MoinMoin: MoinMoinChat/Logs/moin-dev/2005-12-02

2005-12-02T00:25:59  <xorAxAx> ok, fixed a packages bug and this testing bug in main branch
2005-12-02T00:35:08  * xorAxAx fixes another bug related to SyntaxReference
2005-12-02T00:44:56  *** Fabi has quit IRC
2005-12-02T00:54:39  * xorAxAx adds a bunch of tests
2005-12-02T01:00:44  <_keturn> ugh.   now I see why the arch docs say to make sure your tree is clean before you merge from upstream.  That was messy.
2005-12-02T01:02:47  <_keturn> xorAxAx: while you're in the tests, the list comprehension in tests/runtests.run is broken.  It actually does not work if the names have the package path prepended to them, contrary to the documentation in _tests.run.
2005-12-02T01:03:36  <xorAxAx> yeah, suggested fix? :-)
2005-12-02T01:04:14  <_keturn>  names = sys.argv[1:]  Works For Me
2005-12-02T01:07:25  <xorAxAx> ok
2005-12-02T01:16:24  *** starshine_away is now known as starshine
2005-12-02T01:16:27  <xorAxAx> committed
2005-12-02T01:25:48  * xorAxAx tries fabi's stroke fixes for docbook
2005-12-02T01:25:58  <xorAxAx> does not work ;-)
2005-12-02T01:26:02  * xorAxAx fixes them
2005-12-02T01:34:02  <xorAxAx> ThomasWaldmann: i fucked up the nice tla, please fix up the arch (by running this fixup commmand9
2005-12-02T01:34:09  * xorAxAx loves tla
2005-12-02T01:34:22  * xorAxAx goes to bed with a dirty work dir
2005-12-02T01:35:34  <_keturn> Hmm.  Adding OpenID support won't be too hard.  But adding *optional* OpenID support without adding a hard dependency on the OpenID libraries is going to make for some messy code.
2005-12-02T01:36:01  <xorAxAx> why?
2005-12-02T01:37:42  <xorAxAx> _keturn:
2005-12-02T01:39:32  <xorAxAx> ThomasWaldmann: test case is a simple commit to the 1.5 branch btw, it fails
2005-12-02T01:41:17  <xorAxAx> gub
2005-12-02T01:44:50  <_keturn> xorAxAx: because I want to add a branch to UserSettingsHandler.handleData, and it's going to reference my stuff, right?  But it looks like I'll have to add an extra check to handle the case where my stuff  didn't load (probably because of lack of supporting libraries)
2005-12-02T01:45:29  <_keturn> unless there is a hook somewhere I can employ instead of hacking USH.handleData.
2005-12-02T02:19:53  <ThomasWaldmann> xorAxAx: can you try changing some userpref value, save, click away the msg
2005-12-02T02:20:18  <ThomasWaldmann> does it show the full userform afterwards or like when you are logged out?
2005-12-02T02:20:30  <ThomasWaldmann> what does reloading the page show?
2005-12-02T02:40:08  * ThomasWaldmann fixed tla arch
2005-12-02T02:40:21  <ThomasWaldmann> and the fixup cmd doesnt help, btw
2005-12-02T06:33:37  *** starshine is now known as starshine_away
2005-12-02T07:53:21  *** lydon has joined #moin-dev
2005-12-02T08:08:18  *** starshine_away is now known as starshine
2005-12-02T09:14:16  *** starshine is now known as starshine_away
2005-12-02T09:32:59  *** Fabi has joined #moin-dev
2005-12-02T09:40:54  *** moinBot has joined #moin-dev
2005-12-02T09:40:54  *** moinBot` has quit IRC
2005-12-02T09:47:54  <xorAxAx> ThomasWaldmann: it works as expected
2005-12-02T09:55:09  <xorAxAx>  arch_commit: unable to acquire revision lock (404 Not Found) tree: /home/aschremmer/moin--main--1.5 revision: arch@arch.thinkmo.de--2003-archives/moin--main--1.5--patch-269
2005-12-02T09:55:12  <xorAxAx> ThomasWaldmann: doesnt work
2005-12-02T10:00:17  <xorAxAx> ThomasWaldmann: why did you add that auth_method stuff if its just used for debuggin?
2005-12-02T10:27:52  * xorAxAx announces a paper by the pypy team about their development style: http://codespeak.net/pypy/extradoc/talk/22c3/agility.pdf
2005-12-02T11:48:03  <ThomasWaldmann> it is not just used for debugging
2005-12-02T11:50:51  <ThomasWaldmann> xorAxAx: retry
2005-12-02T11:51:55  <ThomasWaldmann> xorAxAx: and that userprefs page doesnt work as expected for me. almost ever it looks like being logged out after clicking away that msg
2005-12-02T11:52:13  <ThomasWaldmann> if you reload the page, it looks like being logged in again
2005-12-02T11:52:53  <ThomasWaldmann> i could not find the reason for that yet, except that obviously moin_cookie returns None in that case
2005-12-02T11:53:29  <ThomasWaldmann> and thus, that final User() after the auth list loop is used.
2005-12-02T12:01:14  <ThomasWaldmann> hmm, cleared browser cache and now i can't reproduce any more. fsck.
