2008-03-20T01:08:19 <TheSheep> johill: no
2008-03-20T01:08:43 <TheSheep> johill: there are some advanced papers on various normal forms for REs
2008-03-20T01:09:03 <TheSheep> johill: but I'm not aware of any implementation
2008-03-20T01:09:19 <TheSheep> johill: (or even of any usecase)
2008-03-20T01:09:53 <johill> heh
2008-03-20T01:10:08 <johill> I was thinking about making sure they won't ever match the same string
2008-03-20T01:14:44 <johill> but I have now realised that the whole idea I had was flawed (where this wasn't really related to the idea, only to the configuration of the idea)
2008-03-20T01:17:59 <johill> eek
2008-03-20T01:18:11 <johill> hwendel's idea of underlay vs. normal is horrible
2008-03-20T01:34:58 <TheSheep> johill: actually I stand corrected. there is a huge library for manipulating finite automata
2008-03-20T01:37:30 <johill> interesting, but no longer relevant :)
2008-03-20T02:00:01 <johill> bleh
2008-03-20T02:00:10 <johill> is excludeunderlay broken for search?
2008-03-20T02:00:43 <johill> obviously. doh
2008-03-20T02:00:45 <johill> no wonder
2008-03-20T02:06:38 <johill> wtf
2008-03-20T02:26:06 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3579:7f478d8a491c 1.7-storage-hwendel/MoinMoin/storage/backends/meta.py: fix bug in iterator-ification
2008-03-20T02:32:55 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3580:b784198533fc 1.7-storage-hwendel/MoinMoin/ (6 files in 4 dirs): allow marking backends as 'underlay' and properly filter for that
2008-03-20T02:32:56 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3581:22344e571e9b 1.7-storage-hwendel/MoinMoin/storage/backends/meta.py: optimise filtering for 'underlay' in the meta backends
2008-03-20T02:47:40 <johill> that really speeds up AbandonedPages :)
2008-03-20T02:49:38 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3582:ea8f3894e339 1.7-storage-hwendel/wiki/data/pages/HelpOnConfiguration(5c)Storage/ (current revisions/00000002): update included docs
2008-03-20T02:53:53 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3583:29987db20b90 1.7-storage-hwendel/wiki/data/pages/ (8 files in 4 dirs): rename and update the doc page
2008-03-20T03:15:17 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3584:2346f5e4d326 1.7-storage-hwendel/MoinMoin/storage/backends/moin17.py: make moin17 backend honour is_underlay keyword
2008-03-20T03:35:56 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3585:7aebf4a5c71b 1.7-storage-hwendel/MoinMoin/PageEditor.py: fix page-editor for items change
2008-03-20T03:51:27 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3586:c1b8b521cb93 1.7-storage-hwendel/MoinMoin/storage/backends/meta.py: bugfix in meta backends
2008-03-20T03:52:18 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3587:25a43870ff2a 1.7-storage-hwendel/MoinMoin/_tests/test_PageEditor.py: fix page editor test
2008-03-20T04:00:18 <jroes> woops. just read some scrollback and realized that I need to show some prior python work! looks like I better write something that isn't just a quick one-off script in python, huh? :)
2008-03-20T09:49:39 <xorAxAx> jroes: at least that eases our choice of the students :)
2008-03-20T11:05:11 <xorAxAx> jroes: also note that we seem to have some timezone divergence here :)
2008-03-20T11:06:07 <lanius> johill: it does exist because the attachement code was not yet modified to use the new storage code
2008-03-20T11:06:13 <lanius> johill: but after that it has to go
2008-03-20T11:12:34 <johill> lanius: 'it'?
2008-03-20T11:19:49 <lanius> the _getPagePath method
2008-03-20T11:21:24 <dreimark> moin
2008-03-20T11:21:38 <dreimark> wb lanius
2008-03-20T11:21:50 <lanius> hi dreimark :-)
2008-03-20T11:22:17 <dreimark> do you get your project as gsoc application too?
2008-03-20T11:22:30 <lanius> me?
2008-03-20T11:22:58 <dreimark> Andreas told me that you would do mentoring in a project if it is accepted
2008-03-20T11:23:21 <lanius> ah yeah, but the dlr was not accepted
2008-03-20T11:24:30 <johill> lanius: oh, right
2008-03-20T11:24:42 <lanius> johill: why did you do that underlay thing?
2008-03-20T11:25:07 <dreimark> oh, do you like to choose us again as gsoc student ?
2008-03-20T11:28:39 <lanius> no i will write my dipl work, but it seems that there is already another student who wants to continue this stuff?
2008-03-20T11:29:53 <johill> hm. xclip is messing up utf-8
2008-03-20T11:30:12 <dreimark> lanius: yeah seems one is interested. Do you like to help a bit mentoring ?
2008-03-20T11:31:05 <lanius> yeah that is no problem
2008-03-20T11:32:22 <dreimark> thanks
2008-03-20T11:32:29 <dreimark> bbl -> lunch
2008-03-20T11:36:07 <johill> lanius: do you think we can make getPageList() an iterator all the way through?
2008-03-20T11:36:24 <johill> with current underlay it seems to use about 2-3 MB memory for that list
2008-03-20T11:41:15 <lanius> johill: would have to take a deeper look, at the moment not
2008-03-20T11:42:34 <johill> I've succeeded mostly
2008-03-20T11:42:41 <johill> but haven't pushed it through to storage
2008-03-20T11:42:53 <johill> i.e. Page.getPageList() will return an iterator
2008-03-20T11:43:50 <johill> but all the set manipulations you do are quite expensive
2008-03-20T11:44:10 <johill> it would be much better to have the backend filter on !deleted
2008-03-20T11:44:59 <johill> most likely, the filter should be a filter function rather than the dict
2008-03-20T11:45:06 <johill> which gets the page name and the meta info
2008-03-20T11:45:15 <johill> and then it's all pretty easy
2008-03-20T11:45:49 <lanius> sounds pretty good
2008-03-20T11:46:51 <lanius> the filter functions has to operate on the metadata dict then
2008-03-20T11:46:59 <johill> yes
2008-03-20T11:47:04 <johill> and the pagename
2008-03-20T11:47:08 <lanius> yes
2008-03-20T11:47:12 <johill> then we can roll the filter passed to getPageList right into it
2008-03-20T11:47:42 <johill> lanius: have you seen the meta backend news() trick I implemented?
2008-03-20T11:47:42 <lanius> right
2008-03-20T11:47:50 <johill> something like that should be done here as well then I suppose
2008-03-20T11:48:10 <lanius> johill: have seen it but not yet taken a look into it
2008-03-20T11:48:55 <johill> basically just returning an iterator that build the sorted list from multiple backends on the fly
2008-03-20T11:49:51 <lanius> external.py: for n in news: ... doesn't that cost too much time to iterate over it again
2008-03-20T11:49:57 <lanius> maybe set a flag in the previous iteration
2008-03-20T11:50:56 <johill> oh, actually the code is jsut wrong, you're right, thanks
2008-03-20T11:52:06 <johill> why not just set _items to None in the loop?
2008-03-20T11:52:22 <lanius> because it will be set on every entry on the loop
2008-03-20T11:52:29 <lanius> i don't think it will cost much though
2008-03-20T11:53:23 <johill> it's just a variable, not a property, so yeah
2008-03-20T11:53:26 <johill> I'll do that
2008-03-20T11:53:51 <lanius> johill: why did you do that underlay thing?
2008-03-20T11:53:57 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3588:0eb62e62cebd 1.7-storage-hwendel/MoinMoin/ (7 files in 4 dirs): remove remaining mentions of cfg.page_backend
2008-03-20T11:53:58 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3589:7cf870c8142e 1.7-storage-hwendel/MoinMoin/packages.py: underlay isn't that easy any more
2008-03-20T11:53:59 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3590:53e0a74a638c 1.7-storage-hwendel/MoinMoin/storage/external.py: fix small storage bug introduced by making news an iterator
2008-03-20T11:54:17 <johill> because I wanted to not special-case underlay as much
2008-03-20T11:54:32 <lanius> johill: now it seems more special then befor
2008-03-20T11:54:53 <johill> how so?
2008-03-20T11:55:23 <lanius> the is_underlay keyword
2008-03-20T11:55:54 <johill> well but that allows marking anything as underlay
2008-03-20T11:56:00 <johill> even multiple if you feel like it
2008-03-20T11:56:16 <lanius> have to think about it ;)
2008-03-20T11:56:26 <johill> the bigger problem with your way though is that you have two ways to access the backend
2008-03-20T11:56:28 <lanius> it's time for lunch
2008-03-20T11:56:32 <johill> one through the layered and one directly
2008-03-20T11:56:41 <johill> which falls over if you change only one of them in the config and not the other
2008-03-20T12:00:45 <marcel__> whois
2008-03-20T12:01:11 <marcel__> whois
2008-03-20T12:01:14 <marcel__> whois marcel__
2008-03-20T12:07:10 <johill> lanius: why does ItemCollection cache the list? wouldn't the backend be able to make a smarter decision about that and have an easier task at invalidating it?
2008-03-20T12:08:18 <ThomasWaldmann> johill: yes, but then you have to implement it in every backend
2008-03-20T12:08:26 <johill> no
2008-03-20T12:08:35 <johill> an sql backend for example might decide not to bother
2008-03-20T12:10:00 <johill> and really, you can alwasy have a 'ItemListCachingBackend' mixing
2008-03-20T12:11:21 <johill> my python processes are growing to >50M resident already
2008-03-20T12:11:52 <ThomasWaldmann> sounds ok, wait if lanius has more comments (I didnt look at that code recently)
2008-03-20T12:13:06 <ThomasWaldmann> to something completely different: did anybody use MultiCall, applyAuthToken and putPage recently?
2008-03-20T12:13:19 <johill> hm, no
2008-03-20T12:13:23 <johill> anything wrong with it?
2008-03-20T12:13:24 <ThomasWaldmann> while the multicall stuff seems to work, putPage seems not to work
2008-03-20T12:14:33 <johill> odd, what happens?
2008-03-20T12:16:24 <ThomasWaldmann> you get a not very useful traceback
2008-03-20T12:17:42 <ThomasWaldmann> i guess i debug it with the test17.moinmo.in wiki
2008-03-20T12:37:14 <ThomasWaldmann> applyAuthToken does not work in 1.7
2008-03-20T12:37:17 <lanius> johill: sounds good
2008-03-20T12:38:01 * ThomasWaldmann tries test16.moinmo.in
2008-03-20T12:39:02 <dreimark> ThomasWaldmann: I do use multicall with 1.6
2008-03-20T12:39:51 <dreimark> http://moinmo.in/ReimarBauer/xmlrpc
2008-03-20T12:40:03 <dreimark> is your putPage usage differnt to mine?
2008-03-20T12:41:04 <jroes> xorAxAx: well, I do recall seeing some other code that was only a short snippet, and I'm actually probably just forgetting something I've worked on
2008-03-20T12:41:36 <jroes> xorAxAx: that was late evening for me, btw (11pm), I'm usually around when this place is bustling :)
2008-03-20T12:43:05 <ThomasWaldmann> dreimark: very similar
2008-03-20T12:44:35 * dreimark tries with 1.7 now too
2008-03-20T12:44:40 <ThomasWaldmann> and for 1.6 applyAuthToken tells CachingError "is a directory"
2008-03-20T12:45:04 <ThomasWaldmann> dreimark: use test17.moinmo.in
2008-03-20T12:45:11 <dreimark> ok
2008-03-20T12:49:47 <dreimark> ThomasWaldmann: xmlrpclib.ProtocolError: <ProtocolError for test17.moinmo.in?action=xmlrpc2: 404 Not found>
2008-03-20T12:51:04 <dreimark> see http://test17.moinmo.in/TestWiki17?action=xmlrpc2
2008-03-20T12:51:10 <dreimark> bbl 10mins
2008-03-20T12:52:15 <ThomasWaldmann> in 1.6 it works. it just gives a silly error for applyAuthToken with auth_token == ''
2008-03-20T12:53:38 <ThomasWaldmann> dreimark: that url is not valid
2008-03-20T12:59:35 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2575:9bffda2e1443 1.6/MoinMoin/xmlrpc/__init__.py: xmlrpc applyAuthToken: give good error msg for empty token, avoid CachingError 'is a directory'
2008-03-20T13:21:50 <dreimark> ThomasWaldmann: no
2008-03-20T13:22:14 <dreimark> it works on localhost without that change
2008-03-20T13:23:13 <dreimark> and on test17 I can now add pages with wrong credentials
2008-03-20T13:23:35 <dreimark> or without using my username
2008-03-20T13:24:36 <ThomasWaldmann> you dont need to be authenticated on test17
2008-03-20T13:24:54 <dreimark> I know but if I authenticate me it is not used
2008-03-20T13:25:28 <ThomasWaldmann> yes, that looks like a bug. that works on test16 btw.
2008-03-20T13:25:50 <dreimark> it works on my localhost too
2008-03-20T13:25:53 <dreimark> in 1.7
2008-03-20T13:27:39 <ThomasWaldmann> i'll try to fix 1.6 first
2008-03-20T13:27:58 <ThomasWaldmann> (i need that for i18n update)
2008-03-20T13:28:04 <dreimark> ok
2008-03-20T13:28:52 <ThomasWaldmann> the remaining error now is an ExpatError no element found
2008-03-20T13:31:11 <dreimark> ThomasWaldmann: test16 does it right for me too
2008-03-20T13:31:24 <dreimark> where did you get this ExpatError ?
2008-03-20T13:31:28 <dreimark> where did you get this ExpatError ?
2008-03-20T13:35:45 <dreimark> 1.6 on localhost too
2008-03-20T13:36:30 <ThomasWaldmann> hmm, seem to be related to the data i want to put
2008-03-20T13:38:35 <dreimark> johill: We have to reword that for 1.7 too
2008-03-20T13:38:43 <dreimark> "To create an account, see the UserPreferences page"
2008-03-20T13:39:09 <dreimark> it should only be something like "here you can create an account"
2008-03-20T13:45:40 <johill> dreimark: I think I changed all those in 1.7
2008-03-20T13:47:04 <dreimark> you are rigth seems I mixed it up with 1.6
2008-03-20T13:47:26 <johill> and 1.6 still has the old stuff, no?
2008-03-20T13:47:45 <dreimark> yes /UserPreferences
2008-03-20T13:48:17 <dreimark> I was switching between both testwikis and missed that I am on a 1.6 one
2008-03-20T13:48:26 <dreimark> sorry, all is ok!
2008-03-20T13:49:35 <johill> np :)
2008-03-20T13:49:39 <johill> I just recently fixed that all up
2008-03-20T13:49:44 <johill> so two days ago you would have been right ;)
2008-03-20T13:51:09 <dreimark> hehe, another question
2008-03-20T13:52:35 <johill> hm. took me an hour and a half for such a trivial patch... http://git.sipsolutions.net/?p=dovecot-antispam.git;a=commitdiff;h=a5941271d9ba14dd198df01a13400af557ecf6ed
2008-03-20T13:52:40 <dreimark> we don't use currently POST on actions like login, logout, edit, info
2008-03-20T13:52:46 <dreimark> why ?
2008-03-20T13:53:08 <johill> what do you mean?
2008-03-20T13:53:16 <johill> in 1.7 I made the "more actions" dropdown use POST
2008-03-20T13:53:33 <dreimark> and the editbar actions not
2008-03-20T13:53:42 <johill> that doesn't matter though
2008-03-20T13:53:55 <johill> why should they use POST?
2008-03-20T13:54:47 <dreimark> because it is constitent it looks a bit strange if you look at the url
2008-03-20T13:54:53 <TheSheep> does any of these action have a side effect when just selected?
2008-03-20T13:55:00 <dreimark> do an info and afterwards print preview
2008-03-20T13:55:08 <TheSheep> dreimark: now try to boomark or reload a page
2008-03-20T13:55:13 <TheSheep> dreimark: or use the back button
2008-03-20T13:55:42 <johill> dreimark: ah
2008-03-20T13:55:50 <dreimark> TheSheep: that was the reason
2008-03-20T13:56:01 <johill> dreimark: funky, I suppose that's a bug in the theme code, will take a look
2008-03-20T13:57:14 <dreimark> TheSheep: thanks for helping and getting the point
2008-03-20T13:57:50 <johill> dreimark: fixed
2008-03-20T13:57:52 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3342:6003cadc0791 1.7/MoinMoin/theme/__init__.py: make theme generate a proper action in the dropdown action list form
2008-03-20T13:58:12 <TheSheep> dreimark: I'm sorry, I guess I need coffee
2008-03-20T13:59:41 <ThomasWaldmann> the xmlrpc stuff behaves random
2008-03-20T14:00:06 <ThomasWaldmann> if i put the same data again and again, it sometimes works, sometimes not
2008-03-20T14:00:45 <ThomasWaldmann> and if i put longer data, it is mostly not working, if i put shorter data, it more often works
2008-03-20T14:00:52 <johill> dreimark: hmm. except for revert everything in the actions drop down has no side effects
2008-03-20T14:00:57 <johill> but subscribe/unsubscribe do
2008-03-20T14:01:00 <johill> I guess we need to do more work here
2008-03-20T14:01:16 <johill> maybe actually implement the revert feature request mockups from MartinBayer ;)
2008-03-20T14:01:21 <TheSheep> johill: maybe the revert action should display a form with confirmation?
2008-03-20T14:01:24 <johill> and then we can switch to using GET for the actions drop down
2008-03-20T14:01:49 <johill> TheSheep: yeah, that's the feature request, and would allow closing another one (the one that asks for a comment)
2008-03-20T14:02:16 <johill> 'add link' also has side effects
2008-03-20T14:02:27 <TheSheep> johill: on the other hand, having buttons on the page history page would be much more convenient
2008-03-20T14:02:52 <johill> I guess page history could actually have a button and an input field
2008-03-20T14:03:01 <johill> [text text text] [-revert to this-]
2008-03-20T14:03:11 <johill> would make it quite wide though
2008-03-20T14:03:15 <dreimark> bbl, some talks ~30mins
2008-03-20T14:03:53 <johill> I would also love a 'diff to previuos' link
2008-03-20T14:06:04 <TheSheep> I had some mockup for oddmuse for this
2008-03-20T14:06:15 <TheSheep> I was rather thinking about a button for every edit
2008-03-20T14:06:22 <TheSheep> with 'revert this edit'
2008-03-20T14:06:50 <TheSheep> but it need special handling at the side of the action
2008-03-20T14:07:36 <ThomasWaldmann> keep in mind that 1.7 release is scheduled rather soon
2008-03-20T14:07:39 <johill> how would you implement that?
2008-03-20T14:07:52 <johill> merge and show the editor?
2008-03-20T14:09:36 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3343:04548b25e042 1.7/MoinMoin/action/info.py:
2008-03-20T14:09:36 <CIA-39> info action: always use recall to allow sending the link
2008-03-20T14:09:36 <CIA-39> (cf. FeatureRequests/UnsortedStuff: Latest version link should include version number as a GET parameter)
2008-03-20T14:10:21 <TheSheep> ThomasWaldmann: how soon?
2008-03-20T14:10:51 <TheSheep> another option for using posts and not breaking the back button is to use 303 responses
2008-03-20T14:11:55 <TheSheep> johill: no, it would also revert all further edits, of course
2008-03-20T14:12:09 <TheSheep> johill: maybe 'revert up to this edit' would be more accurate
2008-03-20T14:12:55 <johill> TheSheep: that's like 'revert to this version', no?
2008-03-20T14:13:17 <TheSheep> johill: yes, just moved one step down
2008-03-20T14:13:31 <ThomasWaldmann> http://moinmo.in/MoinMoinTodo/Release 1.7.0 maybe everyone working on 1.7 subscribes/quicklinks that
2008-03-20T14:13:56 <johill> TheSheep: I think revert to this is clearer
2008-03-20T14:14:11 <TheSheep> johill: I suppose so
2008-03-20T14:14:31 <ThomasWaldmann> and reworking info page is of much lower importance than doing bug fixes
2008-03-20T14:14:34 <TheSheep> johill: but then the top one has no button, and the bottom one needs extra button
2008-03-20T14:14:57 <TheSheep> johill: plus, you don't really *think* like this, you want to revert edits, not 'to version'
2008-03-20T14:15:27 <ThomasWaldmann> TheSheep: you can't revert edits, you only can revert to rev X
2008-03-20T14:15:56 <TheSheep> ThomasWaldmann: nevermind, you are right, more important things to do
2008-03-20T14:16:04 <TheSheep> ThomasWaldmann: this page doeasn't exit :/
2008-03-20T14:16:29 <ThomasWaldmann> hah. ff3 feature :|
2008-03-20T14:16:48 <ThomasWaldmann> add some %20 :)
2008-03-20T14:17:15 <TheSheep> got it
2008-03-20T14:19:52 <ThomasWaldmann> 2008-05-26 is when coding for soc starts
2008-03-20T14:21:03 <TheSheep> 2 months
2008-03-20T14:22:01 <johill> http://benjamin.sipsolutions.net/?action=info
2008-03-20T14:22:10 <johill> anybody opposed to me committing that 'to previous' thing?
2008-03-20T14:24:48 <ThomasWaldmann> no, but i think you have to swap rev1 and rev2
2008-03-20T14:25:31 <johill> hm, it seemed right in my test
2008-03-20T14:25:44 <ThomasWaldmann> because it is intelligent :P
2008-03-20T14:26:03 <johill> oh
2008-03-20T14:26:13 <johill> so you can't get a reverse diff?
2008-03-20T14:26:25 <ThomasWaldmann> no
2008-03-20T14:26:30 <johill> ok, I'll swap it
2008-03-20T14:26:37 <ThomasWaldmann> btw, that page shows another bug
2008-03-20T14:26:48 <johill> yeah
2008-03-20T14:26:50 <johill> I wonder what that is
2008-03-20T14:26:51 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3344:dba9666479a9 1.7/MoinMoin/action/info.py: info action: add 'diff to previous' link
2008-03-20T14:27:03 <ThomasWaldmann> it could select the latest 2 page edits even if there is attachment stuff after it
2008-03-20T14:27:15 <johill> it also has two revisions 11
2008-03-20T14:27:24 <johill> that page is totally messed up
2008-03-20T14:27:31 <johill> probably a result of the conversion going bad or something
2008-03-20T14:27:38 <ThomasWaldmann> look into edit-log
2008-03-20T14:28:21 <ThomasWaldmann> and you need to use page_front_page
2008-03-20T14:29:05 <johill> I think he does use page_front_page
2008-03-20T14:29:08 <johill> just set to FrontPage
2008-03-20T14:29:29 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3345:41e04736a9f1 1.7/MoinMoin/action/info.py: info action: fix buglet with diff defaults
2008-03-20T14:29:39 <ThomasWaldmann> i got StartSeite when using your url
2008-03-20T14:29:43 <johill> oh
2008-03-20T14:29:47 <johill> ok, bugger
2008-03-20T14:29:52 <johill> that bug is fixed anyway
2008-03-20T14:34:00 <ThomasWaldmann> brb
2008-03-20T15:10:24 <dreimark> can we make this Previous and Next button not look like a button by some css ?=
2008-03-20T15:11:35 <dreimark> and there is a bug if I start on the oldest "to previous"
2008-03-20T15:12:24 <dreimark> I can click on Previous and then Next becomes disabled
2008-03-20T15:12:42 <dreimark> and doesn't change
2008-03-20T15:13:33 <dreimark> may be we can have a "button" in the middle of both doing the "revert to this version"
2008-03-20T15:16:55 <johill> I've just seen the buttons for the first time
2008-03-20T15:16:57 <johill> no idea how to fix it :)
2008-03-20T15:20:45 * dreimark has no idea when that was added.
