MoinMoin: MoinMoinChat/Logs/moin-dev/2007-10-28

2007-10-28T00:14:17  <CIA-31> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2263:45b8b2108052 1.6/MoinMoin/filter/EXIF.py: update EXIF.py from SF project, PEP8 fixes, re-applied all our changes
2007-10-28T00:24:35  <TheSheep> heh
2007-10-28T00:24:54  <TheSheep> they don't want the changes pushed upstream?
2007-10-28T00:26:14  <ThomasWaldmann> i wrote them a mail
2007-10-28T00:26:23  <ThomasWaldmann> s/them/him/
2007-10-28T00:36:23  <CIA-31> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2264:2963e2f43297 1.6/MoinMoin/filter/EXIF.py: cosmetic fix: this is NOT octal (ported from 1.7)
2007-10-28T00:37:38  <CIA-31> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2900:7028ac728357 1.7/MoinMoin/filter/EXIF.py: copied updated EXIF.py from 1.6 branch
2007-10-28T03:53:01  *** irc.freenode.net sets mode: +o ChanServ
2007-10-28T08:59:53  <dreimark> moin
2007-10-28T09:00:21  <dreimark> starshine_away: TheSheep if modern theme is changed please change modern_cms too and if possible please create a similiar classic_cms
2007-10-28T09:30:57  <dreimark> and we should both add to the codebase
2007-10-28T09:39:04  <CIA-31> moin: Reimar Bauer <rb.proj AT googlemail DOT com> default * 2901:5cd9d4c041b9 1.7/MoinMoin/converter/text_html_text_moin_wiki.py: text_html_text_moin_wiki: linebreak in table cell should be written into wiki text as <<BR>>
2007-10-28T09:39:07  <CIA-31> moin: Reimar Bauer <rb.proj AT googlemail DOT com> default * 2902:a900b2f20fb7 1.7/MoinMoin/converter/text_html_text_moin_wiki.py: text_html_text_moin_wiki: an empty formatted cell of a pasted table should not be saved as a formatted br
2007-10-28T09:39:08  <CIA-31> moin: Reimar Bauer <rb.proj AT googlemail DOT com> default * 2903:3cbfe3edd7ed 1.7/MoinMoin/converter/_tests/test_text_html_text_moin_wiki.py: test_text_html_text_moin_wiki: tests for table for empty and formatted cells added
2007-10-28T09:39:09  <CIA-31> moin: Reimar Bauer <rb.proj AT googlemail DOT com> default * 2904:1cf088f06286 1.7/MoinMoin/filter/EXIF.py: merged main
2007-10-28T10:31:07  <grzywacz> moin
2007-10-28T10:31:23  <xorAxAx> moin grzywacz
2007-10-28T10:32:17  <grzywacz> What's the time now?
2007-10-28T10:32:20  <grzywacz> 10:32? ;)
2007-10-28T10:33:19  <dreimark> :)
2007-10-28T10:33:25  <dreimark> moin
2007-10-28T10:35:20  <xorAxAx> grzywacz: at least in central europe, yes
2007-10-28T10:35:40  <grzywacz> Good. I got confused by the time change. ;)
2007-10-28T10:36:10  <grzywacz> My computer's clock was 30 minutes ahead when I turned it on...
2007-10-28T10:36:16  <xorAxAx> oh
2007-10-28T10:36:51  <xorAxAx> mathematicists solution to dst problems? just use the average of old and new time
2007-10-28T10:50:09  <grzywacz> apparently some maths prof tempered with gentoo ;]
2007-10-28T11:49:28  <dreimark> cant reach paste.pocoo.org
2007-10-28T11:52:45  <johill> there are so many other pastebins...
2007-10-28T12:11:14  <dreimark> i know, but the one we enter at moin fails
2007-10-28T12:13:26  <dreimark> bbl  ~3h
2007-10-28T12:31:58  <ThomasWaldmann> moin
2007-10-28T12:40:39  <ThomasWaldmann> http://hg.moinmo.in/moin/1.7/rev/a900b2f20fb7 ???
2007-10-28T12:42:48  <TheSheep> I guess that Word users often leave a single br in their 'empty' table cells
2007-10-28T13:22:41  <ThomasWaldmann> and how does this change make sure it is this case?
2007-10-28T13:32:56  <ThomasWaldmann> anyone holding me back from migrating moinmo.in to 1.6dev today? :)
2007-10-28T13:34:11  * TheSheep whistles quietly
2007-10-28T13:37:24  * ThomasWaldmann didn't hear it :)
2007-10-28T13:58:25  <ThomasWaldmann> brb
2007-10-28T15:17:45  <ThomasWaldmann> hi dreimark
2007-10-28T15:18:26  <dreimark> hi
2007-10-28T15:18:33  <ThomasWaldmann> dreimark: http://hg.moinmo.in/moin/1.7/rev/a900b2f20fb7 why does the code there do what the comment says?
2007-10-28T15:20:27  <dreimark> moment
2007-10-28T15:23:48  <dreimark> testTable07 of test_text_html_text_moin_wiki describes that change
2007-10-28T15:24:06  <dreimark> <td><strong><br /></strong></td>
2007-10-28T15:24:20  <dreimark> is not wanted to give ||'''\n'''||
2007-10-28T15:26:07  <dreimark> arrgh sorry, I tried to split a patch in pieces
2007-10-28T15:27:53  <dreimark> I should have splitted the tests too
2007-10-28T15:29:39  <dreimark> I have this only seen by cut and paste from soffice to the gui
2007-10-28T15:29:59  <dreimark> normally we didnt get a formatted br for an empty cell
2007-10-28T15:30:54  <dreimark> but in soffice the empty cells are formatted too
2007-10-28T15:31:54  <dreimark> ThomasWaldmann: ?
2007-10-28T15:33:22  <dreimark> johill: I need some help for wikiutil.get_choice in 1.6
2007-10-28T15:33:53  <johill> ok, what's hte problem?
2007-10-28T15:34:29  <johill> generally, I think you shouldn't even need to use it
2007-10-28T15:34:47  <johill> by using a default tuple of the choices
2007-10-28T15:38:10  <dreimark> I want to give an example in that chapter I am writing
2007-10-28T15:38:18  <dreimark> http://test.wikiwikiweb.de/ReimarBauer/ExampleScript
2007-10-28T15:38:35  <johill> ugh
2007-10-28T15:38:41  <johill> why don't you use the new macro stuff?
