MoinMoin: MoinMoinChat/Logs/moin-dev/2006-09-05

2006-09-05T00:12:59  * xorAxAx filed a bug
2006-09-05T00:45:44  <ThomasWaldmann> mitsuhiko: yes
2006-09-05T00:46:07  <xorAxAx> ThomasWaldmann: as i said, i will fix that again
2006-09-05T00:46:15  <xorAxAx> otherwise you have to many broken external links
2006-09-05T00:47:40  <xorAxAx> btw, i fixed the bug that i filed yesterday
2006-09-05T00:47:48  <ThomasWaldmann> "_" and " " must be different things
2006-09-05T00:47:58  <xorAxAx> ThomasWaldmann: yeah, thats your goal
2006-09-05T00:48:16  <xorAxAx> the main motivation of your current work was the broken normalisation, though
2006-09-05T00:48:23  <xorAxAx> but i dont doubt its usefulness
2006-09-05T00:48:31  <xorAxAx> nevertheless, using _ will be still possible
2006-09-05T00:49:16  <ThomasWaldmann> and you encode _ to %xx then?
2006-09-05T00:49:25  <xorAxAx> in the output? yes
2006-09-05T00:49:41  <xorAxAx> i will just put a "guess_page" function into request
2006-09-05T00:49:58  <ThomasWaldmann> guess?
2006-09-05T00:50:01  <xorAxAx> yes
2006-09-05T00:50:08  <xorAxAx> because _ means underscore normally
2006-09-05T00:50:13  <xorAxAx> but it might mean space
2006-09-05T00:50:28  <xorAxAx> so guess_page(name) will return a page object or page name
2006-09-05T00:52:01  <ThomasWaldmann> and it does what internally? checking is page exists?
2006-09-05T00:52:42  <ThomasWaldmann> ...if...
2006-09-05T00:53:42  <ThomasWaldmann> btw, if you want to fix something there, you have to document how that would works ASAP
2006-09-05T00:54:06  <ThomasWaldmann> and guessing sounds just like other trouble
2006-09-05T00:55:38  <xorAxAx> ASAP? doesnt make sense, i havent started :)
2006-09-05T00:55:52  <xorAxAx> please elaborate on the trouble
2006-09-05T00:56:34  <ThomasWaldmann> url : item must be 1:1 mapping
2006-09-05T00:57:46  <ThomasWaldmann> and it should be possible to have "my page" and "my_page" at the same time
2006-09-05T00:58:06  <xorAxAx> no, 1:1 doesnt make sense
2006-09-05T00:58:13  <xorAxAx> the latter one will be the case of course
2006-09-05T00:59:03  <ThomasWaldmann> 1:1 doesnt mean it must be identity mapping, but bijective
2006-09-05T00:59:25  <xorAxAx> the 1.5 underscore semantics were unclear when it came to normalisation
2006-09-05T00:59:38  <xorAxAx> so it was done well enough nowhere
2006-09-05T01:00:08  <xorAxAx> of course its an acceptable choice to rip it out completly, but breaking all external links harm user experience too much
2006-09-05T01:00:33  <ThomasWaldmann> and it has to work with "filesystem as pages"
2006-09-05T01:00:53  <xorAxAx> ?
2006-09-05T01:01:30  <ThomasWaldmann> some VirtualPages application I need at a customer
2006-09-05T01:02:20  <ThomasWaldmann> FS/a/b/c -> filesystem /mountpoint/a/b/c file
2006-09-05T01:02:30  <xorAxAx> yeah, i dont see any collision there
2006-09-05T01:02:39  <xorAxAx> if page.exists works
2006-09-05T01:04:02  <ThomasWaldmann> even stuff like "my_letter to_xy.doc" should work
2006-09-05T01:04:19  <xorAxAx> if your page layer supports exists, that should be no problem
2006-09-05T01:04:30  <xorAxAx> i guess VP always returns true there, so it works out of the box
2006-09-05T01:07:15  <ThomasWaldmann> i guess it even returns a correct .exists()
2006-09-05T01:10:17  <ThomasWaldmann> btw, using py 2.5 works for 1.5.current?
2006-09-05T01:12:18  <xorAxAx> hmm, no idea
2006-09-05T01:12:32  <xorAxAx> but i cannot think of a reason not to use it
2006-09-05T01:21:27  <ThomasWaldmann> so your changes were in 1.6 I guess
2006-09-05T01:23:49  <xorAxAx> the thread monitor ones?
2006-09-05T01:23:51  <xorAxAx> of course
2006-09-05T01:23:56  <xorAxAx> new features go into 1.6
2006-09-05T01:29:22  <xorAxAx> ThomasWaldmann: are you sure that ldap_login, which has to save the user object, does not write the password to disk by simply saving the object?
