.
2008-07-23T14:26:01 <zenhase> TheSheep: well, a more appropiate term would be 'verb'
2008-07-23T14:26:03 <TheSheep> zenhase: it's derieved from the initials of authors
2008-07-23T14:26:08 <zenhase> you want to have verbs on resources
2008-07-23T14:26:15 <TheSheep> zenhase: no, there are only 4 verrbs in REST
2008-07-23T14:26:39 <TheSheep> zenhase: instead of 'edit action' you want an 'editor resource'
2008-07-23T14:26:49 <byeongweon> xorAxAx: I should going somewhere now. see you later.
2008-07-23T14:27:00 <byeongweon> xorAxAx: is this done?
2008-07-23T14:27:22 <zenhase> TheSheep: no, you're wrong i think. i understood, that there are only 4 verbs in HTTP ... which is an example for something restful
2008-07-23T14:27:41 <TheSheep> zenhase: ok yes but we are talking about http here
2008-07-23T14:27:54 <TheSheep> zenhase: (there are actually more than 4 methods)
2008-07-23T14:28:01 <TheSheep> zenhase: so it was a simplification
2008-07-23T14:28:02 <zenhase> yeah
2008-07-23T14:28:05 <xorAxAx> byeongweon: yes
2008-07-23T14:28:20 <TheSheep> zenhase: but practically we only have 2-3 methods that are useful for us
2008-07-23T14:28:41 <TheSheep> zenhase: and that's all the verbs we can use
2008-07-23T14:28:49 <TheSheep> zenhase: the rest are nouns
2008-07-23T14:29:17 <zenhase> still i think having the editor modelled as a resource a bit ... hmm .. awkward
2008-07-23T14:29:35 <TheSheep> (well, and attributes such like search queries)
2008-07-23T14:30:44 <TheSheep> zenhase: I'm so interested in it because I did something similar recently
2008-07-23T14:30:50 <TheSheep> zenhase: http://dandelion.sheep.art.pl/url%20structure
2008-07-23T14:30:54 <zenhase> hmm ok
2008-07-23T14:31:05 <TheSheep> zenhase: (that wiki uses this strucure, you can test it)
2008-07-23T14:31:54 <xorAxAx> is that restful?
2008-07-23T14:32:03 <TheSheep> xorAxAx: not fully yet
2008-07-23T14:32:03 <zenhase> not really imo
2008-07-23T14:32:08 <xorAxAx> it forces you to duplicate the verb - get and iew
2008-07-23T14:32:10 <xorAxAx> view
2008-07-23T14:32:31 <TheSheep> xorAxAx: view is not a verb here
2008-07-23T14:32:42 <zenhase> i mean, full restful-ness would kind of REQUIRE a js-gui on top
2008-07-23T14:32:51 <TheSheep> zenhase: howso?
2008-07-23T14:33:18 <xorAxAx> TheSheep: but? :)
2008-07-23T14:33:27 <TheSheep> xorAxAx: noun
2008-07-23T14:33:33 <xorAxAx> and the others ones are?
2008-07-23T14:33:34 <TheSheep> xorAxAx: view as in "it's a nice view"
2008-07-23T14:33:45 <zenhase> TheSheep: because with the server you will only have an exchange of semantically meaningful resources in different representations
2008-07-23T14:34:02 <zenhase> but ok, you already mentioned that editor can be a resource itself
2008-07-23T14:34:12 <TheSheep> xorAxAx: default view raw view, editor, history, diff view
2008-07-23T14:34:38 <TheSheep> xorAxAx: these are separate resources
2008-07-23T14:34:58 <zenhase> well
2008-07-23T14:34:58 <zenhase> why?
