MoinMoin: MoinMoinChat/Logs/moin-dev/2009-08-06

2009-08-06T00:21:58  <ThomasWaldmann> dennda: that meal was at the LUG meeting
2009-08-06T00:37:20  *** grzywacz has quit IRC
2009-08-06T00:40:13  *** vinci_ is now known as tpfennig
2009-08-06T02:27:38  <CIA-38> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6419:2f11d8567fb0 2.0-storage/MoinMoin/storage/backends/hg.py: (hg) refactored revision and internal metadata handling
2009-08-06T02:27:39  <CIA-38> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6420:621f3f0a137a 2.0-storage/MoinMoin/storage/backends/hg.py: (hg) refined comments, removed unsafe _open_revision_data code, minor refactoring
2009-08-06T02:46:35  <CIA-38> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6421:d1329703b69b 2.0-storage/MoinMoin/storage/backends/hg.py: (hg) removed unused filter condition
2009-08-06T02:49:12  *** franklin has quit IRC
2009-08-06T02:49:12  *** ThomasWaldmann has quit IRC
2009-08-06T02:52:30  *** franklin has joined #moin-dev
2009-08-06T02:52:30  *** ThomasWaldmann has joined #moin-dev
2009-08-06T03:21:16  <CIA-38> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6422:a79b491ffea6 2.0-storage/MoinMoin/storage/backends/hg.py:
2009-08-06T03:21:16  <CIA-38> (hg) - item index file: renamed from cache, introduced helper
2009-08-06T03:21:16  <CIA-38> - refactored commit, list_revisions
2009-08-06T03:21:16  <CIA-38> - updated docstrings
2009-08-06T03:21:17  <CIA-38> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6423:3ea024cf01a8 2.0-storage/MoinMoin/storage/backends/hg.py: (hg) pep8 fixes
2009-08-06T03:41:01  *** tpfennig has quit IRC
2009-08-06T04:46:37  <CIA-38> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6424:b2919ff383ee 2.0-storage/MoinMoin/storage/ (_tests/test_backends_hg.py backends/hg.py): (hg) implemented item_destroy
2009-08-06T07:49:45  *** SuBmUnDo_ has quit IRC
2009-08-06T07:50:25  *** SuBmUnDo_ has joined #moin-dev
2009-08-06T07:54:06  *** SuBmUnDo_ has quit IRC
2009-08-06T07:55:58  *** SuBmUnDo_ has joined #moin-dev
2009-08-06T07:57:35  *** SuBmUnDo_ has quit IRC
2009-08-06T08:19:03  <ThomasWaldmann> moin
2009-08-06T08:54:23  <ThomasWaldmann> dennda: moin
2009-08-06T08:57:30  <ThomasWaldmann>      1.7 -        timeout = 2.0 # minimal timout
2009-08-06T08:57:30  <ThomasWaldmann>      1.8 +        timeout = 3.0 # minimal timout
2009-08-06T08:57:45  <ThomasWaldmann> dennda: if you change something, do you read it first?
2009-08-06T08:59:00  <dennda> oh, you mean the comment?
2009-08-06T09:07:36  <ThomasWaldmann> sure
2009-08-06T09:15:33  <ThomasWaldmann> dennda: you also have 33 fails right now?
2009-08-06T09:22:38  <dreimark> moin
2009-08-06T09:22:53  <dreimark> those fail because of the test environment change
2009-08-06T09:25:02  <dennda> Yes
2009-08-06T09:25:11  <dennda> due to what reimar said
2009-08-06T09:25:59  <dennda> that's easily backout-able though, and we could do that at the end if time allows (I don't feel there's much gain right now)
2009-08-06T09:28:16  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6425:7f7b331bc141 2.0-storage/MoinMoin/util/_tests/test_lock.py: storage: fix comment
2009-08-06T09:28:16  <dennda> ui PawelPacana nice
2009-08-06T09:28:18  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6426:b008633826a8 2.0-storage/MoinMoin/storage/ (_tests/test_backends_hg.py backends/hg.py): storage: merge remote 2.0-storage
2009-08-06T09:38:40  *** tpfennig has joined #moin-dev
2009-08-06T11:02:04  <ThomasWaldmann> dennda:  * Note: There has been work on Xapian in 2.0 ongoing as part of GSoC 2009. That is not yet merged.
