2008-07-14T00:05:26 <ThomasWaldmann> gn dreimark
2008-07-14T00:06:39 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3965:0ea00cfffdf3 1.8-dom-bblank/MoinMoin/formatter/compatibility.py: Compatibility formatter - Really handle bytestrings in html parser
2008-07-14T00:06:40 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3966:981cde73d877 1.8-dom-bblank/MoinMoin/Page.py: Page - Ignore formatter instances as argument
2008-07-14T00:29:02 * ThomasWaldmann checked that there are no forward ports to do from 1.6 -> 1.7
2008-07-14T02:12:06 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3835:9e960a76d2c8 1.8/MoinMoin/ (parser/_tests/test_unicode.py _tests/test_parser_unicode.py): move parser test to the right place, small cosmetic fixes
2008-07-14T02:25:44 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3749:a37ed69fafed 1.7/MoinMoin/parser/ (text_xslt.py _tests/test_unicode.py):
2008-07-14T02:25:44 <CIA-53> MoinMoinBugs/DoNotConvertUnicodeToUTF8ForXsltParser (backport from 1.8)
2008-07-14T02:25:44 <CIA-53> Initial patch applied and slightly improved.
2008-07-14T02:25:44 <CIA-53> Moved to correct place.
2008-07-14T02:42:13 <birkenfeld> moin
2008-07-14T02:42:30 <emgent> moin
2008-07-14T02:42:50 <xorAxAx> ThomasWaldmann: please refer to dreimark
2008-07-14T02:43:04 <birkenfeld> hey xorAxAx
2008-07-14T02:43:06 <birkenfeld> how are you?
2008-07-14T02:43:12 <xorAxAx> birkenfeld: stinky
2008-07-14T02:43:22 <birkenfeld> smoking too much crack?
2008-07-14T02:43:30 <xorAxAx> yeah
2008-07-14T02:44:29 <ThomasWaldmann> moin :)
2008-07-14T02:44:59 <xorAxAx> ThomasWaldmann: and to cut it short - i didnt manage to talk to her about it
2008-07-14T02:45:20 <ThomasWaldmann> ok
2008-07-14T02:45:29 <birkenfeld> it would have required 5 minutes of a quiet, receptive leslie
2008-07-14T02:45:40 <birkenfeld> which was impossible to find in the last days
2008-07-14T02:46:07 <xorAxAx> ThomasWaldmann: mainly timewise because reimar notified me too late and she was pretty busy that day until 1:30 am
2008-07-14T02:46:21 <xorAxAx> ThomasWaldmann: and that was the last time i saw her
2008-07-14T02:46:38 <ThomasWaldmann> birkenfeld: you were (are) also in mountain view?
2008-07-14T02:47:14 <birkenfeld> yes
2008-07-14T02:47:54 <ThomasWaldmann> nice :) /me just returned from vilnius yesterday. :)
2008-07-14T02:48:54 <birkenfeld> did you meet some interesting people?
2008-07-14T02:49:40 <emgent> night.
2008-07-14T02:50:50 <ThomasWaldmann> birkenfeld: sure :) maybe we do a Stuttgarter PIG now :)
2008-07-14T02:54:33 <birkenfeld> hm, then I wish you better luck than with our Munich one
2008-07-14T02:56:08 <xorAxAx> ThomasWaldmann: the program survey is still missing
2008-07-14T02:56:20 <xorAxAx> ThomasWaldmann: please submit it on time, not many hours left
2008-07-14T02:57:36 <ThomasWaldmann> it is submitted since long
2008-07-14T02:57:51 <ThomasWaldmann> maybe YOU need to submit yours? :)
2008-07-14T02:58:14 <xorAxAx> um :)
2008-07-14T02:58:32 <xorAxAx> i thoiught it was once per project, ok
2008-07-14T03:04:44 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3750:a7db3c488083 1.7/MoinMoin/config/multiconfig.py: prevent CategoryTemplate being listed as a category (it is a Template, but matched also the category regex) (backport from 1.8)
2008-07-14T03:20:38 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3751:e676cbf8b1e3 1.7/MoinMoin/wikiutil.py: wikiutil.importPlugin: support getting whole plugin module object by giving function=None
2008-07-14T03:36:43 <ThomasWaldmann> dreimark: if we can test sendcached enough this week (so that api can be considered stable), we could include it in 1.7.1 next weekend
2008-07-14T07:48:51 <mmihaljevic> moin
2008-07-14T08:21:27 <ThomasWaldmann> moin
2008-07-14T08:33:54 <dreimark> moin
2008-07-14T08:37:46 <dreimark> ThomasWaldmann: ok, great, will do the changes in arnica soon
2008-07-14T08:44:40 <dreimark> http://www.onyame.de/article.php?story=20080708223720231
2008-07-14T09:19:12 * dreimark has the feeling sendcached chould be named itemcache or somehow similiar, because sendcached.put_cache sounds strange
2008-07-14T09:19:26 <dreimark> s/chould/should/
2008-07-14T09:22:25 <dreimark> itemcache.put, itemcache.get_datafile, itemcache.get_header, itemcache.send, itemcache.key, itemcache.is_cached sounds better
2008-07-14T09:40:39 <dreimark> ThomasWaldmann: current cache key can only handle one itemcache file per attachment
2008-07-14T09:42:06 <dreimark> hmm, I can use the secret to destinguish between thumb and webnail
2008-07-14T09:44:47 <mmihaljevic> re and is going to work on request/config problem
2008-07-14T09:46:22 <johill> PawelPacana: there isn't really, but we do allow creating empty revisions so why should we arbitrarily disallow multiple empty revisions in a row?