2007-10-28T15:38:55  <dreimark> johill: 1.6 hasnt it completly
2007-10-28T15:39:27  <johill> ah
2007-10-28T15:39:31  <johill> I thought it did
2007-10-28T15:39:43  <johill> hmm then still you should probably do some boilerplate code that calls it like 1.7 would
2007-10-28T15:39:55  <johill> hm. otoh, that's quite a bit of code
2007-10-28T15:40:22  <johill> actually, we have invoke_extension_function
2007-10-28T15:40:47  <johill> so you should do that
2007-10-28T15:40:50  <johill> mind if I edit your page?
2007-10-28T15:40:55  <dreimark> please do
2007-10-28T15:42:22  <ThomasWaldmann> dreimark: that doesnt answer my question
2007-10-28T15:43:44  <johill> dreimark: done
2007-10-28T15:44:13  <dreimark> ThomasWaldmann: process_inline is called recursive for formattings and it should not be called for \n
2007-10-28T15:46:06  <johill> dreimark: does that help? I haven't tested it of course
2007-10-28T15:46:17  <johill> but the execute() there should be pretty much template code
2007-10-28T15:46:21  <dreimark> looks better, trying currently
2007-10-28T15:46:38  <johill> error reporting might be bad
2007-10-28T15:46:47  <johill> not sure what invoke_extension_function does there, I think it raises exceptions
2007-10-28T15:46:52  <johill> the execute() might need to catch them
2007-10-28T15:47:05  <johill> yeah it raises ValueErrors
2007-10-28T15:48:27  <dreimark> hmm, the new code in 1.7 is too nice
2007-10-28T15:48:47  <johill> :)
2007-10-28T15:49:00  <johill> I just changed your example to have the same error reporting
2007-10-28T15:49:55  <dreimark> yeah, but now its more complicated as before for a newbie
2007-10-28T15:50:07  <johill> yeah, 1.7 does that for you
2007-10-28T15:50:10  <johill> but it's just template code
2007-10-28T15:51:11  <johill> we could probably put that code into 1.6 too without changing the existing plugins
2007-10-28T15:51:15  <johill> it falls back anyway
2007-10-28T15:52:14  <johill> ThomasWaldmann: the captcha stuff could be programmable so you could have simple math exercises or so like "calculate 30-25"
2007-10-28T15:52:25  <ThomasWaldmann> dreimark: the comment talks of an "empty formatted cell". where is the code handling an empty cell (and not a non-empty cell)?
2007-10-28T15:52:44  <dreimark> johill: I would prefer that or a new method (lievetime < 1.7) in wikiutil which handles parsing and choices
2007-10-28T15:53:05  <johill> dreimark: parsing and choices is all in wikiutil.invoke_extension_function
2007-10-28T15:53:07  <ThomasWaldmann> johill: it could be even plugins :) but I didn't want too big changes for 1.6
2007-10-28T15:53:18  <johill> ThomasWaldmann: ok, sure
2007-10-28T15:53:36  <johill> dreimark: only it doesn't handle errors
2007-10-28T15:53:39  <johill> (and can't)
2007-10-28T15:54:12  <johill> dreimark: I suppose you could easily take that execute() there and put it into a function return
2007-10-28T15:54:16  <johill> so you can write
2007-10-28T15:54:25  <johill> execute = MoinMoin.macro.exec_wrapper(macro_Example)
2007-10-28T15:59:18  <ThomasWaldmann> johill: and also, before we make that more flexible, i think we should cleanup the editor
2007-10-28T15:59:40  <ThomasWaldmann> and before that, we should merge storage branch
2007-10-28T15:59:56  <ThomasWaldmann> and before that, we should finish it
2007-10-28T16:00:42  <dreimark> okok, no changes for me now
2007-10-28T16:02:36  <johill> heh
2007-10-28T16:02:52  <johill> ThomasWaldmann: ok
2007-10-28T16:03:09  <johill> dreimark: the template code is easy enough, but I don't remember why we didn't put the new invocation stuff into 1.6
2007-10-28T16:04:31  <dreimark> something must be leftover for 1.7 ;)
2007-10-28T16:05:10  <dreimark> ThomasWaldmann: an empty cell is tested by testTable01
2007-10-28T16:08:22  <dreimark> and at this part of process_inline command is used for formatting
2007-10-28T16:09:41  <dreimark> so only if command and i.localName != 'br' it should recurse
2007-10-28T16:09:54  <dreimark> If I have missed a test please let me know
2007-10-28T16:10:21  <johill> dreimark: anyway, I'm not following everything here really, did that help? did you want any more like put the template code into an extra function?
2007-10-28T16:10:28  <johill> dreimark: and did it work? :)
2007-10-28T16:11:29  <dreimark> johill: still on work, (answered ThomasWaldmann request)
2007-10-28T16:11:55  <johill> ok
2007-10-28T16:32:32  <johill> http://johannes.sipsolutions.net/patches/moin/all/LATEST/001-xmlrpc-putpage-fault.patch
2007-10-28T16:32:42  <johill> any objections? I think not reporting the errors is bad
2007-10-28T16:33:50  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2265:3d30962f71a6 1.6/MoinMoin/xmlrpc/__init__.py: xmlrpc putPage: fix comments
2007-10-28T16:34:10  <xorAxAx> johill: yes of course
2007-10-28T16:35:21  <johill> of course what?
2007-10-28T16:35:31  <xorAxAx> of course its bad
2007-10-28T16:35:53  <johill> but is the fix correct or should we just return false?
2007-10-28T16:36:00  <johill> i.e. do you think a Fault is appropriate
2007-10-28T16:36:19  <xorAxAx> yes, thats right
2007-10-28T16:36:34  <xorAxAx> i mean a fault is ok
2007-10-28T16:36:51  <johill> ok then I'll commit that (well, did a small _debug=True bugfix too now)
2007-10-28T16:37:37  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2266:b1f1f997f624 1.6/MoinMoin/xmlrpc/__init__.py: xmlrpc putPage: report errors as faults
2007-10-28T16:40:23  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2267:b2930137a300 1.6/MoinMoin/user.py:
2007-10-28T16:40:23  <CIA-31> moin: don't store trail for first request (port of 1.7 d234621dbf2f)
2007-10-28T16:40:23  <CIA-31> moin: avoids storing sessions to disk for requests from users
2007-10-28T16:40:23  <CIA-31> moin: that won't send back the session cookie anyway
2007-10-28T16:40:58  <johill> xorAxAx: how about this: http://johannes.sipsolutions.net/patches/moin/all/LATEST/001-xmlrpc-auth-token.patch
2007-10-28T16:41:16  <johill> couldn't decide btween destroy() or shutdown() or so
2007-10-28T16:42:30  <xorAxAx> well, the sync stuff is fine now
2007-10-28T16:44:22  <johill> but..?