2008-03-20T15:21:10 <dreimark> sounds to be an intersting feature
2008-03-20T15:21:22 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3346:38cbacf3cbc4 1.7/MoinMoin/config/multiconfig.py: slightly relaxed default surge protection settings (port from 1.6)
2008-03-20T15:21:22 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3347:9ceb503dd7a6 1.7/ (3 files in 3 dirs): merged main
2008-03-20T15:21:37 * dreimark looks at the info part
2008-03-20T15:21:51 <johill> dreimark: but I think we should allow revert to have a comment
2008-03-20T15:21:57 <johill> dreimark: so just a revert button there wouldn't cut it
2008-03-20T15:22:25 <ThomasWaldmann> don't add a revert GET link to info page
2008-03-20T15:22:48 <johill> ThomasWaldmann: can't do anyway, revert will reject GET
2008-03-20T15:22:56 <ThomasWaldmann> ok :)
2008-03-20T15:24:41 <dreimark> :)
2008-03-20T15:25:04 <johill> I'm just thinking maybe rever tshould behave like delete
2008-03-20T15:25:08 <johill> where you're asked for a comment first
2008-03-20T15:25:15 <johill> and then we could make the actions menu use GET again
2008-03-20T15:26:18 <TheSheep> johill: yes, it would make it work in themes like classic that don't have a form for actions
2008-03-20T15:26:44 <johill> oh. ok so that's actually a bug fix
2008-03-20T15:26:46 <johill> I'll do it later
2008-03-20T15:26:55 <johill> not sure if it can save a comment though
2008-03-20T15:27:51 <johill> will check
2008-03-20T15:27:57 <johill> just preparing a linux kernel patch to post as RFC
2008-03-20T15:30:39 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3348:9d726e040fda 1.7/MoinMoin/xmlrpc/__init__.py: better xmlrpc debugging, show raw request data in case of loads exceptions
2008-03-20T15:35:21 <johill> ok, patch sent
2008-03-20T15:38:51 <ThomasWaldmann> hmm, if i have more than 1543 bytes of xml data in the xmlrpc call, stuff is missing at the end :X
2008-03-20T15:39:24 <ThomasWaldmann> wtf?!
2008-03-20T15:40:21 <johill> weird
2008-03-20T15:40:23 <johill> say
2008-03-20T15:40:27 <johill> do we still want to make actions classes?
2008-03-20T15:40:35 <johill> loads of the internal ones have this stupid class/glue code stuff
2008-03-20T15:40:51 <xorAxAx> ThomasWaldmann: interesting, which server?
2008-03-20T15:41:10 <ThomasWaldmann> apache2 / mod_wsgi / mm1.7
2008-03-20T15:41:22 <ThomasWaldmann> but sometimes i can post 5k successfully
2008-03-20T15:42:09 <ThomasWaldmann> johill: that action base class is helpful for std stuff
2008-03-20T15:43:26 <johill> yeah
2008-03-20T15:47:59 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3349:81697026ee86 1.7/MoinMoin/xmlrpc/__init__.py: better xmlrpc debugging, show length of raw request data in case of loads exceptions
2008-03-20T15:52:06 <ThomasWaldmann> 8000 bytes, 1440 bytes, ... not quite random, but a bit
2008-03-20T15:53:58 <johill> hm
2008-03-20T15:54:07 <johill> now if I got to revert I don't see the old version
2008-03-20T15:56:48 <dreimark> ThomasWaldmann: xmlrpc putpage should check if the content has changes
2008-03-20T15:57:25 <dreimark> i can create new revisions for no differences
2008-03-20T15:57:55 <johill> I thought PageEditor does that
2008-03-20T15:58:11 <dreimark> it should yes but it seems not
2008-03-20T15:58:27 <ThomasWaldmann> dreimark: i have seen the error "not saved because no changes" already today
2008-03-20T15:58:50 <ThomasWaldmann> (dont remember if it was 1.6 or 1.7)
2008-03-20T16:01:24 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3350:7707cd079690 1.7/MoinMoin/action/info.py: info: show only delete options to users with the delete right
2008-03-20T16:01:24 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3351:55102b9ed091 1.7/MoinMoin/action/diff.py: action.diff: fixed disabling of "Next change" button
2008-03-20T16:01:25 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3352:bbc4932e0fa7 1.7/MoinMoin/xmlrpc/__init__.py: merged main
2008-03-20T16:01:54 <dreimark> ThomasWaldmann: I have that in 1.7 currently
2008-03-20T16:02:21 <dreimark> I can put the same content of a page to n revisions
2008-03-20T16:02:44 <ThomasWaldmann> i primarily would like to have that truncation fixed before doing other changes there
2008-03-20T16:02:54 <dreimark> ok
2008-03-20T16:03:29 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3353:8e615ef236e2 1.7/MoinMoin/PageEditor.py: allow saving a comment when reverting
2008-03-20T16:03:37 <ThomasWaldmann> dreimark: you are testing with 1.7 standalone, right?
2008-03-20T16:03:51 <ThomasWaldmann> does it work ok for arbitrary content length?
2008-03-20T16:03:57 <dreimark> yes, I can go to 1.6 too
2008-03-20T16:04:13 <ThomasWaldmann> i am debugging on 1.7 currently
2008-03-20T16:04:23 <ThomasWaldmann> i added some logging
2008-03-20T16:05:04 <ThomasWaldmann> [loggers]
2008-03-20T16:05:04 <ThomasWaldmann> keys=root,xmlrpc
2008-03-20T16:05:22 <TheSheep> hmm.. the AdvancedSearch macro doesn't seem to take any arguments, why is it listed on the 'to convert' list? should I remove it?
2008-03-20T16:05:27 <ThomasWaldmann> [logger_xmlrpc]
2008-03-20T16:05:27 <ThomasWaldmann> level=DEBUG
2008-03-20T16:05:27 <ThomasWaldmann> handlers=logfile
2008-03-20T16:05:27 <ThomasWaldmann> qualname=MoinMoin.xmlrpc
2008-03-20T16:05:27 <ThomasWaldmann> propagate=0
2008-03-20T16:05:51 <ThomasWaldmann> just change/add that and also set the handler to DEBUG
2008-03-20T16:06:06 <ThomasWaldmann> then you get debug stuff just for xmlrpc
2008-03-20T16:06:32 <dreimark> will do
2008-03-20T16:12:58 <dreimark> TheSheep: isn't here a note to wait till bugfix?
2008-03-20T16:13:08 <ThomasWaldmann> gna http://mail.python.org/pipermail/web-sig/2008-January/003180.html
2008-03-20T16:13:47 <TheSheep> dreimark: ah, ok
2008-03-20T16:14:20 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3354:3f2e51a2dd08 1.7/MoinMoin/action/__init__.py: ActionBase: use do_show to honour request.rev
2008-03-20T16:14:24 <TheSheep> ok, then maybe you can tell me what should I do to not step on anybody's toes? :)
2008-03-20T16:16:37 <dreimark> TheSheep: Include, Navigation, NewPage can be converted
2008-03-20T16:16:51 <dreimark> or the xapian bug fixed
2008-03-20T16:17:15 <TheSheep> ok, I'll try the Navigation then
2008-03-20T16:17:24 <dreimark> :)
2008-03-20T16:18:52 <dreimark> ThomasWaldmann: sorry i missed where I have to enter the loggin lines
2008-03-20T16:20:02 <ThomasWaldmann> dreimark: in your logging configuration file (new in 1.7, see the server adaptor script)
2008-03-20T16:20:20 <ThomasWaldmann> there are samples for those in wiki/config/logging/
2008-03-20T16:20:51 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3355:19a7a6efaa21 1.7/MoinMoin/action/__init__.py: ActionBase: allow is_allowed() to return a tuple (allowed, message)
2008-03-20T16:20:55 <ThomasWaldmann> strange that we did not notice that wsgi problem yet
2008-03-20T16:21:14 <ThomasWaldmann> post data with more than 8000 bytes should happen rather often
2008-03-20T16:22:03 <TheSheep> ThomasWaldmann: the problem with NULL in the c string->python string?
2008-03-20T16:22:42 <ThomasWaldmann> ?
2008-03-20T16:23:04 <TheSheep> ThomasWaldmann: nah, I just had some user with such a werid problem with mod_wsgi
2008-03-20T16:23:12 <ThomasWaldmann> (16:13) < ThomasWal> gna http://mail.python.org/pipermail/web-sig/2008-January/003180.html
2008-03-20T16:23:47 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3356:2dd4904cfee4 1.7/MoinMoin/ (PageEditor.py action/revert.py theme/__init__.py):
2008-03-20T16:23:47 <CIA-39> better revert action, make action menu GET again
2008-03-20T16:23:47 <CIA-39> implements FeatureRequests/CommentsForReverts, because
2008-03-20T16:23:47 <CIA-39> it is two-stage now the action menu can be GET again
2008-03-20T16:26:03 <TheSheep> ThomasWaldmann: madness
2008-03-20T16:26:19 * TheSheep reviews his wiki engine to check this part of code
2008-03-20T16:27:47 <ThomasWaldmann> so what do they expect? just loop reading 8000 bytes until you get less than 8000?
2008-03-20T16:28:03 <ThomasWaldmann> (for the case you just want to get everything)
2008-03-20T16:28:27 <xorAxAx> TheSheep: xmlrpc cant contian NULL bytes
2008-03-20T16:28:37 <xorAxAx> for a wellformed request
2008-03-20T16:28:45 <ThomasWaldmann> or will you also get less than 8000 if not more has arrived yet?
2008-03-20T16:29:18 <TheSheep> ThomasWaldmann: I just check the Content-length header and read that much
2008-03-20T16:29:24 <ThomasWaldmann> or shall i just read 9999999999999 bytes? :P
2008-03-20T16:29:48 <TheSheep> ThomasWaldmann: i.e. "the traditional way"
2008-03-20T16:30:08 <ThomasWaldmann> cl is not mandatory
2008-03-20T16:30:17 <TheSheep> ThomasWaldmann: I think it is in posts
2008-03-20T16:30:46 <TheSheep> at least in HTML 1.0
2008-03-20T16:30:58 <TheSheep> HTTP
2008-03-20T16:31:00 <TheSheep> sorry
2008-03-20T16:31:28 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3357:bb04265ce89a 1.7/MoinMoin/action/revert.py: fix revert action. I suck.
2008-03-20T16:31:44 <ThomasWaldmann> well, I rather search an easy fix for read() with wsgi
2008-03-20T16:32:13 <johill> ok, done. bleh, I suck, stupid mistake
2008-03-20T16:32:13 <TheSheep> hmm... 1.1 only says SHOULD
2008-03-20T16:32:27 <ThomasWaldmann> lol
2008-03-20T16:32:36 <dreimark> <Fault 1: '<class \'xml.parsers.expat.ExpatError\'>
2008-03-20T16:32:48 <johill> why does ActionBase default to GET?
2008-03-20T16:32:52 <dreimark> I tried to put CHANGES
2008-03-20T16:33:26 <ThomasWaldmann> dreimark: the server side debug log is more interesting, esp. the end of the raw request data
2008-03-20T16:34:04 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3358:795ee44690a0 1.7/MoinMoin/action/__init__.py: ActionBase: default to POST
2008-03-20T16:34:37 * ThomasWaldmann hopes you reviewed all usage of that
2008-03-20T16:35:19 <johill> yes
2008-03-20T16:35:32 <johill> it's used in a bunch and none of them actually should use GET
2008-03-20T16:36:24 <ThomasWaldmann> and they still work now? :)
2008-03-20T16:36:25 <dreimark> ThomasWaldmann: I have lots of output on the console but the moin.log file has 0 bytes
2008-03-20T16:37:27 <ThomasWaldmann> you can configure that in your logging configuration
2008-03-20T16:37:38 <ThomasWaldmann> if you don't, it just uses some builtin default
2008-03-20T16:41:46 <johill> yes, I did, in fact, test them all
2008-03-20T16:44:30 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3613:ee218809ba7b 1.7-storage-hwendel/ (15 files in 8 dirs): merge with main branch
2008-03-20T16:50:04 <johill> lanius: actually, having both a filter function and a filter dict would be good because the filter dict can optimise some things much easier
2008-03-20T16:52:27 * ThomasWaldmann has a deja heard
2008-03-20T16:52:54 <johill> well I said earlier that a filter fn is better because it allows you to optimise it to an iterator
2008-03-20T16:53:10 <johill> but a filter dict is better because it allows you to skip complete branches of a meta backend
2008-03-20T16:53:30 <johill> so most effective would be both
2008-03-20T16:54:36 <TheSheep> excuse me, but where exactly the new parameter parser is located? I checked the util and the macro dirs...
2008-03-20T16:55:06 <ThomasWaldmann> wikiutil
2008-03-20T16:55:29 <johill> ThomasWaldmann: wikiutil.invoke_extension_function
2008-03-20T16:56:37 <TheSheep> ThomasWaldmann: got it, thanks
2008-03-20T17:03:00 <dreimark> TheSheep: you can look at EmbedObject too it is currently one with most keywords
2008-03-20T17:03:08 <dreimark> and already converted
2008-03-20T17:03:13 <TheSheep> great
2008-03-20T17:05:32 <johill> wow
2008-03-20T17:05:39 <johill> that indexing backend is a pessimisation
2008-03-20T17:05:42 <johill> wasn't it supposed to be an optimisation?
2008-03-20T17:06:10 * dreimark has extracted 100 bytes of CHANGES which can't be send
2008-03-20T17:06:51 <xorAxAx> johill: "indexing backend"?
2008-03-20T17:07:51 <johill> xorAxAx: storage branch, MoinMoin.storage.filesystem.IndexedBackend
2008-03-20T17:08:21 <xorAxAx> the one with bdb?
2008-03-20T17:08:50 <johill> not sure
2008-03-20T17:09:00 <johill> but it first builds a complete list and then ands it with the index
2008-03-20T17:09:17 <johill> not like building a list from the index and then filtering it further would be the right thing to do...
2008-03-20T17:18:31 <dreimark> does xmlrpc has problems to send unicode chars ?
2008-03-20T17:18:31 <dreimark> does xmlrpc has problems to send unicode chars ?
2008-03-20T17:18:34 <xorAxAx> mitsuhiko: you want to mentor for moin? :)
2008-03-20T17:18:34 <xorAxAx> mitsuhiko: you want to mentor for moin? :)
2008-03-20T17:18:43 <mitsuhiko> xorAxAx: yes
2008-03-20T17:18:44 <mitsuhiko> xorAxAx: yes
2008-03-20T17:18:52 <xorAxAx> TheSheep doesnt seem to want to
2008-03-20T17:18:52 <xorAxAx> TheSheep doesnt seem to want to
2008-03-20T17:19:09 <xorAxAx> mitsuhiko: any preferred topic?
2008-03-20T17:19:09 <xorAxAx> mitsuhiko: any preferred topic?
2008-03-20T17:19:14 <TheSheep> xorAxAx: ?
2008-03-20T17:19:14 <TheSheep> xorAxAx: ?
2008-03-20T17:19:25 <xorAxAx> TheSheep: you didnt apply as a mentor
2008-03-20T17:19:25 <xorAxAx> TheSheep: you didnt apply as a mentor
2008-03-20T17:19:36 <dreimark> TheSheep: you have to fill out the mentor form
2008-03-20T17:19:36 <dreimark> TheSheep: you have to fill out the mentor form
2008-03-20T17:19:39 <TheSheep> xorAxAx: I'd like to note that I'm still not accepted by my mentoring organisation as a mentor
2008-03-20T17:19:39 <TheSheep> xorAxAx: I'd like to note that I'm still not accepted by my mentoring organisation as a mentor
2008-03-20T17:19:47 <mitsuhiko> xorAxAx: let's see what the students pick
2008-03-20T17:19:48 <mitsuhiko> xorAxAx: let's see what the students pick
2008-03-20T17:19:50 <TheSheep> xorAxAx: but I did, 2 days ago
2008-03-20T17:19:50 <TheSheep> xorAxAx: but I did, 2 days ago
2008-03-20T17:19:51 <xorAxAx> TheSheep: you didnt apply accourding to the website :)
2008-03-20T17:19:51 <xorAxAx> TheSheep: you didnt apply accourding to the website :)
2008-03-20T17:19:59 <xorAxAx> s/our/or/
2008-03-20T17:19:59 <xorAxAx> s/our/or/
2008-03-20T17:20:28 <xorAxAx> http://code.google.com/soc/mentor_step1.html
2008-03-20T17:20:28 <xorAxAx> http://code.google.com/soc/mentor_step1.html
2008-03-20T17:20:44 <dreimark> you have to select MM it is not preselected
2008-03-20T17:20:44 <dreimark> you have to select MM it is not preselected
2008-03-20T17:21:50 <TheSheep> xorAxAx: when I go there, I see my already filled form
2008-03-20T17:21:50 <TheSheep> xorAxAx: when I go there, I see my already filled form
2008-03-20T17:22:12 <dreimark> TheSheep: do you have selected the MM project
2008-03-20T17:22:12 <dreimark> TheSheep: do you have selected the MM project
2008-03-20T17:22:35 <TheSheep> I have selected it, I'm re-sending it now
2008-03-20T17:22:35 <TheSheep> I have selected it, I'm re-sending it now
2008-03-20T17:22:45 <xorAxAx> now you appeared
2008-03-20T17:22:45 <xorAxAx> now you appeared
2008-03-20T17:22:54 <xorAxAx> feel free to thank the intern who wrote it :)
2008-03-20T17:22:54 <xorAxAx> feel free to thank the intern who wrote it :)
2008-03-20T17:23:15 <TheSheep> %)
2008-03-20T17:23:24 <TheSheep> %)
2008-03-20T17:24:06 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3614:9f8ed5fc5c51 1.7-storage-hwendel/MoinMoin/storage/ (8 files in 3 dirs): allow passing filter function to backend/itemcollection
2008-03-20T17:24:06 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3614:9f8ed5fc5c51 1.7-storage-hwendel/MoinMoin/storage/ (8 files in 3 dirs): allow passing filter function to backend/itemcollection
2008-03-20T17:24:24 <TheSheep> bbl
2008-03-20T17:24:24 <TheSheep> bbl
2008-03-20T17:25:13 <dennda> oh mitsuhiko is a mentor too? nice :)
2008-03-20T17:25:14 <ThomasWaldmann> return self.stdin.read(2000000000) # 2 GB should be enough for everybody
2008-03-20T17:25:14 <ThomasWaldmann> MemoryError
2008-03-20T17:25:14 <dennda> oh mitsuhiko is a mentor too? nice :)
2008-03-20T17:25:29 <ThomasWaldmann> return self.stdin.read(2000000000) # 2 GB should be enough for everybody
2008-03-20T17:25:29 <ThomasWaldmann> MemoryError
2008-03-20T17:26:09 <xorAxAx> ThomasWaldmann: who wrote that? :)
2008-03-20T17:26:10 <xorAxAx> ThomasWaldmann: who wrote that? :)
2008-03-20T17:26:34 <ThomasWaldmann> looks like I have to write a loop, reading tiny pieces and reassembling them :|
2008-03-20T17:26:34 <ThomasWaldmann> looks like I have to write a loop, reading tiny pieces and reassembling them :|
2008-03-20T17:26:48 <ThomasWaldmann> xorAxAx: see the url i posted above
2008-03-20T17:26:48 <ThomasWaldmann> xorAxAx: see the url i posted above
2008-03-20T17:27:04 <ThomasWaldmann> I am just trying to work around a borked api
2008-03-20T17:27:04 <ThomasWaldmann> I am just trying to work around a borked api
2008-03-20T17:27:40 <xorAxAx> i cant find any url that contains that code
2008-03-20T17:27:40 <xorAxAx> i cant find any url that contains that code
2008-03-20T17:28:06 <ThomasWaldmann> the url contains the problem, that code is trying to solve
2008-03-20T17:28:06 <ThomasWaldmann> the url contains the problem, that code is trying to solve
2008-03-20T17:29:17 <xorAxAx> the line of code doesnt look particularly smart :)
2008-03-20T17:29:17 <xorAxAx> the line of code doesnt look particularly smart :)
2008-03-20T17:29:27 <dennda> in which context is *that* used?
2008-03-20T17:29:27 <dennda> in which context is *that* used?
2008-03-20T17:29:39 <dreimark> hey looks to me like a joke
2008-03-20T17:29:39 <dreimark> hey looks to me like a joke
2008-03-20T17:30:04 <xorAxAx> dreimark: well, in enterprisey applications people write such code seriously :-)
2008-03-20T17:30:04 <xorAxAx> dreimark: well, in enterprisey applications people write such code seriously :-)
2008-03-20T17:30:48 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3615:c29cd256dff0 1.7-storage-hwendel/MoinMoin/Page.py: use new filterfn for getPageList
2008-03-20T17:30:48 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3615:c29cd256dff0 1.7-storage-hwendel/MoinMoin/Page.py: use new filterfn for getPageList
2008-03-20T17:31:13 <dennda> xorAxAx: which enterprise applications in particular? :)
2008-03-20T17:31:13 <dennda> xorAxAx: which enterprise applications in particular? :)
2008-03-20T17:31:40 <xorAxAx> dennda: those that are used internally in large enterprises :)
2008-03-20T17:31:40 <xorAxAx> dennda: those that are used internally in large enterprises :)
2008-03-20T17:31:46 <ThomasWaldmann> read http://mail.python.org/pipermail/web-sig/2008-January/003180.html before asking questions
2008-03-20T17:31:46 <ThomasWaldmann> read http://mail.python.org/pipermail/web-sig/2008-January/003180.html before asking questions
2008-03-20T17:35:28 <johill> the common idiom would be to read until it returns 0
2008-03-20T17:35:28 <johill> the common idiom would be to read until it returns 0
2008-03-20T17:35:33 <johill> in say 8kb sized chunks
2008-03-20T17:35:34 <johill> in say 8kb sized chunks
2008-03-20T17:36:54 <dennda> Just curious, is there any benefit in choosing another chunksize? (I am doing something similar at the moment)
2008-03-20T17:36:54 <dennda> Just curious, is there any benefit in choosing another chunksize? (I am doing something similar at the moment)
2008-03-20T17:37:11 <xorAxAx> johill: bzzt, 8 kiB :)
2008-03-20T17:37:12 <xorAxAx> johill: bzzt, 8 kiB :)
2008-03-20T17:37:30 <johill> read xkcd on that
2008-03-20T17:37:30 <johill> read xkcd on that
2008-03-20T17:37:52 <johill> dennda: in the kernel you'd do page-allocation ;)
2008-03-20T17:37:52 <johill> dennda: in the kernel you'd do page-allocation ;)
2008-03-20T17:37:56 <dennda> I chose 1 kilobyte as chunksize according to the sources I will be reading with that (want to update a progressbar)
2008-03-20T17:37:57 <dennda> I chose 1 kilobyte as chunksize according to the sources I will be reading with that (want to update a progressbar)
2008-03-20T17:38:19 <johill> it really depends on your memory allocator
2008-03-20T17:38:19 <johill> it really depends on your memory allocator
2008-03-20T17:40:29 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3359:dd6293254e4f 1.7/docs/CHANGES: CHANGES: fixed Juergens Name
2008-03-20T17:40:29 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3359:dd6293254e4f 1.7/docs/CHANGES: CHANGES: fixed Juergens Name
2008-03-20T17:42:56 <dreimark> ThomasWaldmann: xmlrpc works on localhost with large files too
2008-03-20T17:42:56 <dreimark> ThomasWaldmann: xmlrpc works on localhost with large files too
2008-03-20T17:45:24 <dreimark> but we should think about a page size limit.
2008-03-20T17:45:24 <dreimark> but we should think about a page size limit.
2008-03-20T17:54:15 <ThomasWaldmann> dreimark: in 5 mins we have a wsgi workaround
2008-03-20T17:54:15 <ThomasWaldmann> dreimark: in 5 mins we have a wsgi workaround
2008-03-20T17:54:20 <dreimark> :)
2008-03-20T17:54:21 <dreimark> :)
2008-03-20T17:54:48 <dreimark> in the past there was a limit for the browsers text editor
2008-03-20T17:54:48 <dreimark> in the past there was a limit for the browsers text editor
2008-03-20T17:55:51 <johill> not really
2008-03-20T17:55:51 <johill> not really
2008-03-20T17:55:58 <johill> you can always post with wget or so :)
2008-03-20T17:55:58 <johill> you can always post with wget or so :)
2008-03-20T17:56:56 <dennda> wget supports post operations?
2008-03-20T17:56:56 <dennda> wget supports post operations?
2008-03-20T17:57:23 <johill> of course
2008-03-20T17:57:23 <johill> of course
2008-03-20T17:57:56 <dennda> very comfortable ;)
2008-03-20T17:57:57 <dennda> very comfortable ;)
2008-03-20T17:58:23 <TheSheep> wget supports posts, cookies, etc.