2008-07-23T14:35:15 <zenhase> actually it's the same resource in different representations ;)
2008-07-23T14:35:28 <TheSheep> they can have different mime type, different caching policies
2008-07-23T14:35:47 <TheSheep> zenhase: well, revision history and page view is not
2008-07-23T14:36:02 <TheSheep> zenhase: agreed that maybe default view and raw view are just different representations
2008-07-23T14:36:18 <zenhase> for a single page, raw, html, docbook are representations, diff is something complicated, because it's actually 2 resources compared
2008-07-23T14:36:20 <TheSheep> zenhase: in the same way a diff is not just a different representation
2008-07-23T14:36:31 <zenhase> editor is an interface to change the representation
2008-07-23T14:36:35 <TheSheep> zenhase: because it's not a single revision
2008-07-23T14:36:45 <zenhase> so an editor might be modelled as a resource, ok
2008-07-23T14:36:51 <zenhase> yes
2008-07-23T14:36:56 <zenhase> diff is something else
2008-07-23T14:37:14 <TheSheep> so we have at least diff, history, editor attachements
2008-07-23T14:37:20 <zenhase> something like a metaresource
2008-07-23T14:37:42 <TheSheep> particular revisions should be probably sub-resources of the history page
2008-07-23T14:37:48 <zenhase> comparing 2 different resources (tho from the same page, different point in time)
2008-07-23T14:38:03 <zenhase> history is definitely subresource of a page
2008-07-23T14:38:03 <zenhase> yeah
2008-07-23T14:38:36 <zenhase> hmm
2008-07-23T14:38:51 <zenhase> regarding the subpage problem
2008-07-23T14:38:53 <TheSheep> even some 'action' actions, like delete page or rename page are just different kinds of 'editor resources'
2008-07-23T14:39:42 <zenhase> perhaps just say that by convention there cannot be subpages named history and the like
2008-07-23T14:39:58 <zenhase> because those are reserved subresources
2008-07-23T14:39:58 <TheSheep> that's what I started with
2008-07-23T14:40:09 <zenhase> yeah
2008-07-23T14:40:11 <TheSheep> then I moved actions to the front because of several reasons
2008-07-23T14:40:46 <TheSheep> 1. I can have special pages like RecentChanges, as another resource
2008-07-23T14:40:58 <TheSheep> (this is specific to my wiki though, moin has it as normal page)
2008-07-23T14:41:43 <TheSheep> 2. I can have a /download/style.css and from it use url(image.png) without needed to do some weird url(../image.png/download)
2008-07-23T14:42:35 <TheSheep> 3. I can set http auth for /edit action in apache
2008-07-23T14:43:30 <TheSheep> or block certain actions for certain clients or ips at the web server level
2008-07-23T14:43:44 <TheSheep> or list hem in robots.txt
2008-07-23T14:46:49 <TheSheep> actually I think that the robots.txt was what made me decide on it finally
2008-07-23T14:47:02 <TheSheep> I still need to autogenerate robots.txt
2008-07-23T14:49:36 <zenhase> why do you need robots.txt for actions?
2008-07-23T14:50:35 <zenhase> shouldn't it be sufficient to have a nofollow on those links you don't want to be indexed?
2008-07-23T14:50:35 <TheSheep> zenhase: to prevent web crawles in a standards-compliant way from indexing all revisions of all pages all editors, all diffs, etc.
2008-07-23T14:50:58 <TheSheep> zenhase: nofollow is nons-standard and is not respected by many crawlers
2008-07-23T14:50:58 <zenhase> wait a sec
2008-07-23T14:51:20 <zenhase> revisions and history are resources :)
2008-07-23T14:51:25 <TheSheep> zenhase: I also can't have pragma no-folloow and no-index on pages that are not html
2008-07-23T14:51:41 <zenhase> that's right
2008-07-23T14:52:09 <TheSheep> robots.txt solves this nicely and it works with this url scheme precisely because it was designed for restful apps
2008-07-23T14:53:29 <TheSheep> I'm not arguing that moin should use the same scheme or anything like that
2008-07-23T14:53:41 <TheSheep> I'm just suggesting that the current approach is maybe not optimal :)
2008-07-23T15:00:10 <waldi> it seems that include works :)
2008-07-23T15:00:25 <waldi> no recursion detection yet, but ...
2008-07-23T15:07:00 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4013:bc83940a3343 1.8-dom-bblank/MoinMoin/converter2/html_out.py: HTML output converter - Raise exception on unknown list type
2008-07-23T15:07:03 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4014:1c14c1d65c6b 1.8-dom-bblank/MoinMoin/Page.py: Page - Factor out input conversion into own methods
2008-07-23T15:07:05 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4015:c62a77b5d37c 1.8-dom-bblank/MoinMoin/Page.py: Page - Allow input conversion to be called without data
2008-07-23T15:07:06 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4016:2c4b17fb06e1 1.8-dom-bblank/MoinMoin/converter2/ (_tests/test__wiki_macro.py _wiki_macro.py): Pseudo-macro handling - Support Include
2008-07-23T15:07:07 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4017:442f94b01b29 1.8-dom-bblank/MoinMoin/converter2/ (_tests/test__wiki_macro.py _wiki_macro.py creole_in.py): Macro handling - Add request attribute
2008-07-23T15:07:08 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4018:a131d7b43e7c 1.8-dom-bblank/MoinMoin/util/namespaces.py: Add additional namespaces
2008-07-23T15:07:11 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4019:00f7b762cc3c 1.8-dom-bblank/MoinMoin/converter2/include.py: Add include converter
2008-07-23T15:07:14 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4020:ff1c0504504d 1.8-dom-bblank/MoinMoin/Page.py: Page - Use include converter
2008-07-23T15:28:00 <waldi> hmm, should include adjust headings? the current implementation does not
2008-07-23T15:30:00 <xorAxAx> "adjust"?