2007-10-28T16:44:54  <xorAxAx> i have no idea about your session stuff
2007-10-28T16:44:57  <xorAxAx> you should know best
2007-10-28T16:47:08  <johill> yeah hmm
2007-10-28T16:47:15  <johill> maybe I should assign request.session after all when doing xmlrpc
2007-10-28T16:47:31  <johill> but we don't do much via xmlrpc
2007-10-28T16:49:19  <johill> ideally xmlrpc should be treated as a different request module
2007-10-28T16:50:23  <johill> hmm xmlrpc can actually format pages so I guess I do need to integrate this with the regular session stuff
2007-10-28T17:02:52  <dreimark> johill: I have some work with my converter changes (uncomplete) I try later your suggestions
2007-10-28T17:04:27  <johill> ok
2007-10-28T17:04:30  <johill> let me know
2007-10-28T17:20:23  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2268:55a9ac9cbe3a 1.6/MoinMoin/ (action/SyncPages.py wikisync.py xmlrpc/__init__.py):
2007-10-28T17:20:23  <CIA-31> moin: make xmlrpc use session system and real tokens instead of user IDs
2007-10-28T17:20:23  <CIA-31> moin: This is as far as I know the last piece of code where user IDs could
2007-10-28T17:20:23  <CIA-31> moin: be used to log in as the user, hence removing the problem that user
2007-10-28T17:20:23  <CIA-31> moin: IDs need to be secret.
2007-10-28T17:22:18  <johill> huh
2007-10-28T17:22:24  <johill> didn't I just commit the user trail fix?
2007-10-28T17:22:44  <johill> guess not, odd
2007-10-28T17:23:33  <johill> oh I did, getting confused here with quilt+hg
2007-10-28T17:24:19  <johill> hah, read-only now
2007-10-28T17:24:38  <johill> xorAxAx: what's your take on putPage() and anonymous users?
2007-10-28T17:26:08  <xorAxAx> johill: well, there is a config option, right?
2007-10-28T17:26:31  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2269:7e87ca9ec583 1.6/MoinMoin/xmlrpc/__init__.py: xmlrpc: fix docstring of _generate_auth_token
2007-10-28T17:26:37  <johill> not any more
2007-10-28T17:26:41  <xorAxAx> besides that, xmlrpc should offer the same level of operations as regular http
2007-10-28T17:27:01  <johill> xmlrpc can be turned off, but not unknown users in particular
2007-10-28T17:27:11  <johill> yeah it does, it's equivalent now afaict
2007-10-28T17:30:06  * ThomasWaldmann runs mig scripts
2007-10-28T17:31:17  * johill hopes it works
2007-10-28T17:32:06  <johill> xorAxAx: to handle session stuff properly I'd need another abstraction layer
2007-10-28T17:32:24  <johill> or remove the ability to configure a session handler
2007-10-28T17:32:50  <xorAxAx> johill: is the session system documented somewhere?
2007-10-28T17:33:19  <johill> for 1.7 in MoinMoin/session.py
2007-10-28T17:33:28  <johill> mostly docstring of SessionHandler etc
2007-10-28T17:34:33  <johill> what I was thinking is that session handlers could be overridden to integrate with other systems to override the cookie name
2007-10-28T17:34:47  <johill> e.g. if you have some php app storing session data in a database you can make moin use that too
2007-10-28T17:36:01  <xorAxAx> johill: thats not easily possible
2007-10-28T17:36:10  <xorAxAx> in the general case
2007-10-28T17:36:10  <johill> sure, why not?
2007-10-28T17:36:24  <xorAxAx> because the other system might keep them e.g. in RAM
2007-10-28T17:36:30  <johill> oh of course
2007-10-28T17:36:35  <johill> you need to control both and write a bit of code
2007-10-28T17:36:49  <xorAxAx> well, php sessions are already understood by moin
2007-10-28T17:36:56  <xorAxAx> without any session layers :)
2007-10-28T17:37:16  <johill> well that's different
2007-10-28T17:37:19  <xorAxAx> but coupling on that level would actually allow you to use one session, indeed
2007-10-28T17:37:20  <johill> you authenticate using php session
2007-10-28T17:37:24  <johill> you can't share data
2007-10-28T17:38:04  <johill> which is what the session system allows
2007-10-28T17:38:23  <johill> but decoupling session instantiation from cookie handling is yet another layer
2007-10-28T17:38:33  <johill> but in a way required if we want to unify it across to xmlrpc where cookies aren't used
2007-10-28T17:38:41  <xorAxAx> yes
2007-10-28T17:38:53  <xorAxAx> i would call it layer, though :)
2007-10-28T17:38:58  <xorAxAx> how about dimension?
2007-10-28T17:39:18  <johill> yeah, maybe
2007-10-28T17:39:22  <johill> question is how to design that
2007-10-28T17:39:33  <johill> I suppose the session handler would need to be given a 'cookie storage' class
2007-10-28T17:39:56  <johill> that allows set() and get() or something
2007-10-28T17:40:02  <johill> and is instantiated with a cookie name
2007-10-28T17:40:07  <johill> (which xmlrpc ignores)
2007-10-28T17:41:33  <xorAxAx> hmm, no
2007-10-28T17:42:09  <xorAxAx> the session system has a backing storage that stores data. it handles that bucket-wise. every bucket has an id
2007-10-28T17:42:17  <johill> that's different
2007-10-28T17:42:33  <johill> the backing storage is instantiated by the session handler
2007-10-28T17:42:41  <johill> it can be changed too, it's a separate dimension
2007-10-28T17:42:44  <xorAxAx> and then you can have a class that uses that layer to have an object that is used as a request attrib
2007-10-28T17:42:46  <johill> it just does key->session lookup
2007-10-28T17:42:51  <xorAxAx> yes
2007-10-28T17:43:11  <johill> that I don't understand
2007-10-28T17:43:43  <johill> what do you mean by the request attribute?
2007-10-28T17:43:57  <johill> currently, the session handler will read the cookie, instantiate the storage object and assign it to request.session
2007-10-28T17:44:16  <johill> ThomasWaldmann: pep8 fixes on an imported lib?!
2007-10-28T17:44:31  <johill> ThomasWaldmann: that makes it easy for upstream to pick up changes if they want...
2007-10-28T17:45:14  <xorAxAx> johill: request.session
2007-10-28T17:45:32  <johill> ok yes, request.session is assigned by the session handler
2007-10-28T17:45:38  <xorAxAx> johill: hmm, i am not sure if the storage handler should do that
2007-10-28T17:45:43  <johill> i.e. DefaultSessionHandler
2007-10-28T17:45:50  <xorAxAx> because it will make it nearly impossible to share that code, right?
