MoinMoin: MoinMoinChat/Logs/moin-dev/2007-08-11

2007-08-11T00:25:00  * ThomasWaldmann tests new link markup
2007-08-11T08:49:33  <dreimark> moin
2007-08-11T12:06:09  <ThomasWaldmann> moin
2007-08-11T13:46:24  <ThomasWaldmann> (?P<heading>^\s*(?P<hmarker>=+)\s+(<?P<heading_text>.*?)\s+(?P=hmarker) $)
2007-08-11T13:46:37  <ThomasWaldmann> does someone see why this doesn't work?
2007-08-11T13:47:23  <ThomasWaldmann> (the blank in front of the $ is correct and must be there)
2007-08-11T14:01:13  <ThomasWaldmann> hmm, if i remove the heading_text group name, it matches
2007-08-11T14:17:52  <ThomasWaldmann> strange, now it works
2007-08-11T15:12:18  <dreimark> moin
2007-08-11T15:17:12  <ThomasWaldmann> hi dreimark
2007-08-11T15:17:55  <ThomasWaldmann> back from holidays?
2007-08-11T15:29:39  <dreimark> ThomasWaldmann: is all ok with neagulm?
2007-08-11T15:36:30  <xorAxAx> no :)
2007-08-11T16:10:22  <ThomasWaldmann> that was too much :)
2007-08-11T16:59:15  <ThomasWaldmann> >   error: nothing to repeat
2007-08-11T16:59:29  <ThomasWaldmann> (the joy of sre error msgs 8( )
2007-08-11T17:37:20  <ThomasWaldmann> lanius_: any news on the failing tests?
2007-08-11T18:07:20  <ThomasWaldmann> lanius clones himself :)
2007-08-11T18:27:51  * ThomasWaldmann translates the big ugly one to a bigger, less ugly re.VERBOSE one
2007-08-11T19:44:40  <ThomasWaldmann> s/the big ugly one/the big pretty one/
2007-08-11T19:45:35  <lanius> ThomasWaldmann: , sorry don't had time to look at it, yet, seems to be some raceconditions so
2007-08-11T19:51:13  <lanius> because sometimes they work
2007-08-11T19:58:01  <ThomasWaldmann> =X-)
2007-08-11T19:59:17  <ThomasWaldmann> how can you have race conditions in a single thread?
2007-08-11T19:59:53  <xorAxAx> some lines are faster than others
2007-08-11T20:00:25  <lanius> flushing of things to disk
2007-08-11T20:00:31  <xorAxAx> (you can have concurrency via microthreads/generators, but i am sure that lanius doesnt :))
2007-08-11T20:01:10  <lanius> i guess
2007-08-11T20:02:59  <ThomasWaldmann> maybe file mtime granularity?
2007-08-11T20:05:19  <xorAxAx> maybe you should query the size instead of the mtime :)
2007-08-11T20:05:45  <lanius> sounds like an idea
2007-08-11T20:06:30  <ThomasWaldmann> (see old "news" code :)
2007-08-11T22:30:16  <ThomasWaldmann> moin nir :)
2007-08-11T22:30:25  <nir> moin
2007-08-11T22:30:40  <nir> Does new hierarchical acl solve the problem of renames?
2007-08-11T22:30:54  <nir> for example, a/b/c inherit acl from a
2007-08-11T22:31:07  <nir> not a user rename a/b/c to foo/b/c
2007-08-11T22:31:12  <nir> acl gone :-)
2007-08-11T22:31:46  <xorAxAx> nir: no
2007-08-11T22:31:49  <ThomasWaldmann> well, ygwyd :)
2007-08-11T22:32:09  <xorAxAx> nir: same happens on windows if you move a file with inherited permissions, right?
2007-08-11T22:32:35  <nir> I don't use windows a lot :)
2007-08-11T22:32:37  <xorAxAx> hmm, the rename action could be enhanced to fix this, though
2007-08-11T22:32:40  <xorAxAx> me neither
2007-08-11T22:33:01  <nir> the problem is you may like to rename /a/b/c to /a/b/other
2007-08-11T22:33:22  <nir> the issue is moving the page from the parent
2007-08-11T22:33:32  <xorAxAx> "from the parent"?