2009-08-06T11:06:22  <dennda> is that not true? i thought dimazest was working on that
2009-08-06T11:08:28  <ThomasWaldmann> http://hg.moinmo.in/moin/1.9-xapian-dmilajevs/
2009-08-06T11:09:18  <dreimark> dennda: for 1.9
2009-08-06T11:10:17  <dennda> ahh
2009-08-06T11:10:21  <dennda> well ok
2009-08-06T11:35:54  <ThomasWaldmann> dennda: your test environ test is incomplete / wrong
2009-08-06T11:36:58  <ThomasWaldmann> a) it should not use a well-known systempage for testing what it tests right now, that should be done with any random page name
2009-08-06T11:37:46  <ThomasWaldmann> b) it should test wether preloading the system pages works and whether they are also freshly preloaded for the next tests
2009-08-06T11:38:19  <ThomasWaldmann> please fix that asap
2009-08-06T11:38:45  <dreimark> bbl
2009-08-06T11:39:06  <ThomasWaldmann> c) it should test whether user defined Config class works
2009-08-06T11:39:26  <ThomasWaldmann> (maybe this is already tested with b)
2009-08-06T11:41:14  <ThomasWaldmann> testitems.xml misses the is_systempage (or so) flag
2009-08-06T11:41:46  <ThomasWaldmann> we still need a tagging script, so that metadata can be added after conversion from 1.9 syspages
2009-08-06T11:42:30  <ThomasWaldmann> also, we can use the same script to tag syspages with the version number when we create a release
2009-08-06T11:43:48  <ThomasWaldmann> that script should work with the usual scripting framework: moin maint setmeta ...
2009-08-06T11:44:24  <ThomasWaldmann> dennda: ?
2009-08-06T11:45:53  <dennda> yeah sorry was on the phone with the citi bank. there was some issue related to my gsoc card
2009-08-06T11:47:13  <dennda> ok I will add that
2009-08-06T11:47:21  <dennda> I currently have my editor for HOACL 2.0 open
2009-08-06T11:47:30  <dennda> sorry, that call was rather lengthy...
2009-08-06T11:48:02  <ThomasWaldmann> btw, the testsearch failure looks like preload is not working. your new test (see above) will find that also maybe.
2009-08-06T11:48:45  <ThomasWaldmann> I'll have a look at test_security
2009-08-06T11:49:07  <dennda> that's due to backends being refreshed for every testcase and the preloaded stuff is only done once on the class level (and after that, set to None so it cannot work again)
2009-08-06T11:49:38  <ThomasWaldmann> that's inconsistent
2009-08-06T11:51:32  <dennda> sure
2009-08-06T11:58:02  <ThomasWaldmann> what's about enduser?
2009-08-06T11:59:24  <ThomasWaldmann> it still doesn't have the simple changes we talked about before quite a while and it is still in wrong package
2009-08-06T12:01:42  <ThomasWaldmann> dirties_backend - does it work, is it outdated?
2009-08-06T12:01:57  <dennda> well yes I was fiddling with the test changes that you wanted (of which I'm starting to think that they are not necessary at the moment and a waste of time)
2009-08-06T12:02:01  <dennda> well that depends
2009-08-06T12:02:09  <dennda> see the questions I asked you yesterday evening
2009-08-06T12:04:45  <ThomasWaldmann> can you be a bit more precise?
2009-08-06T12:04:55  <dennda> ok
2009-08-06T12:06:49  <dennda> We have had 0 failing tests. Then you wanted me to make the tests more consistent by having each test run on fresh backends. Finding out how to do that alone took quite some time. Now there's some solution, but of course this breaks our tests again (for almost no gain)
2009-08-06T12:07:00  <dennda> so PawelPacana raised the question yesterday of whether that was actually necessary
2009-08-06T12:07:37  <ThomasWaldmann> that we had 0 failing was by having workarounds, as you should know
2009-08-06T12:07:42  <dennda> yes
2009-08-06T12:08:11  <dennda> but it worked and the tests were not invalidated due to the workarounds
2009-08-06T12:08:16  <ThomasWaldmann> and there is no solution yet and not even a test for that solution yet
2009-08-06T12:08:42  <dennda> ?
2009-08-06T12:08:55  <dennda> as it currently is, every test case (read: method) has fresh backends
2009-08-06T12:09:37  <ThomasWaldmann> empty ones, yes, even the test who want preloaded ones
2009-08-06T12:10:53  <ThomasWaldmann> so that does rather sound like half-of-a-solution
2009-08-06T12:11:31  <dennda> true
2009-08-06T12:11:43  <dennda> but if that is done and tests are adjusted, what do we gain in the end?
2009-08-06T12:12:23  <ThomasWaldmann> that we have (finally) more predictable tests and don't need the workarounds any more?
2009-08-06T12:13:14  <dennda> is that really the most important thing right now? besides, it will slow down tests
2009-08-06T12:15:19  <ThomasWaldmann> well, maybe just try?
2009-08-06T12:15:30  <dennda> i have
2009-08-06T12:15:36  <ThomasWaldmann> so?
2009-08-06T12:16:01  <dennda> it needs further special casing
2009-08-06T12:16:06  <dennda> as I have just found out
2009-08-06T12:16:39  <ThomasWaldmann> ?
2009-08-06T12:16:50  <dennda> well let me just do it, it's not that easy
2009-08-06T12:16:59  <ThomasWaldmann> first add the test
2009-08-06T12:17:23  <ThomasWaldmann> and maybe fiddle a bit less
2009-08-06T12:17:34  <dennda> %-)
2009-08-06T12:17:41  <dennda> what is that supposed to mean?