2008-03-20T17:58:23 <TheSheep> wget supports posts, cookies, etc.
2008-03-20T17:58:50 <dennda> we all love it
2008-03-20T17:58:50 <dennda> we all love it
2008-03-20T17:59:09 <johill> dd if=/dev/null bs=1M seek=100000 of=/tmp/sparse-huge
2008-03-20T17:59:09 <johill> dd if=/dev/null bs=1M seek=100000 of=/tmp/sparse-huge
2008-03-20T17:59:17 <johill> wget --post-file=/tmp/sparse-huge ...
2008-03-20T17:59:17 <johill> wget --post-file=/tmp/sparse-huge ...
2008-03-20T17:59:30 <johill> will probably cause you try allocate that much memory because the Content-Length header says so ;)
2008-03-20T17:59:30 <johill> will probably cause you try allocate that much memory because the Content-Length header says so ;)
2008-03-20T18:00:27 <TheSheep> I think it even supports gzip
2008-03-20T18:00:27 <TheSheep> I think it even supports gzip
2008-03-20T18:00:50 <johill> that wiki markup failing test
2008-03-20T18:00:50 <johill> that wiki markup failing test
2008-03-20T18:00:57 <johill> could anyone bother fixing it?
2008-03-20T18:00:57 <johill> could anyone bother fixing it?
2008-03-20T18:01:04 <ThomasWaldmann> summary: work around unpythonic WSGI 1.0 read() API, fixing broken xmlrpc putPage with mod_wsgi
2008-03-20T18:01:04 <ThomasWaldmann> summary: work around unpythonic WSGI 1.0 read() API, fixing broken xmlrpc putPage with mod_wsgi
2008-03-20T18:01:32 <ThomasWaldmann> http://hg.moinmo.in/moin/1.7/rev/b5b49ff88d0c
2008-03-20T18:01:32 <ThomasWaldmann> http://hg.moinmo.in/moin/1.7/rev/b5b49ff88d0c
2008-03-20T18:01:41 <johill> no bot?
2008-03-20T18:01:41 <johill> no bot?
2008-03-20T18:01:55 <ThomasWaldmann> i didn't use push in that case :)
2008-03-20T18:01:56 <ThomasWaldmann> i didn't use push in that case :)
2008-03-20T18:02:22 <TheSheep> backdoor acces! not fair! :P
2008-03-20T18:02:22 <TheSheep> backdoor acces! not fair! :P
2008-03-20T18:03:01 <johill> rofl
2008-03-20T18:03:01 <johill> rofl
2008-03-20T18:03:07 <johill> "Zentralrat der Juden will YouTube verklagen"
2008-03-20T18:03:07 <johill> "Zentralrat der Juden will YouTube verklagen"
2008-03-20T18:03:32 <dennda> Any non-german speakers in here?
2008-03-20T18:03:32 <dennda> Any non-german speakers in here?
2008-03-20T18:03:38 <johill> probably
2008-03-20T18:03:39 <johill> probably
2008-03-20T18:03:39 <TheSheep> o/
2008-03-20T18:03:39 <TheSheep> o/
2008-03-20T18:04:24 <johill> "<central jew counsel> wants to use YOuTube"
2008-03-20T18:04:24 <johill> "<central jew counsel> wants to use YOuTube"
2008-03-20T18:04:28 <johill> sue
2008-03-20T18:04:28 <johill> sue
2008-03-20T18:05:14 <TheSheep> nihil novi sub sole
2008-03-20T18:05:14 <TheSheep> nihil novi sub sole
2008-03-20T18:05:27 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3616:dbc778342390 1.7-storage-hwendel/MoinMoin/storage/external.py: small bugfix for previous commit
2008-03-20T18:05:27 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3616:dbc778342390 1.7-storage-hwendel/MoinMoin/storage/external.py: small bugfix for previous commit
2008-03-20T18:05:28 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3617:2767513328cb 1.7-storage-hwendel/MoinMoin/ (5 files in 3 dirs): allow list_items to be an iterator, fix ItemCollection's dictmixin
2008-03-20T18:05:28 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3617:2767513328cb 1.7-storage-hwendel/MoinMoin/ (5 files in 3 dirs): allow list_items to be an iterator, fix ItemCollection's dictmixin
2008-03-20T18:05:29 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3618:9df175d2535f 1.7-storage-hwendel/MoinMoin/ (25 files in 10 dirs):
2008-03-20T18:05:29 <CIA-39> make getPageList return an iterable and use the backend's iterable for it,
2008-03-20T18:05:29 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3618:9df175d2535f 1.7-storage-hwendel/MoinMoin/ (25 files in 10 dirs):
2008-03-20T18:05:29 <CIA-39> make getPageList return an iterable and use the backend's iterable for it,
2008-03-20T18:05:31 <CIA-39> fix more bugs with the ItemCollection dict interface
2008-03-20T18:05:31 <CIA-39> fix more bugs with the ItemCollection dict interface
2008-03-20T18:05:37 <dennda> nothing new under the sun?
2008-03-20T18:05:37 <dennda> nothing new under the sun?
2008-03-20T18:05:58 <TheSheep> dennda: yup, nothing
2008-03-20T18:05:58 <TheSheep> dennda: yup, nothing
2008-03-20T18:06:02 <dreimark> some sentences are quite nice. "Haste makes waste."
2008-03-20T18:06:02 <dreimark> some sentences are quite nice. "Haste makes waste."
2008-03-20T18:06:28 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3619:31c001f46b02 1.7-storage-hwendel/MoinMoin/storage/backends/moin16.py: remove stray print statement
2008-03-20T18:06:28 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3619:31c001f46b02 1.7-storage-hwendel/MoinMoin/storage/backends/moin16.py: remove stray print statement
2008-03-20T18:06:39 <TheSheep> haste waste mahts? :P
2008-03-20T18:06:39 <TheSheep> haste waste mahts? :P
2008-03-20T18:10:28 <dreimark> 10MB seems to work for the texteditor but the action is quite slow
2008-03-20T18:10:28 <dreimark> 10MB seems to work for the texteditor but the action is quite slow
2008-03-20T18:11:25 <johill> heh
2008-03-20T18:11:25 <johill> heh
2008-03-20T18:12:04 <johill> ok, done
2008-03-20T18:12:04 <johill> ok, done
2008-03-20T18:12:04 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3620:8b08a50bee5c 1.7-storage-hwendel/MoinMoin/storage/backends/ (filesystem.py moin16.py moin17.py): make moin16/moin17 backends return an iterable instead of a list for list_items
2008-03-20T18:12:04 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3620:8b08a50bee5c 1.7-storage-hwendel/MoinMoin/storage/backends/ (filesystem.py moin16.py moin17.py): make moin16/moin17 backends return an iterable instead of a list for list_items
2008-03-20T18:12:12 <johill> item lists are iterable all the way through now
2008-03-20T18:12:12 <johill> item lists are iterable all the way through now
2008-03-20T18:12:13 <johill> where possible
2008-03-20T18:12:13 <johill> where possible
2008-03-20T18:12:19 <johill> i.e. dict.keys() can't be, it has to be a list
2008-03-20T18:12:20 <johill> i.e. dict.keys() can't be, it has to be a list
2008-03-20T18:12:25 <johill> but dict.iteritems() can be
2008-03-20T18:12:25 <johill> but dict.iteritems() can be
2008-03-20T18:13:08 <johill> oh wait
2008-03-20T18:13:08 <johill> oh wait
2008-03-20T18:13:11 <johill> I think I f'ed that up
2008-03-20T18:13:11 <johill> I think I f'ed that up
2008-03-20T18:15:54 <dreimark> does one tried to revert a page ?
2008-03-20T18:15:54 <dreimark> does one tried to revert a page ?
2008-03-20T18:16:44 <johill> hm?
2008-03-20T18:16:45 <johill> hm?
2008-03-20T18:18:08 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3621:37a2d7082a0b 1.7-storage-hwendel/MoinMoin/ (4 files in 3 dirs): fix iteritems mess, I was thinking something else for a moment...
2008-03-20T18:18:08 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3621:37a2d7082a0b 1.7-storage-hwendel/MoinMoin/ (4 files in 3 dirs): fix iteritems mess, I was thinking something else for a moment...
2008-03-20T18:18:11 <johill> there, that's better
2008-03-20T18:18:11 <johill> there, that's better
2008-03-20T18:21:05 <dreimark> acl_rights_default = u"All:read,write,revert,admin ReimarBauer:read,write,delete,revert,admin"
2008-03-20T18:21:05 <dreimark> acl_rights_default = u"All:read,write,revert,admin ReimarBauer:read,write,delete,revert,admin"
2008-03-20T18:21:14 <dreimark> what is wrong with that line
2008-03-20T18:21:14 <dreimark> what is wrong with that line
2008-03-20T18:21:44 <dreimark> I can revert any page with the default config
2008-03-20T18:21:44 <dreimark> I can revert any page with the default config
2008-03-20T18:23:41 <johill> what happens?
2008-03-20T18:23:41 <johill> what happens?
2008-03-20T18:24:14 <dreimark> I get the message I am not allowed to do it
2008-03-20T18:24:15 <dreimark> I get the message I am not allowed to do it
2008-03-20T18:24:35 <dreimark> ahh
2008-03-20T18:24:35 <dreimark> ahh
2008-03-20T18:24:53 <dreimark> I have to define the rules switched
2008-03-20T18:24:53 <dreimark> I have to define the rules switched
2008-03-20T18:24:59 <dreimark> All at the end works
2008-03-20T18:24:59 <dreimark> All at the end works
2008-03-20T18:25:06 <johill> oh heh yeah
2008-03-20T18:25:06 <johill> oh heh yeah
2008-03-20T18:25:28 <johill> bbl
2008-03-20T18:25:40 <johill> bbl
2008-03-20T18:36:24 <ThomasWaldmann> ok, xmlrpc works now, except that applyAuthToken does not work in 1.7
2008-03-20T18:36:24 <ThomasWaldmann> ok, xmlrpc works now, except that applyAuthToken does not work in 1.7
2008-03-20T18:37:24 <johill> hm
2008-03-20T18:37:24 <johill> hm
2008-03-20T18:38:38 <ThomasWaldmann> applyAuthToken says "SUCCESS' btw :D
2008-03-20T18:38:38 <ThomasWaldmann> applyAuthToken says "SUCCESS' btw :D
2008-03-20T18:39:15 <ThomasWaldmann> [['SUCCESS'], ['You are <unknown user>. valid=0.'], {'faultCode': 1, 'faultString
2008-03-20T18:39:16 <ThomasWaldmann> [['SUCCESS'], ['You are <unknown user>. valid=0.'], {'faultCode': 1, 'faultString
2008-03-20T18:39:18 <ThomasWaldmann> ': 'You are not allowed to edit this page'}, ['SUCCESS']]
2008-03-20T18:39:18 <ThomasWaldmann> ': 'You are not allowed to edit this page'}, ['SUCCESS']]
2008-03-20T18:39:30 <johill> heh
2008-03-20T18:39:30 <johill> heh
2008-03-20T18:39:32 <johill> not sure
2008-03-20T18:39:32 <johill> not sure
2008-03-20T18:39:37 <ThomasWaldmann> applyAuthToken, WhoAmI, putPage, deleteAuthToken
2008-03-20T18:39:37 <ThomasWaldmann> applyAuthToken, WhoAmI, putPage, deleteAuthToken
2008-03-20T18:44:14 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2577:5d63e1ba710e 1.6/MoinMoin/i18n/tools/po2wiki.py: i18n tools: use multicall and getAuthToken instead of http basic auth transport
2008-03-20T18:44:14 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2577:5d63e1ba710e 1.6/MoinMoin/i18n/tools/po2wiki.py: i18n tools: use multicall and getAuthToken instead of http basic auth transport
2008-03-20T18:44:14 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2578:cfec24f33a70 1.6/MoinMoin/request/request_wsgi.py: merged main
2008-03-20T18:44:14 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2578:cfec24f33a70 1.6/MoinMoin/request/request_wsgi.py: merged main
2008-03-20T18:44:15 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3361:8fca4e0c0de5 1.7/MoinMoin/action/diff.py: action.diff: added a revert button
2008-03-20T18:44:15 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3361:8fca4e0c0de5 1.7/MoinMoin/action/diff.py: action.diff: added a revert button
2008-03-20T18:45:30 <johill> I'm stumped
2008-03-20T18:45:30 <johill> I'm stumped
2008-03-20T18:45:38 <johill> the storage test code is too complex for me
2008-03-20T18:45:38 <johill> the storage test code is too complex for me
2008-03-20T18:45:44 <johill> can't figure out why it fails
2008-03-20T18:45:44 <johill> can't figure out why it fails
2008-03-20T18:45:50 <dennda> johill: you are in todays top ten :)
2008-03-20T18:45:51 <dennda> johill: you are in todays top ten :)
2008-03-20T18:46:09 <johill> where?
2008-03-20T18:46:09 <johill> where?
2008-03-20T18:46:17 <dennda> http://cia.vc/
2008-03-20T18:46:17 <dennda> http://cia.vc/
2008-03-20T18:46:23 <johill> oh
2008-03-20T18:46:23 <johill> oh
2008-03-20T18:47:32 <ThomasWaldmann> hehe
2008-03-20T18:47:32 <ThomasWaldmann> hehe
2008-03-20T18:48:13 <johill> hrm
2008-03-20T18:48:13 <johill> hrm
2008-03-20T18:48:21 <johill> the tests seem to rely on the iterator returning sorted values?!?!
2008-03-20T18:48:22 <johill> the tests seem to rely on the iterator returning sorted values?!?!
2008-03-20T18:48:26 <dreimark> and we are currently before Mediawiki
2008-03-20T18:48:26 <dreimark> and we are currently before Mediawiki
2008-03-20T18:49:03 <johill> indeed
2008-03-20T18:49:03 <johill> indeed
2008-03-20T18:49:04 <dennda> dreimark: yes
2008-03-20T18:49:04 <dreimark> does the iconsize depends there on the ranking ?
2008-03-20T18:49:04 <dreimark> does the iconsize depends there on the ranking ?
2008-03-20T18:49:04 <dennda> dreimark: yes
2008-03-20T18:49:04 <johill> wtf
2008-03-20T18:49:04 <johill> wtf
2008-03-20T18:49:07 <johill> those tests are borked
2008-03-20T18:49:07 <johill> those tests are borked
2008-03-20T18:49:16 <dennda> when I opened that page 10 minutes ago both projects had 40 commits
2008-03-20T18:49:16 <dennda> when I opened that page 10 minutes ago both projects had 40 commits
2008-03-20T18:49:41 <dennda> dreimark: looks like, but i doubt it
2008-03-20T18:49:41 <dennda> dreimark: looks like, but i doubt it
2008-03-20T18:50:08 <dennda> gentoo has quite a few commits
2008-03-20T18:50:09 <dennda> gentoo has quite a few commits
2008-03-20T18:50:28 <dreimark> it looks like they do use our favicon.ico and not the logo
2008-03-20T18:50:28 <dreimark> it looks like they do use our favicon.ico and not the logo
2008-03-20T18:51:42 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3622:ab204128d98a 1.7-storage-hwendel/MoinMoin/storage/backends/meta.py: meta backends: make list_items iterator
2008-03-20T18:51:42 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3622:ab204128d98a 1.7-storage-hwendel/MoinMoin/storage/backends/meta.py: meta backends: make list_items iterator
2008-03-20T18:51:47 <johill> now to fix the test... arg
2008-03-20T18:51:47 <johill> now to fix the test... arg
2008-03-20T18:56:30 * dreimark goes shopping bbl
2008-03-20T18:56:30 * dreimark goes shopping bbl
2008-03-20T18:56:35 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3623:a3dffc901021 1.7-storage-hwendel/MoinMoin/storage/_tests/ (__init__.py test_backends_meta.py): fix storage tests (they assume sorting which isn't true)
2008-03-20T18:56:35 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3623:a3dffc901021 1.7-storage-hwendel/MoinMoin/storage/_tests/ (__init__.py test_backends_meta.py): fix storage tests (they assume sorting which isn't true)
2008-03-20T18:56:39 <johill> that's better
2008-03-20T18:56:39 <johill> that's better
2008-03-20T18:56:41 <johill> duh
2008-03-20T18:56:41 <johill> duh
2008-03-20T19:00:00 <johill> it works
2008-03-20T19:00:00 <johill> it works
2008-03-20T19:00:10 <johill> python no longer needs a few MB memory to build the lists
2008-03-20T19:00:10 <johill> python no longer needs a few MB memory to build the lists
2008-03-20T19:00:23 <johill> of course, not all code supports that, some just does a list() on the iterator...
2008-03-20T19:00:23 <johill> of course, not all code supports that, some just does a list() on the iterator...
2008-03-20T19:00:26 <johill> needs to be checked
2008-03-20T19:00:26 <johill> needs to be checked
2008-03-20T19:01:40 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3625:b5b49ff88d0c 1.7-storage-hwendel/MoinMoin/request/request_wsgi.py: work around unpythonic WSGI 1.0 read() API, fixing broken xmlrpc putPage with mod_wsgi
2008-03-20T19:01:40 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3625:b5b49ff88d0c 1.7-storage-hwendel/MoinMoin/request/request_wsgi.py: work around unpythonic WSGI 1.0 read() API, fixing broken xmlrpc putPage with mod_wsgi
2008-03-20T19:01:41 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3627:6d1da41e7c43 1.7-storage-hwendel/ (3 files in 3 dirs): merge main branch
2008-03-20T19:01:41 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3627:6d1da41e7c43 1.7-storage-hwendel/ (3 files in 3 dirs): merge main branch
2008-03-20T19:06:28 <TheSheep> hm, there is no ChoiceArgument, I shpould just use a string?
2008-03-20T19:06:28 <TheSheep> hm, there is no ChoiceArgument, I shpould just use a string?
2008-03-20T19:06:58 <TheSheep> in the argument parser, I mean
2008-03-20T19:06:58 <TheSheep> in the argument parser, I mean
2008-03-20T19:26:13 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2579:aab6ff8746cf 1.6/MoinMoin/i18n/ (12 files in 2 dirs): updated i18n, fixed target url in po2wiki
2008-03-20T19:26:13 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2579:aab6ff8746cf 1.6/MoinMoin/i18n/ (12 files in 2 dirs): updated i18n, fixed target url in po2wiki
2008-03-20T19:26:14 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2580:1497696d587d 1.6/MoinMoin/i18n/mk.MoinMoin.po: i18n: added Macedonian system text translation
2008-03-20T19:26:14 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2580:1497696d587d 1.6/MoinMoin/i18n/mk.MoinMoin.po: i18n: added Macedonian system text translation
2008-03-20T19:32:53 <johill> TheSheep: just use a tuple
2008-03-20T19:32:53 <johill> TheSheep: just use a tuple
2008-03-20T19:33:08 <johill> TheSheep: arg=(u'a', u'b')
2008-03-20T19:33:08 <johill> TheSheep: arg=(u'a', u'b')
2008-03-20T19:34:55 <johill> lanius: "If changing an underlay page it will no longer be copied to the standard pages directory, but [...]"
2008-03-20T19:34:55 <johill> lanius: "If changing an underlay page it will no longer be copied to the standard pages directory, but [...]"
2008-03-20T19:34:58 <TheSheep> johill: ah, great
2008-03-20T19:34:58 <TheSheep> johill: ah, great
2008-03-20T19:35:01 <johill> lanius: is that by design? it feels wrong
2008-03-20T19:35:01 <johill> lanius: is that by design? it feels wrong
2008-03-20T19:36:05 <xorAxAx> johill: see the logs
2008-03-20T19:36:05 <xorAxAx> johill: see the logs
2008-03-20T19:36:12 <xorAxAx> johill: was discussed a few times
2008-03-20T19:36:12 <xorAxAx> johill: was discussed a few times
2008-03-20T19:36:19 <johill> bleh, searching through logs suck
2008-03-20T19:36:19 <johill> bleh, searching through logs suck
2008-03-20T19:36:27 <xorAxAx> not really
2008-03-20T19:36:27 <xorAxAx> not really
2008-03-20T19:36:29 <johill> the design decision should be documented right there in the file
2008-03-20T19:36:29 <johill> the design decision should be documented right there in the file
2008-03-20T19:36:33 <xorAxAx> good point
2008-03-20T19:36:33 <xorAxAx> good point
2008-03-20T19:36:52 <xorAxAx> (moins search is so powerful, should be fun to use it if it worked)
2008-03-20T19:36:52 <xorAxAx> (moins search is so powerful, should be fun to use it if it worked)
2008-03-20T19:39:11 <vpv> there's a problem with moinmaster logins, or at least my login
2008-03-20T19:39:11 <vpv> there's a problem with moinmaster logins, or at least my login
2008-03-20T19:40:06 <johill> xorAxAx: do you remember where I can possibly find the discussion about that?
2008-03-20T19:40:07 <johill> xorAxAx: do you remember where I can possibly find the discussion about that?
2008-03-20T19:40:19 <vpv> if I log in and then go edit http://master.moinmo.in/MoinI18n/fi it logs me out when clicking on "edit" and doesn't let me edit the page at all
2008-03-20T19:40:19 <vpv> if I log in and then go edit http://master.moinmo.in/MoinI18n/fi it logs me out when clicking on "edit" and doesn't let me edit the page at all
2008-03-20T19:41:06 <johill> vpv: I had some trouble with sessions on moinmo.in, try maybe deleting your cookie
2008-03-20T19:41:06 <johill> vpv: I had some trouble with sessions on moinmo.in, try maybe deleting your cookie
2008-03-20T19:42:26 <dreimark> I am not sure if we do want to remove the info that a page change was done by a revert
2008-03-20T19:42:27 <dreimark> I am not sure if we do want to remove the info that a page change was done by a revert
2008-03-20T19:42:46 <johill> dreimark: it's still there
2008-03-20T19:42:46 <johill> dreimark: it's still there
2008-03-20T19:42:54 <johill> dreimark: the edit log has SAVE/REVERT
2008-03-20T19:42:54 <johill> dreimark: the edit log has SAVE/REVERT
2008-03-20T19:42:56 <dreimark> not if you enter a comment
2008-03-20T19:42:56 <dreimark> not if you enter a comment
2008-03-20T19:42:56 <vpv> johill: thanks, that worked
2008-03-20T19:42:56 <vpv> johill: thanks, that worked
2008-03-20T19:43:09 <johill> dreimark: yes, it is, it's just not shown
2008-03-20T19:43:09 <johill> dreimark: yes, it is, it's just not shown
2008-03-20T19:43:16 <johill> dreimark: feel free to change the info action :)
2008-03-20T19:43:16 <johill> dreimark: feel free to change the info action :)
2008-03-20T19:43:31 <dreimark> ok
2008-03-20T19:43:32 <dreimark> ok
2008-03-20T19:43:52 <johill> I considered doing so but the action seemed written as such
2008-03-20T19:43:52 <johill> I considered doing so but the action seemed written as such
2008-03-20T19:44:23 <johill> vpv: it's weird and I have no idea what could be causing that, but then again I don't want to play with 1.6's session code either
2008-03-20T19:44:23 <johill> vpv: it's weird and I have no idea what could be causing that, but then again I don't want to play with 1.6's session code either
2008-03-20T19:44:25 <xorAxAx> johill: http://moinmo.in/?action=fullsearch&context=180&value=t%3Achat+r%3Acopy.%2Aunderlay&fullsearch=Text
2008-03-20T19:44:25 <xorAxAx> johill: http://moinmo.in/?action=fullsearch&context=180&value=t%3Achat+r%3Acopy.%2Aunderlay&fullsearch=Text
2008-03-20T19:44:32 <xorAxAx> johill: call me mr. full text search :)
2008-03-20T19:44:32 <xorAxAx> johill: call me mr. full text search :)
2008-03-20T19:44:47 <xorAxAx> even the excerpt looks nice
2008-03-20T19:44:47 <xorAxAx> even the excerpt looks nice
2008-03-20T19:44:57 <johill> hah I didn't include "copy" and found nothing useful
2008-03-20T19:44:57 <johill> hah I didn't include "copy" and found nothing useful
2008-03-20T19:45:17 <dreimark> I do want to kill Revert to this revision from the more action menu.