2008-07-14T10:12:23 <mmihaljevic> dreimark: I saw you message I seen http://www.web2ldap.de/ already
2008-07-14T10:13:18 <mmihaljevic> dreimark: I done some ldap groupping before I get to work on this now, but I will continue after I solve the request
2008-07-14T10:33:45 <dreimark> mmihaljevic: ok, that was only a hint for a nice tool I found searching for something else
2008-07-14T10:34:08 <mmihaljevic> dreimark: yes it is a great tool
2008-07-14T10:34:38 <dreimark> if it is useful you can mentione it on the description page of the ldap group backend
2008-07-14T10:34:42 <mmihaljevic> and the author of that tool is the person that replays to most of the questions posted on python-ldap-dev list
2008-07-14T10:34:48 <dreimark> :)
2008-07-14T10:35:04 <mmihaljevic> dreimark: yes sutre I'll put the link there
2008-07-14T10:38:14 <mmihaljevic> dreimark: at the moment I'm playing a bit with this http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/86900/index_txt in order to solve request like that somehow
2008-07-14T10:39:19 <mmihaljevic> but also I'm not sure if that's a good idea and I think I shouldn't stick to some implementation rather to a way I think how things need to work
2008-07-14T11:10:33 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3967:dee609a017c8 1.8-dom-bblank/MoinMoin/formatter/compatibility.py: Compatibility formatter - Support rule, nop smiley
2008-07-14T11:10:33 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3968:1b5cbe3dc950 1.8-dom-bblank/MoinMoin/formatter/compatibility.py: Compatibility formatter - Implement parser, nop several methods
2008-07-14T11:10:37 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3969:fca86541a07c 1.8-dom-bblank/MoinMoin/formatter/compatibility.py: Compatibility formatter - Use wiki macro handling code
2008-07-14T11:25:18 <mmihaljevic> dreimark: http://paste.pocoo.org/show/79390/ there is no code in in it only idea and probably some parametres are missing, will try to make some simple example not connected to MM that works
2008-07-14T11:56:51 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3970:6ffbf366bb0f 1.8-dom-bblank/MoinMoin/converter2/ (_tests/test_html_out.py html_out.py): HTML output converter - Copy style attributes into CSS formatted style attribute
2008-07-14T11:56:53 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3971:e4c086d42fe7 1.8-dom-bblank/MoinMoin/formatter/compatibility.py: Compatibility formatter - Support big and small
2008-07-14T12:04:09 <waldi> ThomasWaldmann: even without output caching it is only slightly slower, 10.5 vs. 11.2 request/s
2008-07-14T12:13:30 <waldi> and the second most function is _serialize_xml ...
2008-07-14T12:18:19 <mmihaljevic> dreimark: forget my last pastebin
2008-07-14T12:18:33 <waldi> socket.py:245(flush) <- request_standalone.py:87(flush)(198) 0.001
2008-07-14T12:18:36 <waldi> request_standalone.py:90(finish)(99) 0.011
2008-07-14T12:18:43 <waldi> hmm, i asked ab to do _100_ requests
2008-07-14T12:24:24 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3972:8274929aecda 1.8-dom-bblank/MoinMoin/Page.py: Page - Update cache only if something new is available
2008-07-14T12:24:25 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3973:b81175baeb28 1.8-dom-bblank/MoinMoin/Page.py: Page - Caching was unset by accident, fix it
2008-07-14T12:29:20 <mmihaljevic> dreimark: I think this would work ok http://paste.pocoo.org/show/79397/ (something like this)
2008-07-14T12:31:47 <PawelPacana> finally back home ;)
2008-07-14T12:34:51 <dreimark> PawelPacana: :), we were quite lucky that we got the proposed ICE after the accident in cologne
2008-07-14T12:40:02 <PawelPacana> accident?
2008-07-14T12:43:34 <PawelPacana> i managed to hitchhike up to Poznan, then took a train :>
2008-07-14T12:45:37 <dreimark> a train is derailed
2008-07-14T12:45:45 <dreimark> An axle was broken
2008-07-14T12:48:39 <mmihaljevic> dreimark: when was the accident?
2008-07-14T12:51:26 <dreimark> on wednesday
2008-07-14T12:51:49 <TheSheep> a sheep standing on the track again?
2008-07-14T12:51:57 <dreimark> no
2008-07-14T12:52:24 <dreimark> I don't find an engl. article http://www.spiegel.de/wirtschaft/0,1518,565264,00.html
2008-07-14T12:52:31 <johill> ok...
2008-07-14T12:52:37 <johill> I found why sometimes I get two mails
2008-07-14T12:53:07 <dreimark> johill: why ?
2008-07-14T12:53:24 <johill> it's the create+send email button
2008-07-14T12:53:33 <johill> it does something weird
2008-07-14T12:53:35 <johill> but we should remove it anyway
2008-07-14T12:55:17 <johill> aha yes
2008-07-14T12:55:34 <johill> the issue is that when you click create+email it calls theuser.mailAccountData()
2008-07-14T12:55:44 <johill> on a user object that was just created and not loaded
2008-07-14T12:55:52 <johill> so the issue really is that .save() doesn't reset
2008-07-14T13:01:15 <johill> http://moinmo.in/MoinMoinBugs/DuplicateNewUserNotification
2008-07-14T13:14:52 <ThomasWaldmann> waldi: which page did you use for measuring?
2008-07-14T13:15:44 <johill> cia down again?
2008-07-14T13:15:52 <johill> is there a way to do those notifications non-blockingly?
2008-07-14T13:15:59 <johill> it's annoying that hg push doesn't return
2008-07-14T13:22:17 <mmihaljevic> dreimark: where to define BackendFactory class?
2008-07-14T13:28:03 <ThomasWaldmann> dreimark: if you need 2 cached rendered items per original item, you can just prepend something to the key
2008-07-14T13:28:32 <CIA-53> Johannes Berg <johannes AT sipsolutions DOT net> default * 3845:eb8df08f88f5 1.8/MoinMoin/ (user.py wikiutil.py): merge 1.7
2008-07-14T13:28:32 <CIA-53> Johannes Berg <johannes AT sipsolutions DOT net> default * 3843:42c4d8d6a80c 1.8/ (3 files in 3 dirs): merge some fixes from 1.7
2008-07-14T13:28:32 <CIA-53> Johannes Berg <johannes AT sipsolutions DOT net> default * 3844:5c91b9e2b9fc 1.8/MoinMoin/config/multiconfig.py: no-op merge a7db3c488083 (was a backport from 1.8)
2008-07-14T13:28:57 <CIA-53> Johannes Berg <johannes AT sipsolutions DOT net> default * 3846:92ae3d3f0a31 1.8/MoinMoin/action/newaccount.py: newaccount: remove useless create+email button
2008-07-14T13:29:54 <dreimark> johill: sorry - late, but it is not totally useless.
