MoinMoin: MoinMoinChat/Logs/moin-dev/2009-06-20

2009-06-20T00:07:50  *** amartani has joined #moin-dev
2009-06-20T00:09:23  <amartani> freenode blocked mibbit =/
2009-06-20T00:50:05  <CIA-38> Alexandre Martani <amartani AT gmail DOT com> default * 5936:b60fa5e53772 2.0-storage-editor-amartani/MoinMoin/ (2 files in 2 dirs): Real-time editor: save editors names on metadata.
2009-06-20T00:50:08  <CIA-38> Alexandre Martani <amartani AT gmail DOT com> default * 5937:16b3f3932793 2.0-storage-editor-amartani/MoinMoin/action/ (_tests/test_mobwrite.py mobwrite.py): Real-time editor: add download and upload methods to mobwrite worker and simplify get_editors using them
2009-06-20T01:18:06  <xorAxAx> dreimark: yes
2009-06-20T01:29:00  *** amartani has quit IRC
2009-06-20T09:20:50  *** grzywacz has joined #moin-dev
2009-06-20T11:21:10  <ThomasWaldmann> moin
2009-06-20T13:41:42  <ThomasWaldmann> waldi: your toc-code is generating IDs toc-0, toc-1, ... (not "speaking IDs", derived from the headline contents as the old code did)
2009-06-20T13:43:33  <ThomasWaldmann> that's bad for #AboutFoo link targets
2009-06-20T13:54:25  * dennda works on error handling
2009-06-20T14:21:13  <xorAxAx> sigh, my student posts travel reports on his blog but nothing her
2009-06-20T14:21:14  <xorAxAx> e
2009-06-20T14:22:26  <xorAxAx> "but now I get to relax for the next week or so before the next semester sets in."
2009-06-20T14:22:29  <xorAxAx> uhuh
2009-06-20T14:22:31  <xorAxAx> relax, sure!
2009-06-20T14:22:47  <dennda> not the best strategy in these web2.0y days
2009-06-20T14:23:05  <dennda> (disclosing such private stuff in this situation)
2009-06-20T14:24:50  <xorAxAx> well, it doesnt compromise a lot of his privacy
2009-06-20T14:25:10  <xorAxAx> and its consistent with the stuff we have been seeing until now
2009-06-20T14:28:20  <ThomasWaldmann> xorAxAx: url?
2009-06-20T14:29:49  <waldi> ThomasWaldmann: http://moinmo.in/BastianBlank/TreeOutputFormatter/ToDo
2009-06-20T14:31:56  <ThomasWaldmann> ah, good :)
2009-06-20T14:33:39  <dreimark> ThomasWaldmann: http://blogs.iitk.ac.in/chintal/
2009-06-20T14:40:06  <ThomasWaldmann> xorAxAx: where is that "relax" quote from? url?
2009-06-20T14:41:21  <dreimark> http://devilsadvocate-chs.blogspot.com/2008/07/and-they-all-look-just-same.html
2009-06-20T14:42:02  <dennda> Wednesday, July 16, 2008
2009-06-20T14:42:40  <dreimark> may be there is a newer one, that was the one I found by google
2009-06-20T14:42:42  <ThomasWaldmann> yeah, that's from last year
2009-06-20T14:43:42  <ThomasWaldmann> but the posts from recently also show that he's quite busy on other F/OSS stuff (much more than on his SOC project)
2009-06-20T14:44:08  <dreimark> it looks like a different job
2009-06-20T14:47:06  <ThomasWaldmann> it rather looks like some F/OSS fan giving talks and sessions
2009-06-20T14:48:05  <ThomasWaldmann> (there is nothing bad about this, but it seems like either he has already given up for SoC or he doesn't get priorities right)
2009-06-20T14:49:38  <ThomasWaldmann> devilsadvocate: ^^ as this is about you, you maybe want to give some comment, too.
