2009-06-08T00:40:02 *** grzywacz has quit IRC
2009-06-08T00:48:47 <amartani> moin
2009-06-08T00:51:51 <ThomasWaldmann> hi amartani
2009-06-08T01:22:54 <amartani> ThomasWaldmann: should the editor have a button for saving without getting out of the page?
2009-06-08T01:23:47 <ThomasWaldmann> that depends on your save-strategy, maybe yes
2009-06-08T01:27:11 <ThomasWaldmann> btw, you should make more use of your wiki page, your work time seems to be anti-cyclic to your mentor's
2009-06-08T01:35:26 <amartani> sorry, I'm not very good at writing, esp in english.
2009-06-08T01:41:12 <amartani> about the editor, moinmoin doesn't have any ajax interface for saving items, right?
2009-06-08T01:47:22 <ThomasWaldmann> do you need something else than POSTing the data?
2009-06-08T01:47:53 *** ernestas has quit IRC
2009-06-08T01:50:47 <amartani> a confirmation message maybe..
2009-06-08T03:32:50 *** rkerr has quit IRC
2009-06-08T03:57:04 *** rkerr has joined #moin-dev
2009-06-08T05:26:24 *** amartani has quit IRC
2009-06-08T05:44:57 *** amartani has joined #moin-dev
2009-06-08T05:46:01 <CIA-19> Alexandre Martani <amartani AT gmail DOT com> default * 5925:820f52abd0f1 2.0-storage-editor-amartani/MoinMoin/ (3 files in 3 dirs): Real-time editor: remove editor from list when quit editing page
2009-06-08T05:46:04 <CIA-19> Alexandre Martani <amartani AT gmail DOT com> default * 5926:b222c2c093a6 2.0-storage-editor-amartani/MoinMoin/web/static/htdocs/applets/Mobwrite/mobwrite_moinmoin.js: Real-time editor: Rename javascript functions to mobwrite convention. Retrieve list of editors periodically.
2009-06-08T08:41:02 <ThomasWaldmann> moin
2009-06-08T09:50:18 <dreimark> moin
2009-06-08T10:09:39 <dimazest> moin
2009-06-08T10:23:39 <dimazest> i move wikidicts.py to dicts/__init__.py
2009-06-08T10:24:16 <dimazest> and _tests/test_wikidicts.py to dicts/_tests/test_dicts.py
2009-06-08T10:24:27 <ThomasWaldmann> or just rewrite it there cleanly and kill it at the old place
2009-06-08T10:24:53 <ThomasWaldmann> did you already do the changes we talked about to the group code?
2009-06-08T10:25:01 <dimazest> no
2009-06-08T10:25:05 <ThomasWaldmann> maybe it is less work doing that first
2009-06-08T10:25:11 <dimazest> that's my plan for today
2009-06-08T10:28:53 <CIA-19> Reimar Bauer <rb.proj AT googlemail DOT com> default * 5910:9d0eea117dfc 2.0-storage/MoinMoin/action/ (PackageItems.py PackagePages.py): PackagePages renamed to PackageItems
2009-06-08T10:28:57 <CIA-19> Reimar Bauer <rb.proj AT googlemail DOT com> default * 5911:96781382f06e 2.0-storage/MoinMoin/action/PackageItems.py: PackageItems: refactored to store items instead of pages or attachments
2009-06-08T10:28:59 <CIA-19> Reimar Bauer <rb.proj AT googlemail DOT com> default * 5912:b1278256ce79 2.0-storage/MoinMoin/ (5 files in 4 dirs): packages: refactoring for items and templates
2009-06-08T10:29:00 <CIA-19> Reimar Bauer <rb.proj AT googlemail DOT com> default * 5913:e418abf4de5a 2.0-storage/MoinMoin/action/PackageItems.py: PackageItems: broken include_subitems feature removed
2009-06-08T10:29:02 <CIA-19> Reimar Bauer <rb.proj AT googlemail DOT com> default * 5914:bb58bd43fe2e 2.0-storage/MoinMoin/action/PackageItems.py: PackageItems: small problem with wrong form name fixed also exception handler for not existing item added
2009-06-08T10:29:04 <CIA-19> Reimar Bauer <rb.proj AT googlemail DOT com> default * 5915:678e44ab82f7 2.0-storage/MoinMoin/ (17 files in 9 dirs): branch merge
2009-06-08T10:29:40 <dreimark> dimazest: please do the group changes first. And I think rewriting of the dicts code is easier
2009-06-08T10:30:07 <dimazest> o
2009-06-08T10:30:09 <dimazest> ok
2009-06-08T10:30:19 * dimazest doing group changes
2009-06-08T10:32:42 <dreimark> ThomasWaldmann: tests on 2.0-s don't work for me.
