2007-07-29T03:45:41 <grzywacz> heh, a few days without looking at my code and bugs are so apparent now ;-)
2007-07-29T03:46:00 <grzywacz> Taking a brake actually has its uses...
2007-07-29T03:46:06 <grzywacz> break ;S
2007-07-29T11:48:16 <lanius_> ThomasWaldmann: each backend, but not each filesystem backend
2007-07-29T11:48:35 <ThomasWaldmann> hm?
2007-07-29T11:48:49 <lanius_> this is generic code for a filesystem backend
2007-07-29T11:50:08 <ThomasWaldmann> i thought it was just generic code...
2007-07-29T11:50:16 <lanius_> no
2007-07-29T11:50:31 <lanius_> oh, i named it wrong
2007-07-29T11:50:39 <lanius_> no, i didn't
2007-07-29T11:50:45 <lanius_> fs_storage = filesystem_storage
2007-07-29T11:57:22 <xorAxAx> lanius_: the point is that encoding the filename is something that is backend-dependant
2007-07-29T12:15:49 <ThomasWaldmann> lanius_: AbstractStorage is not very abstract
2007-07-29T12:17:07 <ThomasWaldmann> but otoh it is maybe useful if one wants to do similar fs storage implementations, so another name would be the best
2007-07-29T12:22:31 <ThomasWaldmann> lanius_: AbstractData is using codecs.open and I think this is wrong
2007-07-29T12:26:40 <lanius_> xorAxAx: but i think it will be used for a new filesystem storage as well
2007-07-29T12:26:53 <lanius_> xorAxAx: every other storage is completly different and does not use this code
2007-07-29T12:26:59 <lanius_> xorAxAx: furthermore this option can be disabled
2007-07-29T12:27:22 <lanius_> ThomasWaldmann: AbstractStorage is very abstract since it does not implement all methods of the interface
2007-07-29T12:28:12 <xorAxAx> lanius_: it would be more sensible to offer callbacks
2007-07-29T12:28:20 <xorAxAx> than a boolean flag IMHO
2007-07-29T12:28:29 <xorAxAx> i.e. methods that you can override
2007-07-29T12:28:55 <lanius_> you can always override the methods?
2007-07-29T12:29:06 <xorAxAx> lanius_: then you dont need a flag :)
2007-07-29T12:29:17 <xorAxAx> lanius_: do they start with an underscore?
2007-07-29T12:29:32 <lanius_> ThomasWaldmann: AbstractData uses codecs.open because get_body in the old Page.py uses it as well
2007-07-29T12:29:37 <lanius_> xorAxAx: no they don't start with underscores
2007-07-29T12:29:41 <xorAxAx> ok
2007-07-29T12:30:13 <lanius_> ThomasWaldmann: for the Parser change, look at the complete changeset
2007-07-29T12:46:34 <ThomasWaldmann> lanius_: Pages are text, items are binaries (that includes encoded text and arbitrary files)
2007-07-29T12:49:57 <ThomasWaldmann> lanius_: what Parser change?
2007-07-29T12:50:12 <lanius_> < ThomasWaldmann> lanius: -from MoinMoin.parser.text_moin_wiki import Parser
2007-07-29T12:50:12 <lanius_> 17:50 < ThomasWaldmann> +from MoinMoin.parser.text import Parser
2007-07-29T12:51:12 <ThomasWaldmann> ah
2007-07-29T12:56:22 <ThomasWaldmann> lanius_: and why is it using parser.text (and why are you refactoring this stuff in a changeset commented as "remove more getPagePath"?)