2009-06-20T14:58:39  <xorAxAx> http://blogs.iitk.ac.in/chintal/
2009-06-20T15:07:03  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5941:e41c50daff45 2.0-storage/MoinMoin/ (storage/backends/acl.py storage/error.py wsgiapp.py): storage: Improve AccessDeniedError failure message
2009-06-20T15:07:05  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5942:d37f5e44deda 2.0-storage/MoinMoin/storage/backends/acl.py: storage: Make all thrown ADErrors provide failure information
2009-06-20T15:07:10  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5943:c028e4fcd09e 2.0-storage/MoinMoin/storage/ (backends/acl.py error.py): storage: small fix and code simplification
2009-06-20T15:07:12  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5944:081a48c999e4 2.0-storage/MoinMoin/theme/__init__.py: storage: First indicator for immutable pages
2009-06-20T15:24:01  <dennda> I love pdb
2009-06-20T15:25:35  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5945:2e28e745c9a6 2.0-storage/MoinMoin/Page.py: storage: Fix. Don't freak out completely if only one page that we link to may not be accessed
2009-06-20T15:28:31  <ThomasWaldmann> :)
2009-06-20T15:33:28  <ThomasWaldmann> dennda: s/user/username/
2009-06-20T15:33:49  <dennda> where?
2009-06-20T15:35:56  <dennda> PawelPacana / TheSheep: Has the patch that was necessary to run the new hg backend been applied to mainline hg by now? Because then I'd throw out the old hg backend... (I think someone mentioned that already but can't find the reference)
2009-06-20T15:37:35  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5946:815df9023e85 2.0-storage/create_persistent_dev_wiki.py: storage: PEP8 Fix
2009-06-20T15:38:15  <ThomasWaldmann> dennda: http://hg.moinmo.in/moin/2.0-storage/rev/081a48c999e4 what does the commit comment mean exactly/what's the point of this changeset?
2009-06-20T15:40:22  <dennda> ThomasWaldmann: if not user.may.write: "modify" -> "immutable page"
2009-06-20T15:40:49  <dennda> i'm sick of trying to write to that page just to test if it's writable
2009-06-20T15:41:03  <ThomasWaldmann> ah, and why "First", do you plan more?
2009-06-20T15:41:28  <dennda> ThomasWaldmann: there's still a link to the modify action
2009-06-20T15:41:36  <dennda> i.e. you can still click on immutable page
2009-06-20T15:42:09  <dennda> didn't realize immediately how to change that and postponed it due to it's not-so-importance
2009-06-20T15:43:03  <ThomasWaldmann> ok
2009-06-20T15:43:57  <ThomasWaldmann> dennda: PawelPacana: TheSheep: yes, that should be checked. having hg(old) and hgnew is confusing.
2009-06-20T15:44:30  <dennda> and someone (not me) should fix the failing tests
2009-06-20T16:02:31  <dreimark> xorAxAx: the post travel comment was from 2008
2009-06-20T16:08:20  <dreimark> htg bbl
2009-06-20T16:14:52  <dennda> ThomasWaldmann: Other than the hg failures there is no backend-code specific test case that keeps failing. (There are a bunch of tests that fail because standard things like FrontPage are not in the membackend that is used for testing. how to fix this?)
2009-06-20T16:17:53  <dennda> (which I don't think is a backend-code-based failure)
2009-06-20T16:25:20  <ThomasWaldmann> either load the membackend with them or don't use the membackend
2009-06-20T16:25:42  * ThomasWaldmann plays with weird generator stuff
2009-06-20T16:26:10  <dennda> i.e. pump the necessary files from underlay into the membackend?