2007-10-28T17:46:05  <johill> storage handler == request.session
2007-10-28T17:46:15  <johill> you can tell DefaultSessionHandler which storage handler class to use
2007-10-28T17:47:39  <xorAxAx> so there is a sessionhandler and a storage handler?
2007-10-28T17:47:48  <johill> yes
2007-10-28T17:47:58  <johill> but I used an OO model for storage
2007-10-28T17:48:03  <johill> rather than a procedural
2007-10-28T17:48:12  <johill> i.e. request.session is an instantiation of the storage handler
2007-10-28T17:48:34  <xorAxAx> well, then tell me how you want to share code here :)
2007-10-28T17:48:48  <johill> share between what?
2007-10-28T17:49:19  <johill> the session handler is given the storage handler class (with a sane default)
2007-10-28T17:49:36  <johill> see DefaultSessionHandler.__init__
2007-10-28T17:51:18  <xorAxAx> between two storage systems
2007-10-28T17:51:28  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2905:cd9be78f15db 1.7/MoinMoin/session.py: remove a stray # in a comment
2007-10-28T17:51:46  <johill> oh
2007-10-28T17:51:56  <johill> well they don't really share code anyway
2007-10-28T17:52:26  <johill> what code would they share?
2007-10-28T17:53:26  <xorAxAx> hmm, not much
2007-10-28T17:53:38  <xorAxAx> johill: but i see that there is some cleanup func that is not integrated into this concept
2007-10-28T17:53:47  <johill> yeah I just noticed that I didn't integrate that
2007-10-28T17:53:55  <johill> I should put it into DefaultSessionData
2007-10-28T17:54:09  <ThomasWaldmann> no we have only half as much pagedirs as before :)
2007-10-28T17:54:28  <johill> because of pages that never existed?
2007-10-28T17:55:06  <johill> hm can you have class functions rather than methods?
2007-10-28T17:55:08  <ThomasWaldmann> 2-3k crap, 700 deleted, 800 MoinEditorBackups
2007-10-28T17:55:13  <xorAxAx> johill: sure
2007-10-28T17:55:22  <xorAxAx> johill: classmethod
2007-10-28T17:55:47  <xorAxAx> ThomasWaldmann: the mig scripts dont migrate deleted pages?
2007-10-28T17:55:59  <johill> xorAxAx: do we require 2.4 now?
2007-10-28T17:56:24  <xorAxAx> no
2007-10-28T17:56:30  <johill> ok
2007-10-28T18:01:19  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2906:aaa550567872 1.7/MoinMoin/session.py: integrate DefaultSessionHandler cleanup into DefaultSessionData
2007-10-28T18:01:35  <johill> ok so that integrates that now
2007-10-28T18:01:45  <johill> but that still doesn't solve hte extra dimension problem
2007-10-28T18:01:52  <ThomasWaldmann> xorAxAx: i removed them via cleanpage maint
2007-10-28T18:02:11  <ThomasWaldmann> (i have a backup, just for the case :)
2007-10-28T18:02:25  <johill> xorAxAx: I think I'll go for the persistent-ID storage thing (cookie)
2007-10-28T18:05:11  <xorAxAx> johill: "(cookie)"?
2007-10-28T18:05:53  <johill> well for regular requests the cookie is the persistent ID
2007-10-28T18:06:00  <johill> but for xmlrpc the auth token is
2007-10-28T18:06:07  <johill> so I need to abstract that dimension out
2007-10-28T18:06:45  <johill> so that's what I said before, I'm thinking to have some sort of interface with get(), set() methods and instantiated by some name (cookie name)
2007-10-28T18:06:55  <johill> just not sure yet where to set what
2007-10-28T18:08:03  <johill> i.e. who sets the cookie name
2007-10-28T18:09:07  <xorAxAx> johill: the storage backend
2007-10-28T18:09:41  <johill> no you're thinking of the identifier
2007-10-28T18:09:48  <johill> which is actually determined by the session handler
2007-10-28T18:10:01  <johill> but cookies are key=value and I'm wondering about key, not value
2007-10-28T18:10:51  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2907:947d0fb3afb7 1.7/MoinMoin/session.py: session code: remove unused _delete_cookie
2007-10-28T18:12:27  <xorAxAx> johill: if the session handler determines it, you cannot reuse existing sessions from another backend
2007-10-28T18:13:02  <johill> well you need to write a new session handler for that
2007-10-28T18:13:11  <xorAxAx> ugh
2007-10-28T18:13:11  <johill> which is the whole point
2007-10-28T18:13:30  <johill> I don't think you've understood the API yet
2007-10-28T18:13:39  <xorAxAx> i have
2007-10-28T18:13:48  <xorAxAx> the issue is that you would need to write 2 classes
2007-10-28T18:13:58  <xorAxAx> and cant share any code at all
2007-10-28T18:14:20  <johill> again, what code could you possibly share?
2007-10-28T18:14:33  <xorAxAx> well, the whole code of the session handler
2007-10-28T18:14:45  <xorAxAx> except for a few lines
2007-10-28T18:15:03  <johill> it's 70 lines
2007-10-28T18:15:15  <xorAxAx> ...
2007-10-28T18:15:20  <xorAxAx> sigh
2007-10-28T18:15:39  <johill> look, there isn't anything to share
2007-10-28T18:15:55  <johill> except maybe the auth callling
2007-10-28T18:16:41  <johill> actually that's just a filter which you may or may not need
2007-10-28T18:17:09  <johill> and writing two classes rather than one that is twice as long isn't really a problem, is it?
2007-10-28T18:18:02  <johill> I'd rather have storage abstracted and session handling abstracted than an ad-hoc procedural interfaces that munches them together
2007-10-28T18:19:03  <xorAxAx> no, this means that the chosen abstraction is useless
2007-10-28T18:19:45  <johill> not at all!