2008-03-20T19:45:17 <dreimark> I do want to kill Revert to this revision from the more action menu.
2008-03-20T19:45:24 <dreimark> It is now on the diff page
2008-03-20T19:45:24 <dreimark> It is now on the diff page
2008-03-20T19:45:34 <dreimark> different opnions?
2008-03-20T19:45:34 <dreimark> different opnions?
2008-03-20T19:46:05 <dreimark> ah no forget view from the info
2008-03-20T19:46:05 <dreimark> ah no forget view from the info
2008-03-20T19:46:13 <johill> ?
2008-03-20T19:46:13 <johill> ?
2008-03-20T19:46:22 <dreimark> I thought it is duplicated now
2008-03-20T19:46:22 <dreimark> I thought it is duplicated now
2008-03-20T19:46:46 <dreimark> but you can view on the info and go to this revision by the action in the menu
2008-03-20T19:46:46 <dreimark> but you can view on the info and go to this revision by the action in the menu
2008-03-20T19:46:46 <johill> oh I hate 1.6
2008-03-20T19:46:46 <johill> oh I hate 1.6
2008-03-20T19:47:22 <dreimark> heh, it just born some month ago
2008-03-20T19:47:22 <dreimark> heh, it just born some month ago
2008-03-20T19:48:39 <johill> xorAxAx: I don't think I understand that
2008-03-20T19:48:40 <johill> xorAxAx: I don't think I understand that
2008-03-20T19:49:40 <johill> you seem to be saying that copy-underlay shouldn't be done because of underlay updates?
2008-03-20T19:49:40 <johill> you seem to be saying that copy-underlay shouldn't be done because of underlay updates?
2008-03-20T19:50:28 <xorAxAx> johill: first of all, the question is why the current semantics make sense
2008-03-20T19:50:28 <xorAxAx> johill: first of all, the question is why the current semantics make sense
2008-03-20T19:50:51 <TheSheep> first question is 'do they make sense?' ;)
2008-03-20T19:50:53 <johill> because underlay is not only used for help pages
2008-03-20T19:50:54 <TheSheep> first question is 'do they make sense?' ;)
2008-03-20T19:50:54 <johill> because underlay is not only used for help pages
2008-03-20T19:50:58 <xorAxAx> the disadvantages have been seen - major usability issues like e.g. no way to revert rev 1 in the data dir to rev 1 in the underlay dir
2008-03-20T19:50:58 <xorAxAx> the disadvantages have been seen - major usability issues like e.g. no way to revert rev 1 in the data dir to rev 1 in the underlay dir
2008-03-20T19:51:15 <xorAxAx> johill: please elaborate
2008-03-20T19:51:15 <xorAxAx> johill: please elaborate
2008-03-20T19:51:15 <johill> granted, the use cases you noted with the revert stuff etc are bogus
2008-03-20T19:51:15 <johill> granted, the use cases you noted with the revert stuff etc are bogus
2008-03-20T19:51:32 <johill> currently
2008-03-20T19:51:32 <johill> currently
2008-03-20T19:51:34 <xorAxAx> so we thought about existing use cases and found solution
2008-03-20T19:51:34 <xorAxAx> so we thought about existing use cases and found solution
2008-03-20T19:51:35 <xorAxAx> s
2008-03-20T19:51:35 <xorAxAx> s
2008-03-20T19:51:38 <johill> but they can easily be fixed in the layer backend
2008-03-20T19:51:38 <johill> but they can easily be fixed in the layer backend
2008-03-20T19:51:54 <xorAxAx> johill: no, thats a semantic issue - you would need to show all layers to the users
2008-03-20T19:51:54 <xorAxAx> johill: no, thats a semantic issue - you would need to show all layers to the users
2008-03-20T19:52:17 <xorAxAx> and then we derived solutions that seemed simple and sound
2008-03-20T19:52:17 <xorAxAx> and then we derived solutions that seemed simple and sound
2008-03-20T19:52:18 <johill> no
2008-03-20T19:52:18 <johill> no
2008-03-20T19:52:19 <johill> not layers
2008-03-20T19:52:20 <johill> not layers
2008-03-20T19:52:23 <johill> only the pages throughout the layers
2008-03-20T19:52:23 <johill> only the pages throughout the layers
2008-03-20T19:52:27 <xorAxAx> hmm?
2008-03-20T19:52:28 <xorAxAx> hmm?
2008-03-20T19:52:44 <johill> I mean if you have a page with 10 revisions in the upper and 5 in the lower layer you'd show 15 revisions
2008-03-20T19:52:44 <johill> I mean if you have a page with 10 revisions in the upper and 5 in the lower layer you'd show 15 revisions
2008-03-20T19:53:07 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2581:357a2f83fc65 1.6/docs/CHANGES: updated CHANGES
2008-03-20T19:53:07 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2581:357a2f83fc65 1.6/docs/CHANGES: updated CHANGES
2008-03-20T19:53:19 <xorAxAx> johill: and you can only write to the upper layer?
2008-03-20T19:53:19 <xorAxAx> johill: and you can only write to the upper layer?
2008-03-20T19:53:27 <johill> say I have a wiki which uses the regular underlay
2008-03-20T19:53:27 <johill> say I have a wiki which uses the regular underlay
2008-03-20T19:53:43 <johill> but I want PermissionDeniedPage to tell users to please log in because that's how my wiki is configured
2008-03-20T19:53:43 <johill> but I want PermissionDeniedPage to tell users to please log in because that's how my wiki is configured
2008-03-20T19:53:55 <johill> so I go and edit that page
2008-03-20T19:53:55 <johill> so I go and edit that page
2008-03-20T19:53:58 <xorAxAx> when should that page show up?
2008-03-20T19:53:58 <xorAxAx> when should that page show up?
2008-03-20T19:54:01 <johill> with the current semantics all is fine
2008-03-20T19:54:01 <johill> with the current semantics all is fine
2008-03-20T19:54:20 <johill> xorAxAx: whenever permissions are denied to view a page, I just did that, like MissingPage
2008-03-20T19:54:20 <johill> xorAxAx: whenever permissions are denied to view a page, I just did that, like MissingPage
2008-03-20T19:54:33 <johill> with the new semantics an underlay update would kill my changes
2008-03-20T19:54:33 <johill> with the new semantics an underlay update would kill my changes
2008-03-20T19:54:35 <johill> that's no good
2008-03-20T19:54:35 <johill> that's no good
2008-03-20T19:55:18 <xorAxAx> no, not at all
2008-03-20T19:55:18 <xorAxAx> no, not at all
2008-03-20T19:55:22 <johill> take MissingPage if you wish insteda of PermissionDeniedPage
2008-03-20T19:55:22 <johill> take MissingPage if you wish insteda of PermissionDeniedPage
2008-03-20T19:55:35 <xorAxAx> you can modify your underlay pages and updating them can still work
2008-03-20T19:55:35 <xorAxAx> you can modify your underlay pages and updating them can still work
2008-03-20T19:55:45 <xorAxAx> thats completly unrelated to copy-on-write semantics
2008-03-20T19:55:45 <xorAxAx> thats completly unrelated to copy-on-write semantics
2008-03-20T19:56:06 <johill> sort of
2008-03-20T19:56:06 <johill> sort of
2008-03-20T19:56:07 <xorAxAx> i think i described that there
2008-03-20T19:56:07 <xorAxAx> i think i described that there
2008-03-20T19:56:20 <johill> different use case then: sharing underlay between different wikis
2008-03-20T19:56:20 <johill> different use case then: sharing underlay between different wikis
2008-03-20T19:56:28 <johill> where one wants to override that and another doesn't
2008-03-20T19:56:28 <johill> where one wants to override that and another doesn't
2008-03-20T19:56:41 <xorAxAx> hmm, please explain the last sentence, what is "that"?
2008-03-20T19:56:41 <xorAxAx> hmm, please explain the last sentence, what is "that"?
2008-03-20T19:57:50 <johill> which sentence?
2008-03-20T19:57:50 <johill> which sentence?
2008-03-20T19:58:59 <xorAxAx> 19:56:28 < johill> where one wants to override that and another doesn't
2008-03-20T19:58:59 <xorAxAx> 19:56:28 < johill> where one wants to override that and another doesn't
2008-03-20T19:59:13 <johill> oh
2008-03-20T19:59:13 <johill> oh
2008-03-20T19:59:25 <johill> one of the wikis in my farm edits MissingPage and the other doesn't
2008-03-20T19:59:25 <johill> one of the wikis in my farm edits MissingPage and the other doesn't
2008-03-20T19:59:37 <johill> with the new stuff I can no longer share underlay which sucks a lot because it's huge
2008-03-20T19:59:37 <johill> with the new stuff I can no longer share underlay which sucks a lot because it's huge
2008-03-20T20:00:59 <johill> I rely on that copying feature :)
2008-03-20T20:00:59 <johill> I rely on that copying feature :)
2008-03-20T20:01:11 <xorAxAx> yes, we didnt think about that use case
2008-03-20T20:01:11 <xorAxAx> yes, we didnt think about that use case
2008-03-20T20:01:47 <xorAxAx> well, then it might be necessary, maybe
2008-03-20T20:01:47 <xorAxAx> well, then it might be necessary, maybe
2008-03-20T20:01:50 <johill> I agree that currently with delete and stuff it's rather weirdthe weird current semantics with
2008-03-20T20:01:50 <johill> I agree that currently with delete and stuff it's rather weirdthe weird current semantics with
2008-03-20T20:02:01 <johill> eh, scratch everything after the first weird :)
2008-03-20T20:02:01 <johill> eh, scratch everything after the first weird :)
2008-03-20T20:02:35 <johill> but that can be fixed in the layer backend by making it iterate through, and possibly having a sort of "this page was deleted in upper backend don't show in lower" flag
2008-03-20T20:02:35 <johill> but that can be fixed in the layer backend by making it iterate through, and possibly having a sort of "this page was deleted in upper backend don't show in lower" flag
2008-03-20T20:02:53 <johill> (in fact I think that's how it will be anyhow right now due to the way deleted works)
2008-03-20T20:02:53 <johill> (in fact I think that's how it will be anyhow right now due to the way deleted works)
2008-03-20T20:02:54 <xorAxAx> yes, that needs a bit of design
2008-03-20T20:02:54 <xorAxAx> yes, that needs a bit of design
2008-03-20T20:03:14 <xorAxAx> but hopefully it will be done in a clean and configurable way
2008-03-20T20:03:14 <xorAxAx> but hopefully it will be done in a clean and configurable way
2008-03-20T20:03:38 <johill> I suppose the only configurable thing here is whether to edit in the top layer or not
2008-03-20T20:03:38 <johill> I suppose the only configurable thing here is whether to edit in the top layer or not
2008-03-20T20:03:42 <johill> do you see anything else?
2008-03-20T20:03:42 <johill> do you see anything else?
2008-03-20T20:04:03 <xorAxAx> no idea, i am still not completly sure what the initial intentions were
2008-03-20T20:04:03 <xorAxAx> no idea, i am still not completly sure what the initial intentions were
2008-03-20T20:04:10 <xorAxAx> maybe you can find some documented ones on the wiki
2008-03-20T20:04:10 <xorAxAx> maybe you can find some documented ones on the wiki
2008-03-20T20:04:15 <johill> which initial intentions?
2008-03-20T20:04:15 <johill> which initial intentions?
2008-03-20T20:04:25 <xorAxAx> of the underlay feature
2008-03-20T20:04:25 <xorAxAx> of the underlay feature
2008-03-20T20:04:37 <johill> I don't think that really matters
2008-03-20T20:04:37 <johill> I don't think that really matters
2008-03-20T20:04:46 <johill> we have use cases for both ways
2008-03-20T20:04:46 <johill> we have use cases for both ways
2008-03-20T20:04:53 <xorAxAx> yes
2008-03-20T20:04:53 <xorAxAx> yes
2008-03-20T20:05:51 <johill> and the current behaviour is to copy so one could argue that should be default
2008-03-20T20:05:51 <johill> and the current behaviour is to copy so one could argue that should be default
2008-03-20T20:06:00 <johill> doesn't matter too much to me to be honest
2008-03-20T20:06:00 <johill> doesn't matter too much to me to be honest
2008-03-20T20:06:51 <ThomasWaldmann> the main intention for underlay was to have s/h pages separate, so you can simply replace the underlay dir on upgrade
2008-03-20T20:06:51 <ThomasWaldmann> the main intention for underlay was to have s/h pages separate, so you can simply replace the underlay dir on upgrade
2008-03-20T20:07:12 <ThomasWaldmann> before, s/h pages were within the data_dir, mixed with user's content
2008-03-20T20:07:12 <ThomasWaldmann> before, s/h pages were within the data_dir, mixed with user's content
2008-03-20T20:09:11 <johill> I just realised that I have a bug that sucks and fixing it sucks too :(
2008-03-20T20:09:11 <johill> I just realised that I have a bug that sucks and fixing it sucks too :(
2008-03-20T20:11:09 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3628:44db10af7919 1.7-storage-hwendel/MoinMoin/storage/backends/meta.py: layer backend: don't list a page from multiple layers more than once
2008-03-20T20:11:09 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3628:44db10af7919 1.7-storage-hwendel/MoinMoin/storage/backends/meta.py: layer backend: don't list a page from multiple layers more than once
2008-03-20T20:13:15 <xorAxAx> johill: i dont see a unittest
2008-03-20T20:13:15 <xorAxAx> johill: i dont see a unittest
2008-03-20T20:13:42 <johill> yeah I'm just looking into that
2008-03-20T20:13:42 <johill> yeah I'm just looking into that
2008-03-20T20:13:50 <johill> the tests lanius wrote are quite obscure
2008-03-20T20:13:50 <johill> the tests lanius wrote are quite obscure
2008-03-20T20:14:05 <johill> looks like millions of classes but no concept ;)
2008-03-20T20:14:05 <johill> looks like millions of classes but no concept ;)
2008-03-20T20:14:30 <johill> everything configurable, but that's pretty useless
2008-03-20T20:14:31 <johill> everything configurable, but that's pretty useless
2008-03-20T20:14:35 <johill> I think I'll just write a bunch of new tests
2008-03-20T20:14:35 <johill> I think I'll just write a bunch of new tests
2008-03-20T20:14:54 <xorAxAx> yes, why not :)
2008-03-20T20:14:54 <xorAxAx> yes, why not :)
2008-03-20T20:15:06 <johill> and a memory backend first
2008-03-20T20:15:06 <johill> and a memory backend first
2008-03-20T20:17:17 <johill> actually, let me first refactor the AbstractBackend that really is part abstract and part filesystem-base
2008-03-20T20:17:17 <johill> actually, let me first refactor the AbstractBackend that really is part abstract and part filesystem-base
2008-03-20T20:18:43 <xorAxAx> i would write the tests first, then refactor :)
2008-03-20T20:18:43 <xorAxAx> i would write the tests first, then refactor :)
2008-03-20T20:18:55 <johill> can't write tests without a backend
2008-03-20T20:18:55 <johill> can't write tests without a backend
2008-03-20T20:19:18 <lanius> sure you can ;)
2008-03-20T20:19:19 <lanius> sure you can ;)
2008-03-20T20:19:22 <johill> and I don't want to implement a proper in-core backend without the base class
2008-03-20T20:19:22 <johill> and I don't want to implement a proper in-core backend without the base class
2008-03-20T20:19:26 <lanius> johill: why is returning a sorted list wrong?
2008-03-20T20:19:27 <lanius> johill: why is returning a sorted list wrong?
2008-03-20T20:19:43 <johill> lanius: it isn't, but it's unnecessary
2008-03-20T20:19:43 <johill> lanius: it isn't, but it's unnecessary
2008-03-20T20:20:00 <lanius> johill: you want the page list to be unsorted?
2008-03-20T20:20:00 <lanius> johill: you want the page list to be unsorted?
2008-03-20T20:20:02 <johill> relying on it is wrong (or you should have documented that it will always return a sorted list)
2008-03-20T20:20:02 <johill> relying on it is wrong (or you should have documented that it will always return a sorted list)
2008-03-20T20:20:34 <xorAxAx> i guess johill is thinking of the listdir() function
2008-03-20T20:20:34 <xorAxAx> i guess johill is thinking of the listdir() function
2008-03-20T20:20:37 <johill> lanius: yes. if anyone needs it sorted they can do it, many users don't so it's much more efficient
2008-03-20T20:20:37 <johill> lanius: yes. if anyone needs it sorted they can do it, many users don't so it's much more efficient
2008-03-20T20:22:13 <lanius> johill: there is a concept behind the unit tests
2008-03-20T20:22:13 <lanius> johill: there is a concept behind the unit tests
2008-03-20T20:22:34 <johill> I can see that, but I can't see what it is
2008-03-20T20:22:34 <johill> I can see that, but I can't see what it is
2008-03-20T20:22:50 <lanius> johill: it's that you have the data for your test wiki created automatically
2008-03-20T20:22:50 <lanius> johill: it's that you have the data for your test wiki created automatically
2008-03-20T20:23:07 <lanius> and the abstractbackendtest should work for all backends
2008-03-20T20:23:07 <lanius> and the abstractbackendtest should work for all backends
2008-03-20T20:23:08 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3629:86c3e2c3aed5 1.7-storage-hwendel/MoinMoin/storage/backends/ (filesystem.py moin16.py moin17.py): rename filesystems.AbstractBackend to filesystems.BaseFilesystemBackend
2008-03-20T20:23:08 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3629:86c3e2c3aed5 1.7-storage-hwendel/MoinMoin/storage/backends/ (filesystem.py moin16.py moin17.py): rename filesystems.AbstractBackend to filesystems.BaseFilesystemBackend
2008-03-20T20:23:22 <johill> lanius: yeah but why all the lists all over then?
2008-03-20T20:23:22 <johill> lanius: yeah but why all the lists all over then?
2008-03-20T20:23:29 <lanius> johill: what lists?
2008-03-20T20:23:30 <lanius> johill: what lists?
2008-03-20T20:23:56 <johill> like "new_items[]"
2008-03-20T20:23:56 <johill> like "new_items[]"
2008-03-20T20:24:16 <johill> anyway, it's ok, I'll write some more tests
2008-03-20T20:24:16 <johill> anyway, it's ok, I'll write some more tests
2008-03-20T20:24:33 <johill> if you don't understand my tests and I don't understand yours, then that doesn't matter either ;)
2008-03-20T20:24:33 <johill> if you don't understand my tests and I don't understand yours, then that doesn't matter either ;)
2008-03-20T20:25:00 <lanius> johill: that is the data for the test wiki
2008-03-20T20:25:00 <lanius> johill: that is the data for the test wiki
2008-03-20T20:26:45 <johill> right
2008-03-20T20:26:45 <johill> right
2008-03-20T20:27:20 <lanius> which will automatically be created based on this lists
2008-03-20T20:27:20 <lanius> which will automatically be created based on this lists
2008-03-20T20:28:58 <johill> I think it's just too much magic for me
2008-03-20T20:28:58 <johill> I think it's just too much magic for me
2008-03-20T20:29:25 <lanius> might be ;)
2008-03-20T20:29:25 <lanius> might be ;)
2008-03-20T20:29:52 <lanius> your iterator improvement is not of much use if you have to cache the list in LayerBackend again, is it?
2008-03-20T20:29:52 <lanius> your iterator improvement is not of much use if you have to cache the list in LayerBackend again, is it?
2008-03-20T20:30:45 <johill> sort of, yeah
2008-03-20T20:30:45 <johill> sort of, yeah
2008-03-20T20:31:06 <johill> it still is an improvement if somebody doesn't want all items
2008-03-20T20:31:06 <johill> it still is an improvement if somebody doesn't want all items
2008-03-20T20:32:50 <lanius> your pagelist isn't sorted anymore, is it?
2008-03-20T20:32:51 <lanius> your pagelist isn't sorted anymore, is it?
2008-03-20T20:33:53 <johill> right
2008-03-20T20:33:53 <johill> right
2008-03-20T20:34:42 <TheSheep> funny, every other py.test run returns 5 faileds tests on 1.7
2008-03-20T20:34:42 <TheSheep> funny, every other py.test run returns 5 faileds tests on 1.7
2008-03-20T20:35:13 <TheSheep> even ones have 2 failures
2008-03-20T20:35:13 <TheSheep> even ones have 2 failures
2008-03-20T20:36:07 <johill> lanius: can you explain _parse_value?
2008-03-20T20:36:07 <johill> lanius: can you explain _parse_value?
2008-03-20T20:37:05 <johill> lanius: it makes no sense to me at all especially since the else will always execute
2008-03-20T20:37:05 <johill> lanius: it makes no sense to me at all especially since the else will always execute
2008-03-20T20:37:10 <TheSheep> 666 passed %)
2008-03-20T20:37:10 <TheSheep> 666 passed %)
2008-03-20T20:37:22 <johill> TheSheep: weird I only have two failures
2008-03-20T20:37:22 <johill> TheSheep: weird I only have two failures
2008-03-20T20:37:44 <TheSheep> ====== tests finished: 666 passed, 2 failed, 65 skipped in 64.00 seconds =======
2008-03-20T20:37:44 <TheSheep> ====== tests finished: 666 passed, 2 failed, 65 skipped in 64.00 seconds =======
2008-03-20T20:37:50 <lanius> johill: which module
2008-03-20T20:37:51 <lanius> johill: which module
2008-03-20T20:38:05 <johill> storage.backends.filesystem
2008-03-20T20:38:05 <johill> storage.backends.filesystem
2008-03-20T20:38:47 <TheSheep> johill: nevermind, it's those wikidicts tests
2008-03-20T20:38:47 <TheSheep> johill: nevermind, it's those wikidicts tests
2008-03-20T20:39:16 <johill> lanius: is meta data supposed to be always unicode strings or not?
2008-03-20T20:39:16 <johill> lanius: is meta data supposed to be always unicode strings or not?
2008-03-20T20:39:41 <lanius> johill: no, it can be lists or tuples or dicts as well
2008-03-20T20:39:42 <lanius> johill: no, it can be lists or tuples or dicts as well
2008-03-20T20:39:50 <lanius> johill: see encode/decode list/dict in moin16.py
2008-03-20T20:39:50 <lanius> johill: see encode/decode list/dict in moin16.py
2008-03-20T20:40:13 <johill> ok so why do you set meta[DELETED] = u'True' rather than just boolean True?
2008-03-20T20:40:13 <johill> ok so why do you set meta[DELETED] = u'True' rather than just boolean True?
2008-03-20T20:40:25 <lanius> johill: because boolean is not supported
2008-03-20T20:40:25 <lanius> johill: because boolean is not supported
2008-03-20T20:41:07 <johill> couldn't it just pickle them?
2008-03-20T20:41:21 <johill> couldn't it just pickle them?
2008-03-20T20:41:29 <johill> lanius: but _parse_value needs to be changed a lot then
2008-03-20T20:41:29 <johill> lanius: but _parse_value needs to be changed a lot then
2008-03-20T20:41:39 <johill> not 'dict' but dict etc
2008-03-20T20:41:39 <johill> not 'dict' but dict etc
2008-03-20T20:41:47 <lanius> no
2008-03-20T20:41:47 <lanius> no
2008-03-20T20:42:14 <johill> type doesn't return a string
2008-03-20T20:42:14 <johill> type doesn't return a string
2008-03-20T20:42:40 <johill> >>> type({}) in ['dict']
2008-03-20T20:42:40 <johill> >>> type({}) in ['dict']
2008-03-20T20:42:41 <johill> False
2008-03-20T20:42:41 <johill> False
2008-03-20T20:42:42 <lanius> strange
2008-03-20T20:42:43 <lanius> strange
2008-03-20T20:42:48 <johill> it never did
2008-03-20T20:42:48 <johill> it never did
2008-03-20T20:42:54 <lanius> is there another method?
2008-03-20T20:42:55 <lanius> is there another method?