2009-06-20T16:26:26  <dennda> perhaps fs is easier
2009-06-20T16:27:23  <ThomasWaldmann> if you do not use the membackend, you are likely to fall over the removed nuke_item stuff
2009-06-20T16:28:07  <dennda> hmpf
2009-06-20T16:28:15  <ThomasWaldmann> so maybe have 2 membackends:
2009-06-20T16:28:27  <ThomasWaldmann> a fresh one you do not modify (but load with pages)
2009-06-20T16:28:36  <dennda> i don't want to untar underlay every time i run py.test
2009-06-20T16:28:44  <ThomasWaldmann> and then you clone from that one for every test
2009-06-20T16:29:05  <dennda> the question rather is where to get the data from
2009-06-20T16:29:46  <ThomasWaldmann> the xml unserialize is still being worked on :)
2009-06-20T16:30:10  <dennda> well it's not that important now anyways
2009-06-20T16:30:32  <ThomasWaldmann> having less failing tests is important
2009-06-20T17:15:05  <franklin_> Hello, I wanted to backport the "security fix" in moinmoin 1.8.3 (http://hg.moinmo.in/moin/1.9/rev/8f913a5ac5f1 ) to DebianLenny's moinmoin 1.7, But I wonder if I misunderstood something, or if there is a mistake...
2009-06-20T17:15:11  <franklin_> The page http://master19.moinmo.in/HelpOnAccessControlLists#Hierarchical_ACL_processing explains that the parent ACLs are evaluated the one after the other (inherited)...
2009-06-20T17:15:50  <franklin_> The test (http://hg.moinmo.in/moin/1.8/rev/47c0ada5c1a2 ) seems to expect a different behavior: the ACL for "subitem_4boss" should be "#acl JoeDoe:read,write All:read", shouldn't it?
2009-06-20T17:15:52  <dennda> franklin_: The wording on that page is not optimal. It confused me at first, too
2009-06-20T17:16:40  <dennda> franklin_: Parent ACLs should always be checked if there is no acl at all (not even an empty one) on the current page
2009-06-20T17:17:02  <dennda> according to the spec. where the spec is what ThomasWaldmann told me it was :-)
2009-06-20T17:18:26  <franklin_> But the documentation and the code used to be consistent (even though they were differents from the specs), isn't it?
2009-06-20T17:19:09  <dennda> ThomasWaldmann: ^
2009-06-20T17:21:58  <franklin_> ThomasWaldmann: Am I correct to say if ther is no security issue in 1.7, but the behavior has changed in 1.8.4? [If this is the case, I would leave 1.7 as-is, and put a note about the new behaviour in Debian Squeeze]
2009-06-20T17:25:10  <ThomasWaldmann> franklin_: the security fix happened in 1.8.4, not 1.8.3
2009-06-20T17:25:27  <ThomasWaldmann> i also fixed it in 1.7 btw
2009-06-20T17:26:57  <franklin_> thanks, that would be much much easier to "backport" ;-)
2009-06-20T17:29:38  <franklin_> ThomasWaldmann: But as you know, Debian wants to avoid changing the behavior of application in stable. From your point of view, what is the best way to address this problem?
2009-06-20T17:31:25  <ThomasWaldmann> i read the help page's paragraph about hacls, it is confusing.
2009-06-20T17:31:43  <ThomasWaldmann> i doubt it ever worked as it was described there.
2009-06-20T17:32:03  <ThomasWaldmann> xorAxAx: can you have a look?
2009-06-20T17:32:41  <waldi> ThomasWaldmann: okay, I did some updates to the MimeType page
2009-06-20T17:33:46  <dennda> < ThomasWaldmann> dennda: s/user/username/ <-- where?
2009-06-20T17:35:54  <ThomasWaldmann> waldi: x-moin AND x.moin?
2009-06-20T17:36:11  <ThomasWaldmann> dennda: in the ADE stuff
2009-06-20T17:36:15  <waldi> ThomasWaldmann: hmm?
2009-06-20T17:36:33  <ThomasWaldmann> = Type: x-moin =
2009-06-20T17:36:43  <ThomasWaldmann> == Subtype: x.moin.wiki ==
2009-06-20T17:37:01  <waldi> there is no spec for x. in the types
2009-06-20T17:37:27  <ThomasWaldmann> ah, ok
2009-06-20T17:37:28  <waldi> the tree like x. was specified much later then the original mime spec
2009-06-20T17:37:37  <waldi> and only applies to subtypes
2009-06-20T17:38:04  <dennda> ThomasWaldmann: you mean in order to give the variable a more appropriate name i guess...