2007-08-11T22:34:02  <nir> if it was a folder, you took the file from one folder to another
2007-08-11T22:34:32  <nir> b/c was a child of a, not a child of foo
2007-08-11T22:35:08  * xorAxAx doesnt understand
2007-08-11T22:35:22  <nir> lets simplify
2007-08-11T22:35:26  <xorAxAx> i dont see any foo in your example
2007-08-11T22:35:39  <nir> a/b
2007-08-11T22:35:47  <nir> b inherit from a
2007-08-11T22:35:59  <nir> not you rename a/b to c/b
2007-08-11T22:36:11  <xorAxAx> yes
2007-08-11T22:36:15  <nir> this is a move
2007-08-11T22:36:28  <nir> but a/b -> a/b2 is a rename
2007-08-11T22:36:48  <xorAxAx> well, yes
2007-08-11T22:37:15  <nir> in a file system you need permission to move the file from the folder
2007-08-11T22:37:31  <xorAxAx> same for moin
2007-08-11T22:37:58  <nir> but move and rename are the same rights in moin
2007-08-11T22:38:06  <nir> write + delete
2007-08-11T22:38:15  <xorAxAx> yes
2007-08-11T22:38:28  <xorAxAx> they are on e.g. windows as well AFAIK
2007-08-11T22:39:43  <nir> can be solve if rename operate only on the last component of the page name
2007-08-11T22:40:02  <xorAxAx> well, which problem in the first place?
2007-08-11T22:40:28  <nir> how to allow rename under the parent, and disable move to another parent
2007-08-11T22:40:30  <xorAxAx> that a renamepage action might reveal data to other users without the user doing it knowing about it?
2007-08-11T22:40:38  <nir> probably by mistake
2007-08-11T22:41:06  <nir> imagine a stupid user, get a rename box with a/b/c
2007-08-11T22:41:15  <xorAxAx> well, but that limitation wont help if anybody actually needs the feature
2007-08-11T22:41:19  <nir> then he remove the /a/b and call the page foo
2007-08-11T22:42:26  <nir> the same user would drag and drop files in his GUI file manager
2007-08-11T22:42:45  <nir> he does not use mv on the command line
2007-08-11T22:43:31  <lanius> the mtime is not the problem
2007-08-11T22:45:44  <nir> xorAxAx, try this http://nirs.jot.com/RecentChanges
2007-08-11T22:46:08  <nir> you can rename the page, click on the title
2007-08-11T22:46:22  <nir> but the page remain under WikiHome
2007-08-11T22:47:04  <xorAxAx> and how can i move it?
2007-08-11T22:47:26  <xorAxAx> note that moins idea of hierachies is pretty lax
2007-08-11T22:47:28  <nir> they have admin tools to move pages around
2007-08-11T22:47:41  <xorAxAx> hier. acls being the first step to make it more relevant
2007-08-11T22:47:55  <xorAxAx> admin tools ... sounds un-wiki-like
2007-08-11T22:48:15  <nir> look at the bottom of the page
2007-08-11T22:48:32  <nir> If you are admin, you get some links there
2007-08-11T22:50:09  <nir> I can't find now where is the move page thing :)
2007-08-11T22:51:07  <xorAxAx> wiki wiki is haiwaiian and means "hidden well"
2007-08-11T22:52:36  <nir> I found it - on some pages you have "move this page" under tools
2007-08-11T22:53:49  <xorAxAx> lanius: did you find the real problem?
2007-08-11T22:54:29  <nir> What about the issue of different acl for the page content and page children?
2007-08-11T22:54:39  <nir> there is a page in the wiki about it
2007-08-11T22:55:46  <xorAxAx> well, the evaluation order is clear
2007-08-11T22:55:58  <xorAxAx> somebody said that he wanted to edit my explanations of it
2007-08-11T22:57:52  <nir> xorAxAx, http://moinmo.in/FeatureRequests/SubpagesInheritDifferentAclFromParent
2007-08-11T22:58:55  <xorAxAx> yes, do you spot my nir-like comment on the page? :)
2007-08-11T22:59:10  <nir> The problem is not evaluation but the way to setup the content acl vs children acl
2007-08-11T22:59:23  <xorAxAx> yes
2007-08-11T22:59:48  <nir> both solutions - subacl and localacl are ugly
2007-08-11T23:00:05  <xorAxAx> yes, and the natural workaround is less ugly
2007-08-11T23:00:22  <nir> which is?
2007-08-11T23:01:02  <xorAxAx> moving pages so that you dont need to have a different ACL line on the parent page
2007-08-11T23:01:52  <nir> put the pages under a different parent?
2007-08-11T23:02:04  <xorAxAx> yes
2007-08-11T23:02:53  <nir> Then you have messy urls
2007-08-11T23:07:41  <xorAxAx> maybe
2007-08-11T23:08:24  <nir> It would be nice to have something like "index" magic
2007-08-11T23:08:38  <nir> which is the content of a "folder"
2007-08-11T23:10:58  <xorAxAx> yes, modify missing page to call a third macro :)
2007-08-11T23:12:04  <nir> which call back into send page :P
2007-08-11T23:12:56  <xorAxAx> talking about send_page might trigger neurotic issues for some people in this channel
2007-08-11T23:26:30  <ThomasWaldmann> lol