2009-06-08T10:34:15 <dreimark> bbl
2009-06-08T10:55:23 <dimazest> ThomasWaldmann: is caching done right? http://paste.pocoo.org/show/121737/
2009-06-08T11:15:17 <dreimark> self. member_groups ???
2009-06-08T11:15:58 <ThomasWaldmann> dreimark: some tests and also quite some other code on 2.0 is still broken
2009-06-08T11:16:05 <dreimark> I have a blank in the pastebin after the .
2009-06-08T11:16:33 <dreimark> ThomasWaldmann: I have not any working test I just wonder if that is the current normal state
2009-06-08T11:16:41 <ThomasWaldmann> some of the test failures are related to the memory backend (there are NO pages in there, so looking for them lets the test fail)
2009-06-08T11:17:08 <dreimark> ok
2009-06-08T11:17:12 <ThomasWaldmann> some tests work, about 30(?) failures currently
2009-06-08T11:17:47 <dreimark> hmm, if I run py.test it looks like every test is broken
2009-06-08T11:18:17 <ThomasWaldmann> dimazest: move line1/2 to 6/7
2009-06-08T11:19:38 <ThomasWaldmann> dimazest: you didn't change the caching code afaics
2009-06-08T11:21:09 <ThomasWaldmann> dimazest: the naming could stil improve
2009-06-08T11:22:34 <ThomasWaldmann> maybe members -> members_final, groups -> members_groups
2009-06-08T11:25:06 <ThomasWaldmann> dimazest: btw, in the end (== when it goes into 2.0-storage) I guess we will replace the mtime-based update check by a hash(page content)-based update check
2009-06-08T11:26:25 <ThomasWaldmann> for 1.9, just keep on of the old update check mechanisms, they work in most cases
2009-06-08T11:31:39 <ThomasWaldmann> anyway, use some consistent naming, but avoid calling the direct/final members "users" (it can be anything, including bananas :)
2009-06-08T11:57:44 <xorAxAx> devilsadvocate_: do you still have school these days?
2009-06-08T12:00:41 <xorAxAx> devilsadvocate_: whats your next step?
2009-06-08T16:02:54 *** Rilt has joined #moin-dev
2009-06-08T16:47:03 *** rkerr has quit IRC
2009-06-08T17:00:08 <xorAxAx> woah, socghop app dropped my account
2009-06-08T17:04:53 <ThomasWaldmann> xorAxAx: you are still shown as org admin
2009-06-08T17:07:03 <xorAxAx> now it worked
2009-06-08T17:08:51 <xorAxAx> amartani, devilsadvocate_, dennda, dimazest, dreimark, ThomasWaldmann, waldi: please participate in http://doodle.com/ftbyk8sq26tv7va6
2009-06-08T17:11:02 <ThomasWaldmann> nice examples :P
2009-06-08T17:12:25 <ThomasWaldmann> and maybe everybody mark the time when he USUALLY has time, not only this week
2009-06-08T17:13:06 <dimazest> hard to say about usually, because exams starts next week
2009-06-08T17:14:48 <dreimark> amartani has noted that he has classes until 10:30 CEST
2009-06-08T17:15:21 <xorAxAx> dimazest: you are pretty busy for a soc student :)
2009-06-08T17:16:53 <dimazest> xorAxAx: i'm not sure about Friday and Saturday
2009-06-08T17:21:18 * ThomasWaldmann marked a lot times as free because it is hardly predictably anyway
2009-06-08T17:21:31 <ThomasWaldmann> -y+e
2009-06-08T17:22:27 * dreimark it is difficult if the internet problem is not fixed
2009-06-08T17:25:52 <ThomasWaldmann> you can input the time with network now and edit the table for extended times later
2009-06-08T17:26:18 <dreimark> i have edited already
2009-06-08T17:34:12 <ThomasWaldmann> xorAxAx: you add a link to the GoogleSoc2009 page?