2008-03-20T20:43:02 <TheSheep> isinstance
2008-03-20T20:43:03 <TheSheep> isinstance
2008-03-20T20:43:05 <johill> >>> type({}) in [dict]
2008-03-20T20:43:05 <johill> >>> type({}) in [dict]
2008-03-20T20:43:05 <johill> True
2008-03-20T20:43:06 <johill> True
2008-03-20T20:43:12 <johill> isinstance is better though
2008-03-20T20:43:12 <johill> isinstance is better though
2008-03-20T20:43:14 <johill> imho
2008-03-20T20:43:15 <johill> imho
2008-03-20T20:44:08 <lanius> johill: you could pickle it, but it must be compatible with 1.6 so it isn't
2008-03-20T20:44:09 <lanius> johill: you could pickle it, but it must be compatible with 1.6 so it isn't
2008-03-20T20:44:31 <lanius> furthermore the value is in the page itself so no pickling there, just strings
2008-03-20T20:44:31 <lanius> furthermore the value is in the page itself so no pickling there, just strings
2008-03-20T20:44:33 <johill> ok
2008-03-20T20:44:33 <johill> ok
2008-03-20T20:44:38 <johill> good point
2008-03-20T20:44:38 <johill> good point
2008-03-20T20:44:51 <johill> but deleted is very special
2008-03-20T20:44:51 <johill> but deleted is very special
2008-03-20T20:44:57 <johill> it could still have a bool value, no?
2008-03-20T20:44:57 <johill> it could still have a bool value, no?
2008-03-20T20:45:03 <lanius> yes
2008-03-20T20:45:03 <lanius> yes
2008-03-20T20:45:33 <johill> I'll do that change after I'm done with this
2008-03-20T20:45:33 <johill> I'll do that change after I'm done with this
2008-03-20T20:49:48 <lanius> johill: i acutally tested that the indexed backend is faster then without it
2008-03-20T20:49:48 <lanius> johill: i acutally tested that the indexed backend is faster then without it
2008-03-20T20:50:42 <johill> seriously? hard to believe
2008-03-20T20:50:42 <johill> seriously? hard to believe
2008-03-20T20:50:56 <johill> maybe because fewer tests are done
2008-03-20T20:50:56 <johill> maybe because fewer tests are done
2008-03-20T20:51:04 <johill> if the order is reversed it should be even faster
2008-03-20T20:51:04 <johill> if the order is reversed it should be even faster
2008-03-20T20:51:43 <lanius> johill: reversed order?
2008-03-20T20:51:43 <lanius> johill: reversed order?
2008-03-20T20:52:28 <johill> well not building the whole page list and then using the index but using the index to build the first matches list
2008-03-20T20:52:28 <johill> well not building the whole page list and then using the index but using the index to build the first matches list
2008-03-20T20:53:04 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3630:f28cdd68fe64 1.7-storage-hwendel/MoinMoin/storage/ (5 files in 2 dirs): move BaseFilesystemBackend._filter to a common check_filter function
2008-03-20T20:53:04 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3630:f28cdd68fe64 1.7-storage-hwendel/MoinMoin/storage/ (5 files in 2 dirs): move BaseFilesystemBackend._filter to a common check_filter function
2008-03-20T20:54:00 <lanius> johill: i don't understand yet
2008-03-20T20:54:00 <lanius> johill: i don't understand yet
2008-03-20T20:55:05 <johill> you're getting a whole page list
2008-03-20T20:55:05 <johill> you're getting a whole page list
2008-03-20T20:55:07 <johill> of all pages
2008-03-20T20:55:08 <johill> of all pages
2008-03-20T20:55:18 <johill> and then you subtract whatever you can by the index
2008-03-20T20:55:18 <johill> and then you subtract whatever you can by the index
2008-03-20T20:55:30 <johill> it should be faster to build the initial page list from the index and then filter it further
2008-03-20T20:55:30 <johill> it should be faster to build the initial page list from the index and then filter it further
2008-03-20T20:55:34 <lanius> johill: about what mehtod are you talking
2008-03-20T20:55:35 <lanius> johill: about what mehtod are you talking
2008-03-20T20:56:01 <johill> IndexedBackend.list_items
2008-03-20T20:56:01 <johill> IndexedBackend.list_items
2008-03-20T20:57:54 <lanius> johill: go on
2008-03-20T20:57:54 <lanius> johill: go on
2008-03-20T20:58:24 <johill> that's all
2008-03-20T20:58:24 <johill> that's all
2008-03-20T20:58:31 <lanius> i don't understand your complaint yet
2008-03-20T20:58:31 <lanius> i don't understand your complaint yet
2008-03-20T20:58:37 <lanius> explain it with the sourcecode
2008-03-20T20:58:37 <lanius> explain it with the sourcecode
2008-03-20T20:58:42 <johill> you're building a = (full list) and b = (list of all pages with X from index)
2008-03-20T20:58:42 <johill> you're building a = (full list) and b = (list of all pages with X from index)
2008-03-20T20:58:47 <johill> and then you return a & b
2008-03-20T20:58:47 <johill> and then you return a & b
2008-03-20T20:58:58 <johill> it would be more efficient to just build b and then filter it
2008-03-20T20:58:58 <johill> it would be more efficient to just build b and then filter it
2008-03-20T20:59:03 <lanius> yeah that's just a list operation, not very expensive
2008-03-20T20:59:04 <lanius> yeah that's just a list operation, not very expensive
2008-03-20T20:59:16 <lanius> the problem the indexedbackend solves is file system access
2008-03-20T20:59:17 <lanius> the problem the indexedbackend solves is file system access
2008-03-20T20:59:24 <lanius> which is really slow
2008-03-20T20:59:25 <lanius> which is really slow
2008-03-20T20:59:40 <johill> yeah but that's the thing
2008-03-20T20:59:40 <johill> yeah but that's the thing
2008-03-20T20:59:50 <johill> you get the list of pages which might come from the filesystem
2008-03-20T20:59:50 <johill> you get the list of pages which might come from the filesystem
2008-03-20T21:02:57 <lanius> i think a check for len(other_filters) == 0 is missing
2008-03-20T21:02:58 <lanius> i think a check for len(other_filters) == 0 is missing
2008-03-20T21:03:40 <johill> that's not the point
2008-03-20T21:03:40 <johill> that's not the point
2008-03-20T21:03:47 <johill> building the full page list can be expensive
2008-03-20T21:03:47 <johill> building the full page list can be expensive
2008-03-20T21:03:58 <johill> if you were to load the partial list from the index that'd be much cheaper
2008-03-20T21:03:58 <johill> if you were to load the partial list from the index that'd be much cheaper
2008-03-20T21:04:05 <johill> and the filters might need loading the page etc
2008-03-20T21:04:06 <johill> and the filters might need loading the page etc
2008-03-20T21:04:24 <lanius> i don't wan't to build the whole page list
2008-03-20T21:04:24 <lanius> i don't wan't to build the whole page list
2008-03-20T21:04:35 <lanius> the concept is the following
2008-03-20T21:04:35 <lanius> the concept is the following
2008-03-20T21:04:49 <lanius> there are metadata values which are indexed and some are not
2008-03-20T21:04:49 <lanius> there are metadata values which are indexed and some are not
2008-03-20T21:04:53 <johill> you do build the whole page list!
2008-03-20T21:04:53 <johill> you do build the whole page list!
2008-03-20T21:04:53 <lanius> (cfg.indexes)
2008-03-20T21:04:54 <lanius> (cfg.indexes)
2008-03-20T21:05:12 <lanius> if i have metadata values which are not indexed i need to build the whole page list
2008-03-20T21:05:12 <lanius> if i have metadata values which are not indexed i need to build the whole page list
2008-03-20T21:05:19 <johill> yes
2008-03-20T21:05:19 <johill> yes
2008-03-20T21:05:22 <johill> which is wrong
2008-03-20T21:05:22 <johill> which is wrong
2008-03-20T21:05:24 <lanius> no
2008-03-20T21:05:25 <lanius> no
2008-03-20T21:05:25 <johill> you don't have to
2008-03-20T21:05:25 <johill> you don't have to
2008-03-20T21:05:27 <lanius> why
2008-03-20T21:05:28 <lanius> why
2008-03-20T21:05:34 <johill> because you can use the indexed ones *first*
2008-03-20T21:05:34 <johill> because you can use the indexed ones *first*
2008-03-20T21:05:43 <lanius> but i know that they are not indexed
2008-03-20T21:05:43 <johill> rather than using the non-indexed ones first!
2008-03-20T21:05:43 <johill> rather than using the non-indexed ones first!
2008-03-20T21:05:43 <lanius> but i know that they are not indexed
2008-03-20T21:05:59 <johill> but then you only have to apply the non-indexed ones to a much smaller list
2008-03-20T21:05:59 <johill> but then you only have to apply the non-indexed ones to a much smaller list
2008-03-20T21:06:19 <lanius> but i don't have any list on this level
2008-03-20T21:06:19 <lanius> but i don't have any list on this level
2008-03-20T21:06:37 <lanius> i say self._backend.list_items(other_filters, filterfn)
2008-03-20T21:06:37 <lanius> i say self._backend.list_items(other_filters, filterfn)
2008-03-20T21:06:42 <lanius> no list there
2008-03-20T21:06:43 <lanius> no list there
2008-03-20T21:06:46 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3631:5e59592b711e 1.7-storage-hwendel/MoinMoin/ (6 files in 4 dirs): treat DELETED metadata specially, allow bool
2008-03-20T21:06:46 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3631:5e59592b711e 1.7-storage-hwendel/MoinMoin/ (6 files in 4 dirs): treat DELETED metadata specially, allow bool
2008-03-20T21:06:59 <johill> no you say
2008-03-20T21:06:59 <johill> no you say
2008-03-20T21:07:00 <johill> set(self._backend.list_items(other_filters, filterfn))
2008-03-20T21:07:00 <johill> set(self._backend.list_items(other_filters, filterfn))
2008-03-20T21:07:06 <johill> which builds the whole damn list
2008-03-20T21:07:06 <johill> which builds the whole damn list
2008-03-20T21:07:34 <lanius> sure but i have to do that if i have keys which are not indexed
2008-03-20T21:07:34 <lanius> sure but i have to do that if i have keys which are not indexed
2008-03-20T21:07:39 <johill> NO!
2008-03-20T21:07:39 <johill> NO!
2008-03-20T21:07:47 <johill> I'll just fix the code
2008-03-20T21:07:47 <johill> I'll just fix the code
2008-03-20T21:08:35 <lanius> ok, i'll look into it then
2008-03-20T21:08:35 <lanius> ok, i'll look into it then
2008-03-20T21:08:54 <lanius> but it is faster because if i don't use indexes the backend below has to parse all files
2008-03-20T21:08:54 <lanius> but it is faster because if i don't use indexes the backend below has to parse all files
2008-03-20T21:09:02 <lanius> in this case it only does a listdir
2008-03-20T21:09:03 <lanius> in this case it only does a listdir
2008-03-20T21:09:07 <lanius> that's one io vs. hundreds
2008-03-20T21:09:08 <lanius> that's one io vs. hundreds
2008-03-20T21:09:35 <johill> not if there are other filters
2008-03-20T21:09:36 <johill> not if there are other filters
2008-03-20T21:09:42 <lanius> yes
2008-03-20T21:09:42 <lanius> yes
2008-03-20T21:09:44 <johill> in which case it still has to parse all files
2008-03-20T21:09:44 <johill> in which case it still has to parse all files
2008-03-20T21:09:50 <johill> instead of just those that match the indexed filters
2008-03-20T21:09:50 <johill> instead of just those that match the indexed filters
2008-03-20T21:09:55 <johill> why is that so hard to understand?
2008-03-20T21:09:55 <johill> why is that so hard to understand?
2008-03-20T21:10:20 <lanius> because the _backend.list_items does not have any argument wich you can pass a list to
2008-03-20T21:10:20 <lanius> because the _backend.list_items does not have any argument wich you can pass a list to
2008-03-20T21:10:35 <johill> how does that matter?
2008-03-20T21:10:35 <johill> how does that matter?
2008-03-20T21:10:56 <lanius> so how do you tell the backend to only filter the filtered items?
2008-03-20T21:10:56 <lanius> so how do you tell the backend to only filter the filtered items?
2008-03-20T21:11:11 <johill> you do it yourself of course
2008-03-20T21:11:11 <johill> you do it yourself of course
2008-03-20T21:11:48 <xorAxAx> hopefully johill doesnt break the api separation
2008-03-20T21:11:48 <xorAxAx> hopefully johill doesnt break the api separation
2008-03-20T21:12:05 <lanius> johill: do it myself?
2008-03-20T21:12:05 <lanius> johill: do it myself?
2008-03-20T21:13:48 <johill> you can all show me how wrong I am in a minute
2008-03-20T21:13:48 <johill> you can all show me how wrong I am in a minute
2008-03-20T21:14:04 <lanius> give me a few more, will get something to eat first ;)
2008-03-20T21:14:04 <lanius> give me a few more, will get something to eat first ;)
2008-03-20T21:15:57 <johill> is there a test for indexed?
2008-03-20T21:15:58 <johill> is there a test for indexed?
2008-03-20T21:16:40 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2582:41c7ab4f0b56 1.6/setup.py: setup.py: use own server for download, remove unused settings
2008-03-20T21:16:40 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2582:41c7ab4f0b56 1.6/setup.py: setup.py: use own server for download, remove unused settings
2008-03-20T21:20:05 <TheSheep> johill: can I make a tuple a mandatory argument somehow? the required_arg complaints it's not a valid type
2008-03-20T21:20:05 <TheSheep> johill: can I make a tuple a mandatory argument somehow? the required_arg complaints it's not a valid type
2008-03-20T21:20:55 <dreimark> hmm. seems we don't define msg string in unicode
2008-03-20T21:20:55 <dreimark> hmm. seems we don't define msg string in unicode
2008-03-20T21:21:11 <johill> TheSheep: odd
2008-03-20T21:21:11 <johill> TheSheep: odd
2008-03-20T21:21:18 <johill> TheSheep: must be a bug in required_arg I'll check
2008-03-20T21:21:18 <johill> TheSheep: must be a bug in required_arg I'll check
2008-03-20T21:22:14 <dreimark> johill: is None allowed for a required_arg ?
2008-03-20T21:22:14 <dreimark> johill: is None allowed for a required_arg ?
2008-03-20T21:22:27 <johill> huh what sense would that make?
2008-03-20T21:22:28 <johill> huh what sense would that make?
2008-03-20T21:22:47 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3362:97fe3aadef34 1.7/MoinMoin/wikiutil.py: allow required_arg to take a tuple/list
2008-03-20T21:22:47 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3362:97fe3aadef34 1.7/MoinMoin/wikiutil.py: allow required_arg to take a tuple/list
2008-03-20T21:22:48 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3363:3b5cb4f28020 1.7/ (3 files in 3 dirs): merge with main
2008-03-20T21:22:48 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3363:3b5cb4f28020 1.7/ (3 files in 3 dirs): merge with main
2008-03-20T21:22:56 <johill> TheSheep: there
2008-03-20T21:22:56 <johill> TheSheep: there
2008-03-20T21:22:59 <dreimark> if None: in the macro
2008-03-20T21:22:59 <dreimark> if None: in the macro
2008-03-20T21:23:22 <dreimark> required_arg should tell the user has to enter something
2008-03-20T21:23:22 <dreimark> required_arg should tell the user has to enter something
2008-03-20T21:23:40 <dreimark> of course we can do set one as default
2008-03-20T21:23:40 <dreimark> of course we can do set one as default
2008-03-20T21:24:02 <johill> I don't understand
2008-03-20T21:24:02 <johill> I don't understand
2008-03-20T21:24:08 <johill> if you use required_arg you never get None
2008-03-20T21:24:08 <johill> if you use required_arg you never get None
2008-03-20T21:24:24 <dreimark> see EmbedObject
2008-03-20T21:24:24 <dreimark> see EmbedObject
2008-03-20T21:24:38 <dreimark> target is needed but i can't give a default value
2008-03-20T21:24:38 <dreimark> target is needed but i can't give a default value
2008-03-20T21:24:44 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 2583:06338654a51d 1.6/MoinMoin/wikiutil.py: allow required_arg to take a tuple/list (same as in 1.7)
2008-03-20T21:24:44 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 2583:06338654a51d 1.6/MoinMoin/wikiutil.py: allow required_arg to take a tuple/list (same as in 1.7)
2008-03-20T21:24:46 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 2584:8c384be0f932 1.6/ (25 files in 13 dirs): merge with main
2008-03-20T21:24:46 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 2584:8c384be0f932 1.6/ (25 files in 13 dirs): merge with main
2008-03-20T21:24:50 <dreimark> so it is arequired argument
2008-03-20T21:24:50 <dreimark> so it is arequired argument
2008-03-20T21:24:53 <johill> dreimark: the point of required_arg is that you don't give a default value
2008-03-20T21:24:53 <johill> dreimark: the point of required_arg is that you don't give a default value
2008-03-20T21:25:03 <johill> just use required_arg(u'') then if you want unicode
2008-03-20T21:25:03 <johill> just use required_arg(u'') then if you want unicode
2008-03-20T21:25:12 <johill> or required_arg(1.0) if you want a float
2008-03-20T21:25:12 <johill> or required_arg(1.0) if you want a float
2008-03-20T21:25:13 <dreimark> ah, an other misunderstanding by me
2008-03-20T21:25:13 <dreimark> ah, an other misunderstanding by me
2008-03-20T21:25:13 <johill> etc
2008-03-20T21:25:13 <johill> etc
2008-03-20T21:25:46 <dreimark> ok
2008-03-20T21:25:46 <dreimark> ok
2008-03-20T21:26:13 <johill> I suppose we could extend it to allowe required_arg(float)
2008-03-20T21:26:13 <johill> I suppose we could extend it to allowe required_arg(float)
2008-03-20T21:26:25 <johill> that would be a tad easier to understand maybe?
2008-03-20T21:26:25 <johill> that would be a tad easier to understand maybe?
2008-03-20T21:26:54 <xorAxAx> yes
2008-03-20T21:26:54 <xorAxAx> yes
2008-03-20T21:27:19 <johill> not sure why I did it the other way really
2008-03-20T21:27:19 <johill> not sure why I did it the other way really
2008-03-20T21:28:00 <dreimark> a type is better
2008-03-20T21:28:00 <dreimark> a type is better
2008-03-20T21:28:12 <johill> for a tuple though, you still need the tuple etc
2008-03-20T21:28:12 <johill> for a tuple though, you still need the tuple etc
2008-03-20T21:28:26 <johill> makes the code a bit more complex
2008-03-20T21:28:26 <johill> makes the code a bit more complex
2008-03-20T21:28:31 <johill> I'll review it once I figured out this test failure
2008-03-20T21:28:31 <johill> I'll review it once I figured out this test failure
2008-03-20T21:28:33 <xorAxAx> i dont understand the tuple issue
2008-03-20T21:28:33 <xorAxAx> i dont understand the tuple issue
2008-03-20T21:29:10 <johill> xorAxAx: a tuple means "one of"
2008-03-20T21:29:10 <johill> xorAxAx: a tuple means "one of"
2008-03-20T21:29:19 <johill> ie. arg=('red', 'green') means it has to be red or green
2008-03-20T21:29:19 <johill> ie. arg=('red', 'green') means it has to be red or green
2008-03-20T21:29:52 <johill> (same with a list though I guess we could just disallow a list)
2008-03-20T21:29:52 <johill> (same with a list though I guess we could just disallow a list)
2008-03-20T21:29:56 <TheSheep> johill: thanks!
2008-03-20T21:29:56 <TheSheep> johill: thanks!
2008-03-20T21:30:09 <johill> TheSheep: does it work properly?
2008-03-20T21:30:09 <johill> TheSheep: does it work properly?
2008-03-20T21:30:10 <xorAxAx> johill: ah
2008-03-20T21:30:10 <xorAxAx> johill: ah
2008-03-20T21:30:28 <johill> xorAxAx: with the first entry being default
2008-03-20T21:30:28 <johill> xorAxAx: with the first entry being default
2008-03-20T21:31:36 <TheSheep> johill: no
2008-03-20T21:31:36 <TheSheep> johill: no
2008-03-20T21:31:42 <dreimark> is is already wikiutil.required_arg(unicode)
2008-03-20T21:31:42 <dreimark> is is already wikiutil.required_arg(unicode)
2008-03-20T21:31:52 <TheSheep> johill: raises the same exception still
2008-03-20T21:31:53 <TheSheep> johill: raises the same exception still
2008-03-20T21:31:54 <johill> heh
2008-03-20T21:31:54 <johill> heh
2008-03-20T21:32:00 <johill> ok wait
2008-03-20T21:32:00 <johill> ok wait
2008-03-20T21:34:56 <TheSheep> johill: isinstance would seem to be better there too
2008-03-20T21:34:56 <TheSheep> johill: isinstance would seem to be better there too
2008-03-20T21:35:15 <TheSheep> johill: you can pass a tuple as its second arg
2008-03-20T21:35:15 <TheSheep> johill: you can pass a tuple as its second arg
2008-03-20T21:35:26 <johill> yeah
2008-03-20T21:35:26 <johill> yeah
2008-03-20T21:40:00 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3364:e58d480044c2 1.7/MoinMoin/action/diff.py: action.diff: fixed to the right rev no
2008-03-20T21:40:00 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3364:e58d480044c2 1.7/MoinMoin/action/diff.py: action.diff: fixed to the right rev no
2008-03-20T21:40:01 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3365:e64bb042f178 1.7/MoinMoin/action/diff.py: action.diff: rearanged buttons and title
2008-03-20T21:40:01 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3365:e64bb042f178 1.7/MoinMoin/action/diff.py: action.diff: rearanged buttons and title
2008-03-20T21:40:02 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3366:80ecec6e008b 1.7/MoinMoin/action/revert.py: action.revert: keep revert action in log
2008-03-20T21:40:02 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3366:80ecec6e008b 1.7/MoinMoin/action/revert.py: action.revert: keep revert action in log
2008-03-20T21:40:03 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3367:48acb9e0a097 1.7/MoinMoin/action/revert.py: action.revert: fixed doc string
2008-03-20T21:40:03 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3367:48acb9e0a097 1.7/MoinMoin/action/revert.py: action.revert: fixed doc string
2008-03-20T21:40:05 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3368:000c9fb7efe0 1.7/MoinMoin/wikiutil.py: merged main
2008-03-20T21:40:05 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3368:000c9fb7efe0 1.7/MoinMoin/wikiutil.py: merged main
2008-03-20T21:40:08 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3369:73b64e4085ff 1.7/MoinMoin/macro/EmbedObject.py: macro.EmbedObject: adjusted required_arg removed duplicate msg
2008-03-20T21:40:08 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3369:73b64e4085ff 1.7/MoinMoin/macro/EmbedObject.py: macro.EmbedObject: adjusted required_arg removed duplicate msg
2008-03-20T21:40:11 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3370:fee75398b525 1.7/: merged main
2008-03-20T21:40:11 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3370:fee75398b525 1.7/: merged main
2008-03-20T21:42:57 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3371:8ac95a64f1a8 1.7/MoinMoin/ (_tests/test_wikiutil.py wikiutil.py): really make required_arg take choices (including test)
2008-03-20T21:42:57 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3371:8ac95a64f1a8 1.7/MoinMoin/ (_tests/test_wikiutil.py wikiutil.py): really make required_arg take choices (including test)
2008-03-20T21:45:43 <johill> TheSheep: that better?
2008-03-20T21:45:49 <johill> TheSheep: that better?
2008-03-20T21:45:51 <ThomasWaldmann> dreimark: http://hg.moinmo.in/moin/1.7/rev/80ecec6e008b you should first clean and then use
2008-03-20T21:45:51 <ThomasWaldmann> dreimark: http://hg.moinmo.in/moin/1.7/rev/80ecec6e008b you should first clean and then use
2008-03-20T21:46:22 <TheSheep> I just discarded all my modifications %)
2008-03-20T21:46:22 <TheSheep> I just discarded all my modifications %)
2008-03-20T21:46:32 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 2585:1a362afed96c 1.6/MoinMoin/ (_tests/test_wikiutil.py wikiutil.py): really make required_arg take choices (including test)
2008-03-20T21:46:32 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 2585:1a362afed96c 1.6/MoinMoin/ (_tests/test_wikiutil.py wikiutil.py): really make required_arg take choices (including test)
2008-03-20T21:46:52 <ThomasWaldmann> dreimark: i.e. move the cleaning to the else branch
2008-03-20T21:46:52 <ThomasWaldmann> dreimark: i.e. move the cleaning to the else branch
2008-03-20T21:47:08 <johill> dreimark: please revert that change
2008-03-20T21:47:08 <johill> dreimark: please revert that change
2008-03-20T21:47:23 <johill> dreimark: and instead change the info action to display SAVE/REVERT
2008-03-20T21:47:23 <johill> dreimark: and instead change the info action to display SAVE/REVERT
2008-03-20T21:47:38 <johill> with the right message
2008-03-20T21:47:38 <johill> with the right message
2008-03-20T21:48:02 <johill> no need to save that translated message to the log!