2008-07-23T15:30:05 <xorAxAx> you mean shift?
2008-07-23T15:30:28 <xorAxAx> hmm, how about adding a parameter for that?
2008-07-23T16:00:21 * dreimark has a question about forms disabling a text input field
2008-07-23T16:00:59 <dreimark> if we disable for example name of user preferences we trust the browser do disable the form field
2008-07-23T16:01:07 <dreimark> so noone could enter data?
2008-07-23T16:01:52 <dreimark> and we don't check if one is able to enter data by a broken browser?
2008-07-23T16:02:12 <dreimark> or some other methods ?
2008-07-23T16:02:42 <dreimark> s/do disable/to disable/
2008-07-23T16:03:40 <ThomasWaldmann> re
2008-07-23T16:12:30 <dennda> ThomasWaldmann: If that is the issue, the getPagePath methods are the culprits
2008-07-23T16:12:35 <dennda> I will try to remove them
2008-07-23T16:12:40 <dennda> (regarding the bug)
2008-07-23T16:17:19 <ThomasWaldmann> dennda: you need to be able to handle attachments at least so far as a read-only 1.6 backend needs it, to be able to write a converter attachments > items
2008-07-23T16:20:06 <ThomasWaldmann> zenhase: any news about proxy code configuration?
2008-07-23T16:20:13 <dennda> ThomasWaldmann: What counts as attachment?
2008-07-23T16:20:45 <ThomasWaldmann> everything in a existing 1.6/1.7 pages/PAGENAME/attachments/ dir
2008-07-23T16:21:45 <dennda> When just creating a Page with a little text there is no attachment involved, is there?
2008-07-23T16:22:03 <ThomasWaldmann> [[attachment:foo.png]]
2008-07-23T16:22:17 <ThomasWaldmann> {{attachment:foo.png}}
2008-07-23T16:22:43 <dennda> Ok, so that statement of yours was more of a "by the way" nature?
2008-07-23T16:23:38 <ThomasWaldmann> dennda: you need to write code for that or your stuff won't be able to get productive
2008-07-23T16:24:07 <dennda> Well, sure, I just want to fix this issue first
2008-07-23T16:24:16 <dennda> (And could need some pointers)
2008-07-23T16:26:04 <ThomasWaldmann> what issue?
2008-07-23T16:27:54 <dennda> ThomasWaldmann: Well, run a wikiserver with the clone you got from me, create a page and then switch around between pages.
2008-07-23T16:30:36 <zenhase> ThomasWaldmann: well, i could import wikiconfig early in the WSGI-level code (before any particular moin-code) and check the contained Config-class for a proxy-configuration
2008-07-23T16:30:54 <zenhase> then wrap moins wsgi-app in a middleware that fixes up the environ accordingly
2008-07-23T16:31:35 <ThomasWaldmann> zenhase: no, you can't
2008-07-23T16:32:11 <zenhase> i already thought that this might give problems with farms and stuff :o
2008-07-23T16:34:11 <dennda> ThomasWaldmann: See?
2008-07-23T16:34:12 <zenhase> hmm
2008-07-23T16:35:33 <zenhase> i think i still don't totally grasp the way, how configuration for moin is retrieved
2008-07-23T16:36:02 <zenhase> and what you can put where to have access early on to the options
2008-07-23T16:39:01 <ThomasWaldmann> dennda: i still have lots of exceptions
2008-07-23T16:39:21 <ThomasWaldmann> i suggest you just create a clone yourself and fix the obvious stuff first
2008-07-23T16:39:40 <dennda> Hm, ok
2008-07-23T16:40:16 <dennda> But I just tried a little hack to determine if that's really that methods fault. I created a randomly called subdir in /tmp/ for the stuff and now it works. Thus, I need to get rid of that method and replace it
2008-07-23T16:41:37 <ThomasWaldmann> try to use less "stuff" and "that method" and people will follow you more easily
2008-07-23T16:42:42 <dennda> But I just tried a little hack to determine if that's really getPagePaths fault. I created a randomly called subdir in /tmp/ that getPagePath returns and now it works. Thus, I need to get rid of getPagePath and replace it
2008-07-23T16:43:01 <ThomasWaldmann> zenhase: when a request is processed, it tries to load farmconfig and lookup the correct config for the current url there
2008-07-23T16:43:14 <ThomasWaldmann> if there is no farmconfig, it maps every url to wikiconfig
2008-07-23T16:43:27 <ThomasWaldmann> then it loads a matching config
2008-07-23T16:44:21 <ThomasWaldmann> the problem with proxy stuff configuration is that in the old code, the proxy code had to be executed before the wikiconfig/farmconfig is loaded, so it could not get configured there
2008-07-23T16:44:38 <dennda> johill: Ping
2008-07-23T16:44:39 <ThomasWaldmann> so that stuff needs to be in server configuration, not in wikiconfig
2008-07-23T16:46:54 <ThomasWaldmann> dennda: AttributeError: 'StoredRevision' object has no attribute 'mtime'
2008-07-23T16:47:05 <ThomasWaldmann> File "/home/wedvwiki/moin-1.8-storage/MoinMoin/Page.py", line 453, in mtime
2008-07-23T16:47:31 <dennda> ThomasWaldmann: That shouldn't keep you from creating and saving a page
2008-07-23T16:48:38 <ThomasWaldmann> AttributeError: 'MemoryBackend' object has no attribute 'list_items'
2008-07-23T16:48:43 <ThomasWaldmann> File "/home/wedvwiki/moin-1.8-storage/MoinMoin/storage/external.py", line 149, in __iter__
2008-07-23T16:48:46 <ThomasWaldmann> ...