2009-06-08T17:34:20 <xorAxAx> i can do that
2009-06-08T18:09:49 <dimazest> dreimark: ThomasWaldmann any comments on http://paste.pocoo.org/show/121791/ especially on _contains and _iter
2009-06-08T18:48:34 <ThomasWaldmann> dimazest: move line 73 to the beginning of the else branch
2009-06-08T18:49:11 <ThomasWaldmann> dimazest: nice that you have considered / found and solved that endless recursion problem
2009-06-08T18:49:58 <dimazest> it came to me yesterday night
2009-06-08T18:50:46 <ThomasWaldmann> a similar change is needed for line 108
2009-06-08T18:51:48 <dimazest> done
2009-06-08T18:51:49 <ThomasWaldmann> 105/106 has been -> have been
2009-06-08T18:55:03 <ThomasWaldmann> maybe do s/checked_groups/processed_groups/
2009-06-08T18:55:38 <ThomasWaldmann> checking is ok for _contains, but for _iter it is rather processing. otoh, processing is not wrong for _contains, so you can use it at both places.
2009-06-08T18:56:19 <ThomasWaldmann> seen_groups is maybe an alternative
2009-06-08T18:56:49 <dimazest> i like processed
2009-06-08T18:56:51 <ThomasWaldmann> other than that, code looks pretty good :)
2009-06-08T18:57:08 <ThomasWaldmann> maybe add a __repr__ for better debug output
2009-06-08T18:58:37 <ThomasWaldmann> maybe that "seen check" could be moved a bit. you do not need it at the very beginning of the recursion, as the set will be empty there anyway.
2009-06-08T18:59:21 <ThomasWaldmann> so you could do that check between line 83 and 84 (still referring to the pastebin line nos)
2009-06-08T18:59:44 <dimazest> i usually put it in the beginnig not to forget
2009-06-08T18:59:58 <ThomasWaldmann> it is slightly faster the other way :)
2009-06-08T19:00:57 <ThomasWaldmann> (it can even be significantly faster, if groups[group_name] is expensive
2009-06-08T19:01:01 <ThomasWaldmann> )
2009-06-08T19:01:11 <dimazest> i prefere in the beginnig have working code and then tune it
2009-06-08T19:01:28 <ThomasWaldmann> sure. it did work, right? :)
2009-06-08T19:01:55 <dimazest> it works
2009-06-08T19:02:01 <dreimark> sorry was doing a phone call. thursday is official holiday
2009-06-08T19:02:05 <ThomasWaldmann> so now we can tune it :D
2009-06-08T19:02:22 <dimazest> yes :)
2009-06-08T19:04:32 <dimazest> line 84 now looks like if self.name in processed_groups and groups[group_name]._contains(member, processed_groups):
2009-06-08T19:05:20 <ThomasWaldmann> no
2009-06-08T19:05:33 <dimazest> why?