2008-03-20T21:48:02 <johill> no need to save that translated message to the log!
2008-03-20T21:48:06 <johill> (totally bogus in fact)
2008-03-20T21:48:06 <johill> (totally bogus in fact)
2008-03-20T21:49:46 <ThomasWaldmann> yeah, that's even better. we have the old revno in the extra field?
2008-03-20T21:49:46 <ThomasWaldmann> yeah, that's even better. we have the old revno in the extra field?
2008-03-20T21:50:13 <dreimark> for storing yes but how is then the comment shown in RC
2008-03-20T21:50:13 <dreimark> for storing yes but how is then the comment shown in RC
2008-03-20T21:50:24 <johill> then make the same change in RC too
2008-03-20T21:50:24 <johill> then make the same change in RC too
2008-03-20T21:50:39 <johill> where's the problem?
2008-03-20T21:50:39 <johill> where's the problem?
2008-03-20T21:51:03 <johill> ThomasWaldmann: of course, how else could it have worked before :)
2008-03-20T21:51:03 <johill> ThomasWaldmann: of course, how else could it have worked before :)
2008-03-20T21:51:17 <dreimark> ok
2008-03-20T21:51:17 <dreimark> ok
2008-03-20T21:51:28 <johill> (before my patch it the description was always empty and done by the info action)
2008-03-20T21:51:28 <johill> (before my patch it the description was always empty and done by the info action)
2008-03-20T21:51:41 <johill> I suppose RC also has the same logic if it displayed "revert to..." before
2008-03-20T21:51:41 <johill> I suppose RC also has the same logic if it displayed "revert to..." before
2008-03-20T21:51:49 <dreimark> rigth
2008-03-20T21:51:49 <dreimark> rigth
2008-03-20T21:52:05 <dreimark> but if one enters a comment where should it be recognized
2008-03-20T21:52:05 <dreimark> but if one enters a comment where should it be recognized
2008-03-20T21:52:19 <johill> well you just make the info action display both
2008-03-20T21:52:19 <johill> well you just make the info action display both
2008-03-20T21:52:56 <dreimark> usually you lok first in RC to get changes
2008-03-20T21:52:56 <dreimark> usually you lok first in RC to get changes
2008-03-20T21:53:19 <dreimark> a message someone reverts something is enough there
2008-03-20T21:53:19 <dreimark> a message someone reverts something is enough there
2008-03-20T21:53:53 <johill> RC has a message now, right?
2008-03-20T21:53:53 <johill> RC has a message now, right?
2008-03-20T21:53:59 <dreimark> yes
2008-03-20T21:53:59 <dreimark> yes
2008-03-20T21:54:21 <johill> elif not comment:
2008-03-20T21:54:21 <johill> elif not comment:
2008-03-20T21:54:21 <johill> if '/REVERT' in action:
2008-03-20T21:54:21 <johill> rev = int(line.extra)
2008-03-20T21:54:21 <johill> comment = _("Revert to revision %(rev)d.") % {'rev': rev}
2008-03-20T21:54:21 <johill> if '/REVERT' in action:
2008-03-20T21:54:21 <johill> rev = int(line.extra)
2008-03-20T21:54:21 <johill> comment = _("Revert to revision %(rev)d.") % {'rev': rev}
2008-03-20T21:54:24 <johill> just change that to
2008-03-20T21:54:24 <johill> just change that to
2008-03-20T21:54:27 <johill> something like
2008-03-20T21:54:27 <johill> something like
2008-03-20T21:54:37 <johill> if '/REVERT' in action:
2008-03-20T21:54:37 <johill> if '/REVERT' in action:
2008-03-20T21:54:38 <johill> rev = ...
2008-03-20T21:54:39 <johill> rev = ...
2008-03-20T21:54:56 <johill> comment = (_("Revert to ...") % ... ) + comment
2008-03-20T21:54:56 <johill> comment = (_("Revert to ...") % ... ) + comment
2008-03-20T21:55:13 <johill> the fact that it is reverted doesn't need to be stored in the message
2008-03-20T21:55:13 <johill> the fact that it is reverted doesn't need to be stored in the message
2008-03-20T21:55:17 <johill> it's already stored in the action description
2008-03-20T21:55:17 <johill> it's already stored in the action description
2008-03-20T21:55:25 <dreimark> yeah
2008-03-20T21:55:25 <dreimark> yeah
2008-03-20T21:55:38 <johill> and your original patch is bogus because it stores the translated version
2008-03-20T21:55:38 <johill> and your original patch is bogus because it stores the translated version
2008-03-20T21:55:41 <dreimark> decreases the file size
2008-03-20T21:55:41 <dreimark> decreases the file size
2008-03-20T21:56:10 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3632:baec403eef65 1.7-storage-hwendel/MoinMoin/storage/backends/filesystem.py: make IndexedBackend use indexes before regular filters
2008-03-20T21:56:10 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3632:baec403eef65 1.7-storage-hwendel/MoinMoin/storage/backends/filesystem.py: make IndexedBackend use indexes before regular filters
2008-03-20T21:56:13 <johill> lanius: there
2008-03-20T21:56:13 <johill> lanius: there
2008-03-20T21:56:32 <TheSheep> johill: works perfectly now, thanks!
2008-03-20T21:56:32 <TheSheep> johill: works perfectly now, thanks!
2008-03-20T21:56:37 <johill> TheSheep: :)
2008-03-20T21:56:37 <johill> TheSheep: :)
2008-03-20T21:57:02 <johill> dreimark: do you have anything using the UnitArgument? we need a few tests for it
2008-03-20T21:57:02 <johill> dreimark: do you have anything using the UnitArgument? we need a few tests for it
2008-03-20T21:57:10 <dreimark> yep, that was because of don't I like lang mixup. But it is clear now havent thouht on that aspect
2008-03-20T21:57:10 <dreimark> yep, that was because of don't I like lang mixup. But it is clear now havent thouht on that aspect
2008-03-20T21:57:28 <dreimark> EmbedObject does
2008-03-20T21:57:28 <dreimark> EmbedObject does
2008-03-20T21:58:31 <dreimark> we have to add it to the parser for translusion but only if we have a converter too
2008-03-20T21:58:31 <dreimark> we have to add it to the parser for translusion but only if we have a converter too
2008-03-20T21:59:08 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3633:5edac19e599c 1.7-storage-hwendel/MoinMoin/storage/_tests/test_external.py: fix another storage test
2008-03-20T21:59:08 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3633:5edac19e599c 1.7-storage-hwendel/MoinMoin/storage/_tests/test_external.py: fix another storage test
2008-03-20T21:59:19 <johill> dreimark: if you have a minute, can you write some tests?
2008-03-20T21:59:19 <johill> dreimark: if you have a minute, can you write some tests?
2008-03-20T22:00:56 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3644:6b74eea122bf 1.7-storage-hwendel/MoinMoin/ (5 files in 4 dirs): merge with main
2008-03-20T22:00:56 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3644:6b74eea122bf 1.7-storage-hwendel/MoinMoin/ (5 files in 4 dirs): merge with main
2008-03-20T22:01:09 <johill> that way we can push up our changeset number ;)
2008-03-20T22:01:09 <johill> that way we can push up our changeset number ;)
2008-03-20T22:04:18 <dreimark> johill: hehe, merged main does it too :)
2008-03-20T22:04:19 <dreimark> johill: hehe, merged main does it too :)
2008-03-20T22:05:05 <dreimark> but I won't do a commit for each test
2008-03-20T22:05:05 <dreimark> but I won't do a commit for each test
2008-03-20T22:05:12 <dreimark> as some others did on vc
2008-03-20T22:05:12 <dreimark> as some others did on vc
2008-03-20T22:05:29 <johill> heh
2008-03-20T22:05:29 <johill> heh
2008-03-20T22:06:29 <dreimark> e.g. http://cia.vc/stats/project/mediawiki
2008-03-20T22:06:29 <dreimark> e.g. http://cia.vc/stats/project/mediawiki
2008-03-20T22:09:39 <johill> "fix syntax error"
2008-03-20T22:09:39 <johill> "fix syntax error"
2008-03-20T22:09:44 <johill> oh the joys of PHP ;)
2008-03-20T22:09:44 <johill> oh the joys of PHP ;)
2008-03-20T22:15:02 <CIA-39> Radomir Dopieralski <moindev@sheep.art.pl> default * 3372:55b787c6c083 1.7/MoinMoin/macro/Navigation.py: macro Navigation converted to the new argument parser
2008-03-20T22:15:02 <CIA-39> Radomir Dopieralski <moindev@sheep.art.pl> default * 3372:55b787c6c083 1.7/MoinMoin/macro/Navigation.py: macro Navigation converted to the new argument parser
2008-03-20T22:15:28 <TheSheep> hopefully the rest will be faster :)
2008-03-20T22:15:28 <TheSheep> hopefully the rest will be faster :)
2008-03-20T22:15:29 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3645:8439d26e080d 1.7-storage-hwendel/MoinMoin/storage/backends/ (moin16.py moin17.py): add back unquoting I forgot in commit f28cdd68fe64
2008-03-20T22:15:29 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3645:8439d26e080d 1.7-storage-hwendel/MoinMoin/storage/backends/ (moin16.py moin17.py): add back unquoting I forgot in commit f28cdd68fe64
2008-03-20T22:15:30 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3646:f7af5a7367a2 1.7-storage-hwendel/MoinMoin/storage/backends/filesystem.py:
2008-03-20T22:15:30 <CIA-39> IndexedBackend: DELETED index is only built for True
2008-03-20T22:15:30 <CIA-39> Therefore, we need to ignore it when looking for NOT deleted pages.
2008-03-20T22:15:30 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3646:f7af5a7367a2 1.7-storage-hwendel/MoinMoin/storage/backends/filesystem.py:
2008-03-20T22:15:30 <CIA-39> IndexedBackend: DELETED index is only built for True
2008-03-20T22:15:30 <CIA-39> Therefore, we need to ignore it when looking for NOT deleted pages.
2008-03-20T22:19:12 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3373:c20aa5b0158e 1.7/MoinMoin/action/revert.py: action.revert: reverted to bb04265ce89a
2008-03-20T22:19:12 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3373:c20aa5b0158e 1.7/MoinMoin/action/revert.py: action.revert: reverted to bb04265ce89a
2008-03-20T22:19:13 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3374:043689a3b04c 1.7/MoinMoin/action/revert.py: action.revert: fixed doc string
2008-03-20T22:19:13 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3374:043689a3b04c 1.7/MoinMoin/action/revert.py: action.revert: fixed doc string
2008-03-20T22:19:14 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3375:50fe2fa62f30 1.7/MoinMoin/macro/RecentChanges.py: macro.RecentChanges: shows revert and rename action with comment
2008-03-20T22:19:14 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3375:50fe2fa62f30 1.7/MoinMoin/macro/RecentChanges.py: macro.RecentChanges: shows revert and rename action with comment
2008-03-20T22:19:15 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3376:0215dd9fa6ee 1.7/MoinMoin/macro/Navigation.py: merged main
2008-03-20T22:19:15 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3376:0215dd9fa6ee 1.7/MoinMoin/macro/Navigation.py: merged main
2008-03-20T22:20:51 <johill> I hope you didn't actually revert *to* bb04... but only that one
2008-03-20T22:20:51 <johill> I hope you didn't actually revert *to* bb04... but only that one
2008-03-20T22:22:51 <dreimark> I have given the filename
2008-03-20T22:22:51 <dreimark> I have given the filename
2008-03-20T22:23:12 <dreimark> http://hg.moinmo.in/moin/1.7/rev/c20aa5b0158e
2008-03-20T22:23:12 <dreimark> http://hg.moinmo.in/moin/1.7/rev/c20aa5b0158e
2008-03-20T22:23:36 <johill> oh ok, yeah
2008-03-20T22:23:36 <johill> oh ok, yeah
2008-03-20T22:24:52 <dreimark> ThomasWaldmann: btw. can you configure hg that it enters the url to the revision here
2008-03-20T22:24:52 <dreimark> ThomasWaldmann: btw. can you configure hg that it enters the url to the revision here
2008-03-20T22:25:09 <dreimark> later when time
2008-03-20T22:25:09 <dreimark> later when time
2008-03-20T22:26:29 <johill> that'd be a CIA config, no?
2008-03-20T22:26:29 <johill> that'd be a CIA config, no?
2008-03-20T22:27:17 <xorAxAx> dreimark: you can configure that
2008-03-20T22:27:17 <xorAxAx> dreimark: you can configure that
2008-03-20T22:27:26 <xorAxAx> dreimark: see cia.vc
2008-03-20T22:27:26 <xorAxAx> dreimark: see cia.vc
2008-03-20T22:27:27 <dreimark> me how ?
2008-03-20T22:27:27 <dreimark> me how ?
2008-03-20T22:27:32 <dreimark> ah ok
2008-03-20T22:27:32 <dreimark> ah ok
2008-03-20T22:27:38 <xorAxAx> then edit the bot
2008-03-20T22:27:38 <xorAxAx> then edit the bot
2008-03-20T22:27:46 <xorAxAx> and write a custom filter
2008-03-20T22:27:46 <xorAxAx> and write a custom filter
2008-03-20T22:28:06 <xorAxAx> i hope that it allows to append such messages
2008-03-20T22:28:06 <xorAxAx> i hope that it allows to append such messages
2008-03-20T22:28:09 <johill> I think that wasn't meant literally. ThomasWaldmann can because he runs the bot
2008-03-20T22:28:09 <johill> I think that wasn't meant literally. ThomasWaldmann can because he runs the bot
2008-03-20T22:28:30 <dreimark> http://cia.vc/clients/hg/hgcia.py
2008-03-20T22:28:31 <dreimark> http://cia.vc/clients/hg/hgcia.py
2008-03-20T22:28:41 <xorAxAx> johill: he doesnt
2008-03-20T22:28:41 <xorAxAx> johill: he doesnt
2008-03-20T22:28:53 <johill> baseurl = http://server/path/to/repo
2008-03-20T22:28:53 <johill> baseurl = http://server/path/to/repo
2008-03-20T22:29:00 <johill> who runs the bot then?
2008-03-20T22:29:00 <johill> who runs the bot then?
2008-03-20T22:29:03 <xorAxAx> well, he could hack the notifier script (which is not a bot) to append an http url to the desc
2008-03-20T22:29:03 <xorAxAx> well, he could hack the notifier script (which is not a bot) to append an http url to the desc
2008-03-20T22:29:11 <xorAxAx> johill: there is no band^Wbot
2008-03-20T22:29:11 <xorAxAx> johill: there is no band^Wbot
2008-03-20T22:29:13 <johill> yeah ok
2008-03-20T22:29:13 <johill> yeah ok
2008-03-20T22:29:21 <johill> the hgcia commit script has a config for t hat
2008-03-20T22:29:21 <johill> the hgcia commit script has a config for t hat
2008-03-20T22:29:29 <xorAxAx> to append urls?
2008-03-20T22:29:29 <xorAxAx> to append urls?
2008-03-20T22:29:30 <xorAxAx> ok
2008-03-20T22:29:31 <xorAxAx> ok
2008-03-20T22:29:34 <johill> [web]
2008-03-20T22:29:34 <johill> [web]
2008-03-20T22:29:36 <johill> baseurl = ...
2008-03-20T22:29:36 <johill> baseurl = ...
2008-03-20T22:29:43 <johill> as lesat the one dreimark just pasted a link to
2008-03-20T22:29:43 <johill> as lesat the one dreimark just pasted a link to
2008-03-20T22:31:26 <dreimark> I do use baseurl too for mail notification of our repos. nice
2008-03-20T22:31:26 <dreimark> I do use baseurl too for mail notification of our repos. nice
2008-03-20T22:31:37 <dreimark> but irc is nicer
2008-03-20T22:31:37 <dreimark> but irc is nicer
2008-03-20T22:35:10 <dreimark> any idea how we can speed up Hits and other log related functions ?
2008-03-20T22:35:10 <dreimark> any idea how we can speed up Hits and other log related functions ?
2008-03-20T22:35:39 <johill> storage might be faster
2008-03-20T22:35:40 <johill> storage might be faster
2008-03-20T22:40:29 <dreimark> in the past reading myself the file was much faster for Hits as now
2008-03-20T22:40:29 <dreimark> in the past reading myself the file was much faster for Hits as now
2008-03-20T22:40:36 <johill> huh I'm confused about python's variable binding rules vs. lambda
2008-03-20T22:40:36 <johill> huh I'm confused about python's variable binding rules vs. lambda
2008-03-20T22:41:12 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3647:1736e984f5d3 1.7-storage-hwendel/MoinMoin/Page.py: fix getPageList filter function bug
2008-03-20T22:41:12 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3647:1736e984f5d3 1.7-storage-hwendel/MoinMoin/Page.py: fix getPageList filter function bug
2008-03-20T22:41:51 <dreimark> I hate ogg not beeing an audio mimetype
2008-03-20T22:41:51 <dreimark> I hate ogg not beeing an audio mimetype
2008-03-20T22:42:10 <johill> could contain video too
2008-03-20T22:42:10 <johill> could contain video too
2008-03-20T22:42:16 <lanius> johill: the deleted stuff looks strange
2008-03-20T22:42:16 <lanius> johill: the deleted stuff looks strange
2008-03-20T22:42:26 <lanius> johill: it does not belong into filesystem.py
2008-03-20T22:42:26 <lanius> johill: it does not belong into filesystem.py
2008-03-20T22:43:33 <johill> explain
2008-03-20T22:43:33 <johill> explain
2008-03-20T22:43:58 <lanius> filesystem.py should not know anything about a special metadata key
2008-03-20T22:43:58 <lanius> filesystem.py should not know anything about a special metadata key
2008-03-20T22:44:03 <lanius> it should work for everything
2008-03-20T22:44:03 <lanius> it should work for everything
2008-03-20T22:44:14 <johill> ??
2008-03-20T22:44:14 <johill> ??
2008-03-20T22:44:26 <johill> you use that special metadata key throughout
2008-03-20T22:44:26 <johill> you use that special metadata key throughout
2008-03-20T22:44:30 <johill> like searching for it in Page.py
2008-03-20T22:44:30 <johill> like searching for it in Page.py
2008-03-20T22:44:32 <ThomasWaldmann> [web]
2008-03-20T22:44:32 <ThomasWaldmann> [web]
2008-03-20T22:44:33 <ThomasWaldmann> # If you want hyperlinks (optional)
2008-03-20T22:44:33 <ThomasWaldmann> baseurl = http://hg.moinmo.in/moin/1.7
2008-03-20T22:44:33 <ThomasWaldmann> # If you want hyperlinks (optional)
2008-03-20T22:44:33 <ThomasWaldmann> baseurl = http://hg.moinmo.in/moin/1.7
2008-03-20T22:44:37 <johill> how can filesystem.py be agnostic then?
2008-03-20T22:44:37 <johill> how can filesystem.py be agnostic then?
2008-03-20T22:44:37 <lanius> johill: Page.py is okey
2008-03-20T22:44:38 <lanius> johill: Page.py is okey
2008-03-20T22:44:56 <lanius> johill: because it is just a metadata key on this level
2008-03-20T22:44:56 <lanius> johill: because it is just a metadata key on this level
2008-03-20T22:44:57 <lanius> like any other
2008-03-20T22:44:58 <lanius> like any other
2008-03-20T22:45:19 <johill> then we have to fix all backends to set metadata[DELETED] = False for all pages
2008-03-20T22:45:19 <johill> then we have to fix all backends to set metadata[DELETED] = False for all pages
2008-03-20T22:45:25 <johill> which they don't currently do
2008-03-20T22:45:25 <johill> which they don't currently do
2008-03-20T22:45:39 <lanius> i decided against that
2008-03-20T22:45:39 <lanius> i decided against that
2008-03-20T22:45:58 <lanius> because a new backend might really save it as metadata key (no emulation like moin16)
2008-03-20T22:45:58 <lanius> because a new backend might really save it as metadata key (no emulation like moin16)
2008-03-20T22:46:00 <johill> then filesystem.py has to know about it
2008-03-20T22:46:00 <johill> then filesystem.py has to know about it
2008-03-20T22:46:18 <lanius> and in this case it will just not be there if it is not set
2008-03-20T22:46:18 <lanius> and in this case it will just not be there if it is not set
2008-03-20T22:46:37 <johill> then you can't filter for {DELETED: False}
2008-03-20T22:46:37 <johill> then you can't filter for {DELETED: False}
2008-03-20T22:46:39 <johill> which would suck
2008-03-20T22:46:40 <johill> which would suck
2008-03-20T22:46:45 <lanius> then use the filter function stuff
2008-03-20T22:46:45 <lanius> then use the filter function stuff
2008-03-20T22:46:52 <johill> then you can't use the index for it
2008-03-20T22:46:52 <johill> then you can't use the index for it
2008-03-20T22:47:08 <lanius> maybe it can be implemented on that level
2008-03-20T22:47:08 <lanius> maybe it can be implemented on that level
2008-03-20T22:47:11 <johill> nor could a backend optimise elsewhere
2008-03-20T22:47:11 <johill> nor could a backend optimise elsewhere
2008-03-20T22:47:18 <johill> I mean, another backend optimise it at all
2008-03-20T22:47:18 <johill> I mean, another backend optimise it at all
2008-03-20T22:47:40 <lanius> that's why i did it the other way around until now
2008-03-20T22:47:40 <lanius> that's why i did it the other way around until now
2008-03-20T22:48:38 <johill> I don't know how you can save it's ok for Page.py but not for filesystem.py
2008-03-20T22:48:38 <johill> I don't know how you can save it's ok for Page.py but not for filesystem.py
2008-03-20T22:48:50 <johill> either Page.py can use it, then it must be "the way" to indicate with metadata whether a page is deleted or not
2008-03-20T22:48:50 <johill> either Page.py can use it, then it must be "the way" to indicate with metadata whether a page is deleted or not
2008-03-20T22:48:56 <johill> then filesystem.py can use it too
2008-03-20T22:48:56 <johill> then filesystem.py can use it too
2008-03-20T22:49:00 <lanius> it is just an arbitrary metadata key
2008-03-20T22:49:00 <lanius> it is just an arbitrary metadata key
2008-03-20T22:49:08 <lanius> and page.py may use any metadata key
2008-03-20T22:49:09 <lanius> and page.py may use any metadata key
2008-03-20T22:49:12 <johill> yeah
2008-03-20T22:49:12 <johill> yeah
2008-03-20T22:49:16 <lanius> but filesystem.py may not any special metadata key
2008-03-20T22:49:17 <lanius> but filesystem.py may not any special metadata key
2008-03-20T22:49:20 <johill> but what if the backend doesn't use DELETED
2008-03-20T22:49:20 <johill> but what if the backend doesn't use DELETED
2008-03-20T22:49:20 <lanius> but handle all keys the same
2008-03-20T22:49:21 <lanius> but handle all keys the same
2008-03-20T22:49:24 <johill> but something else?
2008-03-20T22:49:24 <johill> but something else?
2008-03-20T22:49:31 <johill> you're not making sense but I'm on the phone right now
2008-03-20T22:49:31 <johill> you're not making sense but I'm on the phone right now
2008-03-20T22:49:46 <lanius> any new backend won't use DELETED
2008-03-20T22:49:46 <lanius> any new backend won't use DELETED
2008-03-20T22:49:50 <lanius> it will use what it gets from page.py
2008-03-20T22:49:51 <lanius> it will use what it gets from page.py
2008-03-20T22:49:56 <johill> that's the point
2008-03-20T22:49:56 <johill> that's the point
2008-03-20T22:49:57 <lanius> only the current backends use deleted to make it compatible
2008-03-20T22:49:58 <lanius> only the current backends use deleted to make it compatible
2008-03-20T22:50:00 <johill> so Page.py can't use DELETED either
2008-03-20T22:50:00 <johill> so Page.py can't use DELETED either
2008-03-20T22:50:03 <lanius> sure it can
2008-03-20T22:50:04 <lanius> sure it can
2008-03-20T22:50:06 <johill> NO!