2009-06-20T17:38:13  <ThomasWaldmann> yes
2009-06-20T17:38:33  <ThomasWaldmann> user would be rather a user object, but you give user.name
2009-06-20T17:38:38  <dennda> jup
2009-06-20T17:39:23  <ThomasWaldmann> waldi: i guess I'll change all the x-unidentified-wiki-format, text/moin-wiki, ... crap later in 2.0-st
2009-06-20T17:39:42  <ThomasWaldmann> (later=later today)
2009-06-20T17:39:47  <waldi> ThomasWaldmann: yeah. please add that to the page
2009-06-20T17:40:01  <ThomasWaldmann> add what?
2009-06-20T17:40:26  <waldi> i don't know, what you want to change
2009-06-20T17:40:44  <ThomasWaldmann> i want to clean up and remove the invalid / inconsistent crap
2009-06-20T17:40:45  <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5947:ab6f7c91e07d 2.0-storage/MoinMoin/storage/ (backends/acl.py error.py): storage: use more appropriate variable name
2009-06-20T17:41:06  <waldi> okay
2009-06-20T17:41:30  <ThomasWaldmann> btw, we need a mimetype for moin creole
2009-06-20T17:42:22  <ThomasWaldmann> it is a variation of creole
2009-06-20T17:42:44  <ThomasWaldmann> so text/x.moin.creole ?
2009-06-20T17:42:50  <waldi> x.moin.wiki.creole?
2009-06-20T17:43:16  <ThomasWaldmann> x.moin.wiki is moin's wiki markup
2009-06-20T17:43:40  <ThomasWaldmann> it should be parallel to that, not below
2009-06-20T17:44:56  <ThomasWaldmann> is it allowed that we default to utf-8 encoding if no charset is given?
2009-06-20T17:44:57  <ThomasWaldmann> * text/html explicitely define this
2009-06-20T17:45:06  <ThomasWaldmann> that's unclear to me
2009-06-20T17:45:27  <ThomasWaldmann> is it allowed to have this default or does one need to explicitely give it?
2009-06-20T17:45:51  <ThomasWaldmann> waldi: ^^
2009-06-20T17:46:06  <waldi> i think it is bad style
2009-06-20T17:47:24  <ThomasWaldmann> ok, so we always give it. what do we do if somebody edits metadata and does not give it?
2009-06-20T17:47:48  <waldi> because it breaks the assumption, that every unknown text subtype can be handled as text/plain
2009-06-20T17:47:58  <ThomasWaldmann> a fallback to ascii would be a bit unexpected because moin content usually is utf-8 since some years
2009-06-20T17:48:35  <ThomasWaldmann> i am not talking about unknown subtypes, just about the charset
2009-06-20T17:48:43  <waldi> just add it if necessary. do we want to support saving in non-utf8?
2009-06-20T17:48:56  <ThomasWaldmann> i guess we need to
2009-06-20T17:48:57  <waldi> this is part of the text media type definition
2009-06-20T17:49:19  <ThomasWaldmann> e.g. if somebody uploads a source code file, it could be also some other encoding
2009-06-20T17:49:20  <waldi> http://tools.ietf.org/html/rfc2046#section-4.1.4
2009-06-20T17:50:17  <waldi> sure, but we can always convert the charset if we want during the save operation
2009-06-20T17:50:47  <ThomasWaldmann> and what if the subtype is known and the charset not?
2009-06-20T17:51:06  <waldi> it is ascii
2009-06-20T17:51:44  <ThomasWaldmann> that sounds like many tracebacks :=)
2009-06-20T17:53:10  <waldi> "As with all multipart MIME types, each part has an optional "Content-Type" header that defaults to "text/plain". User agents should supply the "Content-Type" header, accompanied by a "charset" parameter."
2009-06-20T17:53:20  <waldi> (html spec ...)