2009-06-08T19:05:49 <ThomasWaldmann> you want to check group_name
2009-06-08T19:06:20 <ThomasWaldmann> (not self.name)
2009-06-08T19:06:54 <dimazest> you are right, thats because of headache
2009-06-08T19:07:12 <dreimark> self.name = group_name may be should not use an other wording
2009-06-08T19:08:00 <dimazest> dreimark: so name or group_name?
2009-06-08T19:08:22 <dreimark> the latter one. otherwise it will confuse later
2009-06-08T19:09:06 <dennda> meh
2009-06-08T19:09:15 <dennda> Now I marked the times when I have time *this* week ;)
2009-06-08T19:09:28 <dennda> But as dimazest already said, that's difficult to tell anyway
2009-06-08T19:10:44 <dreimark> for me it becomes next week better
2009-06-08T19:11:05 <dennda> For me it's totally inconsistent
2009-06-08T19:11:16 <dimazest> ThomasWaldmann: acltually it should be if group_name *not* in processed_groups and groups[group_name]._contains(member, processed_groups):
2009-06-08T19:12:40 <dreimark> that can become a famous tongue twister
2009-06-08T19:13:48 <dreimark> htg
2009-06-08T19:15:19 *** Rilt has quit IRC
2009-06-08T19:15:52 * dennda adjusted his entry to reflect when he usually has time, which doesn't say at all when he has time this week
2009-06-08T19:16:53 <ThomasWaldmann> dimazest: yes :)
2009-06-08T19:18:20 <dimazest> ThomasWaldmann: similar can be done for _iter
2009-06-08T19:22:18 <ThomasWaldmann> sure
2009-06-08T19:28:55 <dimazest> is it good for repr? '<Group group_name=SomeTestGroup members=set([]) member_groups=set([])>'
2009-06-08T19:30:09 <ThomasWaldmann> great
2009-06-08T19:35:29 <ThomasWaldmann> dimazest: btw, that code is rather generic for any group backend supporting groups-in-groups
2009-06-08T19:45:16 <dimazest> we will see when there will be several backends
2009-06-08T19:45:23 <dimazest> as i told yesterday
2009-06-08T19:46:12 <dimazest> because now code is changing so often and drammatically
2009-06-08T19:55:09 *** grzywacz has joined #moin-dev
2009-06-08T20:04:14 *** ernestas has joined #moin-dev
2009-06-08T20:05:23 <ernestas> HelpOnImages: "upload the image file (using the 'upload file' link)" ehm.. there is no such link
2009-06-08T20:05:44 <ernestas> maybe "Attachments" link?
2009-06-08T20:08:15 <ernestas> and I get "You can't change ACLs on this page since you have no admin rights on it!" when I try to save my translation. Should I not add the #acl line like?
2009-06-08T20:09:06 <ernestas> ok I have to go for a moment, I'll be back for answers shortly!
2009-06-08T20:09:12 *** ernestas has quit IRC
2009-06-08T20:16:54 <ThomasWaldmann> dimazest: if your stuff is in a more or less stable state, commit it
2009-06-08T20:18:11 <ThomasWaldmann> and push
2009-06-08T20:18:43 * dimazest is doing final checks
2009-06-08T20:19:12 <ThomasWaldmann> run the tests :)
2009-06-08T20:19:31 <dimazest> that includes final checks
2009-06-08T20:24:51 *** cdyson37 has joined #moin-dev
2009-06-08T20:28:13 <CIA-19> Dmitrijs Milajevs <dimazest@gmail.com> default * 4772:cab105f9fd55 1.9-groups-dmilajevs/MoinMoin/groups/backends/ (_tests/test_wiki_group.py wiki_group.py):
2009-06-08T20:28:13 <CIA-19> Groups2009: wiki_page.Group doesn't expand itself during initialization, instead it iterates over subgroups during __iter__ and checks subgroups during __contains__.
2009-06-08T20:28:13 <CIA-19> __repr__ for Group.
2009-06-08T20:28:13 <CIA-19> Tests for recursive groups.