2008-03-20T22:50:06 <johill> NO!
2008-03-20T22:50:08 <lanius> why not
2008-03-20T22:50:09 <lanius> why not
2008-03-20T22:50:11 <johill> not with any sane semantics anyway
2008-03-20T22:50:11 <johill> not with any sane semantics anyway
2008-03-20T22:50:19 <lanius> i don't understand
2008-03-20T22:50:19 <lanius> i don't understand
2008-03-20T22:50:26 <johill> hrm
2008-03-20T22:50:26 <johill> hrm
2008-03-20T22:50:26 <johill> ok
2008-03-20T22:50:26 <johill> ok
2008-03-20T22:50:35 <johill> if a backend doesn't use DELETED metadata key
2008-03-20T22:50:35 <johill> if a backend doesn't use DELETED metadata key
2008-03-20T22:50:43 <johill> and Page.py wants to get a list of all deleted pages
2008-03-20T22:50:43 <johill> and Page.py wants to get a list of all deleted pages
2008-03-20T22:50:45 <johill> what should it do?
2008-03-20T22:50:45 <johill> what should it do?
2008-03-20T22:50:48 <lanius> filter for it
2008-03-20T22:50:48 <lanius> filter for it
2008-03-20T22:50:52 <johill> how?
2008-03-20T22:50:52 <johill> how?
2008-03-20T22:50:53 <lanius> with filterfuncs
2008-03-20T22:50:53 <lanius> with filterfuncs
2008-03-20T22:50:58 <johill> that's crappy
2008-03-20T22:50:58 <johill> that's crappy
2008-03-20T22:51:05 <CIA-39> Radomir Dopieralski <moindev@sheep.art.pl> default * 3377:4aeb96183e04 1.7/MoinMoin/macro/NewPage.py: macro NewPage converted to the new arg parser
2008-03-20T22:51:05 <CIA-39> Radomir Dopieralski <moindev@sheep.art.pl> default * 3377:4aeb96183e04 1.7/MoinMoin/macro/NewPage.py: macro NewPage converted to the new arg parser
2008-03-20T22:51:06 <CIA-39> Radomir Dopieralski <moindev@sheep.art.pl> default * 3378:dd7df6fe929a 1.7/MoinMoin/ (action/revert.py macro/RecentChanges.py): merge main
2008-03-20T22:51:06 <CIA-39> Radomir Dopieralski <moindev@sheep.art.pl> default * 3378:dd7df6fe929a 1.7/MoinMoin/ (action/revert.py macro/RecentChanges.py): merge main
2008-03-20T22:51:07 <lanius> you just added this functionality
2008-03-20T22:51:07 <lanius> you just added this functionality
2008-03-20T22:51:19 <lanius> so make it not be crappy
2008-03-20T22:51:20 <lanius> so make it not be crappy
2008-03-20T22:51:24 <johill> I cannot
2008-03-20T22:51:24 <johill> I cannot
2008-03-20T22:51:31 <lanius> then of what use is it
2008-03-20T22:51:32 <lanius> then of what use is it
2008-03-20T22:51:32 <johill> at least not your way
2008-03-20T22:51:32 <johill> at least not your way
2008-03-20T22:51:37 <johill> it's for complex functions
2008-03-20T22:51:37 <johill> it's for complex functions
2008-03-20T22:51:43 <johill> not something like deleted
2008-03-20T22:51:43 <johill> not something like deleted
2008-03-20T22:51:48 <johill> say you have an sql backend
2008-03-20T22:51:49 <johill> say you have an sql backend
2008-03-20T22:51:54 <johill> with a column saying "deleted"
2008-03-20T22:51:54 <johill> with a column saying "deleted"
2008-03-20T22:51:57 <lanius> no
2008-03-20T22:51:57 <lanius> no
2008-03-20T22:52:00 <lanius> there is no deleted column
2008-03-20T22:52:00 <lanius> there is no deleted column
2008-03-20T22:52:05 <lanius> there is a metadata key->value table
2008-03-20T22:52:05 <lanius> there is a metadata key->value table
2008-03-20T22:52:15 <johill> so it has DELETED metadata
2008-03-20T22:52:15 <johill> so it has DELETED metadata
2008-03-20T22:52:23 <lanius> but only if it is deleted
2008-03-20T22:52:23 <lanius> but only if it is deleted
2008-03-20T22:52:29 <johill> but you just said not every backend needs to have DELETED metadata
2008-03-20T22:52:29 <johill> but you just said not every backend needs to have DELETED metadata
2008-03-20T22:52:34 <lanius> no
2008-03-20T22:52:34 <lanius> no
2008-03-20T22:52:37 <lanius> it must have metadata
2008-03-20T22:52:37 <lanius> it must have metadata
2008-03-20T22:52:42 <lanius> and page.py tells him to have deleted metadata
2008-03-20T22:52:42 <lanius> and page.py tells him to have deleted metadata
2008-03-20T22:52:50 <johill> then what's wrong with using it in filessytem.h
2008-03-20T22:52:50 <johill> then what's wrong with using it in filessytem.h
2008-03-20T22:52:56 <xorAxAx> haha
2008-03-20T22:52:56 <xorAxAx> haha
2008-03-20T22:53:01 <lanius> but the backend doesn't care if this metadata is named deleted or random or foo or ba
2008-03-20T22:53:01 <lanius> but the backend doesn't care if this metadata is named deleted or random or foo or ba
2008-03-20T22:53:33 <lanius> besonen_mobile2: filesystem.py knows only about metadata in general, not about any special key
2008-03-20T22:53:33 <lanius> besonen_mobile2: filesystem.py knows only about metadata in general, not about any special key
2008-03-20T22:53:35 <johill> ok so we need to extend the filter dict stuff
2008-03-20T22:53:36 <johill> ok so we need to extend the filter dict stuff
2008-03-20T22:53:43 <lanius> to handle the not case
2008-03-20T22:53:44 <lanius> to handle the not case
2008-03-20T22:53:50 <johill> to be able to filter for "not present" or something
2008-03-20T22:53:50 <johill> to be able to filter for "not present" or something
2008-03-20T22:53:53 <lanius> right
2008-03-20T22:53:53 <lanius> right
2008-03-20T22:54:10 <johill> *and* make page.py *delete* the DELETED metadata for undeleted pages
2008-03-20T22:54:10 <johill> *and* make page.py *delete* the DELETED metadata for undeleted pages
2008-03-20T22:54:15 <johill> rather than setting it to False
2008-03-20T22:54:15 <johill> rather than setting it to False
2008-03-20T22:54:34 <johill> you've made it set it to False which is why I initially thought every page had it always
2008-03-20T22:54:34 <johill> you've made it set it to False which is why I initially thought every page had it always
2008-03-20T22:54:45 <johill> you're not being consistent :)
2008-03-20T22:54:45 <johill> you're not being consistent :)
2008-03-20T22:55:03 <lanius> ok, that's my fault
2008-03-20T22:55:03 <lanius> ok, that's my fault
2008-03-20T22:55:14 <johill> so ok
2008-03-20T22:55:14 <johill> so ok
2008-03-20T22:56:02 <johill> we need a filter-not-present then
2008-03-20T22:56:02 <johill> we need a filter-not-present then
2008-03-20T22:56:07 <lanius> yes
2008-03-20T22:56:07 <lanius> yes
2008-03-20T22:56:09 <johill> how about using {'meta': None}
2008-03-20T22:56:09 <johill> how about using {'meta': None}
2008-03-20T22:56:19 <lanius> sounds good
2008-03-20T22:56:19 <lanius> sounds good
2008-03-20T22:56:45 <johill> then filesystem.py only needs to see if any of the indexed things is None and not use the index for it then or something
2008-03-20T22:56:45 <johill> then filesystem.py only needs to see if any of the indexed things is None and not use the index for it then or something
2008-03-20T22:56:49 <lanius> right
2008-03-20T22:56:49 <lanius> right
2008-03-20T22:57:22 <johill> but PageEditor or whatever has to unset the key then
2008-03-20T22:57:22 <johill> but PageEditor or whatever has to unset the key then
2008-03-20T22:57:36 <lanius> yes
2008-03-20T22:57:36 <lanius> yes
2008-03-20T22:58:02 <lanius> and the filterfn stuff is wrong because it won't be aplied if index_filters is set
2008-03-20T22:58:02 <lanius> and the filterfn stuff is wrong because it won't be aplied if index_filters is set
2008-03-20T22:58:05 <lanius> ah no
2008-03-20T22:58:05 <lanius> ah no
2008-03-20T22:58:08 <lanius> sorry
2008-03-20T22:58:08 <lanius> sorry
2008-03-20T23:00:13 <lanius> and is there really a use case for filterfn then?
2008-03-20T23:00:14 <lanius> and is there really a use case for filterfn then?
2008-03-20T23:02:36 <johill> yes
2008-03-20T23:02:36 <johill> yes
2008-03-20T23:02:38 <lanius> johill: i think check_filters belongs into filesystem.py because other backends like sql can do filtering more efficien
2008-03-20T23:02:39 <lanius> johill: i think check_filters belongs into filesystem.py because other backends like sql can do filtering more efficien
2008-03-20T23:02:42 <lanius> t
2008-03-20T23:02:42 <lanius> t
2008-03-20T23:02:51 <johill> lanius: yeah, good point
2008-03-20T23:02:51 <johill> lanius: yeah, good point
2008-03-20T23:03:05 <johill> the use case for filterfn is say doing a TitleSearch
2008-03-20T23:03:05 <johill> the use case for filterfn is say doing a TitleSearch
2008-03-20T23:03:45 <lanius> which could be optimized in a sql backend again, but not if you use a filterfn
2008-03-20T23:03:45 <lanius> which could be optimized in a sql backend again, but not if you use a filterfn
2008-03-20T23:04:54 <johill> how would you want to express it then?
2008-03-20T23:04:54 <johill> how would you want to express it then?
2008-03-20T23:05:09 <johill> give an arbitrarily complex filter language that the backend can parse out?
2008-03-20T23:05:09 <johill> give an arbitrarily complex filter language that the backend can parse out?
2008-03-20T23:05:28 <lanius> yes
2008-03-20T23:05:28 <lanius> yes
2008-03-20T23:06:39 <johill> otoh, check_filter can be there as (a) a reference implementation and (b) for backends that don't want to/care to/cannot optimise
2008-03-20T23:06:39 <johill> otoh, check_filter can be there as (a) a reference implementation and (b) for backends that don't want to/care to/cannot optimise
2008-03-20T23:09:43 <johill> we'll probably always need a filter function anyway
2008-03-20T23:09:43 <johill> we'll probably always need a filter function anyway
2008-03-20T23:10:37 <lanius> my idea was to begin with some easy filter syntax which i implemented and then extend it to the needs as they come
2008-03-20T23:10:38 <lanius> my idea was to begin with some easy filter syntax which i implemented and then extend it to the needs as they come
2008-03-20T23:11:03 <johill> good idea in principle
2008-03-20T23:11:03 <johill> good idea in principle
2008-03-20T23:11:09 <johill> but the exectuion is a bit lacking
2008-03-20T23:11:09 <johill> but the exectuion is a bit lacking
2008-03-20T23:11:16 <johill> since you only allowed a single dict and no actual filter terms
2008-03-20T23:11:16 <johill> since you only allowed a single dict and no actual filter terms
2008-03-20T23:11:23 <johill> and that dict is tied to metadata
2008-03-20T23:11:23 <johill> and that dict is tied to metadata
2008-03-20T23:11:52 <lanius> you could say that the pagename is part of the metadata
2008-03-20T23:11:53 <lanius> you could say that the pagename is part of the metadata
2008-03-20T23:12:03 <johill> that would be rather weird
2008-03-20T23:12:03 <johill> that would be rather weird
2008-03-20T23:12:07 <lanius> no at all
2008-03-20T23:12:07 <lanius> no at all
2008-03-20T23:12:16 <lanius> your pages are just a random collection of ids
2008-03-20T23:12:16 <lanius> your pages are just a random collection of ids
2008-03-20T23:12:24 <johill> and wouldn't help
2008-03-20T23:12:24 <johill> and wouldn't help
2008-03-20T23:12:28 <lanius> which have names which are represented by the metadata
2008-03-20T23:12:28 <lanius> which have names which are represented by the metadata
2008-03-20T23:12:37 <johill> because you only exact metadata matches
2008-03-20T23:12:37 <johill> because you only exact metadata matches
2008-03-20T23:12:44 <johill> no that's not going to fly
2008-03-20T23:12:45 <johill> no that's not going to fly
2008-03-20T23:12:49 <lanius> because i started with the simple case
2008-03-20T23:12:49 <lanius> because i started with the simple case
2008-03-20T23:13:02 <johill> the item list is *everywhere* assumed to be consisting of the actual page names
2008-03-20T23:13:03 <johill> the item list is *everywhere* assumed to be consisting of the actual page names
2008-03-20T23:13:03 <lanius> there are no performance issues on a real database with that
2008-03-20T23:13:03 <lanius> there are no performance issues on a real database with that
2008-03-20T23:13:34 <johill> you don't do for p in getPageList(): request.write(Page(p).page_name)
2008-03-20T23:13:34 <johill> you don't do for p in getPageList(): request.write(Page(p).page_name)
2008-03-20T23:13:41 <xorAxAx> ?
2008-03-20T23:13:41 <xorAxAx> ?
2008-03-20T23:13:41 <johill> you do for p in getPageList(): request.write(p)
2008-03-20T23:13:41 <johill> you do for p in getPageList(): request.write(p)
2008-03-20T23:14:17 <lanius> ?
2008-03-20T23:14:17 <lanius> ?
2008-03-20T23:14:24 <johill> ?
2008-03-20T23:14:24 <johill> ?
2008-03-20T23:14:32 <johill> that's how the code works
2008-03-20T23:14:32 <johill> that's how the code works
2008-03-20T23:14:33 <lanius> i don't understand what you want to say with that
2008-03-20T23:14:33 <johill> trust me
2008-03-20T23:14:34 <johill> trust me
2008-03-20T23:14:34 <lanius> i don't understand what you want to say with that
2008-03-20T23:14:38 <xorAxAx> me neither
2008-03-20T23:14:38 <xorAxAx> me neither
2008-03-20T23:14:42 <johill> you were saying "random IDs"
2008-03-20T23:14:42 <johill> you were saying "random IDs"
2008-03-20T23:14:51 <xorAxAx> currently the page name is the primary key
2008-03-20T23:14:51 <xorAxAx> currently the page name is the primary key
2008-03-20T23:14:56 <lanius> sure you have to change a lot
2008-03-20T23:14:56 <lanius> sure you have to change a lot
2008-03-20T23:14:59 <xorAxAx> of the entity "item"
2008-03-20T23:14:59 <xorAxAx> of the entity "item"
2008-03-20T23:15:02 <johill> yes, the page name is the primary key
2008-03-20T23:15:02 <johill> yes, the page name is the primary key
2008-03-20T23:15:10 <johill> so it shouldn't also be contained in the metadata
2008-03-20T23:15:10 <johill> so it shouldn't also be contained in the metadata
2008-03-20T23:15:22 <johill> because you can change metadata freely but not the primary key
2008-03-20T23:15:22 <johill> because you can change metadata freely but not the primary key
2008-03-20T23:15:30 <xorAxAx> thats not true :)
2008-03-20T23:15:30 <xorAxAx> thats not true :)
2008-03-20T23:15:34 <lanius> renaming a page would only be changing the metadata key
2008-03-20T23:15:35 <lanius> renaming a page would only be changing the metadata key
2008-03-20T23:15:38 <lanius> sounds cool, doesn't it?
2008-03-20T23:15:39 <lanius> sounds cool, doesn't it?
2008-03-20T23:15:41 <xorAxAx> but the duplication issue is a good point
2008-03-20T23:15:41 <xorAxAx> but the duplication issue is a good point
2008-03-20T23:15:43 <johill> not really
2008-03-20T23:15:44 <johill> not really
2008-03-20T23:15:48 <xorAxAx> lanius: no, thats bad
2008-03-20T23:15:48 <xorAxAx> lanius: no, thats bad
2008-03-20T23:15:50 <johill> don't ever duplicate data
2008-03-20T23:15:50 <johill> don't ever duplicate data
2008-03-20T23:15:51 <lanius> why
2008-03-20T23:15:51 <lanius> why
2008-03-20T23:16:06 <johill> also
2008-03-20T23:16:07 <johill> also
2008-03-20T23:16:10 <lanius> yes you have to look for duplicates
2008-03-20T23:16:10 <lanius> yes you have to look for duplicates
2008-03-20T23:16:17 <johill> then you have a rather weird way of looking for duplicates
2008-03-20T23:16:17 <johill> then you have a rather weird way of looking for duplicates
2008-03-20T23:16:24 <xorAxAx> lanius: because the backend uses the item id as the lookup key, no?
2008-03-20T23:16:24 <xorAxAx> lanius: because the backend uses the item id as the lookup key, no?
2008-03-20T23:16:26 <johill> and doing a query for a pagename over the metadata column
2008-03-20T23:16:26 <johill> and doing a query for a pagename over the metadata column
2008-03-20T23:16:30 <johill> seriously
2008-03-20T23:16:30 <johill> seriously
2008-03-20T23:16:34 <xorAxAx> its not related to duplicates ...
2008-03-20T23:16:34 <xorAxAx> its not related to duplicates ...
2008-03-20T23:16:37 <johill> we aren't designing an arbitrary database
2008-03-20T23:16:37 <johill> we aren't designing an arbitrary database
2008-03-20T23:16:45 <johill> we're designing something that is for storing wiki items
2008-03-20T23:16:45 <johill> we're designing something that is for storing wiki items
2008-03-20T23:16:51 <xorAxAx> lanius: and if you just change the metadata, that wont work
2008-03-20T23:16:51 <xorAxAx> lanius: and if you just change the metadata, that wont work
2008-03-20T23:16:58 <johill> and those *naturally* have a name as primary key
2008-03-20T23:16:59 <johill> and those *naturally* have a name as primary key
2008-03-20T23:17:19 <lanius> ok
2008-03-20T23:17:19 <lanius> ok
2008-03-20T23:17:21 <xorAxAx> lanius: do the backends implement any notion of a rename operation?
2008-03-20T23:17:21 <xorAxAx> lanius: do the backends implement any notion of a rename operation?
2008-03-20T23:17:22 <lanius> back to the topic
2008-03-20T23:17:22 <lanius> back to the topic
2008-03-20T23:17:31 <lanius> the query language could be extended to get the pagename
2008-03-20T23:17:31 <lanius> the query language could be extended to get the pagename
2008-03-20T23:17:41 <lanius> xorAxAx: any notion?
2008-03-20T23:17:41 <lanius> xorAxAx: any notion?
2008-03-20T23:17:49 <johill> xorAxAx: yes
2008-03-20T23:17:49 <johill> xorAxAx: yes
2008-03-20T23:17:52 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2586:5fc5500376b9 1.6/Makefile: fix pathes for underlay creation on new server
2008-03-20T23:17:52 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2586:5fc5500376b9 1.6/Makefile: fix pathes for underlay creation on new server
2008-03-20T23:17:53 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2587:29679befa166 1.6/wiki/ (underlay.tar underlay.tar.bz2): updated underlay, remove underlay.tar.bz2, use underlay.tar from now on
2008-03-20T23:17:53 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2587:29679befa166 1.6/wiki/ (underlay.tar underlay.tar.bz2): updated underlay, remove underlay.tar.bz2, use underlay.tar from now on
2008-03-20T23:18:17 <johill> def rename_item(self, name, newname):
2008-03-20T23:18:17 <johill> def rename_item(self, name, newname):
2008-03-20T23:18:18 <johill> ...
2008-03-20T23:18:18 <johill> ...
2008-03-20T23:18:41 <xorAxAx> johill: and that does what?
2008-03-20T23:18:41 <xorAxAx> johill: and that does what?
2008-03-20T23:18:49 <xorAxAx> (my dear source code reader)
2008-03-20T23:18:49 <xorAxAx> (my dear source code reader)
2008-03-20T23:18:57 <johill> what do you want to know? that's the generic interface definition
2008-03-20T23:18:57 <johill> what do you want to know? that's the generic interface definition
2008-03-20T23:19:05 <xorAxAx> i guess it does more than metadata changes
2008-03-20T23:19:05 <johill> it really does self._backend.rename_item(name, newname)
2008-03-20T23:19:05 <xorAxAx> i guess it does more than metadata changes
2008-03-20T23:19:05 <johill> it really does self._backend.rename_item(name, newname)
2008-03-20T23:19:16 <johill> which in the current backend will rename the dir I suppose
2008-03-20T23:19:16 <johill> which in the current backend will rename the dir I suppose
2008-03-20T23:19:18 <xorAxAx> so, back to the real topic
2008-03-20T23:19:18 <xorAxAx> so, back to the real topic
2008-03-20T23:19:21 <johill> yeah
2008-03-20T23:19:21 <johill> yeah
2008-03-20T23:19:23 <johill> query language
2008-03-20T23:19:23 <johill> query language
2008-03-20T23:19:34 <johill> let me summarise
2008-03-20T23:19:34 <johill> let me summarise
2008-03-20T23:19:51 <johill> original: {metakey: metavalue}, exact match required, no notion of "absent" values
2008-03-20T23:19:51 <johill> original: {metakey: metavalue}, exact match required, no notion of "absent" values
2008-03-20T23:20:10 <johill> my addition: filter function that gets item name, metadata
2008-03-20T23:20:10 <johill> my addition: filter function that gets item name, metadata
2008-03-20T23:20:25 <johill> planned addition: notion of "absent" values, i.e. checking "doesn't have metadata key X"
2008-03-20T23:20:25 <johill> planned addition: notion of "absent" values, i.e. checking "doesn't have metadata key X"
2008-03-20T23:20:26 <lanius> problem: can't be optimized
2008-03-20T23:20:26 <lanius> problem: can't be optimized
2008-03-20T23:20:38 <johill> filter function can't be optimised, yes
2008-03-20T23:20:38 <johill> filter function can't be optimised, yes
2008-03-20T23:21:03 <lanius> so you could just do this filtering some levels above
2008-03-20T23:21:03 <lanius> so you could just do this filtering some levels above
2008-03-20T23:21:06 <lanius> same performance
2008-03-20T23:21:06 <lanius> same performance
2008-03-20T23:21:08 <johill> no
2008-03-20T23:21:08 <johill> no
2008-03-20T23:21:13 <johill> much more memory use
2008-03-20T23:21:13 <johill> much more memory use
2008-03-20T23:21:43 <lanius> much more = how many mb
2008-03-20T23:21:43 <lanius> much more = how many mb
2008-03-20T23:21:49 <johill> if I, say, want to count all pages that re.compile('^Help.*Ab.*').match
2008-03-20T23:21:49 <johill> if I, say, want to count all pages that re.compile('^Help.*Ab.*').match
2008-03-20T23:22:15 <lanius> yes
2008-03-20T23:22:16 <lanius> yes
2008-03-20T23:22:17 <johill> quite a lot
2008-03-20T23:22:17 <johill> quite a lot
2008-03-20T23:22:20 <lanius> numbers
2008-03-20T23:22:20 <lanius> numbers
2008-03-20T23:22:27 <johill> maybe 2-3 MiB RSS
2008-03-20T23:22:27 <johill> maybe 2-3 MiB RSS
2008-03-20T23:22:37 <lanius> why
2008-03-20T23:22:37 <lanius> why
2008-03-20T23:22:48 <johill> dunno. page list is huge? not sure
2008-03-20T23:22:48 <johill> dunno. page list is huge? not sure
2008-03-20T23:22:57 <lanius> but you implemented iterators for that now
2008-03-20T23:22:58 <lanius> but you implemented iterators for that now
2008-03-20T23:23:03 <johill> yes
2008-03-20T23:23:03 <johill> yes
2008-03-20T23:23:05 <johill> that solves that
2008-03-20T23:23:05 <johill> that solves that
2008-03-20T23:23:12 <lanius> so
2008-03-20T23:23:12 <lanius> so
2008-03-20T23:23:18 <johill> ah, I see what you mena
2008-03-20T23:23:18 <johill> ah, I see what you mena
2008-03-20T23:23:20 <johill> mean
2008-03-20T23:23:20 <johill> mean
2008-03-20T23:23:32 <johill> sorry, misunderstood you, yes
2008-03-20T23:23:32 <johill> sorry, misunderstood you, yes
2008-03-20T23:23:41 <johill> you're right, the filter function doesn't need to be pushed through
2008-03-20T23:23:41 <johill> you're right, the filter function doesn't need to be pushed through
2008-03-20T23:23:47 <lanius> ok
2008-03-20T23:23:48 <lanius> ok
2008-03-20T23:23:57 <johill> I'll remove that again
2008-03-20T23:23:57 <johill> I'll remove that again
2008-03-20T23:24:15 <johill> (although it is actually a bit more efficient if you have a filter function that wants metadata)
2008-03-20T23:24:15 <johill> (although it is actually a bit more efficient if you have a filter function that wants metadata)
2008-03-20T23:25:11 <lanius> then the underlay thing
2008-03-20T23:25:11 <lanius> then the underlay thing
2008-03-20T23:25:20 <lanius> the idea was to remove the idea of an underlay
2008-03-20T23:25:20 <lanius> the idea was to remove the idea of an underlay
2008-03-20T23:25:27 <johill> lets actually discuss the filter more
2008-03-20T23:25:27 <johill> lets actually discuss the filter more
2008-03-20T23:25:31 <lanius> ok
2008-03-20T23:25:32 <lanius> ok
2008-03-20T23:25:32 <johill> first
2008-03-20T23:25:32 <johill> first
2008-03-20T23:25:49 <johill> if I remove the filter function we can only filter on metadata
2008-03-20T23:25:50 <johill> if I remove the filter function we can only filter on metadata
2008-03-20T23:26:14 <johill> imho we should solve that before it gets into wider use
2008-03-20T23:26:14 <johill> imho we should solve that before it gets into wider use
2008-03-20T23:26:18 <lanius> no, you can filter on the pagename as well on the higher level
2008-03-20T23:26:19 <lanius> no, you can filter on the pagename as well on the higher level
2008-03-20T23:26:27 <johill> yeah, but not ask the backend for that
2008-03-20T23:26:27 <johill> yeah, but not ask the backend for that
2008-03-20T23:26:37 <lanius> but we already that that it won't be more efficient
2008-03-20T23:26:37 <lanius> but we already that that it won't be more efficient
2008-03-20T23:26:43 <lanius> s/that/said/
2008-03-20T23:26:43 <lanius> s/that/said/
2008-03-20T23:26:51 <johill> it will if you have say pagename.startswith("Help")
2008-03-20T23:26:51 <johill> it will if you have say pagename.startswith("Help")
2008-03-20T23:26:56 <johill> because then you can do it in sql
2008-03-20T23:26:56 <johill> because then you can do it in sql
2008-03-20T23:27:12 <lanius> how that
2008-03-20T23:27:12 <lanius> how that
2008-03-20T23:27:25 <lanius> you can't translate a python filter func to sql, can you?