2009-08-06T12:18:02  <ThomasWaldmann> that sometimes thinking before doing saves a lot of time
2009-08-06T12:18:15  <ThomasWaldmann> and changesets
2009-08-06T12:18:51  <dennda> i have tried to find a less hacky solution, seriously
2009-08-06T12:19:02  <dennda> but the docs are poor and py.test is magic
2009-08-06T12:19:18  <ThomasWaldmann> btw, a vague idea that just came to my mind is: how about just recreating the cfg instance?
2009-08-06T12:20:27  <dennda> is that idea equivalent to just doing self.request = init_test_request(given_config) before each testcase runs?
2009-08-06T12:20:36  <ThomasWaldmann> as everything is initialized there, that would be maybe much easier
2009-08-06T12:21:14  <ThomasWaldmann> isn't given_config always the same instance?
2009-08-06T12:21:51  <dennda> I am a bit confused by the term "instance"
2009-08-06T12:22:22  <dennda> you are not calling __init__ anywhere... it's just from MoinMoin._tests import wikiconfig and then wikiconfig.Config is  used
2009-08-06T12:22:23  <ThomasWaldmann> "instance of a class"?
2009-08-06T12:23:12  <ThomasWaldmann> if that's all, that's maybe wrong anyway
2009-08-06T12:23:19  <ThomasWaldmann> look at multiconfig.py
2009-08-06T12:25:02  <ThomasWaldmann> _makeConfig e.g.
2009-08-06T12:31:24  <dennda> hm perhaps only the class is passed as a reference and the instance is created later. at least MoinMoin.web.BaseContext does that
2009-08-06T12:33:25  <dennda> hm
2009-08-06T12:40:33  <ThomasWaldmann>         cfg = configClass(name)
2009-08-06T12:40:42  <ThomasWaldmann> so what do you think is this line doing?
2009-08-06T12:42:43  <dennda> so it is not always the same instance and hence it should work. is that the hypothesis?
2009-08-06T12:46:52  <ThomasWaldmann> see getConfig
2009-08-06T12:47:14  <ThomasWaldmann> (when moin runs normally, it is the same instance)
2009-08-06T12:47:57  <ThomasWaldmann> but if one wants everything fresh, one could create it per test (like cgi creates it per request)
2009-08-06T12:48:40  <dennda> hm where's the entry point where that cfg is created? I'm still searching that
2009-08-06T12:48:55  <dennda> it just says request.given_config = given_config and then a miracle occurrs
2009-08-06T12:49:48  <ThomasWaldmann> web.contexts:128 usually
2009-08-06T12:50:02  <dennda> ah yes
2009-08-06T12:50:06  <dennda> just found that
2009-08-06T12:50:18  <dennda> actually found that before but didn't see how that applies
2009-08-06T12:50:33  <dennda> ok
2009-08-06T12:50:49  <dennda> so caching seems to be a bit of a problem here
2009-08-06T12:50:59  <dennda> or did I get that wrong?
2009-08-06T12:51:39  <ThomasWaldmann> -v
2009-08-06T12:54:15  <dennda> ok what I did now is this: I've thrown out provide_fresh_backends and inject new requests created with init_test_request into each testcase, each with given_config
2009-08-06T12:54:49  <dennda> but still, the second of the test_test_environ test fails
2009-08-06T12:55:10  <dennda> because there's already that Item in the backend, which shouldn't be there
2009-08-06T12:56:00  <dennda> and since backends are created in the config (and I do not have a better explanation) my guess is that the second test is running on a new request object, BUT with the old config and hence the old backend
2009-08-06T12:56:20  <dennda> due to caching
2009-08-06T12:56:23  <dennda> does that sound reasonable?