2008-07-23T16:48:52 <zenhase> hmm
2008-07-23T16:49:24 <dennda> that's not the latest changeset, is it?
2008-07-23T16:49:43 <ThomasWaldmann> it is
2008-07-23T16:50:12 <dennda> What did you do?
2008-07-23T16:50:37 <ThomasWaldmann> cloned, changed wikiconfig_local from your pastebin, nothing else
2008-07-23T16:51:44 <johill> dennda: pong, in talk tho
2008-07-23T16:51:46 <dennda> How can I reproduce the latter?
2008-07-23T16:51:59 <dennda> johill: How did you want me to remove getPagePath?
2008-07-23T16:52:07 <johill> well remove the users?
2008-07-23T16:52:09 <dennda> johill: It is used for caching as far as I can tell
2008-07-23T16:52:56 <dennda> johill: Stupidly delete them?
2008-07-23T16:56:49 <johill> it's used for attachments and caching
2008-07-23T16:57:13 <johill> as for caching, we'll have to come up with something else and cache maybe based on t he sha1(pagename) in the normal caching dir
2008-07-23T16:57:34 <dennda> and attachments will become items
2008-07-23T16:57:34 <zenhase> i'll be back later
2008-07-23T16:58:22 <dennda> johill: So indeed just delete the method and the calls, right?
2008-07-23T17:01:24 <johill> yeah once attachments are items and you do the caching slightly differently, you can indeed delete it
2008-07-23T17:01:28 <johill> anyhow, talk over, gtg
2008-07-23T17:02:06 <dennda> ok, will delete it then
2008-07-23T17:08:54 <dreimark> bbl
2008-07-23T17:09:12 <ThomasWaldmann> johill: do you think sha1(pn) is faster that quoteWikinameFS?
2008-07-23T17:10:26 <TheSheep> that's easy to check
2008-07-23T17:12:42 <ThomasWaldmann> (otherwise I would just use some attachment_dir / quoteWikinameFS(pn) / cache / ... as a replacement
2008-07-23T17:13:41 <dennda> Is it ok if I keep my debug-print-statements within the code and remove them afterwards?
2008-07-23T17:13:49 <dennda> Makes life much easier
2008-07-23T17:16:56 <ThomasWaldmann> use logging.debug()
2008-07-23T17:21:27 <johill> ThomasWaldmann: no, but it's saner
2008-07-23T17:21:42 <johill> ThomasWaldmann: problem with quoteWikinameFS is that you run into length restrictions
2008-07-23T17:22:22 <ThomasWaldmann> that doesn't mean we want to do expensive stuff for accessing the cache if we can avoid it
2008-07-23T17:25:54 <johill> I doubt it's more expensive, and besides, correctness is more important
2008-07-23T17:26:01 <johill> if it becomes a problem we can optimise it later
2008-07-23T17:26:13 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station PERIOD com> default * 4350:d95ff6cadd84 1.8-storage-cdenter/MoinMoin/ (Page.py PageEditor.py caching.py): storage: Page.py: Fixing AttributeError. Code cleanup
2008-07-23T17:44:16 <ThomasWaldmann> dennda: action=info crashes
2008-07-23T17:45:34 <dennda> I know
2008-07-23T17:46:34 <dennda> Can I somehow disable caching completely as a configuration setting?
