MoinMoin: MoinMoinChat/Logs/moin-dev/2009-06-05

2009-06-05T00:08:23  *** adg` has joined #moin-dev
2009-06-05T00:25:43  *** adg has quit IRC
2009-06-05T00:32:20  <dennda> hierarchic is missing
2009-06-05T00:37:32  <dennda> or at least I think that's where it's failing
2009-06-05T00:57:12  <dennda> actually... hmmm
2009-06-05T01:08:13  <dennda> think i got it
2009-06-05T01:11:12  <ThomasWaldmann> dennda: there's also a history() todo still
2009-06-05T01:12:07  <dennda> down to 1 failing test now
2009-06-05T01:12:10  <dennda> (security test)
2009-06-05T01:13:46  <ThomasWaldmann> cool :)
2009-06-05T01:14:31  <dennda> guess whose fault it was again
2009-06-05T01:14:32  <ThomasWaldmann> "Kill-Bill" Bill is dead.
2009-06-05T01:14:44  <dennda> oh
2009-06-05T01:14:45  <dennda> sad
2009-06-05T01:14:53  <dennda> recently watched those films again
2009-06-05T01:20:00  <ThomasWaldmann> """        In order to store an ACL on a page you must have the ADMIN privilege.
2009-06-05T01:20:03  <ThomasWaldmann>         We must allow storing the preceeding revision's ACL in the new revision
2009-06-05T01:20:08  <ThomasWaldmann>         (i.e., keeping it), though."""
2009-06-05T01:20:17  <ThomasWaldmann> admin means being able to CHANGE acls.
2009-06-05T01:20:52  <ThomasWaldmann> (that can also be to remove an acl, no just to store one)
2009-06-05T01:23:06  <ThomasWaldmann>             acl_changed = not (value == last_acl)
2009-06-05T01:23:13  <ThomasWaldmann> please simplify
2009-06-05T01:24:05  <ThomasWaldmann> please add a \n at EOF
2009-06-05T01:24:34  <ThomasWaldmann> please add class docstrings
2009-06-05T01:24:58  <ThomasWaldmann> please add the usual logging import
2009-06-05T01:25:10  <dennda> usual logging import?
2009-06-05T01:26:26  <ThomasWaldmann> see many other files
2009-06-05T01:28:53  <dennda> i summarized the rest as "documentation" on the wiki page :P
2009-06-05T01:29:03  <dennda> thanks for outlining
2009-06-05T01:29:14  <dennda> I'm gonna get some sleep, good night
2009-06-05T01:30:45  <ThomasWaldmann> gn
2009-06-05T05:58:02  <CIA-19> Alexandre Martani <amartani AT gmail DOT com> default * 5893:c4c4707818d1 2.0-storage-editor-amartani/MoinMoin/action/ (_tests/test_mobwrite.py mobwrite.py): Real-time editor: Add methods to register who is editing some page
2009-06-05T05:58:04  <CIA-19> Alexandre Martani <amartani AT gmail DOT com> default * 5894:0678af9c4d85 2.0-storage-editor-amartani/MoinMoin/web/static/htdocs/applets/Mobwrite/compressed_form.js: Real-time editor: License on mobwrite compressed form script
2009-06-05T05:58:06  <CIA-19> Alexandre Martani <amartani AT gmail DOT com> default * 5895:b81bd7e12b74 2.0-storage-editor-amartani/MoinMoin/action/ (_tests/test_mobwrite.py mobwrite.py): pep8 fix
2009-06-05T06:03:10  *** amartani has quit IRC
2009-06-05T08:57:43  <ThomasWaldmann> moin
2009-06-05T11:44:22  *** TheSheep has quit IRC
2009-06-05T11:44:25  *** TheSheep has joined #moin-dev
2009-06-05T13:01:54  <ThomasWaldmann> devilsadvocate_: another week has gone by...
2009-06-05T13:15:26  <dennda> turns out the tests failure is weirder than I thought
2009-06-05T13:27:58  <dennda> is 'JoeDoe: JaneDoe:read,write' a valid acl string?
2009-06-05T13:29:03  *** franklin has quit IRC
2009-06-05T13:30:48  *** franklin has joined #moin-dev
2009-06-05T13:37:48  * ThomasWaldmann tries virtualbox 2.2.4 on jaunty
2009-06-05T13:45:03  *** franklin has quit IRC
2009-06-05T13:45:55  *** franklin has joined #moin-dev
2009-06-05T13:53:26  *** franklin has quit IRC
2009-06-05T13:55:20  *** franklin has joined #moin-dev
2009-06-05T15:42:37  <devilsadvocate_> moin
2009-06-05T15:54:56  *** devilsadvocate_ is now known as devilsadvocate-a
2009-06-05T15:55:03  *** devilsadvocate-a is now known as dadvocate-afk
2009-06-05T16:20:38  <dennda> ThomasWaldmann: can you do me a big favor and help to fix the security tests? i've been sitting in front of that now for several hours and I think I'm missing something
2009-06-05T16:21:34  <CIA-19> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5886:1ac22538337b 2.0-storage/MoinMoin/ (security/_tests/test_security.py user.py): storage: Provide May class in order to rewrite current user.may.right calls to data_backend._may. Just in order to clean things up for now.