2009-08-06T12:57:10  <dennda> hm
2009-08-06T12:58:08  <dennda> doesn't seem to be the case though. changed getConfig to always return new configs with _makeConfig and the error persists
2009-08-06T13:03:34  * dennda fires up the debugger
2009-08-06T13:11:57  <ThomasWaldmann> print repr(cfg)
2009-08-06T13:12:34  <ThomasWaldmann> and if you just would read and understand the code I pointed you to, it would be all clear
2009-08-06T13:14:52  <ThomasWaldmann> (at least for the normal moin behaviour, when it is running productively)
2009-08-06T13:17:11  <dennda> ok, let me recheck
2009-08-06T13:21:00  <dennda> yes it doesn't look like that code is even used in the tests
2009-08-06T13:22:14  <dennda> btw: it's different config objects but the same backends %-)
2009-08-06T13:22:22  <dennda> (with my local code)
2009-08-06T13:22:49  <dennda> argh
2009-08-06T13:23:04  <dennda> because that backend is created on the module level and then each instance of the config has that exact backend
2009-08-06T13:26:32  <dennda> anyway, will check that after lunch
2009-08-06T13:27:48  <ThomasWaldmann> hey, finally some insights :)
2009-08-06T13:27:53  <dennda> s/module level/class level/
2009-08-06T13:28:11  <dennda> don't tell me you knew that all the time, I won't believe you :-)
2009-08-06T13:28:12  <dennda> brb
2009-08-06T13:28:22  <ThomasWaldmann> btw, another test I would like to see in that test env test is whether cfg is a class or a instance of a class
2009-08-06T13:38:32  <ThomasWaldmann> dennda: please note I want to see some "added some failing testing env tests" commit BEFORE you fix them in subsequent commits
2009-08-06T13:39:08  <ThomasWaldmann> brb
2009-08-06T13:42:28  <dennda> so that people from the outside think that we are doing proper TDD? :D
2009-08-06T13:59:28  <xorAxAx> in TDD, you often have a no failing tests policy which would be violated in above procedure
2009-08-06T14:00:02  <dennda> then we have a conflict here
2009-08-06T14:00:11  <dennda> because in TDD you write tests before you implement the tested stuff
2009-08-06T14:00:38  <dennda> so both would need to go into the same changeset, which isn't perfect either :-)
2009-08-06T14:03:21  <dennda> ThomasWaldmann: btw, testitems.xml has both syspages flags. i added them
2009-08-06T14:23:30  <dennda> ok so in order to fix the abovely described brokenness I will reimplement enduser stuff and make the test config use that (zwei fliegen mit einer klappe)
2009-08-06T14:23:50  <dennda> at least if my push would finally get through...
2009-08-06T14:30:57  <dennda> still pushing....
2009-08-06T14:35:18  <xorAxAx> you have to push harder
2009-08-06T14:35:41  <xorAxAx> checking in failing tests is really a bad habbit, esp. if there is no changeset that makes them work
2009-08-06T14:35:51  <xorAxAx> thomas does so sometimes
2009-08-06T14:35:53  <dennda> GRARHG! Doesn't come through!
2009-08-06T14:36:27  <dennda> hm why is that a bad habit? if it outlines failure that's there anyways?
2009-08-06T14:36:37  <dennda> so at least you know then that it's broken
2009-08-06T14:36:51  <dennda> (there'll be a fix for these failing testcases though)
2009-08-06T14:37:05  <dennda> hm
2009-08-06T14:37:09  <dennda> the repo seems kinda borked
2009-08-06T14:37:15  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6427:855073eedac0 2.0-storage/MoinMoin/_tests/test_test_environ.py: storage: testenviron: don't use well-known pagename
2009-08-06T14:37:18  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6428:351081d78777 2.0-storage/MoinMoin/_tests/test_test_environ.py: storage: further test_environ tests
2009-08-06T14:37:20  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6429:e451be9f4ca8 2.0-storage/MoinMoin/_tests/test_test_environ.py: storage: test whether request.cfg is instance of the test config
2009-08-06T14:37:21  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6430:97db3b2babf5 2.0-storage/MoinMoin/ (3 files in 3 dirs): storage: get rid of provide_fresh_backends
2009-08-06T14:37:23  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6431:f31e9d52cfa9 2.0-storage/MoinMoin/_tests/test_test_environ.py: storage: fix for test
2009-08-06T14:37:28  <dennda> ^C for teh win
2009-08-06T14:38:46  <xorAxAx> dennda: because tests are useful to see if you broke something
2009-08-06T14:38:51  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6432:44230630c60e 2.0-storage/MoinMoin/conftest.py: storage: get rid of dirties backend. only used by test_PageEditor which is gonna go away anyway
2009-08-06T14:39:03  <xorAxAx> dennda: if they are broken, you might think that you or a changeset did. but in fact they never really worked
2009-08-06T14:39:13  <dennda> xorAxAx: i see
2009-08-06T14:39:53  <dennda> xorAxAx: but if there's some bug in the code that you're unable to fix and that is not yet covered by a testcase, what do you do? have trust that your means of communication with your coworkers are good?
2009-08-06T14:40:26  <xorAxAx> check in a skipped test
2009-08-06T14:40:31  <xorAxAx> that shows the problem
2009-08-06T14:40:34  <dennda> ah
2009-08-06T14:40:44  <dennda> and before you release there should be a policy to check every skipped test
2009-08-06T14:40:45  <dennda> i see
2009-08-06T14:47:48  *** devilsadvocate has joined #moin-dev
2009-08-06T15:16:59  <dimazest> moin
2009-08-06T15:29:49  <dimazest> dreimark: http://paste.pocoo.org/show/132853/ is it how you wnt to change wiki_groups and wiki_dicts?
2009-08-06T15:29:50  *** tpfennig has quit IRC
2009-08-06T15:30:10  <dimazest> dreimark: may be in 2.0 branch there is a better way to do this?