2007-10-28T18:19:59  <johill> you could decide to use the default session handler (moin cookie) but store data into a db
2007-10-28T18:20:08  <johill> if the cache file based storage doesn't work out for some reason
2007-10-28T18:20:27  <xorAxAx> no, not if you need to reeuse existing sessions
2007-10-28T18:20:29  <xorAxAx> hmm
2007-10-28T18:21:50  <johill> and why should the storage backend decide on the ID? it's a simple id->dict storage
2007-10-28T18:22:05  <johill> anyway, back later, I'll have some dinner
2007-10-28T18:22:10  <xorAxAx> because there might be an existing session id
2007-10-28T18:22:14  <xorAxAx> in the other system
2007-10-28T18:22:35  <johill> well the separation is to support other storage, not other systems
2007-10-28T18:22:45  <johill> other systems need a different session handler and a different storage
2007-10-28T18:22:56  <xorAxAx> hmm, yeah
2007-10-28T18:29:01  <ThomasWaldmann> xorAxAx: johill misses your _ magic
2007-10-28T18:37:20  <ThomasWaldmann> someone has shrinked my quicklinsk X)
2007-10-28T18:50:14  <johill> ThomasWaldmann: when can we edit again? I want to remove the xmlrpc todo
2007-10-28T18:51:36  <ThomasWaldmann> johill: did you check the conversion result?
2007-10-28T18:52:44  <CIA-31> moin: Reimar Bauer <rb.proj AT googlemail DOT com> default * 2908:0bc6871c65de 1.7/MoinMoin/converter/text_html_text_moin_wiki.py: text_html_text_moin_wiki.process_inline: linebreaks as <<BR>> and not \n
2007-10-28T18:52:46  <CIA-31> moin: Reimar Bauer <rb.proj AT googlemail DOT com> default * 2909:c39426a679b4 1.7/MoinMoin/converter/_tests/test_text_html_text_moin_wiki.py: test_text_html_text_moin_wiki: further tests for table cells and <br \>
2007-10-28T18:52:47  <CIA-31> moin: Reimar Bauer <rb.proj AT googlemail DOT com> default * 2910:9dd063760883 1.7/MoinMoin/session.py: merged main
2007-10-28T18:53:29  <johill> what did you want checked?
2007-10-28T18:54:04  <dreimark> usually if we miss something
2007-10-28T18:54:04  <ThomasWaldmann> if everything works :P
2007-10-28T18:54:19  <ThomasWaldmann> page count should be ok
2007-10-28T18:54:27  * ThomasWaldmann found 1 bug
2007-10-28T18:55:22  <johill> markup?
2007-10-28T18:57:17  <dreimark> http://moinmo.in/action/recall/FeatureRequests/PackagesNeedsAddAttachmentAndMore
2007-10-28T18:57:42  <dreimark> inline is not inline guess mimetype ?
2007-10-28T18:58:19  <ThomasWaldmann> quicklinks are not renamed correctly
2007-10-28T18:58:41  <johill> oh ok
2007-10-28T19:00:56  <johill> is the old version still available?
2007-10-28T19:00:57  <CIA-31> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2270:2100e8d8d481 1.6/MoinMoin/request/__init__.py: fix AttributeError for timing_log
2007-10-28T19:00:58  <CIA-31> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2271:335246dfca9a 1.6/MoinMoin/ (action/SyncPages.py user.py wikisync.py xmlrpc/__init__.py): merged main
2007-10-28T19:01:10  <ThomasWaldmann> johill: moinmoin.wikiwikiweb.de
2007-10-28T19:01:19  <johill> that also says 'converted'
2007-10-28T19:01:34  <johill> ah I see
2007-10-28T19:02:12  <ThomasWaldmann> the red thing is the old (unchanged) wiki, the green is the new (with converted data)
2007-10-28T19:02:19  <johill> I'm confused by http://moinmoin.wikiwikiweb.de/FeatureRequests/MetaTagForUnderlayPages
2007-10-28T19:02:28  <johill> I used to think that two indented paragraphs would give two indented paragraphs
2007-10-28T19:02:44  <johill> but that's shown as a single one
2007-10-28T19:03:27  <xorAxAx> no, they have to be collapsed
2007-10-28T19:03:38  <xorAxAx> if you need two of them, you need to use thomas' new dot-syntax
2007-10-28T19:03:56  <johill> huh
2007-10-28T19:03:58  <johill> ok
2007-10-28T19:04:08  <johill> why do they have to be collapsed though?
2007-10-28T19:04:23  <ThomasWaldmann> btw, we are not searching bugs in the 1.5 code now
2007-10-28T19:04:28  <johill> an empty line would start a new paragraph most of the time
2007-10-28T19:04:31  <xorAxAx> hm
2007-10-28T19:04:38  <xorAxAx> ok, maybe :)
2007-10-28T19:04:42  <johill> yeah, I thought it was a new bug in 1.6 and then 1.5 having it that way confused me
2007-10-28T19:13:58  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2272:050731364b65 1.6/MoinMoin/user.py: revert b2930137a300, bad port from 1.7
2007-10-28T19:14:09  <xorAxAx> johill: please use backout :)
2007-10-28T19:14:22  <johill> ahh
2007-10-28T19:14:29  <johill> I was looking at revert and that didn't do it
2007-10-28T19:17:44  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2273:06eac3ce7f9a 1.6/MoinMoin/user.py: fix user.saveTrail() comment
2007-10-28T19:17:56  <johill> but hg just confuses me all the time since I'm used to git
2007-10-28T19:18:10  <johill> hg log doesn't pipe through less automatically, there's no 'git show' equivalent, ..
2007-10-28T19:18:45  <johill> (well hg export is a bit like git show)
2007-10-28T19:18:57  <ThomasWaldmann> johill: finished?
2007-10-28T19:19:03  <xorAxAx> well, write your command wrapper, johill :)
2007-10-28T19:19:05  <johill> well I reverted it
2007-10-28T19:19:14  <johill> I don't see a good way to not create trails for the first request
2007-10-28T19:19:16  <xorAxAx> or search for one
2007-10-28T19:19:21  <johill> but we should, and I'm s till looking for a way
2007-10-28T19:19:50  <ThomasWaldmann> ok, moin restarted
2007-10-28T19:22:17  <johill> the problem is that the API is badly abstracted in 1.6
2007-10-28T19:22:22  <johill> with auth handlers doing session stuff too
2007-10-28T19:22:30  <johill> and thus it's harder
2007-10-28T19:33:09  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2911:50e0afbdc870 1.7/MoinMoin/session.py: session handling: fix bug in docstring
2007-10-28T19:35:39  <johill> xorAxAx: here's an idea
2007-10-28T19:35:54  <johill> that new dimension, the persistent-ID handling, makes the session handling superfluous
2007-10-28T19:36:04  <johill> or rather makes it generic
2007-10-28T19:36:28  <johill> integration with a different session system would only need a different storage class and a different persistent-ID handler (different cookie or whatever)
2007-10-28T20:12:26  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2912:4be125bafc8a 1.7/MoinMoin/ (config/multiconfig.py request/__init__.py session.py): session handling: separate out an extra dimension for persistent IDs
2007-10-28T20:13:47  <johill> grzywacz: I think request.handle_jid_auth is wrong
2007-10-28T20:13:50  <johill> why is that in request?