2009-06-08T20:40:40 * ThomasWaldmann pulls
2009-06-08T20:42:44 <ThomasWaldmann> dimazest: E raise KeyError("There is no such group %s" % group_name)
2009-06-08T20:42:53 <ThomasWaldmann> > KeyError: u'There is no such group ExampleProject/AdminGroup'
2009-06-08T20:50:50 <ThomasWaldmann> dimazest: maybe put line 45+ in a separate method called from __init__
2009-06-08T20:52:01 *** ernestas has joined #moin-dev
2009-06-08T20:55:43 <ThomasWaldmann> dimazest: is line 105 needed?
2009-06-08T20:57:42 <ThomasWaldmann> ernestas: i guess (for 1.8/1.9) he means to click on Attachments, select a file and then click that "upload" button
2009-06-08T20:58:04 <dimazest> processed_groups.add(group_name) ?
2009-06-08T20:58:11 <ThomasWaldmann> ernestas: you don't need to add acls, they are added automatically, just keep the ##acl ...
2009-06-08T20:58:15 <ThomasWaldmann> dimazest: yes
2009-06-08T20:58:41 <ernestas> ThomasWaldmann: ok, thanks
2009-06-08T21:00:36 <dimazest> i'm not sure about that line
2009-06-08T21:00:49 <dimazest> and i'll extract taht lines
2009-06-08T21:01:05 <ThomasWaldmann> you could add a assert group_name not in processed_groups before it :)
2009-06-08T21:03:12 <dimazest> ThomasWaldmann: wasit a joke or suggestion?
2009-06-08T21:04:00 <ThomasWaldmann> well, I guess you would find that it always fails with assertionerror
2009-06-08T21:04:16 <ThomasWaldmann> you know why?
2009-06-08T21:04:41 <dimazest> ThomasWaldmann: i need to go now, can we discuss it tomorrow?
2009-06-08T21:05:28 <ThomasWaldmann> ok
2009-06-08T21:07:36 <ThomasWaldmann> btw, "member" in _contains should be rather "name" (you don't know whether it is a member)
2009-06-08T21:07:42 *** cdyson37 has quit IRC
2009-06-08T21:09:59 <ThomasWaldmann> dimazest: same as for 105 is likely true for 139
2009-06-08T21:12:00 <dimazest> ThomasWaldmann: thank you for comments
2009-06-08T21:13:22 <ThomasWaldmann> and don't forget to process all the comments, including those from the past days :)
2009-06-08T21:13:39 <ThomasWaldmann> code will get much simpler then
2009-06-08T21:17:28 <dimazest> ok, see you tomorrow, ciao!
2009-06-08T21:18:52 <ThomasWaldmann> dennda: ok, back to storage. how's it going?
2009-06-08T21:32:11 <dennda> I'll push some commits later this evening; Preparing for an exam in parallel
2009-06-08T22:06:04 *** devilsadvocate_ has quit IRC
2009-06-08T22:06:18 *** devilsadvocate has joined #moin-dev
2009-06-08T22:30:01 <xorAxAx> devilsadvocate: another day passed, no checkins, no replies to my questions. do you have time on wednesday for a small chat? please suggest a time
2009-06-08T22:46:03 <xorAxAx> dennda: you seem even more busy than dimazest :) why is that?
2009-06-08T22:46:09 <xorAxAx> (regarding the doodle)
2009-06-08T22:47:19 <dennda> xorAxAx: Well that's because I'm in the fifth of six semesters :-)
2009-06-08T22:47:44 <xorAxAx> thats not a reason, ask your profs to switch to american term schedule :-P
2009-06-08T22:47:55 <xorAxAx> like i would have nearly done when i was a soc student!