2009-08-06T15:30:45  <dimazest> if not it would be nice to get rid of the code repetiotion
2009-08-06T15:35:37  <dimazest> in the patch import math is missing in wiki_dicts
2009-08-06T15:35:47  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6433:9532d99708f0 2.0-storage/ (3 files in 2 dirs): storage: Throw out get_enduser_backend and introduce create_simple_mapping
2009-08-06T15:43:16  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6434:60b1fec65d8e 2.0-storage/MoinMoin/ (_tests/wikiconfig.py conftest.py): storage: Fresh backends for each testcase, finally.
2009-08-06T15:43:17  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6435:bd447f37bd74 2.0-storage/MoinMoin/security/_tests/test_security.py: storage: Throw out provide_fresh_backend leftover
2009-08-06T15:54:21  <dennda> hm, ok that was faulty
2009-08-06T15:59:32  <dimazest> does search "language:en" returns anything? it seems, that it doesn't work
2009-08-06T16:00:44  *** _peck_ has quit IRC
2009-08-06T16:02:34  *** peck has joined #moin-dev
2009-08-06T16:08:20  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6436:d9753413e8eb 2.0-storage/MoinMoin/_tests/test_test_environ.py: storage: Add test for test environ
2009-08-06T16:21:40  <dennda> hm
2009-08-06T16:21:47  <dennda> I am unsure how to fix that test properly
2009-08-06T16:22:02  <dennda> (fixing that should also make test_security work)
2009-08-06T16:34:35  <dimazest> dreimark: http://hg.moinmo.in/moin/1.9-xapian-dmilajevs/file/2171281b6d79/MoinMoin/search/queryparser.py#l597
2009-08-06T16:36:10  <dimazest> dreimark: isn't there wrong indentation?
2009-08-06T16:44:20  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6437:70b5d5edddeb 2.0-storage/MoinMoin/ (4 files in 3 dirs): storage: Fix for recently added failing test environ test
2009-08-06T16:49:18  <dennda> didn't fix test_security for some reason
2009-08-06T16:54:09  *** franklin_ has joined #moin-dev
2009-08-06T17:09:51  <dennda> ThomasWaldmann: If you still want to take a look at it, go ahead
2009-08-06T17:10:42  *** franklin has quit IRC
2009-08-06T17:15:23  <dennda> about that tagging script: didn't we agree that it is rather pointless to automatically invoke that when migrating from 1.9?
2009-08-06T18:10:07  *** devilsadvocate has quit IRC
2009-08-06T18:12:12  *** devilsadvocate has joined #moin-dev
2009-08-06T18:17:11  <dreimark> dimazest: test_sourcecode tells for me moin-1.9-xapian-dmilajevs/MoinMoin/search/builtin.py:650:52: E231 missing whitespace after ',' 'revision': r.data.get('revision',[0])[0]}
2009-08-06T18:19:57  <CIA-38> Dmitrijs Milajevs <dimazest@gmail.com> default * 4936:0b5aaec043cd 1.9-xapian-dmilajevs/MoinMoin/search/_tests/test_search.py: Xapian2009: Tests for mimetype, language and domain. Test classes are new style classes.
2009-08-06T18:19:59  <CIA-38> Dmitrijs Milajevs <dimazest@gmail.com> default * 4937:d2cda24a5475 1.9-xapian-dmilajevs/MoinMoin/search/ (Xapian.py queryparser.py):
2009-08-06T18:19:59  <CIA-38> Xapian2009: MoinSearchConnection class was introduced which inherits from
2009-08-06T18:19:59  <CIA-38> xappy.SearchConnection and provides get_all_documents method. Various cahnges to
2009-08-06T18:20:00  <CIA-38> the queryparser.py now tests fail on asserts, but not in the errors in the code.
2009-08-06T18:20:02  <CIA-38> Dmitrijs Milajevs <dimazest@gmail.com> default * 4938:d43af366f10a 1.9-xapian-dmilajevs/MoinMoin/support/xapwrap/ (__init__.py document.py index.py): Xapian2009: xapwrap was deleted from MoinMoin/support.
2009-08-06T18:20:07  <CIA-38> Dmitrijs Milajevs <dimazest@gmail.com> default * 4939:f1d1d8105d52 1.9-xapian-dmilajevs/MoinMoin/search/ (6 files in 2 dirs): Xapian2009: pep8 fixes.
2009-08-06T18:20:26  <dimazest> dreimark: it is fixed
2009-08-06T18:20:48  <dreimark> was there an indenting problem, I haven't seen it
2009-08-06T18:22:55  <dimazest> ok, this else is for loop
2009-08-06T18:24:20  <dreimark> ah
2009-08-06T18:24:41  <dreimark> can you push the 1.9 fix for wiki_groups too
2009-08-06T18:25:23  <dimazest> yes, but is there a better way for 2.0 for caching?