2008-07-14T13:30:06 <dreimark> mmihaljevic: which options do you have
2008-07-14T13:31:00 <mmihaljevic> dreimark: mean where can I put them or?
2008-07-14T13:31:00 <dreimark> johill: without the recovery token it makes sense to tell on which wiki you have created your account
2008-07-14T13:31:05 <dreimark> mmihaljevic: yeah
2008-07-14T13:31:46 <mmihaljevic> well, I think I can put it somewhere in backend code or put somewhere in request.__init__
2008-07-14T13:31:50 <mmihaljevic> I think
2008-07-14T13:32:03 <mmihaljevic> dreimark:
2008-07-14T13:32:14 <dreimark> backends.__init__
2008-07-14T13:32:18 <dreimark> is a nice place
2008-07-14T13:32:32 <mmihaljevic> dreimark: ok will make there and test it if it works
2008-07-14T13:32:33 <mmihaljevic> thy
2008-07-14T13:32:36 <mmihaljevic> thx
2008-07-14T13:33:48 <dreimark> btw. I do like singular name for backend dir. (not plural)
2008-07-14T13:34:04 <dreimark> we don't have converters, parsers ... as dir
2008-07-14T13:34:16 <mmihaljevic> dreimark: ok
2008-07-14T13:34:40 <dreimark> do this by a changeset whitoot any other change "hg rename"
2008-07-14T13:40:43 <ThomasWaldmann> xorAxAx: dennda: a full week no commits?
2008-07-14T13:41:42 <johill> ThomasWaldmann: did you mean me? ;)
2008-07-14T13:42:19 <ThomasWaldmann> no, i was trying to address the students, but moon is not here
2008-07-14T13:42:27 <johill> ah heh
2008-07-14T13:42:40 <johill> anyway dennda has an exam today or tomorrow or something
2008-07-14T13:42:41 <ThomasWaldmann> but as a mentor, you can also feel addressed :)
2008-07-14T13:42:56 <johill> and he has some uncommitted stuff *grmbl*
2008-07-14T13:43:08 <johill> anyway, gtg
2008-07-14T13:43:15 <mmihaljevic> dreimark: ok I done this
2008-07-14T13:43:21 <waldi> ThomasWaldmann: hmm, not sure
2008-07-14T13:43:49 <ThomasWaldmann> waldi: maybe the difference depends on length and complexity of page
2008-07-14T13:45:12 <waldi> ThomasWaldmann: i redid it with the same, StartSeite from the underlay. 14 vs 16.5
2008-07-14T13:45:59 <ThomasWaldmann> does action=content or =print work?
2008-07-14T13:46:16 <ThomasWaldmann> so you get rid of the theme (and the time needed for it)
2008-07-14T13:47:58 <waldi> HelpOnMoinWikiSyntax needs more time ...
2008-07-14T13:48:05 <ThomasWaldmann> (for the old code, rendering the theme often was more expensive than the content)
2008-07-14T13:48:58 <waldi> 1.9 vs 2.9
2008-07-14T13:49:22 <ThomasWaldmann> without theme?
2008-07-14T13:49:27 <waldi> yes
2008-07-14T13:49:57 <ThomasWaldmann> which number is for what?
2008-07-14T13:50:30 <waldi> 1.9 for my code, 2.9 for the current 1.8 branch
2008-07-14T13:51:23 <ThomasWaldmann> nice
2008-07-14T13:51:32 <ThomasWaldmann> your code with caching I assume
2008-07-14T13:51:56 <waldi> both variants with caching
2008-07-14T13:52:38 <ThomasWaldmann> hmm, why is it faster?
2008-07-14T13:53:33 <ThomasWaldmann> ehrm, no, new code is a bit slower (req/s)
2008-07-14T13:54:40 <waldi> it is slower. lets see what the profiler wants to tell us
2008-07-14T13:55:52 <waldi> loading 5MiB of hotshot output takes a lot of time ...
2008-07-14T13:56:37 <waldi> ups, its 40MiB input ...
2008-07-14T13:57:45 <ThomasWaldmann> you're testing on lo interface?
2008-07-14T13:57:50 <waldi> yes
2008-07-14T13:58:37 <ThomasWaldmann> ok, then streaming vs. spilling out all at once maybe should not make a big difference
2008-07-14T13:59:08 <ThomasWaldmann> (that'll get worse with real transmission speeds maybe)
2008-07-14T13:59:42 <waldi> 328581 11.120 0.000 11.120 0.000 socket.py:245(flush)
2008-07-14T13:59:42 <waldi> 147609/99 5.015 0.000 27.097 0.274 ElementTree.py:766(_serialize_xml)
2008-07-14T14:00:10 <ThomasWaldmann> titles?
2008-07-14T14:00:18 <waldi> ncalls tottime percall cumtime percall filename:lineno(function)
2008-07-14T14:00:39 <waldi> the first number is weird, 300k flushs for nothing
2008-07-14T14:01:02 <waldi> Function was called by...
2008-07-14T14:01:02 <waldi> socket.py:245(flush) <- request_standalone.py:90(finish)(99) 0.015
2008-07-14T14:01:05 <waldi> socket.py:254(write)(328482) 13.410
2008-07-14T14:02:08 <ThomasWaldmann> socket write calls flush?