2007-10-28T20:15:53  <dreimark> ThomasWaldmann: navigation on page title is broken
2007-10-28T20:15:56  <dreimark> http://moinmo.in/FeatureRequests/SubscribeUser
2007-10-28T20:16:16  <dreimark> its  only text no link
2007-10-28T20:18:59  <ThomasWaldmann> works for me
2007-10-28T20:19:48  <dreimark> you can click on Feature Requests?
2007-10-28T20:21:13  <ThomasWaldmann> yes
2007-10-28T20:22:09  <dreimark> strange, Help On Languages is text too
2007-10-28T20:22:43  <dreimark> bug
2007-10-28T20:23:13  <dreimark> "Add spaces to displayed wiki names" in preferences on and its text
2007-10-28T20:23:16  <dreimark> and no link
2007-10-28T20:23:52  * ThomasWaldmann wanted to kill that feature for a long time
2007-10-28T20:27:06  * ThomasWaldmann guesses the time has come :)
2007-10-28T20:30:36  <dreimark> mig script changes for users too?
2007-10-28T20:33:27  <ThomasWaldmann> ?
2007-10-28T20:36:51  <johill> huh
2007-10-28T20:36:57  <johill> does hg not delete files that are delete from the repo\?
2007-10-28T20:37:42  <johill> haha
2007-10-28T20:37:43  <johill> $ hg commit MoinMoin/session.py
2007-10-28T20:37:43  <johill> abort: No space left on device
2007-10-28T20:37:45  <johill> fun
2007-10-28T20:38:53  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2913:99d5fca40c1e 1.7/MoinMoin/session.py: fix session expiry update bug
2007-10-28T20:39:32  <dreimark> ThomasWaldmann: if you remove that flag from UserPref needs it be killed from user data ?
2007-10-28T20:39:42  <johill> dreimark: not really
2007-10-28T20:43:48  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2914:25ee42ee758a 1.7/MoinMoin/xmlrpc/__init__.py: tie xmlrpc into regular session/auth
2007-10-28T20:44:59  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2915:05d90815b0f9 1.7/MoinMoin/xmlrpc/__init__.py: xmlrpc: fault on save errors
2007-10-28T20:45:13  <dreimark> all the MoinMoin links could be changed to the new url
2007-10-28T20:45:56  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2916:06547d7df727 1.7/MoinMoin/xmlrpc/__init__.py: fix putPage comment
2007-10-28T20:48:59  <CIA-31> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2274:092c10f4c7b6 1.6/MoinMoin/theme/__init__.py: fix pagename navigation for wikiname_add_spaces == True
2007-10-28T20:49:47  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2917:093ffb70b26f 1.7/MoinMoin/events/emailnotify.py: fix email notification bug
2007-10-28T20:49:56  <johill> grzywacz: grrrr
2007-10-28T20:52:45  <dreimark> subminor problem
2007-10-28T20:53:20  <dreimark> I have a link to that page DeletePageUsesInsecure
2007-10-28T20:53:45  <dreimark> if I go to that page http://moinmo.in/MoinMoinBugs/DeletePageUsesInsecure
2007-10-28T20:54:24  <dreimark> it does not exist (the name name has to end with Validation)
2007-10-28T20:55:19  <dreimark> while on wikiwikiweb I get the editbar and could remove the link I cant do this directly on moinmo.in
2007-10-28T20:56:08  <dreimark> but if I enter an unknown action to the url i get the editbar
2007-10-28T20:56:28  <ThomasWaldmann> some people had the opinion that editbar should not be shown for nonexisting pages
2007-10-28T20:56:51  <johill> I get the edit bar on both
2007-10-28T20:57:21  <CIA-31> moin: Johannes Berg <johannes AT sipsolutions DOT net> default * 2918:404b1887af6c 1.7/MoinMoin/events/emailnotify.py: repair the email notification bugfix
2007-10-28T20:57:31  <dreimark> now I get it too
2007-10-28T20:57:41  <ThomasWaldmann> %)
2007-10-28T21:00:42  <johill> is that an option? :)
2007-10-28T21:02:03  <dreimark> I can tell how you get the edit bar with a new page and I think thats the reason why I gt
2007-10-28T21:02:11  <dreimark> get the editbar with the non existing page
2007-10-28T21:03:14  <dreimark> If you enter action=refresh on a nonexisting page I guess a cache file is written
2007-10-28T21:03:37  <johill> huh
2007-10-28T21:03:51  <ThomasWaldmann> ok, so what's the conclusion? do we live and fix the minor problems or go back to 1.5 and try again later?
2007-10-28T21:04:11  <ThomasWaldmann> (everybody would have to fix his quicklinks manually)
2007-10-28T21:04:11  <johill> most of the problems so far seem to be in the code
2007-10-28T21:04:23  <johill> what's with quicklinks?
2007-10-28T21:04:40  <ThomasWaldmann> they didn't get renamed if the page got renamed
2007-10-28T21:04:52  <johill> you mean due to spaces
2007-10-28T21:05:11  <ThomasWaldmann> any rename I did, mostly spaces
2007-10-28T21:05:38  <dreimark> we need this platform for further tests so I think we use it with 1.6
2007-10-28T21:05:49  <ThomasWaldmann> usually it is fixed by s/_/ /
2007-10-28T21:05:59  <johill> were there any other renames?
2007-10-28T21:06:05  <johill> under what circumstances?
2007-10-28T21:06:20  <johill> in any case, I think it's totally obvious for users
2007-10-28T21:06:25  <ThomasWaldmann> i fixed some typos also
2007-10-28T21:06:33  <johill> since their links just go dead and they're given hints on what could be the right page
2007-10-28T21:07:20  <johill> I guess we should add it to the converter though
2007-10-28T21:07:46  <ThomasWaldmann> iirc it has code for that, but maybe a bug
2007-10-28T21:07:54  <johill> ah
2007-10-28T21:08:04  <ThomasWaldmann> ehrm, no, my fault
2007-10-28T21:08:43  <ThomasWaldmann> I'll fix it by running converter again, with correct interwiki name in config X)
2007-10-28T21:08:55  <johill> heh
2007-10-28T21:09:39  <ThomasWaldmann> keep testing meanwhile :)
2007-10-28T21:09:54  <johill> :)
2007-10-28T21:10:07  <johill> I haven't found anything so far except for that search thing
2007-10-28T21:20:55  <dreimark> is it somewhere configurable how much item in search results are listed?