2009-08-06T18:38:17  <CIA-38> Dmitrijs Milajevs <dimazest@gmail.com> default * 4927:27d5d0b875a9 1.9/MoinMoin/datastruct/backends/ (wiki_dicts.py wiki_groups.py): Groups2009: wiki_groups and wiki_dicts caching fix.
2009-08-06T18:51:38  <CIA-38> Dmitrijs Milajevs <dimazest@gmail.com> default * 4940:b78a768f142b 1.9-xapian-dmilajevs/MoinMoin/search/ (Xapian.py builtin.py): Get rid of unnecessary head
2009-08-06T18:51:40  <CIA-38> Dmitrijs Milajevs <dimazest@gmail.com> default * 4944:ad551d80e556 1.9-xapian-dmilajevs/MoinMoin/ (36 files in 4 dirs): Automated merge with http://hg.moinmo.in/moin/1.9
2009-08-06T19:11:57  <dennda> how do you properly invoke a maintenance script that inherits from MoinScript?
2009-08-06T19:25:43  <dreimark> dimazest: currently not. 1.9 will be merged in 2.0
2009-08-06T19:25:50  <dreimark> bbl
2009-08-06T19:27:59  <dennda> thomas once suggested that storage could be used for caching
2009-08-06T19:35:12  <dennda> dreimark: do you know that by chance?
2009-08-06T19:35:19  <dennda> ah now he's gone as well :-)
2009-08-06T19:37:55  <dennda> hm MoinMoin/script/moin.py I guess
2009-08-06T19:44:03  *** grzywacz has joined #moin-dev
2009-08-06T20:26:58  <ThomasWaldmann> re
2009-08-06T20:27:37  *** amartani has joined #moin-dev
2009-08-06T20:30:36  <dennda> wb :-)
2009-08-06T20:36:42  <ThomasWaldmann> after a bunch of blue screens at other windows crazyness, I need to relax a bit with fine python code now. :D
2009-08-06T20:37:09  <dennda> hehe
2009-08-06T20:39:18  <dennda> ThomasWaldmann: In case you havn't noticed, there's some user seeking your advise on #moin -- I pointed out a few things already but you need to make sure I was right
2009-08-06T21:01:26  <dreimark> dennda: I hope we get a storage backend for caching
2009-08-06T21:01:30  <ThomasWaldmann>      1.7 -        timeout = 3.0 # minimal timout
2009-08-06T21:01:30  <ThomasWaldmann>      1.8 +        timeout = 3.0 # minimal timout is 2.0
2009-08-06T21:01:40  <ThomasWaldmann> dennda: you're kidding?
2009-08-06T21:01:54  <dreimark> that solves a lot of these timing problems
2009-08-06T21:02:06  <dreimark> ThomasWaldmann: why is there a typo?
2009-08-06T21:02:35  <ThomasWaldmann> i mean creating 2 changesets for that line, without fixing an obvious typo?
2009-08-06T21:04:25  <dennda> lol
2009-08-06T21:05:00  <dennda> my usual procedure is hg status, hg diff, hg commit, hg out, hg push
2009-08-06T21:05:14  <dennda> and I always thought I was quite talented at spotting typos
2009-08-06T21:05:40  <dennda> I should think again whether that still applies :P
2009-08-06T21:07:49  * dreimark had watched today a presentation of a java cms
2009-08-06T21:07:52  <dennda> i will fix that silently in some other changeset in order not to create to much noise
2009-08-06T21:09:52  <ThomasWaldmann>     1.39 +            assert rev[SYSPAGE_VERSION] == 1
2009-08-06T21:10:13  <ThomasWaldmann> i maybe rather not check that
2009-08-06T21:10:18  <dennda> ok
2009-08-06T21:10:27  <dreimark> dimazest: so xapwrap is gone now or is there something left
2009-08-06T21:10:34  <ThomasWaldmann> because the value will change and is expected only for release stuff
2009-08-06T21:12:11  <ThomasWaldmann>      1.7 +        assert isinstance(self.request.cfg, wikiconfig.Config)
2009-08-06T21:12:22  <ThomasWaldmann> how is that related to fresh BACKENDS?
2009-08-06T21:12:37  <dennda> 13:28 < ThomasWaldmann> btw, another test I would like to see in that test env test is whether cfg is a class or a instance of a class
2009-08-06T21:13:06  <dennda> i guess you want it in a seperate test method
2009-08-06T21:13:33  <dennda> will do that. am currently fixing the migrate thing
2009-08-06T21:14:36  <dreimark> ThomasWaldmann: do you've seen that xml issue by oliver?
2009-08-06T21:15:23  * ThomasWaldmann reads heise, too
2009-08-06T21:15:50  <ThomasWaldmann> and slashdot, and planet.python.org and blog.fefe.de and lots of other stuff :)
2009-08-06T21:16:14  <dennda> BTW: Technically I don't think we need to rip out revert and delete acl permissions as they will be ignored anyway (they are not in acl_rights:valid). Do we want to do it nevertheless? (That'd require some clone-hack.)