2008-07-14T14:02:49 <waldi> if (self._wbufsize == 0 or
2008-07-14T14:02:49 <waldi> self._wbufsize == 1 and '\n' in data or
2008-07-14T14:02:49 <waldi> self._get_wbuf_len() >= self._wbufsize):
2008-07-14T14:02:49 <waldi> self.flush()
2008-07-14T14:04:22 <ThomasWaldmann> looks like some self._wbufsize tuning is needed :)
2008-07-14T14:04:36 <waldi> lets see where this setting comes from
2008-07-14T14:06:18 <ThomasWaldmann> is that in a loop within write()?
2008-07-14T14:06:48 <ThomasWaldmann> because if you have 100 requests only and you do only 100 writes, how does it do 328000 flushes?
2008-07-14T14:07:11 <ThomasWaldmann> brb
2008-07-14T14:12:33 <waldi> zsh: segmentation fault PYTHONPATH=/home/bastian/moin/emeraldtree ./wikiserver.py
2008-07-14T14:12:38 <waldi> ui
2008-07-14T14:22:05 <ThomasWaldmann> oops
2008-07-14T14:23:02 <ThomasWaldmann> dreimark: i guess i rename action=sendcached to just action=cache,
2008-07-14T14:23:21 <dreimark> :)
2008-07-14T14:23:30 <ThomasWaldmann> and add some subactions like get, del, ...
2008-07-14T14:23:38 <dreimark> yeah
2008-07-14T14:24:27 <dreimark> currently i try to figure out how I can get a file like object from pil without saving it somewhere
2008-07-14T14:24:52 <waldi> ThomasWaldmann: clear cache should clean all cache keys
2008-07-14T14:26:49 <ThomasWaldmann> for that caching module needs something like clear_arena()
2008-07-14T14:26:56 <waldi> yep
2008-07-14T14:28:33 <waldi> hmm, is bla[x] faster than bla.__getitem__(x)?
2008-07-14T14:29:57 <johill> waldi: hmm, no, if anything it's probably slower, but it's far cleaner too :)
2008-07-14T14:32:41 <waldi> ThomasWaldmann: if i set the buffer higher, it even gets slower because socket appends strings
2008-07-14T14:36:31 <ThomasWaldmann> hmpf
2008-07-14T14:41:20 <dreimark> ThomasWaldmann: I'd prefer in cache_key an additional parameter instead of adding a string to the result. Because of escaping and it can be added to the hmac and will not tell something about the functionality it is used for.
2008-07-14T14:47:07 <ThomasWaldmann> i don't think we should add that subkey to the hmac input
2008-07-14T14:47:45 <ThomasWaldmann> because then you have to calculate the hmac for each cache object that comes from same source object
2008-07-14T14:48:48 <ThomasWaldmann> (but nevertheless I can add a optional prefix='' to cache_key that will be just prefixed to the result and must not contain invalid chars for urls
2008-07-14T14:48:51 <ThomasWaldmann> )
2008-07-14T14:52:12 <dreimark> ok, and hexdigest from prefix ?
2008-07-14T14:54:12 <waldi> ThomasWaldmann: if i remove link and macro handling, it gets a lot faster *hide*
2008-07-14T14:54:54 <ThomasWaldmann> dreimark: hexdigest is a method of hmac object
2008-07-14T15:00:26 <gizmach> dreimark: http://paste.pocoo.org/show/79412/ but I need to see if this works like this and to give better names and change where is needed
2008-07-14T15:05:50 <CIA-53> Johannes Berg <johannes AT sipsolutions DOT net> default * 3847:b13f6b68fe43 1.8/MoinMoin/auth/openidrp.py: openid RP: support idselector.com
2008-07-14T15:06:33 <waldi> ThomasWaldmann: 30% faster :)
2008-07-14T15:10:58 <ThomasWaldmann> :)
2008-07-14T15:13:33 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3974:ba4203bcdbd9 1.8-dom-bblank/MoinMoin/converter2/ (link.py macro.py): Link and Macro converter - Don't check for ElementTree.Node in recursor
2008-07-14T15:24:05 <waldi> and this was a really bad idea ...
2008-07-14T15:29:44 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3975:af15f4e08559 1.8-dom-bblank/MoinMoin/converter2/_wiki_macro.py: Pseudo-macro handling - Add Include and TableOfContents
2008-07-14T15:32:00 <gizmach> dreimark: I think this will work, just need to test a bit
2008-07-14T15:39:50 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 3976:c0cca5356b57 1.8-dom-bblank/MoinMoin/converter2/ (link.py macro.py): Revert 3974:ba4203bcdbd9: Link and Macro converter - Don't check for ElementTree.Node in recursor
2008-07-14T15:48:04 <gizmach> dreimark: it works with changes
2008-07-14T15:48:05 <gizmach> :)
2008-07-14T15:50:11 <ThomasWaldmann> ah, TOC and Include, now it gets interesting :)
2008-07-14T15:50:55 <dreimark> gizmach: diff ?
2008-07-14T15:51:06 <gizmach> dreimark: sec
2008-07-14T15:52:02 <dreimark> ThomasWaldmann: does it make also sense for other cache files to create only the meta file if the data file can be created directly?
2008-07-14T15:53:01 <dreimark> thumbf = cache_thumb._filename()
2008-07-14T15:53:20 <dreimark> im_obj.save(thumbf, image_type)
2008-07-14T15:54:23 <gizmach> http://paste.pocoo.org/show/79419/ but I will change names and comments
2008-07-14T15:54:44 <ThomasWaldmann> dreimark: not sure what you mean
2008-07-14T15:55:36 <dennda> ThomasWaldmann: Yeah, got an exam last thursday, last friday and today
2008-07-14T15:55:41 <dennda> but that's over now
2008-07-14T15:55:42 <dennda> wohooo
2008-07-14T15:56:09 <dennda> already got a headache from learning or from the sun or probably from seeing the mess I created around me in the last few days
2008-07-14T15:56:22 <johill> heh
2008-07-14T15:57:51 <ThomasWaldmann> dennda: and how did the exams go?