2009-06-08T22:47:56 <xorAxAx> ;-)
2009-06-08T22:48:04 <dennda> No; Ask google to call it Winter of Code instead
2009-06-08T22:48:06 <dennda> :-)
2009-06-08T22:48:14 <xorAxAx> but seriously, you cannot have 9 hours of lectures daily
2009-06-08T22:48:17 <dennda> or rather spring / autumn
2009-06-08T22:48:21 <xorAxAx> and when do you want to code ...
2009-06-08T22:50:41 <dennda> Well I can add more yellow boxes, but afaict the idea of the poll is to find a time that suits best which I think is not achieved by saying 'I could make it anytime... somehow'
2009-06-08T22:51:28 <xorAxAx> well, we'll see :)
2009-06-08T22:51:42 <dennda> Yes
2009-06-08T23:02:54 <dennda> ThomasWaldmann: still with us?
2009-06-08T23:09:56 <dennda> xorAxAx: Maybe you can help as well. If a page's acls deny access to some user and the after acls allow it, is the user allowed to access the page or not?
2009-06-08T23:10:29 <xorAxAx> no, he is not. acl processing currently stops at the first match
2009-06-08T23:10:40 <dennda> ok
2009-06-08T23:11:29 <dennda> Then I think thomas' test case is wrong or the comment is very misleading
2009-06-08T23:13:54 <dennda> We have an item with 'JoeDoe: ' (no access) and a subitem with '' (no access). JoeDoe wants to access the subitem (hierarchically). The default acl is not used because there's an empty acl (that says no rights for anyone). So the subitem's acl is the first match, isn't it?
2009-06-08T23:14:12 <dennda> Then, the after_acl shouldn't be considered. (Which is 'All:read')
2009-06-08T23:14:52 <dennda> xorAxAx: Do you think that's correct?
2009-06-08T23:15:48 <xorAxAx> isnt the first item's acl considered first?
2009-06-08T23:15:59 <xorAxAx> i dont remember the hierarchical rules
2009-06-08T23:16:11 <ThomasWaldmann> dennda: yes
2009-06-08T23:16:19 <dennda> As far as I can tell from the HelpOnAcl page, subitem is checked, then parent, then the parent's parent and so on
2009-06-08T23:16:38 <dennda> ThomasWaldmann: Is that a 'yes Im still here' or 'yes you are correct'? :-)
2009-06-08T23:17:23 <ThomasWaldmann> first
2009-06-08T23:18:10 <ThomasWaldmann> dennda: if there is an empty acl, there is nothing to "match"
2009-06-08T23:18:33 <dennda> ThomasWaldmann: Then what is the definition of a 'match'?
2009-06-08T23:18:50 <ThomasWaldmann> the username matching the left side of an acl entry
2009-06-08T23:19:00 <ThomasWaldmann> what else should it be?
2009-06-08T23:19:24 <dennda> And 'All:' doesn't count as username I guess
2009-06-08T23:19:48 <ThomasWaldmann> that is the left side of an acl entry
2009-06-08T23:20:03 <dennda> ThomasWaldmann: To be precise here, I'm talking about this test case:
2009-06-08T23:20:15 <dennda> (True, self.subitem2_name, u'JoeDoe', ['read']), # by after acl
2009-06-08T23:21:00 <ThomasWaldmann> you need to look at before, item/[parent/]default, after acl
2009-06-08T23:21:05 <dennda> From what you've just said the mainitem's acls would be checked before acl_rights_after
2009-06-08T23:21:12 <dennda> yes
2009-06-08T23:21:21 <dennda> and on the parent there is 'JoeDoe:'
2009-06-08T23:21:31 <dennda> Which I consider to be the first match
2009-06-08T23:21:55 <ThomasWaldmann> what's on the subitem?
2009-06-08T23:22:02 <dennda> ''
2009-06-08T23:22:35 <ThomasWaldmann> ok, as it has an acl (an empty one to be exact) on "self", it doesn't look at parent or default acl
2009-06-08T23:23:10 <ThomasWaldmann> so it is before + '' + after
2009-06-08T23:23:14 <dennda> Why does it not look on parent?