2009-08-06T21:16:32  <dennda> fefe once burped into my face
2009-08-06T21:16:34  <dennda> story of my life
2009-08-06T21:17:06  <dennda> I don't exactly like him anymore now :-)
2009-08-06T21:18:30  <ThomasWaldmann> hehe, you are famous now, too :))
2009-08-06T21:19:01  <dennda> yah I felt the spirit kinda free itself
2009-08-06T21:19:07  <dennda> (literally)
2009-08-06T21:21:35  <dreimark> dennda: burped means (don't find it using leo)?
2009-08-06T21:22:39  <ThomasWaldmann> storage: get rid of dirties backend. only used by test_PageEditor which is gonna go away anyway
2009-08-06T21:22:52  <dreimark> dimazest: I have only 5 failing tests
2009-08-06T21:22:56  <ThomasWaldmann> well, you could've removed it there, to, nevertheless
2009-08-06T21:23:07  <dennda> dreimark: ruelpsen
2009-08-06T21:23:13  <dreimark> ah
2009-08-06T21:32:20  <dimazest> dreimark: yes, today i changed code
2009-08-06T21:32:59  <dimazest> but these 5 failing tests need more attention, because xapian returns nothing
2009-08-06T21:34:05  * dreimark looks too
2009-08-06T21:39:43  <dreimark> dimazest: I see tests finished: 839 passed, 5 failed, 72 skipped in 132.38 seconds  but only MoinMoin/search/_tests/test_search.py[19] ............FF.F...
2009-08-06T21:39:56  <dreimark> may be I am blind but where are the 2 others?
2009-08-06T21:43:22  * ThomasWaldmann runs tests, too
2009-08-06T21:48:39  <ThomasWaldmann> dennda:         should_be_there = ("FrontPage", "HelpOnLinking", "HelpOnMoinWikiSyntax", )
2009-08-06T21:49:00  <ThomasWaldmann> that should be rather wikiconfig.Config._test_items_names or so
2009-08-06T21:49:35  <dennda> k
2009-08-06T21:52:42  <CIA-38> Alexandre Martani <amartani AT gmail DOT com> default * 6546:4ec2791d65b1 2.0-storage-editor-amartani/MoinMoin/ (action/mobwrite.py items/__init__.py items/lock.py): pep8 fix
2009-08-06T21:54:46  <dimazest> dreimark: i have more tests: MoinMoin/search/_tests/test_search.py ...................F........F....F...F..FFF.FF.F...
2009-08-06T21:57:18  <dreimark> hmm
2009-08-06T21:57:55  <dreimark> dimazest: in queryparser.pagefilter is builtin filter function overwritten
2009-08-06T21:58:42  <dennda> ThomasWaldmann: http://paste.pocoo.org/show/132924/
2009-08-06T22:00:04  <dimazest> dreimark: i do not know, i think i have not touch it
2009-08-06T22:00:18  <ThomasWaldmann> cool, test_environ works :)
2009-08-06T22:00:25  <dennda> sure it does
2009-08-06T22:00:32  <dennda> :D
2009-08-06T22:01:40  <dennda> just found out that we already had a migration script %-)
2009-08-06T22:01:54  <dreimark> ocument it
2009-08-06T22:01:57  <dreimark> +d
2009-08-06T22:02:15  <dimazest> dreimark: i need to do my German, so i won't answer immediately
2009-08-06T22:02:22  <dimazest> dreimark: the filter function?
2009-08-06T22:02:44  <dreimark> dimazest: I think the filter function should be renamed
2009-08-06T22:02:59  <dreimark> I meant with the documenting dennda
2009-08-06T22:03:46  <dennda> I'm fixing it
2009-08-06T22:04:00  <dennda> will be piece of cake to convert an old wiki
2009-08-06T22:05:41  <dreimark> hehe you could offer a cradle to the 1.3.4 one on #moin
2009-08-06T22:05:49  <ThomasWaldmann> dennda: maybe problem with history()?
2009-08-06T22:07:05  <dennda> havn't checked that any further
2009-08-06T22:10:13  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6438:b29e49ff5700 2.0-storage/MoinMoin/ (script/migration/backend.py util/_tests/test_lock.py): storage: Update and simplify migration script.
2009-08-06T22:10:14  *** amartani has quit IRC
2009-08-06T22:10:15  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6439:6ae33567eccf 2.0-storage/migrate_old_wiki_data.py: storage: get rid of migrate_old_wiki_data.py script
2009-08-06T22:10:16  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6440:52d70f7d40e1 2.0-storage/MoinMoin/storage/__init__.py: storage: add missing imports
2009-08-06T22:10:34  <dennda> Assuming a proper config, migration is now as easy as:
2009-08-06T22:10:50  <dennda> python MoinMoin/script/moin.py --config-dir=. migration backend
2009-08-06T22:11:02  <dennda> or moin --config-dir=. migration backend
2009-08-06T22:12:51  <dennda> ripping out revert/delete can be done by some hack in the migration script, IF we want that
2009-08-06T22:15:01  <ThomasWaldmann> would be cleaner.