2009-06-20T17:53:47  <waldi> i think everyone except ie < 7 or so gets this right
2009-06-20T17:54:04  <ThomasWaldmann> ok, so when someone saves an item, we try to decode using the given charset first
2009-06-20T17:54:20  <ThomasWaldmann> when it fails, we reject the save?
2009-06-20T17:55:08  <ThomasWaldmann> i guess that would be the only way not to get invalid/inconsistent item revisions
2009-06-20T17:55:25  <waldi> for text/*, yes. everything else is handled as binary
2009-06-20T17:55:52  <ThomasWaldmann> yes, if course. there is no charset for non-text stuff.
2009-06-20T17:57:23  <ThomasWaldmann> later, we could maybe also try to autodetect at least ascii and utf-8
2009-06-20T18:00:03  *** amartani has joined #moin-dev
2009-06-20T18:06:29  <ThomasWaldmann> waldi: see Charset handling
2009-06-20T18:06:38  <ThomasWaldmann> (on the wiki page)
2009-06-20T18:07:47  <amartani> hi..
2009-06-20T18:07:57  <ThomasWaldmann> hi amartani
2009-06-20T18:09:55  <amartani> freenode blocked mibbit... is there any other decent irc web client?
2009-06-20T18:10:36  <dennda> they have their own one afaik
2009-06-20T18:12:17  <amartani> dennda: yes, I'm using it.. but it is much worse..
2009-06-20T18:13:38  <amartani> my connection is a bit unstable, and this client doesn't give any kind of warning then the connection is lost, nor it tries to reconnect..
2009-06-20T18:17:47  * ThomasWaldmann uses irssi within "screen" by ssh on a permanently connected machine
2009-06-20T18:18:07  <ThomasWaldmann> http://pastebin.com/f307cb0c2 < can someone explain the XXX?
2009-06-20T18:18:53  <waldi> yield have a return value?
2009-06-20T18:19:30  <ThomasWaldmann> yes, it is for generator.send()
2009-06-20T18:19:50  <dennda> d = yield? Oo
2009-06-20T18:20:04  <ThomasWaldmann> i am just wondering about that the initial .next() call does not send None to it
2009-06-20T18:20:16  <dennda> ah for send
2009-06-20T18:20:18  <dennda> i guess
2009-06-20T18:20:28  <ThomasWaldmann> that stuff is also new to me, never played with that yet
2009-06-20T18:20:49  <dennda> I took a look at it once but never practically used it
2009-06-20T18:20:53  <ThomasWaldmann> and IIRC that requires python 2.5
2009-06-20T18:21:24  <dennda> yeah it's a feature that was added recently i think
2009-06-20T18:21:36  <ThomasWaldmann> i am playing with it after the guy on #python suggested doing it
2009-06-20T18:22:17  <ThomasWaldmann> i even improved it a bit, I think the generator stack in the contenthandler is not needed, it can be done like I show it in the pastebin
2009-06-20T18:22:46  <waldi> ah. i think this was added for the email pacjage
2009-06-20T18:23:26  <ThomasWaldmann> the strange thing is, if you remove that dummy .next() call, it complains about getting something not None for the first call
2009-06-20T18:23:46  <ThomasWaldmann> but if you do that dummy .next() call, it never gets that None... X)
2009-06-20T18:23:55  <ThomasWaldmann> thus, I am wondering...
2009-06-20T18:24:09  <waldi> ThomasWaldmann: the None is catched by the first yield
2009-06-20T18:25:12  <ThomasWaldmann> waldi: but shouldn't the yield expression return it? as it does with all subsequent stuff?
2009-06-20T18:26:03  <waldi> http://docs.python.org/reference/expressions.html#grammar-token-yield_expression
2009-06-20T18:26:12  <waldi> the first one is needed to start the generator
2009-06-20T19:11:06  <amartani> when users are not logged in, they receive a different user id on every request?