2005-12-02T12:02:49  <ThomasWaldmann> btw, how about returning x from auth method with x in (False, True, userobj)
2005-12-02T12:03:27  <ThomasWaldmann> if it is a bool it just tells whether to continue with other auth methods
2005-12-02T12:05:36  <ThomasWaldmann> hmm, return continueflag, user  is cleaner
2005-12-02T12:15:16  <xorAxAx> ThomasWaldmann: still doesnt work arch_commit: unable to acquire revision lock (404 Not Found) tree: /home/aschremmer/moin--main--1.5 revision: arch@arch.thinkmo.de--2003-archives/moin--main--1.5--patch-269
2005-12-02T12:20:55  <ThomasWaldmann> retry
2005-12-02T12:21:22  <ThomasWaldmann> must ask waldi if he changed acls there
2005-12-02T12:22:38  <xorAxAx> ThomasWaldmann: still doesnt work arch_commit: unable to acquire revision lock (404 Not Found) tree: /home/aschremmer/moin--main--1.5 revision: arch@arch.thinkmo.de--2003-archives/moin--main--1.5--patch-270
2005-12-02T12:38:28  <ThomasWaldmann> retry
2005-12-02T12:38:49  <lydon> ThomasWaldmann: is moin_dump.py still considered experimental?
2005-12-02T12:38:53  <ThomasWaldmann> I'm asking myself why i suddenly need to set mode/user manually....
2005-12-02T12:38:58  <xorAxAx> works
2005-12-02T12:39:06  <ThomasWaldmann> lydon: no, it should work
2005-12-02T12:39:14  <lydon> ok, then I have some fixes
2005-12-02T12:40:48  <xorAxAx> ThomasWaldmann: i canceled the transaction-safe commit with CTRL-C
2005-12-02T12:40:57  <xorAxAx> ThomasWaldmann: but such a state should never happen anyway
2005-12-02T12:42:34  <lydon> but it seems it will be hard to commit them, currently :)
2005-12-02T12:47:12  <xorAxAx> no, works again
2005-12-02T12:48:20  <lydon> oki.
2005-12-02T12:48:27  <lydon> ok. its a fix and an enhancement
2005-12-02T12:48:41  <lydon> the fix is to load a theme, cause without stuff like smiley break
2005-12-02T12:49:04  <lydon> the enhancement is a filter parameter for dumping only pages starting with a certain prefix
2005-12-02T12:49:28  <xorAxAx> lydon: theme loading has been in for a few days, please check with current
2005-12-02T12:49:51  <xorAxAx> and dont add it again to moin_dump, i moved it from there yesterday
2005-12-02T12:50:32  <lydon> yeah, my installation is about two weeks old, I guess
2005-12-02T12:51:03  <ThomasWaldmann> lydon: i already fixed some theme cli req issue
2005-12-02T12:52:02  <ThomasWaldmann> and we already have some method to dump a single(?) page
2005-12-02T12:52:07  <lydon> its a pity that I don't can spare more time in the last months
2005-12-02T12:52:18  <ThomasWaldmann> maybe that simply should be some regex being matched
2005-12-02T12:52:22  <lydon> yep.
2005-12-02T12:53:20  <lydon> I have some pseudo subwikis on devilsclaw.de. and to carry one of them with me thats a nice function to save space on my mobile
2005-12-02T12:54:21  <lydon> currently I do it with a simple startswith filter map on pages, but a regex is not much harder and perhaps of use, too
2005-12-02T12:59:01  <ThomasWaldmann> a regex patch would be nice
2005-12-02T12:59:45  <ThomasWaldmann> and please do some tests, i dont think I will have time to test moin_dump again
2005-12-02T13:00:20  <ThomasWaldmann> and include a doc update for your stuff
2005-12-02T13:00:43  <ThomasWaldmann> CHANGES and wiki page(s)
2005-12-02T13:06:00  <xorAxAx> the solution to that is writing tests :-)
2005-12-02T13:07:16  <ThomasWaldmann> yeah, best are too restrictive tests you will have to fix all the time, too
2005-12-02T13:12:30  <xorAxAx> yeah, then they have to be trashed like the code they test
2005-12-02T13:12:42  <xorAxAx> e.g. kensanata tests all xml things using xpath
2005-12-02T13:12:56  <xorAxAx> much less pain if the formatter changes ...
2005-12-02T13:13:38  <ThomasWaldmann> gtg/bbl
2005-12-02T14:01:02  * xorAxAx corrects himself and notes that -ize might be correct british english in many cases
2005-12-02T14:02:15  <xorAxAx> at least OED says that -ize is more correct because its nearer to latin
2005-12-02T15:11:30  *** nwp_ has quit IRC
2005-12-02T15:26:11  *** nwp_ has joined #moin-dev
2005-12-02T15:28:34  *** nwp_ has quit IRC
2005-12-02T15:48:03  *** nwp_ has joined #moin-dev
2005-12-02T16:10:25  *** lydon has quit IRC
2005-12-02T23:47:04  <ThomasWaldmann> nwp_: i committed your request change
2005-12-02T23:52:17  <ThomasWaldmann> please all developers test the request change in patch-276 in non-Twisted setups