2009-08-06T22:15:40  <dennda> ok
2009-08-06T22:17:41  <dennda> that's gonna be ugly though :-)
2009-08-06T22:22:06  <dennda> ThomasWaldmann: would it be ok if clone had some enable_migration_hack parameter set to False by default that removes revert and delete from acls? thats ugly as hell but I think the smoothest solution
2009-08-06T22:24:34  <CIA-38> Alexandre Martani <amartani AT gmail DOT com> default * 6547:2204ff8f9b73 2.0-storage-editor-amartani/MoinMoin/ (4 files in 2 dirs): Real-time editor: update mobwrite to r75
2009-08-06T22:25:09  <ThomasWaldmann> why should it have that param?
2009-08-06T22:25:48  <dennda> because then you can remove revert and delete during conversion and don't need to touch anything again
2009-08-06T22:27:05  *** amartani has joined #moin-dev
2009-08-06T22:31:04  <dennda> ThomasWaldmann: put differently, how would you remove revert and delete?
2009-08-06T22:31:41  <ThomasWaldmann> i rather meant "why should it be false" and not ever true
2009-08-06T22:32:24  <dennda> well if it was always true we could just as well always remove revert/delete from metadata
2009-08-06T22:32:34  <dennda> when copying an item
2009-08-06T22:32:38  <ThomasWaldmann> yeah, and?
2009-08-06T22:32:48  <dennda> so we do that instead?
2009-08-06T22:33:07  <dennda> ok
2009-08-06T22:33:14  <ThomasWaldmann> ehrm
2009-08-06T22:33:36  <ThomasWaldmann> i think this is rather something for the fs19 backend
2009-08-06T22:33:46  <dennda> true
2009-08-06T22:33:49  <dennda> stupid
2009-08-06T22:38:36  <xorAxAx> amartani: could you reproduce the sequence that triggered data loss for me?
2009-08-06T22:41:53  <amartani> xorAxAx: haven't tried it, I'll bring the merge code from 1.9 and look at these cases
2009-08-06T22:45:44  <amartani> but it should only lose data now if you delete something on edits, as it does a last sync before saving
2009-08-06T22:51:36  <dreimark> dennda: in 1.9 tests are influenced by the wikiconfig_local.py. Is that gone in 2.0?
2009-08-06T22:52:11  <dennda> dreimark: what does "influenced" mean?
2009-08-06T22:52:16  <dennda> in what way?
2009-08-06T22:52:33  <dreimark> it loads parameters from that file and changesd the configuration
2009-08-06T22:53:08  <dennda> that shouldn't happen in 2.0 as 2.0 uses an own config for tests
2009-08-06T22:54:04  <dreimark> 1.9 has also an own config but it seems to load more
2009-08-06T22:54:08  *** tpfennig has joined #moin-dev
2009-08-06T22:54:18  <dreimark> anyway I am tired
2009-08-06T22:54:20  <dreimark> gn
2009-08-06T22:54:22  <dennda> gn
2009-08-06T23:23:21  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6441:8ea55f047e26 2.0-storage/MoinMoin/storage/backends/fs19.py: storage: When converting from 1.x to 2.0, don't copy the 'revert' and 'deleted' metadata values.
2009-08-06T23:25:29  <ThomasWaldmann> dennda: wrong
2009-08-06T23:25:43  <dennda> -v
2009-08-06T23:26:34  <ThomasWaldmann> JoePrevert:read
2009-08-06T23:27:24  <dennda> X)
2009-08-06T23:27:39  <dennda> ok
2009-08-06T23:30:21  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6442:a6df0af80413 2.0-storage/MoinMoin/storage/backends/fs19.py: storage: fix for last changeset. Make JoePrevert happy.
2009-08-06T23:31:18  <dennda> unless you know people with colons or commas in their surname, this should work now :-)
2009-08-06T23:48:57  * ThomasWaldmann fixed the security tests
2009-08-06T23:49:03  <dennda> great
2009-08-06T23:49:10  <dennda> i was just about to run them again
2009-08-06T23:49:23  <ThomasWaldmann> the groups/dicts stuff is strange
2009-08-06T23:50:43  <dennda> $ py.tst
2009-08-06T23:50:43  <dennda> No command 'py.tst' found, did you mean: Command 'py.test' from package 'python-codespeak-lib' (universe)
2009-08-06T23:50:49  <dennda> Oo
2009-08-06T23:51:11  <dennda> it's getting more and more intelligent
2009-08-06T23:54:37  <ThomasWaldmann> ah, found something
2009-08-06T23:55:05  <ThomasWaldmann> run winword :P
2009-08-06T23:55:17  <dennda> no thanks
2009-08-06T23:55:33  <dennda> mir ist schon schlecht