2007-10-28T21:24:15  <dreimark> I have less quicklinks
2007-10-28T21:28:15  <dreimark> no they are shorter written
2007-10-28T21:28:28  <ThomasWaldmann> dreimark: yeah, i noticed that also
2007-10-28T21:29:43  <dreimark> ThomasWaldmann: looks like exchanged with trail
2007-10-28T21:29:57  <dreimark> those are complete now
2007-10-28T21:31:11  <dreimark> or no, the ones I looked werent not long enough
2007-10-28T21:31:52  <ThomasWaldmann> btw, not every change is a bug :)
2007-10-28T21:33:32  <dreimark> yeah
2007-10-28T21:43:49  <dreimark> ThomasWaldmann: http://moinmo.in/CategoryFeatureRequest is empty
2007-10-28T21:44:38  <xorAxAx> well, that might be still the search bug
2007-10-28T21:44:43  <xorAxAx> is there any negation in it?
2007-10-28T21:45:34  <dreimark> yes -title:regex:^Category
2007-10-28T21:47:31  <xorAxAx> then thats the issue
2007-10-28T21:47:51  <xorAxAx> hmm, i wonder when we should the search bug anniversary into the calendar
2007-10-28T21:47:54  <xorAxAx> +place
2007-10-28T21:54:55  <johill> ok so I have many instances of ["asdf" defg]
2007-10-28T21:54:58  <johill> how can I convert those?
2007-10-28T22:24:49  <johill> hm ok I got it
2007-10-28T22:36:39  <dreimark> gn
2007-10-28T22:42:13  <ThomasWaldmann> ok, 2nd converter run finished, look at your quicklinks now
2007-10-28T22:43:02  <johill> I don't have any
2007-10-28T22:45:08  <johill> hmm
2007-10-28T22:45:15  <johill> I thought --wiki-url should use farmconfig then?
2007-10-28T22:48:45  <ThomasWaldmann> it uses whatever config there normally is used
2007-10-28T22:48:54  <ThomasWaldmann> maybe you need --config-dir=...
2007-10-28T22:49:40  <johill> yeah
2007-10-28T22:50:11  <johill> I thought '.' would be the default config dir
2007-10-28T22:51:03  <CIA-31> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> default * 2919:4cd85d841bbd 1.7/MoinMoin/session.py: Remove trailing whitespace.
2007-10-28T22:52:59  <ThomasWaldmann> ok, if nobody objects, I'll switch moinmo.in to rw now
2007-10-28T22:53:12  <johill> fine with me
2007-10-28T22:59:16  <johill> any ocnfig option to log errors to file?
2007-10-28T23:01:46  <ThomasWaldmann> johill: see the Config class of your server adaptor
2007-10-28T23:02:07  <ThomasWaldmann> ok, mm site is rw again
2007-10-28T23:02:20  <TheSheep> good bye 1.5.x
2007-10-28T23:06:03  <johill> :)
2007-10-28T23:06:09  <ThomasWaldmann> (and goodbye moinmoin.wikiwikiweb.de)
2007-10-28T23:07:00  <johill> moin relies on india now :P
2007-10-28T23:07:45  <grzywacz> :D
2007-10-28T23:12:15  <xorAxAx> not sure if that is a good idea
2007-10-28T23:12:31  <xorAxAx> maybe a redirect would be better?
2007-10-28T23:12:36  <xorAxAx> a permanent one
2007-10-28T23:12:45  <xorAxAx> in order to get bookmarks and spiders updatd
2007-10-28T23:14:02  <TheSheep> a small '301 moved permanently'
2007-10-28T23:14:19  <xorAxAx> yes
2007-10-28T23:14:40  <johill> wow
2007-10-28T23:14:46  <johill> linuxwireless.org event log is 85M
2007-10-28T23:14:53  <TheSheep> hehe, '402 payment required' :D
2007-10-28T23:15:08  <xorAxAx> TheSheep: thats when moin will have gone web 3.0
2007-10-28T23:15:15  <xorAxAx> every wiki edit 1 ct.
2007-10-28T23:15:43  <TheSheep> nah, views 1eur, edits -1eur
2007-10-28T23:16:16  <xorAxAx> hmm
2007-10-28T23:16:35  <TheSheep> hey, you could throw a 406 on unicode decode erros in urls! :)
2007-10-28T23:16:50  <johill> 406?
2007-10-28T23:16:53  <TheSheep> or do I read this wrong?
2007-10-28T23:17:06  <TheSheep> 406: ('Not Acceptable', 'URI not available in preferred format.')
2007-10-28T23:17:18  <johill> I think that's when you have accept-* headers
2007-10-28T23:17:24  <TheSheep> ah, no, it's for content format negotiation
2007-10-28T23:17:30  <TheSheep> yup
2007-10-28T23:18:20  <xorAxAx> mathematicians talking about http result codes ... :)
2007-10-28T23:19:02  <TheSheep> xorAxAx: 417
2007-10-28T23:19:53  <xorAxAx> haha
2007-10-28T23:21:04  <johill> ThomasWaldmann: if there are no pages, no rename1.txt is created
2007-10-28T23:21:07  <johill> that's a bug, no?
2007-10-28T23:23:31  <ThomasWaldmann> ehrm, well, i guess it better should be 0 bytes then
2007-10-28T23:23:50  <johill> yeah, I've often had empty files but once or twice no files now
2007-10-28T23:23:57  <johill> not sure under what circumstances exaclty
2007-10-28T23:24:42  <johill> huh
2007-10-28T23:24:45  <johill> converter takes much mem
2007-10-28T23:24:49  <johill> 253m 196m
2007-10-28T23:26:26  * ThomasWaldmann .oO(johill does not want to see my first tries and their memory consumption)
2007-10-28T23:26:53  <johill> is this normal?
2007-10-28T23:27:06  <ThomasWaldmann> maybe depends on your pagecount
2007-10-28T23:27:13  <johill> huh ok why that?
2007-10-28T23:27:26  <ThomasWaldmann> because it keeps names in memory
2007-10-28T23:27:29  <johill> only has about 250 pages
2007-10-28T23:27:41  <ThomasWaldmann> (and pathes)
2007-10-28T23:27:44  <xorAxAx> it calculates scrabble boards and solves them
2007-10-28T23:28:00  <johill> and on the second to last it starts using more and more mem
2007-10-28T23:28:02  <ThomasWaldmann> it doesnt keep content in memory :)
2007-10-28T23:28:17  <xorAxAx> ThomasWaldmann: then how does that some up to 250 MB?
2007-10-28T23:28:20  <xorAxAx> 250 kb would be ok
2007-10-28T23:28:41  <ThomasWaldmann> no idea
2007-10-28T23:28:55  <xorAxAx> ...