2008-07-14T15:58:16 <gizmach> dreimark: need to do some more traces I think some maybe not work
2008-07-14T15:59:20 <dreimark> ThomasWaldmann: http://paste.pocoo.org/show/79420/
2008-07-14T15:59:52 <dreimark> I like to use only once im_obj.save
2008-07-14T16:00:45 <dreimark> sendcached.put_cache(self.request, webnail... overwrites the image but creates the correct meta file
2008-07-14T16:02:14 <dreimark> gizmach: do you have done already a snippet example ?
2008-07-14T16:02:40 <dennda> ThomasWaldmann: 1,0 in the first, the second wasn't so good due to 5 hours of preparation only and the one I wrote today went fairly well (hopefully)
2008-07-14T16:04:50 <dreimark> gizmach: please pull 1.8
2008-07-14T16:08:40 <dreimark> ThomasWaldmann: should we have in headers the attachment filename which belongs to the cache file?
2008-07-14T16:09:04 <dreimark> which the cache is based on
2008-07-14T16:22:29 <gizmach> dreimark: ok I will
2008-07-14T16:22:54 <gizmach> dreimark: I done an example but didn't write it in snippet
2008-07-14T16:23:22 <gizmach> dreimark: and also I need to make config group backend to work with new code
2008-07-14T16:23:35 <gizmach> need to change that snippet
2008-07-14T16:26:01 <gizmach> http://paste.pocoo.org/show/79421/ dreimark
2008-07-14T16:30:05 <gizmach> dreimark: I tried on wiki and with SystemAdmin works ok, but i need to trace the acl rights checking
2008-07-14T16:39:04 <dreimark> typo in" definig
2008-07-14T16:39:49 <dreimark> and the for in "# for definig backend to use"
2008-07-14T16:39:53 <dreimark> is superfluos
2008-07-14T16:40:24 <dreimark> " to use" should be more verbose
2008-07-14T16:40:32 <dreimark> gizmach:
2008-07-14T16:41:10 <gizmach> dreimark: ok I will make that more verbose
2008-07-14T16:41:48 <ThomasWaldmann> dreimark: vv
2008-07-14T16:42:53 <waldi> "<span /><span /><h1>WikiName Wiki</h1><span /><span /><p><<TableOfContents(2)>> <span /></p><span />"
2008-07-14T16:44:13 <dennda> uh
2008-07-14T16:44:35 <waldi> <ns0:span ns0:id="line-8" /><ns0:span ns0:id="line-9" />
2008-07-14T16:44:36 <dennda> bloated xhtml
2008-07-14T16:48:53 <waldi> ThomasWaldmann: the moin wiki parser puts every macro in a paragraph. the new code checks that and refuses to add a TOC there
2008-07-14T16:51:16 <gizmach> dreimark: wupi it works with acls
2008-07-14T16:51:51 <gizmach> dreimark: I tried when I added acls and than I loged out and try with user not allowed to see the page
2008-07-14T16:54:17 <ThomasWaldmann> dreimark: see 1.8 main repo for recent commits. cia seems to be out of order.
2008-07-14T16:59:44 <ThomasWaldmann> dreimark: if you are calling stat() just for giving content_length, you can just not give it and it will be autodetected
2008-07-14T16:59:54 <ThomasWaldmann> and why are you giving 'dummy' for data?
2008-07-14T17:01:46 <dreimark> f*ck cia.vc
2008-07-14T17:02:38 <waldi> hihi, TOC works
2008-07-14T17:04:06 <dreimark> ThomasWaldmann: the test needs to be changed too
2008-07-14T17:05:27 <ThomasWaldmann> yeah :)
2008-07-14T17:05:33 <ThomasWaldmann> doing that now
2008-07-14T17:17:34 <dreimark> ThomasWaldmann: I can save the image directly on the cache arena location with the correct mimetype, but then I have no meta file
2008-07-14T17:17:57 <dreimark> so I create afterwards a dummy data file, and overwrite it again
2008-07-14T17:18:19 <dreimark> gizmach: check also your tests
2008-07-14T17:18:50 <gizmach> dreimark: yes I need to revrite someone and change in multiconfi
2008-07-14T17:22:57 * dreimark has pushed some arnica changes. currently it works only with recent 1.8
2008-07-14T17:23:30 <gizmach> dreimark: I'll commit the code and after changed the tests will push everything
2008-07-14T17:23:51 <gizmach> dreimark: and also need to change config snippet
2008-07-14T17:24:28 <gizmach> and probably the code because it does not need the request for initialization
2008-07-14T17:34:33 <ThomasWaldmann> dreimark: sounds dirty :)
2008-07-14T17:49:15 <ThomasWaldmann> dreimark: you can use a file-like object instead of the filename after we have file-like api for caching
2008-07-14T17:49:57 <ThomasWaldmann> thomas pfaff was working on that in the sprints
2008-07-14T17:50:59 <dreimark> ThomasWaldmann: sounds much better
2008-07-14T18:17:16 <dreimark> gizmach: while cia is not working please tell yourself when you've pushed something
2008-07-14T18:17:35 <gizmach> dreimark: ok I will
2008-07-14T18:17:40 <gizmach> I'm still changing tests
2008-07-14T18:17:49 <gizmach> dreimark: I changed options in multiconfig
2008-07-14T18:18:58 <ThomasWaldmann> gizmach: some complete hg diff in a pastebin would be a good idea before committing
2008-07-14T18:19:21 <gizmach> ThomasWaldmann: I was putting some pastebins
2008-07-14T18:19:47 <gizmach> ThomasWaldmann: and I commited several things but didn't pushed them, also I tested them on a real wiki
2008-07-14T18:20:00 <ThomasWaldmann> ok, give me url of a complete thing
2008-07-14T18:21:39 <gizmach> http://paste.pocoo.org/show/79419/ ThomasWaldmann
2008-07-14T18:23:04 <ThomasWaldmann> no
2008-07-14T18:24:56 <ThomasWaldmann> gizmach: you are getting slightly closer, but you have not arrived at the solution yet
2008-07-14T18:25:13 <gizmach> ThomasWaldmann: ok
2008-07-14T18:25:30 <ThomasWaldmann> and you are not doing things at the right place
2008-07-14T18:27:07 <ThomasWaldmann> e.g. the GroupManager needs to be initialized in the config, because that is the way to let the wiki admin choose the list of backends he wants to use (and the order of them).