2008-07-23T18:06:27 <ThomasWaldmann> no
2008-07-23T18:17:54 <dennda> Hm, OK. It's just not easy to know where to start
2008-07-23T18:25:50 * dennda reads http://moinmo.in/MoinCaching
2008-07-23T18:28:08 <johill> you need to get rid of the page-arena
2008-07-23T18:32:26 <ThomasWaldmann> or just put it somewhere else
2008-07-23T18:33:17 <johill> yeah
2008-07-23T18:44:50 <dennda> What is the page-arena and what do you mean with "put it somewhere else"?
2008-07-23T18:45:03 <dennda> I figure it's some common place where pages are cached. Is that true?
2008-07-23T18:46:00 <ThomasWaldmann> for the 1.6 storage, that is a cache directory under the page directory
2008-07-23T18:46:13 <ThomasWaldmann> per page
2008-07-23T19:09:59 <dennda> And you mean I shall put it elsewhere so the cached files aren't found and thus no cache is used?
2008-07-23T19:11:27 <dreimark> re
2008-07-23T20:03:25 <dennda> Or did I misinterpret you there, ThomasWaldmann?
2008-07-23T20:17:00 <dreimark> does one heard about raucherclub mensch kultur kneipe ?
2008-07-23T20:17:22 <dreimark> quite strange I have to fill out a form to sit in a restaurant
2008-07-23T20:18:53 <xorAxAx> http://www.dehoga-nrw.de/mensch_kultur_kneipe.html
2008-07-23T20:51:06 <dreimark> xorAxAx: that is somehow stupid, do you have seen a form there to become a member for every restaurant in germany or europe
2008-07-23T20:51:17 <xorAxAx> hehe
2008-07-23T20:51:53 * dreimark dislikes to fill out a form first in every restaurant I like to have dinner
2008-07-23T20:57:39 <johill> just leave
2008-07-23T21:03:08 <gizmach> moin
2008-07-23T21:03:16 * gizmach has connection problems
2008-07-23T21:05:24 <dreimark> gizmach: do you have to also fill out a form in a restaurant if you like to have a drink or something to eat
2008-07-23T21:05:47 <dreimark> because of some people who also want to smoke in the same room
2008-07-23T21:09:23 <dennda> dreimark: Do you need to sign that thing to allow the others to smoke?
2008-07-23T21:09:35 <johill> no, to be served
2008-07-23T21:09:40 <johill> they are legally clubs
2008-07-23T21:09:44 <johill> so they cannot allow non-members in
2008-07-23T21:09:53 <johill> you can't even go in the door without signing up for the club
2008-07-23T21:09:56 <dennda> Ok, I agree: Just leave
2008-07-23T21:10:16 <dennda> Don't know what keeps you there
2008-07-23T21:12:04 <dreimark> dennda: well it is one of the nice restaurants with free wifi
2008-07-23T21:12:14 <johill> don't support them
2008-07-23T21:12:28 <gizmach> dreimark: ?
2008-07-23T21:13:02 <gizmach> I didn't understand
2008-07-23T21:14:05 <gizmach> dreimark: I'm at my parents house
2008-07-23T21:16:44 <gizmach> dreimark: I also written some ldap test for acl rights and it works (but I need to optimise some methods) and I hope I will push these days when I finaly got connection on linux
2008-07-23T21:16:59 <gizmach> also done a multiple bacend using test
2008-07-23T21:17:23 <gizmach> and figure that I need first defined non wiki backend if they are used and than wikibackends
2008-07-23T21:18:09 <dreimark> johill: I found out that in principle it is enough to have the form on the table
2008-07-23T21:18:36 <dreimark> and to show that you can read the page xorAxAx showed
2008-07-23T21:20:29 <dreimark> gizmach: why? normally it should be free of the users choice
2008-07-23T21:25:10 <dennda> johill: How is the conference?
2008-07-23T21:25:20 <gizmach> dreimark: because when it has FooBAr group it creates wikibackend always not depending if that is config group or something else
2008-07-23T21:25:46 <johill> somewhat interesting, the wireless minisummit yesterday was productive, interesting and stressful
2008-07-23T21:26:45 <dennda> Any popular faces around?
2008-07-23T21:27:05 <johill> popular? hmm
2008-07-23T21:27:20 <dennda> linus, andrew, greg...? :)
2008-07-23T21:27:37 <zenhase> alan turing
2008-07-23T21:27:37 <zenhase> charles babbage
2008-07-23T21:27:37 <johill> haven't seen any of them yet but I'm sure gregkh is there
2008-07-23T21:27:46 <johill> alan cox I've seen
2008-07-23T21:27:52 <dennda> cool too
2008-07-23T21:28:12 <dennda> much more interesting than meeting britney spears, is it?