2009-06-08T23:23:27 <ThomasWaldmann> because acls on self override parent acl of course
2009-06-08T23:23:33 <dennda> Didn't you just say that with '' there's nothing to match?
2009-06-08T23:23:53 <dennda> I thought you'd check the parent's if there *was no match*, not if there *was no acl at all*
2009-06-08T23:24:01 <ThomasWaldmann> that's no contradiction
2009-06-08T23:24:09 <ThomasWaldmann> no
2009-06-08T23:24:21 <ThomasWaldmann> it's not THAT complicated
2009-06-08T23:24:45 <ThomasWaldmann> hierarchical acls are quite similar to normal non-h. procedure
2009-06-08T23:25:09 <ThomasWaldmann> non-h. is just before + pageacl + after (if there is a pageacl), or
2009-06-08T23:25:21 <ThomasWaldmann> before + default + after (if there is no pageacl)
2009-06-08T23:25:23 <dennda> I got that
2009-06-08T23:26:16 <ThomasWaldmann> and the only thing different for hierarchical is that if it has no own acl, it doesn't just use the default, but first searches upwards in the hierarchy if there is some page acl
2009-06-08T23:26:28 <dennda> aha
2009-06-08T23:27:02 <dennda> I was under the impression that it checks the parents only if there's no match (there may be an acl, though)
2009-06-08T23:27:19 <dennda> ok
2009-06-08T23:27:35 <dennda> Then I get your testcase
2009-06-08T23:27:48 <ThomasWaldmann> you can read HelpOnaccessControlLists now and add what's missing :)
2009-06-08T23:28:19 <dennda> I had that page open all the time :-)
2009-06-08T23:28:30 <dennda> In a nutshell, if a permission is not resolved by the current page, then the parent page's ACL is checked, and then its parent, and so on until there are no parent pages.
2009-06-08T23:28:35 <dennda> That's the sentence I misinterpreted
2009-06-08T23:28:43 <dennda> "is not resolved"
2009-06-08T23:28:58 <dennda> ThomasWaldmann: That is an Immutable Page
2009-06-08T23:29:38 <dennda> AssertionError: User dennda may read page HelpOnAccessControlLists
2009-06-08T23:29:46 <dennda> AssertionError: User dennda may write page HelpOnAccessControlLists
2009-06-08T23:29:51 <dennda> this way ;)
2009-06-08T23:34:46 <dennda> ok fixed
2009-06-08T23:36:45 <ThomasWaldmann> dennda may read EditingOnMoinMaster :)
2009-06-08T23:37:16 <CIA-19> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5916:81edaac7d1a4 2.0-storage/MoinMoin/storage/backends/acl.py: storage: Fix for hierarchic resolving of permissions. Fixes the two recently added failing security tests.
2009-06-08T23:38:42 <dennda> ThomasWaldmann: looky looky ^
2009-06-08T23:39:16 * ThomasWaldmann looks
2009-06-08T23:43:22 <ThomasWaldmann> dennda: yes, that's correct. :)
2009-06-08T23:43:32 <dennda> That was a nice fix
2009-06-08T23:43:43 <dennda> Thanks for the explanation
2009-06-08T23:43:56 <ThomasWaldmann> now we have to check in 1.9 :)
2009-06-08T23:44:14 <dennda> Well since you copy pasted...