2008-07-14T18:27:43 <ThomasWaldmann> and be more careful with names
2008-07-14T18:27:51 <gizmach> ok.. hm..
2008-07-14T18:27:58 <ThomasWaldmann> + backend = request.cfg.backend_factory < no good
2008-07-14T18:28:48 <ThomasWaldmann> you are just confusing yourself even more by doing bad name choice
2008-07-14T18:28:58 <gizmach> ThomasWaldmann: well that's true
2008-07-14T18:29:28 <ThomasWaldmann> you are also still not complying to pep8.
2008-07-14T18:29:44 <ThomasWaldmann> is it that hard to put a blank after a comma?
2008-07-14T18:30:04 <gizmach> but with the initiaizing in the config, shouldn't it be good only to give the class, because in that way I give na order, or not?
2008-07-14T18:30:12 <gizmach> I fixed pep8 errors
2008-07-14T18:30:31 <ThomasWaldmann> + # take backend
2008-07-14T18:30:31 <ThomasWaldmann> + backend = request.cfg.backend_factory
2008-07-14T18:30:36 <gizmach> ThomasWaldmann: it's not hard, my mistake
2008-07-14T18:30:47 <ThomasWaldmann> comments like that are worse than no comments
2008-07-14T18:30:56 <gizmach> true
2008-07-14T18:32:48 <ThomasWaldmann> ok, maybe work on a single line intended for wikiconfig that shows that you understood how the factory pattern could work for you
2008-07-14T18:35:45 <ThomasWaldmann> you could start by a single line that needs request object, knowing that it can't be done like that.
2008-07-14T18:35:57 <ThomasWaldmann> and then use the pattern to solve that problem
2008-07-14T18:41:48 <ThomasWaldmann> xorAxAx: johill: btw, we currently have 2 secrets in Config: one for mail importing, one for jabber stuff
2008-07-14T18:42:21 <ThomasWaldmann> cache action stuff also needs a secret for hmac, so I'd like to clean up a bit.
2008-07-14T18:42:58 <ThomasWaldmann> either we could just use one secret for everything (maybe this has security implications),
2008-07-14T18:43:22 <ThomasWaldmann> or some dict of secrets with the purpose as the dict key
2008-07-14T18:43:48 <ThomasWaldmann> opinions?
2008-07-14T18:45:40 <starshine> how easy is the secret to break or delete or reset?
2008-07-14T18:45:55 <starshine> is there any special need to reset one without resetting the others?
2008-07-14T18:46:08 <ThomasWaldmann> the secrets are usually just some random string invented by wiki admin
2008-07-14T18:46:37 <ThomasWaldmann> it could maybe happen that one secret is disclosed somehow
2008-07-14T18:47:22 <starshine> well, invented by admin == could be all to easy to break, some people do not make lots of efort at randomness.
2008-07-14T18:47:32 <ThomasWaldmann> (likely not by looking into wikiconfig, but in some other config also using this (shared) secret)
2008-07-14T18:47:56 <starshine> if it's hard to reset, then a cleanup if one is broken is harder to do in a hurry, and combining them loses an advantage.
2008-07-14T18:48:16 <ThomasWaldmann> then, if we use multiple secrets, damage would be limited to one area
2008-07-14T18:48:55 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3849:d9a525a1450e 1.8/MoinMoin/action/cache.py: cache action: s/sendcached/cache/
2008-07-14T18:49:02 <ThomasWaldmann> changing a secret usually involves editing of 2 config files
2008-07-14T18:49:20 <starshine> under some people's security ideas, detecting that one has been lost and shutting off usage is a possible answer
2008-07-14T18:49:28 <ThomasWaldmann> for the cache secret, it would basically invalidate the complete cache (because nothing would get found any more)
2008-07-14T18:49:33 <starshine> locking the doors to check the premises, as it were
2008-07-14T18:49:45 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3850:0324aa222029 1.8/MoinMoin/action/cache.py: action cache: rename some methods, add 'del' subaction
2008-07-14T18:49:45 <CIA-53> Reimar Bauer <rb.proj AT googlemail DOT com> default * 163:2676d2cbfaea 1.7-extensions/data/plugin/parser/text_x_arnica.py: text_x_arnica: comment fixed
2008-07-14T18:49:51 <ThomasWaldmann> lol
2008-07-14T18:50:05 <ThomasWaldmann> cia is really fast today :P
2008-07-14T18:50:12 <starshine> :)
2008-07-14T18:50:45 <CIA-53> Reimar Bauer <rb.proj AT googlemail DOT com> default * 164:db98e843b83c 1.7-extensions/data/plugin/ (action/arnica_slides.py parser/text_x_arnica.py):
2008-07-14T18:50:45 <CIA-53> text_x_arnica, arnica_slides: implemented sendcached from 1.8 repository
2008-07-14T18:50:45 <CIA-53> (this version currently needs the 1.8 dev repository)
2008-07-14T18:51:21 <starshine> is there a need for a fast and easy way to wipe the whole cache?
2008-07-14T18:51:46 <ThomasWaldmann> well, changing the secret does not wipe the cache
2008-07-14T18:51:48 <starshine> I'd think it would only be useful if one was about to move a site, and even then, it might be preferable to leave it in
2008-07-14T18:51:55 <starshine> just makes it useless. hrm
2008-07-14T18:52:20 <ThomasWaldmann> it will just make everything useless and trigger recreation of all caches
2008-07-14T18:53:01 <starshine> iow not the whole set at once, but pages as they get hit?
2008-07-14T18:53:19 <ThomasWaldmann> yes
2008-07-14T18:53:29 <ThomasWaldmann> (or images, or ...)