2009-06-05T16:21:36  <CIA-19> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5887:cf8b6c0e9874 2.0-storage/MoinMoin/ (_tests/__init__.py security/_tests/test_security.py): storage: security tests stored acls the old way in the content. That's why it broke, they just weren't applied. Fixed create_page for that. Down to 1 failing security test
2009-06-05T16:21:39  <CIA-19> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5888:40a6536e8bf1 2.0-storage/MoinMoin/storage/backends/acl.py: storage: use request.user.name directly
2009-06-05T16:21:41  <CIA-19> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5889:3f3ecd1ce600 2.0-storage/MoinMoin/ (security/_tests/test_security.py user.py): storage: fixing security tests is weird. outlined the actual problem by an additional assert
2009-06-05T16:44:16  <dennda> ThomasWaldmann: you just ripped the old _may out, right? it should still work, right?
2009-06-05T16:46:41  * dennda is too frustrated for now... :-)
2009-06-05T17:01:14  <dennda> There's some major name fuckup
2009-06-05T17:01:52  <dennda> http://paste.pocoo.org/show/121179/
2009-06-05T17:23:58  <dennda> hmmmm
2009-06-05T17:33:02  <dennda> I think I can say that this is wrong: http://hg.moinmo.in/moin/2.0-storage/file/3f3ecd1ce600/MoinMoin/storage/backends/acl.py#l91
2009-06-05T17:33:11  *** dadvocate-afk is now known as devilsadvocate
2009-06-05T17:33:19  <dennda> especially when considering the comment
2009-06-05T17:37:15  <dennda> ThomasWaldmann: you wrote that; can you take a look at it? too complex for me now
2009-06-05T17:44:55  *** grzywacz has joined #moin-dev
2009-06-05T18:19:08  <ThomasWaldmann> dennda: that assert is nonsense
2009-06-05T18:19:57  <ThomasWaldmann> because by simply creating a User object, you just have that: a user object
2009-06-05T18:20:15  <ThomasWaldmann> that doesn't change the state of the request object or of the backend
2009-06-05T18:24:45  <ThomasWaldmann> and that line 91 is the thing I pointed out some days ago: it has to be checked. until otherwise proven, it is assumed to be correct.
2009-06-05T18:35:38  <ThomasWaldmann> dennda: why are you providing a "May" class that does approx. what "SecurityPolicy" did until now?
2009-06-05T18:59:08  <devilsadvocate> ThomasWaldmann: quick pm?
2009-06-05T18:59:52  <ThomasWaldmann> ?
2009-06-05T19:32:18  <dreimark> bbl
2009-06-05T20:05:21  <CIA-19> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5890:b1734cb4bea2 2.0-storage/MoinMoin/ (_tests/__init__.py security/_tests/test_security.py): security tests: fixed tests (still failing -> real bug)
2009-06-05T20:53:07  <CIA-19> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5891:f01219566d80 2.0-storage/MoinMoin/ (security/__init__.py storage/backends/acl.py): make acl code easier to understand
2009-06-05T20:58:34  <dennda> I could *swear* that I tried that several times myself
2009-06-05T21:02:44  <dennda> I mean, it's obvious that that's necessary
2009-06-05T21:03:09  <dennda> (self.request.user = u)
2009-06-05T21:09:19  <CIA-19> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5892:1a1e0be71a3b 2.0-storage/MoinMoin/security/_tests/test_security.py:
2009-06-05T21:09:19  <CIA-19> only cosmetic changes (see below)
2009-06-05T21:09:19  <CIA-19> removed unused imports
2009-06-05T21:09:19  <CIA-19> s/page/item/g
2009-06-05T21:09:19  <CIA-19> s/Page/Item/g
2009-06-05T21:09:21  <CIA-19> s/iter/acl_iter/g (iter is a builtin)
2009-06-05T21:09:51  <ThomasWaldmann> dennda: you can help finding the real bugs now
2009-06-05T21:10:11  <ThomasWaldmann> hierarchic ACL's acl inheritance does not work
2009-06-05T21:10:23  <ThomasWaldmann> this is why those 3 tests fail
2009-06-05T21:10:43  <ThomasWaldmann> it uses the default ACL, not the inherited ACL
2009-06-05T21:12:04  <dennda> give me some minutes to catch up; just got home; eating now
2009-06-05T21:26:37  <ThomasWaldmann> gotcha
2009-06-05T21:28:43  <dennda> actually I fail to see where the allowing ACLs for those three cases should be
2009-06-05T21:28:49  <dennda> e.g.