2008-07-23T21:28:13 <zenhase> not quite the same as turing tho
2008-07-23T21:28:18 <johill> heh
2008-07-23T21:28:30 <zenhase> dennda: who knows
2008-07-23T21:29:16 <dennda> I'd rather meet alan cox than britney spears
2008-07-23T21:29:55 <gizmach> dennda: I met alan cox this year
2008-07-23T21:29:59 <dennda> johill: Can you enlighten me about how to get rid of page-arena in a sane way?
2008-07-23T21:30:04 <dennda> gizmach: Oh, did you? Where?
2008-07-23T21:30:05 <zenhase> i am not that interested in meeting any of those :)
2008-07-23T21:30:10 <gizmach> coatia
2008-07-23T21:30:16 <gizmach> dors cluc
2008-07-23T21:30:21 <gizmach> linux conference
2008-07-23T21:30:37 <gizmach> were Harald Welte, Cox and some other
2008-07-23T21:31:34 <zenhase> in germany it's quite common to meet harald welte, i suppose ;)
2008-07-23T21:31:42 <johill> he's here too
2008-07-23T21:32:09 <gizmach> :)
2008-07-23T21:32:39 <xorAxAx> you could ask him whether he has bought a freerunner already
2008-07-23T21:32:50 <dreimark> johill: greetings from me
2008-07-23T21:32:51 <johill> heh
2008-07-23T21:33:08 <gizmach> well this meeting is always in april, you are al invited it's great fun
2008-07-23T21:33:08 <zenhase> kind of a strange question, isn't it?
2008-07-23T21:33:19 <xorAxAx> why is it strange, zenhase? :)
2008-07-23T21:33:57 <zenhase> xorAxAx: isn't it somehow like asking linus if he's running a linux server anywhere?
2008-07-23T21:34:04 <zenhase> ok, not exactly the same
2008-07-23T21:34:49 <zenhase> but you would guess that he's one of the first persons to receive one from the company
2008-07-23T21:35:17 <xorAxAx> linus didnt quit linux, harald quit openmoko -- big difference :)
2008-07-23T21:35:22 <TheSheep> receive!=buy
2008-07-23T21:35:52 <zenhase> a friend of mine did some microcontroller programming for a firmware in midi-devices by a taiwan-based company
2008-07-23T21:36:07 <zenhase> they flooded him with their lowcost devices
2008-07-23T21:36:12 <zenhase> xorAxAx: oh
2008-07-23T21:36:27 <zenhase> xorAxAx: i did not know that
2008-07-23T21:36:37 <zenhase> i ain't reading much OSS news
2008-07-23T21:36:38 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4021:6c4548d34afd 1.8-dom-bblank/MoinMoin/converter2/html_out.py: HTML output converter - Factor out TOC generator, add stack of TOCs
2008-07-23T21:36:39 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4022:83902eb0edaa 1.8-dom-bblank/MoinMoin/converter2/html_out.py: HTML output converter - Implement TOC per included page snippet
2008-07-23T21:36:40 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4023:e094a29f7752 1.8-dom-bblank/MoinMoin/converter2/html_out.py: HTML output converter - Fix minimal level handling of TOC
2008-07-23T21:38:16 <xorAxAx> zenhase: me neither, but happened many months ago
2008-07-23T21:38:26 <zenhase> i see
2008-07-23T21:38:34 <waldi> okay, toc and include works together
2008-07-23T21:38:49 <xorAxAx> early 2007
2008-07-23T21:41:31 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4351:fdaf7be4c0f4 1.8-storage-cdenter/MoinMoin/user.py: storage: user.py: Trivial Fix in order to get the 'switch users' function to work
2008-07-23T21:42:14 <dennda> It is quite handy to already have two different backends
2008-07-23T21:55:06 <dennda> johill: No enlightenment? Or got no time? I can also try to figure it out myself, but it isn't quite easier and I may be faster if you gave a few tips :)
2008-07-23T22:12:43 <dennda> johill: By the way, your fix back then ( http://paste.pocoo.org/show/79709/ ) didn't sort out all issues. The same error still occurs if you try to create a user with an email adress that is already taken or try to change a users preferences
2008-07-23T22:13:21 <dennda> (An Item instance being sent the backend.has_item() check rather than a name)
2008-07-23T22:41:49 <gizmach> gn
2008-07-23T22:44:25 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4024:d81ff5a2e041 1.8-dom-bblank/MoinMoin/Page.py: Page - Use StringIO to buffer output
2008-07-23T22:48:19 <waldi> yeah, StartSeite is about two times as fast
2008-07-23T22:49:26 <xorAxAx> 5 instead of 10 seconds? :)
2008-07-23T22:51:31 <waldi> 0,035 instead of 0,08
2008-07-23T22:53:01 <xorAxAx> slow
2008-07-23T22:58:29 <dennda> What shall I do with the edit-log? Are we going to keep that or the information it used to keep somewhere else?