2008-07-14T18:54:02 <ThomasWaldmann> maybe something like secrets = {
2008-07-14T18:54:13 <ThomasWaldmann> 'email': '12131313123123413123123',
2008-07-14T18:54:25 <ThomasWaldmann> 'jabber': 'bsgragbsdraeafewfrgtf',
2008-07-14T18:54:40 <ThomasWaldmann> 'sendcache': '34gwegq3w4g45gegvawqg5hw4h4eg',
2008-07-14T18:54:42 <ThomasWaldmann> }
2008-07-14T18:58:32 <CIA-53> Reimar Bauer <rb.proj AT googlemail DOT com> default * 168:cf3fef32a87b 1.7-extensions/data/plugin/action/arnica_slides.py: arnica_slides: link aliastext/filenam to attachment file
2008-07-14T19:00:35 <ThomasWaldmann> dreimark: the secret param is rather for testing than for application usage
2008-07-14T19:02:11 <starshine> maybe it shouldnt be called secrets then
2008-07-14T19:02:11 <starshine> more like salt
2008-07-14T19:02:11 <starshine> or hash
2008-07-14T19:02:24 <dreimark> better
2008-07-14T19:02:28 <ThomasWaldmann> dreimark: you are calculating the same key twice
2008-07-14T19:02:30 <starshine> .o( why are interesting code tidbits named after food?
2008-07-14T19:03:00 <ThomasWaldmann> starshine: because it should be secret :d
2008-07-14T19:03:03 <ThomasWaldmann> :D
2008-07-14T19:03:05 <dreimark> ThomasWaldmann: I know, waiting for the prefix
2008-07-14T19:03:29 <ThomasWaldmann> dreimark: it is better not to do that
2008-07-14T19:03:45 * dreimark is removing it
2008-07-14T19:04:10 <ThomasWaldmann> just do key = cache.key(...) ; thumb_key = 'T' + key ; web_key = 'W' + key
2008-07-14T19:04:51 <starshine_away> ThomasWaldmann: ok :)
2008-07-14T19:05:18 <ThomasWaldmann> dreimark: and please use better names
2008-07-14T19:05:59 <dreimark> yeah there are currently some strange names used, will do rename them soon
2008-07-14T19:08:14 * ThomasWaldmann waits for some mail from Thomas Pfaff with the caching file-like api
2008-07-14T19:10:07 <gizmach> ThomasWaldmann: sorry I was at the phone
2008-07-14T19:13:02 * gizmach read ThomasWaldmann comments
2008-07-14T19:14:44 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3853:3b37036d466d 1.8/MoinMoin/action/cache.py: cache action: add remove() function
2008-07-14T19:27:34 <gizmach> if I commited something (I have more than one commit) how to revert this (I didn't push the code) .
2008-07-14T19:29:00 <ThomasWaldmann> clone your repo to another one up to revision X
2008-07-14T19:31:45 <gizmach> ThomasWaldmann: ok thx
2008-07-14T19:40:19 <dreimark> cia hangs again
2008-07-14T19:41:30 <dreimark> ThomasWaldmann: is there a reason for not using: action_name = __name__.split('.')[-1] in cache ?
2008-07-14T19:42:19 <CIA-53> Reimar Bauer <rb.proj AT googlemail DOT com> default * 169:5ad37eeecc76 1.7-extensions/data/plugin/parser/text_x_arnica.py: text_x_arnica: removed keyword secret
2008-07-14T19:45:01 <ThomasWaldmann> dreimark: no
2008-07-14T19:47:03 <CIA-53> Reimar Bauer <rb.proj AT googlemail DOT com> default * 170:a489fd432404 1.7-extensions/data/plugin/parser/text_x_arnica.py: text_x_arnica: create key only once
2008-07-14T19:47:43 <CIA-53> MelitaMihaljevic default * 3880:566673863e59 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/__init__.py: removed unused @parm
2008-07-14T19:48:12 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3854:30240e801a55 1.8/MoinMoin/action/cache.py: cache action: add some hint, cosmetic changes
2008-07-14T19:49:32 <CIA-53> MelitaMihaljevic default * 3881:059092399fac 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/ (12 files in 4 dirs): renamed backands/ to backand
2008-07-14T19:51:58 <CIA-53> Reimar Bauer <rb.proj AT googlemail DOT com> default * 171:6c5af220a85a 1.7-extensions/data/plugin/action/refresh.py: action.refresh: refactored for using of cache.remove
2008-07-14T19:59:39 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3855:5e35dd32656f 1.8/MoinMoin/action/cache.py: cache action: calculate action_name, more comments
2008-07-14T20:00:06 <CIA-53> MelitaMihaljevic default * 3882:197dbe9899ea 1.8-ldapgroups-mmihaljevic/MoinMoin/ (3 files in 2 dirs): changed backends to backend in the code related to datastruct backend
2008-07-14T20:03:07 * ThomasWaldmann will kill someone some time that repeats the same information that is in the path/filename in the module level docstring and does not give some usable information.
2008-07-14T20:04:02 <ThomasWaldmann> s/that/who/
2008-07-14T20:06:46 <gizmach> ups
2008-07-14T20:06:54 <ThomasWaldmann> gizmach: you did not use hg mv
2008-07-14T20:07:57 <gizmach> ThomasWaldmann: I did hg rename
2008-07-14T20:08:11 <ThomasWaldmann> for what?
2008-07-14T20:08:37 <gizmach> for /backends -> /backend
2008-07-14T20:09:05 <ThomasWaldmann> http://hg.moinmo.in/moin/1.8-ldapgroups-mmihaljevic/log/059092399fac/MoinMoin/datastruct/backend/_tests/test_wikigroup.py
2008-07-14T20:09:42 <ThomasWaldmann> then why is it only showing revisions since the rename? that usually happens if you rename just on the filesystem (not using hg mv)
2008-07-14T20:10:49 <gizmach> dono, but I didnt do mv at all (just hg rename)
2008-07-14T20:12:23 <ThomasWaldmann> ah, my fault. one needs to click on "base" link to get the pre-rename history
2008-07-14T20:12:52 <ThomasWaldmann> (did they change UI for this, I thought it has been different before, listing all revs at once)
2008-07-14T20:13:07 <gizmach> uh
2008-07-14T20:14:01 <ThomasWaldmann> mv is the same as rename, btw
2008-07-14T20:14:18 <gizmach> ThomasWaldmann: ok I will remember that
2008-07-14T20:21:51 <dreimark> ThomasWaldmann: the UI was always like this, I'd prefer a complete list too
2008-07-14T20:50:35 <dreimark> gizmach: how is it going on ?