2009-06-05T21:28:55  <dennda> >       AssertionError: u'JoeDoe' may not write u'AclTestMainItem/SubItem' (hierarchic)
2009-06-05T21:29:24  <dennda> for hierarchic = True:             (True,  self.subitem_name, u'JoeDoe', []), # by inherited acl from main item
2009-06-05T21:29:43  <dennda> ok, so let's look at the main item
2009-06-05T21:29:51  <dennda>         (mainitem_name, u'JoeDoe: JaneDoe:read,write', u'Foo!'),
2009-06-05T21:29:58  <dennda> no access for JoeDoe at all
2009-06-05T21:30:02  <dennda> am I misreading this?
2009-06-05T21:31:35  <dennda> hm subitem has None as acl, so default acl
2009-06-05T21:31:53  <dennda> which is All:read,write
2009-06-05T21:32:08  <dennda> is that it?
2009-06-05T21:34:45  <dennda> umm
2009-06-05T21:35:11  <dennda> disregard that
2009-06-05T21:35:44  <dennda> it first checks subitem: no acl. check mainitem: permission denied. no default acl checked. misread it...
2009-06-05T21:36:18  <dennda> ok, got the problem. sounds like you already have a fix
2009-06-05T21:40:52  <ThomasWaldmann> yes
2009-06-05T21:41:00  <ThomasWaldmann> just a minute...
2009-06-05T21:41:21  <dennda> hm
2009-06-05T21:41:35  <dennda> it jumped into the not-hierarchic else branch here
2009-06-05T21:43:48  <ThomasWaldmann> the acl wrapper remembers the acl hierarchic state and ignores us changing the cfg
2009-06-05T21:44:19  <dennda> ah
2009-06-05T21:44:58  <dennda> is that actually a problem you'd encounter in the real-world? (i.e., not in those made-up tests)
2009-06-05T21:45:10  <ThomasWaldmann> maybe not
2009-06-05T21:45:18  <ThomasWaldmann> but it was easy to fix
2009-06-05T21:45:32  <ThomasWaldmann> maybe have a look now at May vs. SecurityPolicy
2009-06-05T21:45:42  <ThomasWaldmann> that's duplicated more or less
2009-06-05T21:45:55  <dennda> ok
2009-06-05T21:47:03  <ThomasWaldmann> and btw, some other code derives from SecPol
2009-06-05T21:47:33  <ThomasWaldmann> (derived? I killed some of them)
2009-06-05T21:49:13  <dennda> telling from grep's output it's "derived"
2009-06-05T21:49:55  <CIA-19> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5893:6424a0a85b2a 2.0-storage/MoinMoin/storage/backends/acl.py: storage: acl wrapper middleware: do not store cfg attr state inside the wrapper, the tests modify cfg and expect changed behaviour
2009-06-05T21:50:22  <ThomasWaldmann> it still stores the backend, btw
2009-06-05T21:53:15  <ThomasWaldmann> and after you fix that, please remove the duplicate code and the unused code
2009-06-05T21:56:24  *** amartani has joined #moin-dev
2009-06-05T21:57:30  <ThomasWaldmann> dennda: btw, do you also see these failures in the hg backend?
2009-06-05T22:01:30  <dennda> ThomasWaldmann: 4 fails for hg
2009-06-05T22:09:26  <ThomasWaldmann> PawelPac1na: in case you find some time, could you have a look? moin/2.0-storage, 4 fails for hg backend.
2009-06-05T22:16:28  <dennda> thanks for the fixes
2009-06-05T22:16:47  <dennda> don't know what I did wrong
2009-06-05T22:21:42  *** amartani has quit IRC
2009-06-05T22:23:53  <dennda> ThomasWaldmann: what problem do you see with storing the backend like that?
2009-06-05T22:24:32  <dennda> i doubt there's a test that wants to swap the backend
2009-06-05T22:25:07  *** amartani has joined #moin-dev
2009-06-05T22:29:06  <ThomasWaldmann> not yet :)
2009-06-05T22:54:49  <CIA-19> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5894:01ec0110f44f 2.0-storage/MoinMoin/ (security/__init__.py user.py): storage: Throw out redundant May class. Use Permission instead. Make Permission use AMW._may and throw out redundant acl _check function
2009-06-05T22:54:51  <CIA-19> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5895:ec429a714d48 2.0-storage/MoinMoin/storage/backends/acl.py: storage: small fix if there's been no acl on an item yet
2009-06-05T23:09:15  <CIA-19> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5896:3c49756356f1 2.0-storage/MoinMoin/storage/backends/flatfile.py: storage: clean up flatfile backend (small fixes, cosmetic changes)
2009-06-05T23:15:45  <CIA-19> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5897:c235fe5db895 2.0-storage/MoinMoin/storage/backends/flatfile.py: storage: flatfile backend: use wikiutil for (un)quoting names