2008-07-23T23:01:51 <xorAxAx> hmm?
2008-07-23T23:01:54 <xorAxAx> which level ...
2008-07-23T23:01:59 <xorAxAx> are you talking about? :)
2008-07-23T23:03:44 <dennda> Look at the code in MoinMoin/logfile/editlog.py
2008-07-23T23:04:40 <dennda> It uses lanius' old API design and I want to rewrite it
2008-07-23T23:05:07 <xorAxAx> yes. have you checked who is using those classes?
2008-07-23T23:05:28 <dennda> But I wonder if we actually want a file on the disk for that or rather store that information edit-log keeps somewhere in our backend
2008-07-23T23:05:40 <xorAxAx> lanius doesnt store it in a file on disk ...
2008-07-23T23:05:46 <xorAxAx> why do you think so?
2008-07-23T23:06:02 <dennda> 132-134
2008-07-23T23:06:14 <xorAxAx> ah, localeditlog
2008-07-23T23:06:33 <xorAxAx> hmm, that looks broken
2008-07-23T23:06:57 <dennda> What exactly looks broken and how?
2008-07-23T23:07:06 <dennda> You mean by design?
2008-07-23T23:07:10 <xorAxAx> ah
2008-07-23T23:07:17 <xorAxAx> he is never reading from that file, is he?
2008-07-23T23:07:28 <dennda> How do I find out?
2008-07-23T23:07:40 <xorAxAx> that looks like he tries to keep a backend with write support
2008-07-23T23:07:57 <xorAxAx> (and kept the code on this level because its not completly refactored yet)
2008-07-23T23:08:09 <xorAxAx> well, the idea is that this is wrong of course :)
2008-07-23T23:08:15 <xorAxAx> just grep for edit-log
2008-07-23T23:08:49 <dennda> after saving something I assume, ok just a second
2008-07-23T23:08:57 <xorAxAx> its the only place i think
2008-07-23T23:09:14 <xorAxAx> so removing that method is fine if you think that the storage api completly maps its semantics
2008-07-23T23:09:31 <xorAxAx> you will need to check whether the callers need to be modified, though
2008-07-23T23:09:47 <xorAxAx> because the method call conveys meta data that hasnt probably been sent to the backend yet
2008-07-23T23:10:02 <dennda> No, don't find anything
2008-07-23T23:10:56 <dennda> Which method are we talking about now?
2008-07-23T23:11:00 <xorAxAx> add
2008-07-23T23:11:21 <dennda> And with remove you mean rewrite, using the new api?
2008-07-23T23:11:45 <xorAxAx> i figure that Global* and Local* already broke the 1.{6,7,8} api editlog wise
2008-07-23T23:11:55 <xorAxAx> so you can probably remove the whole module
2008-07-23T23:12:05 <xorAxAx> no, i remove delete
2008-07-23T23:12:10 <xorAxAx> s/remove/mean/
2008-07-23T23:12:20 <xorAxAx> .oO(omg, weird word replacement syndrom)
2008-07-23T23:12:24 <dennda> heh
2008-07-23T23:12:35 <dennda> rm editlog.py, yes?
2008-07-23T23:12:42 <xorAxAx> look at the callers first
2008-07-23T23:12:52 <xorAxAx> to get a feeling how to rewrite their calls into the module
2008-07-23T23:13:03 <xorAxAx> then you can remove it :)
2008-07-23T23:13:18 <dennda> So rewrite the callers to use the storage for logging directly?
2008-07-23T23:13:24 <xorAxAx> so one question is how the global edit log works after the storage api changes
2008-07-23T23:13:26 <dennda> (As with that other method we discussed yesterday)
2008-07-23T23:13:56 <xorAxAx> yes, if it doesnt look too bad in the end (then you might need to insert layers to rectify the bad storage api design :))
2008-07-23T23:14:24 <xorAxAx> and you will probably realise that the current layer is nice enough, so removing should be done after you have solved the question:
2008-07-23T23:14:28 <xorAxAx> 23:13:23 < xorAxAx > so one question is how the global edit log works after the storage api changes
2008-07-23T23:14:42 <xorAxAx> after ... == now, in your branch
2008-07-23T23:14:55 <dennda> That's a design decision, isn't it?
2008-07-23T23:15:04 <xorAxAx> well, didnt you think about it?
2008-07-23T23:15:12 <dennda> I am doing right now
2008-07-23T23:15:14 <xorAxAx> its one of the most elementary operations in a wiki :)
2008-07-23T23:21:35 <dennda> Sigh, documentation is a bit poor. Or is there any hidden resource I didn't discover yet?