2009-06-20T19:17:34  <ThomasWaldmann> amartani: could be
2009-06-20T19:18:07  <ThomasWaldmann> if you have anon sessions, maybe they receive the same (invalid) uid as long as the session lasts
2009-06-20T19:21:40  <amartani> ThomasWaldmann: is there any way to identify them then? I'm storing the users who are editing a page, but if they are anonymous, every sync request will look like comming from a new editor.
2009-06-20T19:22:25  <waldi> why don't you save editor instances in the browser?
2009-06-20T19:23:01  <waldi> or what do you want to do with this information?
2009-06-20T19:25:13  <amartani> waldi: i want to save the list of users editing a page, so you can see who contributed for each revision then the page is saved.
2009-06-20T19:25:53  <waldi> amartani: and why do you want to handle this different to the normal editor?
2009-06-20T19:28:57  <amartani> waldi: there will be many users editing the same page at the same time, and in the normal editor only the one who saved would get the credits..
2009-06-20T19:31:48  <waldi> i fail to see your point. a change is credited either with a username or in the case of an anonymous user with the ip
2009-06-20T19:35:46  <amartani> if A and B is editing the page, than A makes some change, the change is syncronized, and then B saves the page, than only B will get the credit for the change..
2009-06-20T20:03:21  <waldi> ThomasWaldmann: application/x-unknown needs to be moved as well. a non existent item have no valid type at all
2009-06-20T20:04:03  <ThomasWaldmann> waldi: that's just a hack to find the handler class
2009-06-20T20:05:32  <waldi> i know
2009-06-20T20:06:05  <waldi> "No module named plugin."
2009-06-20T20:06:06  <waldi> yeah
2009-06-20T20:06:42  <waldi> Could not import plugin package "/home/bastian/nobackup/moin/2.0-storage-dom-bblank/MoinMoin/_tests/wiki/data/plugin" because of ImportError:
2009-06-20T20:07:35  <waldi> 2.0-storage is not able to run any test
2009-06-20T20:12:54  <dennda> you need to create the folders it moans about
2009-06-20T20:12:58  <dennda> don't ask me why though
2009-06-20T20:13:20  <dennda> havn't investigated it. it works then
2009-06-20T20:14:20  <waldi> does not work#
2009-06-20T20:14:41  <dennda> then u r doin it rong
2009-06-20T20:14:45  <dennda> :-)
2009-06-20T20:14:49  <dennda> it moans twice
2009-06-20T20:14:53  <dennda> that bloody thing
2009-06-20T20:14:59  <waldi> then fix it
2009-06-20T20:15:36  <dennda> well, just create the folder manually for now
2009-06-20T20:15:57  <waldi> does not work. it wants a "module"
2009-06-20T20:16:10  <dennda> touch ....../__init__.py
2009-06-20T20:16:50  <dennda> module as in python module
2009-06-20T20:30:47  <waldi> I intend to use r'''(?:([-_0-9A-Za-z]+)=)?(?:(\w+)|"(.*?)(?<!\\)"|'(.*?)(?<!\\)')''' for the macro arguments for now. its rather greedy and accepts any non-word-character as delimiter between arguments
2009-06-20T21:23:57  <ThomasWaldmann> please use a re.X version
2009-06-20T22:02:43  <ThomasWaldmann> http://pastebin.com/f26f14960
2009-06-20T22:57:42  <waldi> sure
2009-06-20T23:18:11  <ThomasWaldmann> gna, xml.sax.handler.ContentHandler = old style
2009-06-20T23:34:10  <waldi> wht is new style?
2009-06-20T23:35:16  <ThomasWaldmann> hm?
2009-06-20T23:47:27  <CIA-38> Alexandre Martani <amartani AT gmail DOT com> default * 5938:21159e5bffed 2.0-storage-editor-amartani/MoinMoin/ (3 files in 3 dirs):
2009-06-20T23:47:27  <CIA-38> Real-time editor: - Store list of editors on cache
2009-06-20T23:47:27  <CIA-38> - List also anonymous editors
2009-06-20T23:47:27  <CIA-38> - Store full information of editors