2006-09-05T01:29:49  <xorAxAx> u.create_or_update(True)
2006-09-05T01:29:55  <xorAxAx> i think that will write it to disk
2006-09-05T01:30:01  <ThomasWaldmann> look into ldap_login to see {SHA}NotStored
2006-09-05T01:30:19  <xorAxAx> ah
2006-09-05T01:30:35  <xorAxAx> cool
2006-09-05T01:30:55  <ThomasWaldmann> that was the reason to split moin_cookie into 2 parts interacting this way
2006-09-05T01:39:42  <ThomasWaldmann> xorAxAx: http://moin16.wikiwikiweb.de/SyntaxReference
2006-09-05T01:40:00  <xorAxAx> hmm, it works there
2006-09-05T01:40:00  <ThomasWaldmann> see last edit there
2006-09-05T01:40:38  <xorAxAx> i dont see any edit
2006-09-05T01:40:43  <xorAxAx> just 1 rev
2006-09-05T01:42:22  <ThomasWaldmann> i just have added a space to the }}} inside the pre
2006-09-05T01:42:48  <ThomasWaldmann> otherwise it is expected that the pre ends there
2006-09-05T01:43:45  <xorAxAx> ThomasWaldmann: you will get an edit by martin onto your homepage, lets bet :)
2006-09-05T01:48:07  <ThomasWaldmann> the question there is rather why that worked in 1.5 for that case
2006-09-05T01:53:36  <ThomasWaldmann> xorAxAx: the memory consumption, how much is it lower?
2006-09-05T01:54:14  <xorAxAx> ThomasWaldmann: i got it over 100 MB on 2.4, on 2.5 its at 10 MB
2006-09-05T01:55:53  <ThomasWaldmann> oh, that sounds like a fixed py memory manager :)
2006-09-05T01:58:07  <ThomasWaldmann> The release date is planned for  12 September 2006.
2006-09-05T01:58:22  <ThomasWaldmann> that sounds OK :)
2006-09-05T02:00:34  <ThomasWaldmann> ASCII became the default coding
2006-09-05T02:00:54  <ThomasWaldmann> wasnt it like that ever?
2006-09-05T02:02:24  <xorAxAx> umm, yeah
2006-09-05T02:02:32  <xorAxAx> but not of files
2006-09-05T02:07:00  <xorAxAx> hmm, MoinMoin.search.results is full of illegal formatter usage
2006-09-05T02:07:03  <xorAxAx> fpletz:
2006-09-05T02:16:20  <ThomasWaldmann> fpletz: and shadowing "list" builtin is a bad idea, too
2006-09-05T02:16:57  <ThomasWaldmann> (line 313)
2006-09-05T02:19:06  <xorAxAx> i am currently debugging a really bad bug in the search code
2006-09-05T02:23:35  <ThomasWaldmann> fpletz: same for "type" (125 and many other places)
2006-09-05T02:24:11  <xorAxAx> i wonder how one can write such non-reentrant code (it was written before franz)
2006-09-05T02:24:27  * xorAxAx is in the mood for a blame :)
2006-09-05T02:24:28  <ThomasWaldmann> (sometimes it is good to look at vim's colouring for such cases)
2006-09-05T02:25:41  <ThomasWaldmann> i guess you will find edits by me and fabi :)
2006-09-05T02:26:09  <xorAxAx> it was NIR!
2006-09-05T02:26:20  <xorAxAx> Sat Nov 06 20:33:14 2004 +0000 (22 months ago)
2006-09-05T02:26:30  <ThomasWaldmann> oh, really? :)
2006-09-05T02:28:53  <xorAxAx> but its not the root cause
2006-09-05T02:32:09  <xorAxAx> > c:\dokumente und einstellungen\administrator\eigene dateien\progra\python\moin
2006-09-05T02:32:12  <xorAxAx> moin\moin-1.6-sync\moinmoin\search\results.py(782)getvalue()
2006-09-05T02:32:15  <xorAxAx> thats bad
2006-09-05T02:32:17  <xorAxAx> -> self.formatter.div(1, css_class='searchresults'),
2006-09-05T02:32:20  <xorAxAx> (Pdb) self.formatter
2006-09-05T02:32:23  <xorAxAx> <MoinMoin.formatter.pagelinks.Formatter instance at 0x010F1EB8>
2006-09-05T02:34:17  <xorAxAx> and the offending line is Page.py:1057
2006-09-05T02:34:25  <xorAxAx> i think its quite old
2006-09-05T02:34:32  <xorAxAx> but it conflicted with another change by nir
2006-09-05T02:37:41  <xorAxAx> wee, fixed
2006-09-05T02:38:36  <ThomasWaldmann> did you use any configure options except --enable-unicode=ucs4?
2006-09-05T02:39:04  <ThomasWaldmann> (for py25)
2006-09-05T02:40:28  <xorAxAx> i used the rc1 build
2006-09-05T02:40:41  <xorAxAx> and i think ucs2 is standard on windows
2006-09-05T03:44:06  <ThomasWaldmann> shit, xapian bindings fail on 2.5
2006-09-05T04:22:37  <ThomasWaldmann> i notified richard about it, let's see if it gets fixed before 2.5 release
2006-09-05T12:21:31  <Kepplar> xorAxAx: are you able to cotribute to the diagram and design since TW and others cant?
2006-09-05T12:21:50  <xorAxAx> Kepplar: no, as i said, i suggested another way to do this
2006-09-05T12:21:59  <xorAxAx> Kepplar: we need an overview over the old design first
2006-09-05T12:22:30  <Kepplar> i was talking about the old design :)
2006-09-05T12:22:41  <Kepplar> sorry
2006-09-05T12:23:28  <xorAxAx> re that i said that you should do the first step :)
2006-09-05T12:24:30  <Kepplar> xorAxAx: only bit i know is <disk> <----> <page> and user etc etc and some request object getting parsed around and thats it, but ok ill do that if it helps that much :P
2006-09-05T13:41:47  <xorAxAx> ThomasWaldmann: print action is broken in 1.6
2006-09-05T15:35:32  <ThomasWaldmann> xorAxAx: fixed