2009-06-08T23:46:03 <dennda> hmpf were did you copy that *from* again
2009-06-08T23:46:38 <dennda> ah
2009-06-08T23:46:51 <ThomasWaldmann> uhoh
2009-06-08T23:47:03 <dennda> it's wrong there
2009-06-08T23:47:06 <ThomasWaldmann> now let's check in 1.7
2009-06-08T23:47:07 <dennda> from what I can see
2009-06-08T23:47:37 <dennda> hah hah
2009-06-08T23:47:50 <dennda> looks the same to me
2009-06-08T23:48:20 <ThomasWaldmann> that was ever broken
2009-06-08T23:48:26 <dennda> yes
2009-06-08T23:48:31 <dennda> and no one noticed
2009-06-08T23:49:14 <ThomasWaldmann> ok, I'll do the backporting tomorrow
2009-06-08T23:49:15 <dennda> so much as for the importance of what we do here; nobody cares anyway :D
2009-06-08T23:49:50 <dennda> I can do that as well if you like
2009-06-08T23:49:56 <ThomasWaldmann> not noticing and not caring are 2 different things
2009-06-08T23:50:04 <dennda> sure, I'm just kidding ;)
2009-06-08T23:50:22 <ThomasWaldmann> the tests also need to get backported
2009-06-08T23:50:28 <dennda> ok
2009-06-08T23:50:32 <dennda> your job ;)
2009-06-08T23:51:14 <ThomasWaldmann> and I'll have to look at the implications, whether that means a maintenance release
2009-06-08T23:51:43 <dennda> i can see the headline
2009-06-08T23:52:20 <dennda> """Rockstar alike GSoC student fixed severe security bug in famous wiki software"""
2009-06-08T23:52:32 <ThomasWaldmann> hehehe
2009-06-08T23:52:39 <dennda> You'll also be mentioned as the one who made the stone roll
2009-06-08T23:52:45 <dennda> but i'm gonna be the super hero
2009-06-08T23:53:04 <dennda> ;)
2009-06-08T23:53:19 <dennda> Let's think about it
2009-06-08T23:53:27 <dennda> What's the worst that could've happened?
2009-06-08T23:54:02 <ThomasWaldmann> it continues the for loop, that means it climbs upwards in page hierarchy
2009-06-08T23:54:08 <dennda> yes
2009-06-08T23:54:17 <dennda> that's clear
2009-06-08T23:54:33 <TheSheep> rockstar alike?
2009-06-08T23:54:34 <ThomasWaldmann> so, if the upper acl is less restrictive than the lower...
2009-06-08T23:54:57 <dennda> TheSheep: yeah I tend to exaggerate
2009-06-08T23:55:07 <TheSheep> dennda: which rockstar?
2009-06-08T23:55:15 <dennda> TheSheep: Do you want to be the first person to receive a signature?
2009-06-08T23:55:18 <dennda> TheSheep: me of course
2009-06-08T23:55:31 <TheSheep> alike to whom?
2009-06-08T23:55:40 <dennda> some rockstar
2009-06-08T23:55:43 <dennda> doesn't matter
2009-06-08T23:55:59 <dennda> Did I misuse the term?
2009-06-08T23:56:11 <TheSheep> ah, *the* rockstar :D
2009-06-08T23:56:57 <ThomasWaldmann> hmm, it only climbs if there is no match
2009-06-08T23:57:21 <ThomasWaldmann> subpage: Boss:r,w
2009-06-08T23:57:35 <ThomasWaldmann> page: All:r,w
2009-06-08T23:57:45 <ThomasWaldmann> JoeDoe accesses subpage
2009-06-08T23:58:17 <dennda> and is allowed to read and write
2009-06-08T23:58:27 <dennda> but he shouldn't be since he's not mentioned on subpage's acl
2009-06-08T23:58:37 <dennda> am I correct?
2009-06-08T23:58:49 <ThomasWaldmann> yeah. can you add such a test? just to check if that really happens.
2009-06-08T23:59:07 <dennda> Which version?
2009-06-08T23:59:17 <ThomasWaldmann> all 2.0-st for now
2009-06-08T23:59:32 <dennda> hm?
2009-06-08T23:59:38 <dennda> it certainly won't climb in 2.0 anymore
2009-06-08T23:59:58 <dennda> if has_acl: break MoinMoin: MoinMoinChat/Logs/moin-dev/2009-06-08 (last edited 2009-06-07 22:45:02 by IrcLogImporter)