2008-07-14T20:52:20 <gizmach> dreimark: well I'm still playing with inicializing the GroupManager trying to figure how to fix what I done wrong
2008-07-14T20:53:35 <waldi> hmm, is it documented, that moinmoin must not use a shared group in default config?
2008-07-14T20:54:46 <waldi> a umask of 07 creates the pyc files with group write permissions
2008-07-14T20:59:32 * dreimark walks home bbl
2008-07-14T21:06:43 <ThomasWaldmann> waldi: maybe not
2008-07-14T21:30:10 <zenhase> i'am back from no-net-town
2008-07-14T21:31:03 <zenhase> well ok, the town perhaps had net. but since my parents are currently moving they didn't have any internet
2008-07-14T21:33:54 <zenhase> hmm, the last times i pushed to hg.moinmo.in my hg somehow always hangs o_O
2008-07-14T21:34:46 <zenhase> the changesets are there, as the website states, but my 'hg push' process still says 'searching for changes' and doesn't finish
2008-07-14T21:36:38 <gizmach> zenhase: today?
2008-07-14T21:36:45 <gizmach> for me also tooks a looong
2008-07-14T21:36:49 <gizmach> to push
2008-07-14T21:36:56 <gizmach> for more dan 5-10 minutes
2008-07-14T21:37:08 <gizmach> and it says searching for changes too
2008-07-14T21:37:15 <gizmach> but it pushes finaly
2008-07-14T21:37:15 <zenhase> ah ok
2008-07-14T21:37:48 <zenhase> so it seems to be non-local (meaning it's not my PC alone ;D)
2008-07-14T21:38:32 <gizmach> hihi not
2008-07-14T21:44:25 <zenhase> hmm, no CIA?
2008-07-14T21:44:39 * zenhase nudges CIA a little bit
2008-07-14T21:44:55 <zenhase> nudge nudge, push push, wink wink?
2008-07-14T21:53:15 <gizmach> no
2008-07-14T21:53:17 <gizmach> nottin
2008-07-14T21:57:56 <zenhase> but my push is finally done :)
2008-07-14T21:59:39 <zenhase> wikiwiki editedit ... doubling words kind of puts me into a silly mood :)
2008-07-14T22:00:05 <gizmach> ThomasWaldmann: I'm still not sure why backend_factory = backend.BackendFactory(group_wiki.GroupManagerWikiBackend) wouldn't be enough in the config for admin
2008-07-14T22:01:20 <gizmach> and with that defined my code I previuos sent (with names changed) will work
2008-07-14T22:01:44 <gizmach> ThomasWaldmann: I tested that and I really can't figure the other way how to define it in config
2008-07-14T22:03:02 <zenhase> poor FlorianFesti ... still is subscribed to my moinmo.in-homepage and subpages from last GSoC
2008-07-14T22:03:19 <zenhase> he gets my edit spam delivered straight via mail :o
2008-07-14T22:04:05 <gizmach> or dreimark
2008-07-14T22:05:11 <zenhase> hmm?
2008-07-14T22:05:20 <gizmach> ThomasWaldmann: I tried like that with wiki and it works, and I can day use wikibackend, ldapbackend, configbackend
2008-07-14T22:05:27 <gizmach> zenhase: ? was that to me or?
2008-07-14T22:05:33 <zenhase> ah, this wasn't about the edit-spam ;)
2008-07-14T22:05:41 <gizmach> zenhase: no
2008-07-14T22:05:46 <gizmach> I have my own problems
2008-07-14T22:05:47 <gizmach> :D
2008-07-14T22:05:50 <zenhase> :o
2008-07-14T22:06:03 <zenhase> *hug*
2008-07-14T22:06:07 <gizmach> zenhase: this was a joke
2008-07-14T22:06:12 <gizmach> hehe thx
2008-07-14T22:06:13 <gizmach> :)
2008-07-14T22:28:39 <ThomasWaldmann> gizmach: but you want to have a groupmanager in config, managing multiple backends
2008-07-14T22:29:21 <gizmach> ThomasWaldmann: urgh you are right
2008-07-14T22:31:58 <gizmach> deffinetly wrong way
2008-07-14T22:41:57 <zenhase> which reminds me:
2008-07-14T22:42:39 <zenhase> i find the new MoinMoin.config.multiconfig a bit hard to read
2008-07-14T22:43:19 <zenhase> it wasn't very readable before, but i actually don't find it much of an improvement
2008-07-14T22:46:40 <gizmach> zenhase: it's ok after a while and it makes sence and the most things " are comments
2008-07-14T22:48:17 <zenhase> i would like some scheme for configuration in a similar fashion like trac
2008-07-14T22:49:08 <zenhase> where plugins, modules and components declare their configuration-options in a declarative way
2008-07-14T22:50:50 <zenhase> bbl
2008-07-14T22:51:09 <zenhase> then i will finish off the wiki pages i started
2008-07-14T22:59:42 * gizmach will continue after some hours of sleep bbl
2008-07-14T23:05:21 <ThomasWaldmann> zenhase: don't forget the docs
2008-07-14T23:20:35 <dreimark> re
2008-07-14T23:28:46 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3858:5079d2246367 1.8/MoinMoin/ (search/Xapian.py wikiutil.py): merged moin/1.7
2008-07-14T23:38:54 <dreimark> gn MoinMoin: MoinMoinChat/Logs/moin-dev/2008-07-14 (last edited 2008-07-13 22:15:02 by IrcLogImporter)