2007-10-28T23:29:20  <johill> ohh
2007-10-28T23:29:29  <johill> event log
2007-10-28T23:29:36  <johill> but it's only 56M
2007-10-28T23:29:52  <xorAxAx> well, there are two choices - either it has a lot of cyclic refs and the gc is doing badly and too seldom
2007-10-28T23:29:56  <xorAxAx> or its a bug :)
2007-10-28T23:30:24  <johill> well it looks like it uses a lot of mem to convert the event log
2007-10-28T23:30:47  <johill> about 5x event log size I guess
2007-10-28T23:31:07  <xorAxAx> hmm, easily verifyable by trimming it and doing it again in a scratch dir :)
2007-10-28T23:32:15  <johill> I'm just going to hope it'll finish
2007-10-28T23:32:44  <johill> but I doubit it
2007-10-28T23:32:45  <johill> it wants 264M mem
2007-10-28T23:32:47  <johill> mory and I have 256
2007-10-28T23:33:25  * johill kills again
2007-10-28T23:34:23  <johill> can't it be converted without reading into mem?
2007-10-28T23:35:25  <xorAxAx> moin is enterprisey, less memory usage would not fit into the marketing strategy, johill
2007-10-28T23:35:36  <johill> oh sorry
2007-10-28T23:36:07  <xorAxAx> well, ThomasWaldmann says that he doesnt know why it needs so much mem
2007-10-28T23:36:18  <xorAxAx> and he said that it doesnt read the file contents
2007-10-28T23:38:11  <johill> huh?
2007-10-28T23:38:17  <johill> it reads the whole event log into memory
2007-10-28T23:38:23  <xorAxAx> 23:28:02 < ThomasWaldmann> it doesnt keep content in memory :)
2007-10-28T23:38:35  <xorAxAx> hmm, keep is suspicious
2007-10-28T23:38:43  <ThomasWaldmann> xorAxAx: with "content" i meant wiki (page) content
2007-10-28T23:39:07  <TheSheep> nobody expects the Spanish Inquisition!
2007-10-28T23:39:13  <johill> well, I'm tryihng to make it work in place
2007-10-28T23:39:29  <xorAxAx> johill: it should work linewise of course
2007-10-28T23:39:44  <johill> yes that's what I mean
2007-10-28T23:40:18  <xorAxAx> so it would need 6 GB for the debian wiki?
2007-10-28T23:40:42  <johill> hmm
2007-10-28T23:40:46  <johill> this doesn't seem to have worked
2007-10-28T23:40:48  <johill> or the gc sucks
2007-10-28T23:41:45  <xorAxAx> if its just dealing with strings, the GC is working pretty fine
2007-10-28T23:41:50  <johill> or hmm
2007-10-28T23:41:52  <xorAxAx> because strings cant have any cyclic refs
2007-10-28T23:41:58  <johill> does 'for line in some_file:' read?
2007-10-28T23:42:02  <johill> linewise?
2007-10-28T23:42:04  <xorAxAx> yes
2007-10-28T23:42:08  <TheSheep> "this sentence is false"
2007-10-28T23:42:17  <johill> hmm
2007-10-28T23:42:47  <johill> I only see reads when I sstrace
2007-10-28T23:42:48  <johill> no writes
2007-10-28T23:42:51  <johill> and my loop does both
2007-10-28T23:42:53  <johill> are you sure?
2007-10-28T23:42:57  <CIA-31> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2275:41f93027964f 1.6/ (25 files in 13 dirs): replace wikiwikiweb.de by moinmo.in, replace contact Juergen Hermann by Thomas Waldmann, cleanup wikiconfig
2007-10-28T23:43:05  <TheSheep> johill: cache?
2007-10-28T23:43:11  <johill> TheSheep: ?
2007-10-28T23:43:29  <TheSheep> johill: do you read and write to the same file at the same time?
2007-10-28T23:43:43  <johill> no of course not
2007-10-28T23:43:49  <xorAxAx> johill: well, its bufferin internally
2007-10-28T23:43:56  <johill> how much? :)
2007-10-28T23:43:59  <xorAxAx> but they should be bounded
2007-10-28T23:43:59  <johill> more than 100M
2007-10-28T23:44:28  <johill> ahh wait I applied the patch to the wrong code :)
2007-10-28T23:44:36  <TheSheep> (:
2007-10-28T23:46:08  <johill> yup, takes lots of CPU but only 10M memory now
2007-10-28T23:47:17  <ThomasWaldmann> johill: there is an easy solution for event-log X)
2007-10-28T23:47:49  <johill> http://johannes.sipsolutions.net/patches/moin/conv-event-log.patch
2007-10-28T23:51:50  <ThomasWaldmann> johill: commit it after you have tested it
2007-10-28T23:52:05  <johill> well it converted the wiki with the 56M event log
2007-10-28T23:52:31  <ThomasWaldmann> did you test a wiki without event-log also?
2007-10-28T23:52:40  <johill> no
2007-10-28T23:53:56  <johill> hmm
2007-10-28T23:53:58  <johill> found a conversion bug
2007-10-28T23:54:01  <johill> [#red-square You can also jump to a red square below.]
2007-10-28T23:56:01  <johill> or was that new syntax in 1.6-pre
2007-10-28T23:56:05  <johill> now I got a bug
2007-10-28T23:56:11  <johill> UnicodeDecodeError: 'utf8' codec can't decode byte 0x99 in position 3610: unexpected code byte
2007-10-28T23:56:18  <xorAxAx> oh
2007-10-28T23:56:59  <ThomasWaldmann> johill: add it to the 1.6.0 todo
2007-10-28T23:57:18  <johill> the [#?
2007-10-28T23:57:29  <johill> hmm
2007-10-28T23:57:30  <ThomasWaldmann> johill: (yes) do you have iso content in your wiki
2007-10-28T23:57:36  <johill> the unicode error is in the event log
2007-10-28T23:57:37  <johill> weird
2007-10-28T23:58:07  <ThomasWaldmann> is there non-ascii stuff?
2007-10-28T23:58:19  <johill> hmmmm
2007-10-28T23:58:24  <johill> now I redid it and it worked?!
2007-10-28T23:58:39  <johill> what is not coipied into data.pre160?
2007-10-28T23:59:14  <ThomasWaldmann> the pre160 dir is just a rename of "data"
2007-10-28T23:59:20  <johill> oh wait n/m
2007-10-28T23:59:34  <ThomasWaldmann> (and then it outputs the conversion result to a new "data" dir)
2007-10-28T23:59:48  <johill> wrong wiki