2008-03-20T23:27:25 <lanius> you can't translate a python filter func to sql, can you?
2008-03-20T23:27:40 <johill> select itemname from items where itemname like 'Help%' :)
2008-03-20T23:27:40 <johill> select itemname from items where itemname like 'Help%' :)
2008-03-20T23:27:44 <johill> oh
2008-03-20T23:27:44 <johill> oh
2008-03-20T23:27:45 <johill> that
2008-03-20T23:27:45 <johill> that
2008-03-20T23:27:55 <johill> well yes, look how the search expressions are implemented
2008-03-20T23:27:55 <johill> well yes, look how the search expressions are implemented
2008-03-20T23:28:00 <johill> we could just use that
2008-03-20T23:28:00 <johill> we could just use that
2008-03-20T23:28:21 <lanius> yes something like that
2008-03-20T23:28:21 <lanius> yes something like that
2008-03-20T23:28:22 <lanius> or
2008-03-20T23:28:22 <lanius> or
2008-03-20T23:28:25 <lanius> just a sec
2008-03-20T23:28:25 <lanius> just a sec
2008-03-20T23:28:38 <johill> MoinMoin/search/queryparser.py
2008-03-20T23:28:38 <johill> MoinMoin/search/queryparser.py
2008-03-20T23:30:15 <lanius> use pyparsing: http://www.slideshare.net/schlauch/tech-talk-pyparsing?src=embed
2008-03-20T23:30:16 <lanius> use pyparsing: http://www.slideshare.net/schlauch/tech-talk-pyparsing?src=embed
2008-03-20T23:30:25 <lanius> to define your query language
2008-03-20T23:30:25 <lanius> to define your query language
2008-03-20T23:31:59 <johill> it doesn't need to be a language
2008-03-20T23:31:59 <johill> it doesn't need to be a language
2008-03-20T23:32:03 <johill> it bcan be built fromm objects
2008-03-20T23:32:03 <johill> it bcan be built fromm objects
2008-03-20T23:32:05 <johill> like
2008-03-20T23:32:05 <johill> like
2008-03-20T23:32:45 <johill> filter.AND([filter.MetaDataMatch('key', 'value'), filter.NOT(filter.HasMetaData('key'))])
2008-03-20T23:32:45 <johill> filter.AND([filter.MetaDataMatch('key', 'value'), filter.NOT(filter.HasMetaData('key'))])
2008-03-20T23:33:18 <TheSheep> does anybody happen to remember if the Include macro gets called by some other code too?
2008-03-20T23:33:18 <TheSheep> does anybody happen to remember if the Include macro gets called by some other code too?
2008-03-20T23:33:36 <lanius> johill: think you can do such things with pyparsing
2008-03-20T23:33:36 <lanius> johill: think you can do such things with pyparsing
2008-03-20T23:33:38 <johill> TheSheep: grep?
2008-03-20T23:33:39 <johill> TheSheep: grep?
2008-03-20T23:33:50 <johill> lanius: no need to pull it in though
2008-03-20T23:33:50 <johill> lanius: no need to pull it in though
2008-03-20T23:34:16 <TheSheep> grep seems clear, I just wanted to make sure :)
2008-03-20T23:34:16 <TheSheep> grep seems clear, I just wanted to make sure :)
2008-03-20T23:34:40 <lanius> johill: ok
2008-03-20T23:34:40 <lanius> johill: ok
2008-03-20T23:34:47 <johill> TheSheep: hopefully tests? ;)
2008-03-20T23:34:47 <johill> TheSheep: hopefully tests? ;)
2008-03-20T23:35:05 <johill> lanius: for now we'd only have to define AND and MetaDataMatch and HasMetaData and maybe NOT
2008-03-20T23:35:05 <johill> lanius: for now we'd only have to define AND and MetaDataMatch and HasMetaData and maybe NOT
2008-03-20T23:35:59 <lanius> and something for pagename
2008-03-20T23:35:59 <lanius> and something for pagename
2008-03-20T23:36:17 <lanius> ok not needed atm
2008-03-20T23:36:17 <lanius> ok not needed atm
2008-03-20T23:36:21 <TheSheep> the thing is that it's getting some default args that macros usually don't get, so I figured it must be used somewhere else in addition
2008-03-20T23:36:21 <TheSheep> the thing is that it's getting some default args that macros usually don't get, so I figured it must be used somewhere else in addition
2008-03-20T23:36:21 <lanius> but for the future then
2008-03-20T23:36:21 <lanius> but for the future then
2008-03-20T23:36:36 <johill> lanius: yeah
2008-03-20T23:36:36 <johill> lanius: yeah
2008-03-20T23:36:50 <johill> and then each of those things has an evaluate function the backend can use if it doesn't want to optimise it
2008-03-20T23:36:50 <johill> and then each of those things has an evaluate function the backend can use if it doesn't want to optimise it
2008-03-20T23:36:55 <lanius> yes
2008-03-20T23:36:55 <lanius> yes
2008-03-20T23:37:29 <johill> I'm just thinking we could use the QueryParser stuff
2008-03-20T23:37:29 <johill> I'm just thinking we could use the QueryParser stuff
2008-03-20T23:37:45 <lanius> i haven't looked into it yet, so i don't know
2008-03-20T23:37:45 <lanius> i haven't looked into it yet, so i don't know
2008-03-20T23:38:01 <lanius> maybe ask thomas and xorAxAx
2008-03-20T23:38:02 <lanius> maybe ask thomas and xorAxAx
2008-03-20T23:38:24 <johill> but it does some interesting xapian things
2008-03-20T23:38:24 <johill> but it does some interesting xapian things
2008-03-20T23:38:29 <johill> maybe we should unify that with storage
2008-03-20T23:38:29 <johill> maybe we should unify that with storage
2008-03-20T23:38:58 <johill> so we just have a single entry point to find pages
2008-03-20T23:38:58 <johill> so we just have a single entry point to find pages
2008-03-20T23:39:01 <johill> backend.list_items()
2008-03-20T23:39:01 <johill> backend.list_items()
2008-03-20T23:39:07 <lanius> yes
2008-03-20T23:39:08 <lanius> yes
2008-03-20T23:39:19 <lanius> (query)
2008-03-20T23:39:19 <lanius> (query)
2008-03-20T23:39:21 <johill> and you can have a XapianSearchBackend(myotherbackend)
2008-03-20T23:39:21 <johill> and you can have a XapianSearchBackend(myotherbackend)
2008-03-20T23:39:44 <johill> xorAxAx: thoughts on that?
2008-03-20T23:39:44 <johill> xorAxAx: thoughts on that?
2008-03-20T23:39:49 <lanius> i don't know if that makes sense
2008-03-20T23:39:50 <lanius> i don't know if that makes sense
2008-03-20T23:39:58 <lanius> to put a xapian around a database?
2008-03-20T23:39:59 <lanius> to put a xapian around a database?
2008-03-20T23:40:05 <xorAxAx> once upon a time, the search's parser worked
2008-03-20T23:40:05 <xorAxAx> once upon a time, the search's parser worked
2008-03-20T23:40:17 <xorAxAx> what do you want to do?
2008-03-20T23:40:18 <xorAxAx> what do you want to do?
2008-03-20T23:40:30 <johill> xorAxAx: integrate search and storage
2008-03-20T23:40:30 <johill> xorAxAx: integrate search and storage
2008-03-20T23:40:55 <johill> lanius: of course it makes sense, a database doesn't necessarily index the blob contents
2008-03-20T23:40:55 <johill> lanius: of course it makes sense, a database doesn't necessarily index the blob contents
2008-03-20T23:41:05 <xorAxAx> generally, i would see problems here
2008-03-20T23:41:05 <xorAxAx> generally, i would see problems here
2008-03-20T23:41:09 <johill> where?
2008-03-20T23:41:09 <johill> where?
2008-03-20T23:41:22 <johill> it seems rather elegant to me :)
2008-03-20T23:41:22 <johill> it seems rather elegant to me :)
2008-03-20T23:41:35 <xorAxAx> as long as you dont depend on xapian :)
2008-03-20T23:41:35 <xorAxAx> as long as you dont depend on xapian :)
2008-03-20T23:41:38 <xorAxAx> in any case
2008-03-20T23:41:38 <xorAxAx> in any case
2008-03-20T23:41:43 <johill> heh no, of course not
2008-03-20T23:41:43 <johill> heh no, of course not
2008-03-20T23:41:54 <xorAxAx> also, the search syntax cant quote "
2008-03-20T23:41:54 <xorAxAx> also, the search syntax cant quote "
2008-03-20T23:42:10 <xorAxAx> and i am not sure, but negated queries had been broken for years
2008-03-20T23:42:10 <xorAxAx> and i am not sure, but negated queries had been broken for years
2008-03-20T23:42:19 <johill> that's both just bugs
2008-03-20T23:42:19 <xorAxAx> (because no unittests check that)
2008-03-20T23:42:19 <johill> that's both just bugs
2008-03-20T23:42:19 <xorAxAx> (because no unittests check that)
2008-03-20T23:42:22 <johill> nothing fundamental
2008-03-20T23:42:22 <johill> nothing fundamental
2008-03-20T23:42:36 <lanius> xorAxAx: in xapian itself or in moinmoin-xapian
2008-03-20T23:42:36 <lanius> xorAxAx: in xapian itself or in moinmoin-xapian
2008-03-20T23:43:24 <xorAxAx> in moinmoin
2008-03-20T23:43:24 <xorAxAx> in moinmoin
2008-03-20T23:43:26 <xorAxAx> johill: yep
2008-03-20T23:43:26 <xorAxAx> johill: yep
2008-03-20T23:43:40 <johill> xorAxAx: it seems that being able to ask a backend to find pages for an arbitrarily complex search expression based on metadata, page name and body makes sense
2008-03-20T23:43:40 <johill> xorAxAx: it seems that being able to ask a backend to find pages for an arbitrarily complex search expression based on metadata, page name and body makes sense
2008-03-20T23:43:42 <xorAxAx> well, the point is that the quality of those components decreased in last years soc
2008-03-20T23:43:42 <xorAxAx> well, the point is that the quality of those components decreased in last years soc
2008-03-20T23:43:57 <xorAxAx> and before reusing them for something even more central, some care should taken to increase it again
2008-03-20T23:43:57 <xorAxAx> and before reusing them for something even more central, some care should taken to increase it again
2008-03-20T23:44:11 <johill> the query parser isn't fundamental here
2008-03-20T23:44:11 <johill> the query parser isn't fundamental here
2008-03-20T23:44:17 <johill> the Expression tree is
2008-03-20T23:44:17 <johill> the Expression tree is
2008-03-20T23:44:20 <xorAxAx> its not only the parser
2008-03-20T23:44:20 <xorAxAx> its not only the parser
2008-03-20T23:44:30 <xorAxAx> the bug was/is not in the parser i think
2008-03-20T23:44:30 <xorAxAx> the bug was/is not in the parser i think
2008-03-20T23:44:53 <johill> and with that idea the xapian stuff would mostly have to be rewritten anyway
2008-03-20T23:44:53 <johill> and with that idea the xapian stuff would mostly have to be rewritten anyway
2008-03-20T23:44:56 <johill> or copied over
2008-03-20T23:44:56 <johill> or copied over
2008-03-20T23:44:59 <johill> to be a layered backend
2008-03-20T23:44:59 <johill> to be a layered backend
2008-03-20T23:45:15 <lanius> to replace the indexedbackend?
2008-03-20T23:45:16 <lanius> to replace the indexedbackend?
2008-03-20T23:45:23 <johill> no
2008-03-20T23:45:23 <johill> no
2008-03-20T23:45:42 <johill> possibly live above that to index the bodies, not just meta data
2008-03-20T23:45:42 <johill> possibly live above that to index the bodies, not just meta data
2008-03-20T23:45:49 <johill> or if xapian indexes meta data too, then replace it in that case
2008-03-20T23:45:49 <johill> or if xapian indexes meta data too, then replace it in that case
2008-03-20T23:45:52 <xorAxAx> johill: i guess weaving would be necessary, not sure why rewriting is a must
2008-03-20T23:45:52 <xorAxAx> johill: i guess weaving would be necessary, not sure why rewriting is a must
2008-03-20T23:46:02 <johill> xorAxAx: yeah, not too much rewrite
2008-03-20T23:46:02 <johill> xorAxAx: yeah, not too much rewrite
2008-03-20T23:46:24 <xorAxAx> xapian has another problem - the twisted guys dont maintain xapwrap
2008-03-20T23:46:24 <xorAxAx> xapian has another problem - the twisted guys dont maintain xapwrap
2008-03-20T23:46:30 <xorAxAx> but there are still other users, like openend :)
2008-03-20T23:46:30 <xorAxAx> but there are still other users, like openend :)
2008-03-20T23:46:33 <johill> heh
2008-03-20T23:46:33 <johill> heh
2008-03-20T23:46:41 <johill> I thought that was windows only ;)
2008-03-20T23:46:41 <johill> I thought that was windows only ;)
2008-03-20T23:47:01 <xorAxAx> generally, divmod seems to be bad when it comes to the upstreaming encouragement part of a maintainer's responsibilities :)
2008-03-20T23:47:01 <xorAxAx> generally, divmod seems to be bad when it comes to the upstreaming encouragement part of a maintainer's responsibilities :)
2008-03-20T23:47:16 <xorAxAx> windows is a problem because xapwrap needs a few hacks to support it
2008-03-20T23:47:16 <xorAxAx> windows is a problem because xapwrap needs a few hacks to support it
2008-03-20T23:47:22 <johill> ah ok
2008-03-20T23:47:22 <johill> ah ok
2008-03-20T23:47:32 <johill> well tbh, I don't care too much about xapian
2008-03-20T23:47:32 <johill> well tbh, I don't care too much about xapian
2008-03-20T23:47:32 <xorAxAx> but getting xapian working on windows is a large task on its own
2008-03-20T23:47:32 <xorAxAx> but getting xapian working on windows is a large task on its own
2008-03-20T23:47:48 <johill> but I could probably be convinced to make it at least somewhat work
2008-03-20T23:47:48 <johill> but I could probably be convinced to make it at least somewhat work
2008-03-20T23:47:52 <xorAxAx> about 3 hours of work if its your first time. thats how conslutants earn money :)
2008-03-20T23:47:52 <xorAxAx> about 3 hours of work if its your first time. thats how conslutants earn money :)
2008-03-20T23:47:57 <johill> if I were to put the request language in there
2008-03-20T23:47:57 <johill> if I were to put the request language in there
2008-03-20T23:48:16 <johill> xorAxAx: no, I don't own windows machines nor machines that can even run it
2008-03-20T23:48:16 <johill> xorAxAx: no, I don't own windows machines nor machines that can even run it
2008-03-20T23:48:33 <xorAxAx> johill: so? :)
2008-03-20T23:48:33 <xorAxAx> johill: so? :)
2008-03-20T23:48:35 <johill> or rther, that it could run on :)
2008-03-20T23:48:35 <johill> or rther, that it could run on :)
2008-03-20T23:48:47 <lanius> johill: a big mistake
2008-03-20T23:48:47 <lanius> johill: a big mistake
2008-03-20T23:48:51 <lanius> :)
2008-03-20T23:48:51 <lanius> :)
2008-03-20T23:49:30 <johill> likely. very.
2008-03-20T23:49:30 <johill> likely. very.
2008-03-20T23:50:21 <johill> wow the xapian stuff is crappy
2008-03-20T23:50:21 <johill> wow the xapian stuff is crappy
2008-03-20T23:51:36 <johill> all the special cases in generic code :(
2008-03-20T23:51:36 <johill> all the special cases in generic code :(
2008-03-20T23:52:29 <xorAxAx> which xapian stuff?
2008-03-20T23:52:29 <xorAxAx> which xapian stuff?
2008-03-20T23:52:31 <xorAxAx> moin's?
2008-03-20T23:52:31 <xorAxAx> moin's?
2008-03-20T23:52:39 <johill> yes
2008-03-20T23:52:39 <johill> yes
2008-03-20T23:53:26 <dreimark> we should have a prefilter for the log file which combines same messages
2008-03-20T23:53:26 <dreimark> we should have a prefilter for the log file which combines same messages
2008-03-20T23:53:34 <johill> dreimark: huh?
2008-03-20T23:53:34 <johill> dreimark: huh?
2008-03-20T23:53:35 <xorAxAx> dreimark: ?
2008-03-20T23:53:35 <xorAxAx> dreimark: ?
2008-03-20T23:53:42 <johill> can I have a branch?
2008-03-20T23:53:42 <johill> can I have a branch?
2008-03-20T23:53:44 <johill> :)
2008-03-20T23:53:44 <johill> :)
2008-03-20T23:54:09 <johill> I want to try pushing search through to backends
2008-03-20T23:54:09 <johill> I want to try pushing search through to backends
2008-03-20T23:54:40 <xorAxAx> reminds me at happy tree friends
2008-03-20T23:54:40 <xorAxAx> reminds me at happy tree friends
2008-03-20T23:54:49 <lanius> can i say something about the underlay stuff before i go to sleep ;)
2008-03-20T23:54:49 <lanius> can i say something about the underlay stuff before i go to sleep ;)
2008-03-20T23:54:56 <johill> lanius: sure
2008-03-20T23:54:56 <johill> lanius: sure
2008-03-20T23:55:17 <lanius> the idea was that the new storage doesn't know anything about the underlay stuff
2008-03-20T23:55:17 <lanius> the idea was that the new storage doesn't know anything about the underlay stuff
2008-03-20T23:55:24 <lanius> to throw away the underlay idea
2008-03-20T23:55:24 <lanius> to throw away the underlay idea
2008-03-20T23:55:27 <dreimark> if you do 20000 entries on the same page by xmlrpc
2008-03-20T23:55:27 <dreimark> if you do 20000 entries on the same page by xmlrpc
2008-03-20T23:55:28 <lanius> and replace it by the layers
2008-03-20T23:55:28 <lanius> and replace it by the layers
2008-03-20T23:55:39 <johill> lanius: yes
2008-03-20T23:55:39 <johill> lanius: yes
2008-03-20T23:55:46 <lanius> johill: to be compatible the page.py still needs the underlay though
2008-03-20T23:55:46 <lanius> johill: to be compatible the page.py still needs the underlay though
2008-03-20T23:55:57 <dreimark> for the statistic it is enought to know 20000 and not 20000 x 1 entry
2008-03-20T23:55:57 <dreimark> for the statistic it is enought to know 20000 and not 20000 x 1 entry
2008-03-20T23:56:02 <johill> lanius: but how do you express "want to search only non-underlay"?
2008-03-20T23:56:02 <johill> lanius: but how do you express "want to search only non-underlay"?
2008-03-20T23:56:02 <dreimark> to loop
2008-03-20T23:56:02 <dreimark> to loop
2008-03-20T23:56:10 <ThomasWaldmann> the new download button has a i18n problem
2008-03-20T23:56:10 <ThomasWaldmann> the new download button has a i18n problem
2008-03-20T23:56:10 <lanius> johill: just like i did it
2008-03-20T23:56:11 <lanius> johill: just like i did it
2008-03-20T23:56:25 <ThomasWaldmann> for german, the string is longer than the box
2008-03-20T23:56:25 <ThomasWaldmann> for german, the string is longer than the box
2008-03-20T23:56:46 <johill> lanius: that's no good though, having two ways to get at the same thing
2008-03-20T23:56:46 <johill> lanius: that's no good though, having two ways to get at the same thing
2008-03-20T23:56:56 <lanius> johill: the same thing?
2008-03-20T23:56:56 <lanius> johill: the same thing?
2008-03-20T23:57:08 <johill> either through layered or directly as 'underlay_storage'
2008-03-20T23:57:09 <johill> either through layered or directly as 'underlay_storage'
2008-03-20T23:57:12 <johill> or whatever it was called
2008-03-20T23:57:12 <johill> or whatever it was called
2008-03-20T23:57:26 <lanius> johill: that is just to be backwards compatible
2008-03-20T23:57:26 <lanius> johill: that is just to be backwards compatible
2008-03-20T23:57:53 <lanius> johill: because there are those is_underlay calls in page.py
2008-03-20T23:57:54 <lanius> johill: because there are those is_underlay calls in page.py
2008-03-20T23:58:08 <johill> think a level higher
2008-03-20T23:58:08 <johill> think a level higher
2008-03-20T23:58:29 <johill> and you will need to be able to distinguish which layer something came from
2008-03-20T23:58:30 <johill> and you will need to be able to distinguish which layer something came from
2008-03-20T23:58:40 <lanius> then this must be implemented in the layered backend
2008-03-20T23:58:40 <lanius> then this must be implemented in the layered backend
2008-03-20T23:59:06 <johill> via metadata
2008-03-20T23:59:06 <johill> via metadata
2008-03-20T23:59:16 <lanius> no
2008-03-20T23:59:16 <lanius> no
2008-03-20T23:59:27 <johill> how then
2008-03-20T23:59:27 <johill> how then
2008-03-20T23:59:41 <lanius> i don't think metadata is right
2008-03-20T23:59:41 <lanius> i don't think metadata is right
MoinMoin: MoinMoinChat/Logs/moin-dev/2008-03-20 (last edited 2008-03-20 00:15:02 by IrcLogImporter)