2008-07-23T23:25:47 <xorAxAx> docs about what?
2008-07-23T23:27:58 <dennda> almost anything
2008-07-23T23:28:18 <dennda> I remember having had a conversation about this with TW, let me dig throug my logs
2008-07-23T23:29:22 <xorAxAx> ah
2008-07-23T23:29:31 <xorAxAx> well, there is basically only MoinDev and the code
2008-07-23T23:29:39 <xorAxAx> and the wiki ...
2008-07-23T23:30:34 <dennda> Yes, which doesn't make it any easier for me to learn the concepts :)
2008-07-23T23:30:54 <dennda> I will just store that edit-log stuff in an items revisions metadata
2008-07-23T23:31:24 <xorAxAx> umm
2008-07-23T23:31:41 <xorAxAx> how does lanius do it?
2008-07-23T23:32:48 <dennda> His revisions seem to have attributes (?!?) keeping the information
2008-07-23T23:32:52 <dennda> judging from the code
2008-07-23T23:33:41 <dennda> ah yes
2008-07-23T23:33:46 <dennda> and some __getattr__ hack
2008-07-23T23:34:00 <dennda> returning it from revision metadata
2008-07-23T23:34:45 <dennda> So basically, same idea
2008-07-23T23:35:02 <xorAxAx> dennda: it is using backend.news()
2008-07-23T23:35:08 <xorAxAx> thats the central idea of the old concept
2008-07-23T23:35:12 <dennda> which line?
2008-07-23T23:35:15 <xorAxAx> does the new api rev still have news?
2008-07-23T23:35:20 <xorAxAx> just look for news() calls
2008-07-23T23:35:30 <xorAxAx> in logfile
2008-07-23T23:36:59 <dennda> that method is going to be implemented, yes
2008-07-23T23:37:17 <dennda> I didn't yet, and johill only did a fs-implementation based on reading the logfile
2008-07-23T23:38:10 <dennda> So let me look at the callers...
2008-07-23T23:41:07 <dennda> oops
2008-07-23T23:41:22 <ThomasWaldmann> re
2008-07-23T23:43:17 <dennda> wb, ThomasWaldmann
2008-07-23T23:46:24 <dreimark> gizmach: it is important to solve the order problem, or to find a very good reason why it should be possible that a wikigroup can overwrite any other group definition
2008-07-23T23:46:53 * dreimark thinks solving it is the better solution
2008-07-23T23:47:12 <johill> dennda: I'll take a look next week
2008-07-23T23:47:34 <johill> if you want any answers from me, email me, or put up a wiki page
2008-07-23T23:48:23 <dreimark> johill: seems one was able in a 1.6.x wiki to rename his user account to an other name while the name field was ro
2008-07-23T23:48:31 <dennda> johill: Ok
2008-07-23T23:49:03 <dreimark> and as long the cookie was valid he doesn't recognized something was quite completly wrong
2008-07-23T23:50:09 <ThomasWaldmann> dennda: did you look into storage.external?
2008-07-23T23:50:45 <dreimark> so it looks like currently (sorry can't describe it better) that there were two people using the same uid
2008-07-23T23:51:18 <dennda> ThomasWaldmann: In what context? I do often, yes
2008-07-23T23:51:34 <dreimark> we can discuss this next week me hopes to understand better how that could happen
2008-07-23T23:51:48 <ThomasWaldmann> the save method there stores "edit log" data into metadata
2008-07-23T23:53:19 <dreimark> http://moinmo.in/MoinMoinBugs/PreFormatterBreaksOnHashBangLine-Sample -> traceback
2008-07-23T23:55:09 <ThomasWaldmann> dreimark: sure, it is a sample of a bug
2008-07-23T23:55:13 <dennda> ThomasWaldmann: Yeah, but that's the wrong layer imho
2008-07-23T23:56:04 <dreimark> got it now too ( should have readed the other pages )
2008-07-23T23:56:21 <ThomasWaldmann> dennda: i guess you need to create a wikipage about that
2008-07-23T23:56:29 <dennda> ThomasWaldmann: Why?
2008-07-23T23:57:04 <ThomasWaldmann> to formulate the problem you want to solve and ask async questions
2008-07-23T23:57:35 <dennda> async questions?
2008-07-23T23:57:56 <ThomasWaldmann> obviously your mentor is in a different tz now
2008-07-23T23:58:11 <dennda> Ah, that was a geek way of saying that :)
2008-07-23T23:58:18 <johill> dreimark: really, I'm not here for all intents and purposes, give me a link and I'll look at some point MoinMoin: MoinMoinChat/Logs/moin-dev/2008-07-23 (last edited 2008-07-22 22:15:02 by IrcLogImporter)