2007-07-29T12:57:29 <lanius_> look at the line after that, it reimports Parser from MoinMoin.parser.text
2007-07-29T12:57:42 <ThomasWaldmann> that does not answer the question
2007-07-29T12:58:04 <lanius_> it answers the first one
2007-07-29T12:58:13 <ThomasWaldmann> no
2007-07-29T12:58:17 <lanius_> why not
2007-07-29T12:58:47 <ThomasWaldmann> parser.text is an extremely simple thing meant as a fallback for text/* mimetypes
2007-07-29T12:59:05 <ThomasWaldmann> why anyone would use it for macro testing is very unclear for me
2007-07-29T12:59:22 <ThomasWaldmann> (because that parser can't do macros)
2007-07-29T12:59:38 <lanius_> i don't know, but look at the code:
2007-07-29T12:59:42 <lanius_> first you do
2007-07-29T12:59:44 <lanius_> from MoinMoin.parser.text_moin_wiki import Parser
2007-07-29T12:59:45 <ThomasWaldmann> so my current assumption is that you refactored something wrong to be even more wrong
2007-07-29T12:59:51 <lanius_> then you do:
2007-07-29T12:59:52 <lanius_> from MoinMoin.parser.text import Parser
2007-07-29T12:59:56 <lanius_> and then use use Parser
2007-07-29T13:00:02 <lanius_> so you can skip the first one, can't you?
2007-07-29T13:00:16 <lanius_> ThomasWaldmann: and the unit test still works, so i can assume that i did correct things
2007-07-29T13:00:23 <ThomasWaldmann> if the second one is correct, then yes, else no
2007-07-29T13:00:49 <ThomasWaldmann> anyway, refactoring such stuff outside of storage should not be in such a changeset
2007-07-29T13:00:54 <lanius_> yes
2007-07-29T13:01:14 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2724:462ebf262bdc 1.7-storage-hwendel/MoinMoin/ (packages.py storage/fs_moin16.py): remove more getPagePath
2007-07-29T13:04:49 <lanius_> xorAxAx: can you take a look at the getPagePath in wikisync now?
2007-07-29T13:05:23 <xorAxAx> well, you can as well :)
2007-07-29T13:05:31 <xorAxAx> and i will review and test your changes later
2007-07-29T13:05:46 <xorAxAx> i have explained the usage 3 times already
2007-07-29T13:05:52 <xorAxAx> you know, the wiki sync tags ...
2007-07-29T13:06:51 <lanius_> i know, but you know the code better then me and it would be a good test for the current storage stuff if you just test it
2007-07-29T13:07:39 <xorAxAx> "just test it"? you mean the API?
2007-07-29T13:07:44 <lanius_> right
2007-07-29T13:07:56 <xorAxAx> well, i am pretty busy these days
2007-07-29T13:08:25 <ThomasWaldmann> lanius_: the parser.text change works and is correct. this is because it doesnt use the parser at all, because it directly instantiates the macro. but the original code definitely misses a comment about this "hack".
2007-07-29T13:08:50 <lanius_> ThomasWaldmann: ok, i didn't write the original code ;)
2007-07-29T13:12:06 <xorAxAx> lanius_: can you check whether the " # Clear caches" part is still correct in packages.py?
2007-07-29T13:12:35 <xorAxAx> besides that, it looks ok
2007-07-29T13:13:17 <lanius_> i don't know what it does, but it can't hurt ;)
2007-07-29T13:13:40 <xorAxAx> theoretically, it shouldnt be necessary
2007-07-29T13:13:49 <xorAxAx> hmm, where do you handle ACLs currently?
2007-07-29T13:14:32 <lanius_> acls for packages.py?
2007-07-29T13:15:10 <xorAxAx> no, generally
2007-07-29T13:15:21 <xorAxAx> in the storage layer
2007-07-29T13:15:45 <lanius_> currently you still have to do user.may.read(...)
2007-07-29T13:15:57 <lanius_> but external.Item does contain a acl property
2007-07-29T13:16:11 <lanius_> and checking against this property can be implemented there
2007-07-29T13:16:41 <xorAxAx> *sigh*
2007-07-29T13:16:42 <lanius_> it's just not done yet because i didn't wan't to change all user.may into catching exceptions
2007-07-29T13:16:50 <xorAxAx> didnt we say that the layer should do it transparently?
2007-07-29T13:17:01 <xorAxAx> ok
2007-07-29T13:17:02 <lanius_> sure and there is no problem implementing it
2007-07-29T13:36:34 <ThomasWaldmann> lanius_: + from MoinMoin.action.AttachFile import getAttachDir
2007-07-29T13:36:59 <ThomasWaldmann> maybe better globally import AttachFile and then use AttachFile.getAttachDir
2007-07-29T13:37:10 <lanius_> not possible, because of recursive imports
2007-07-29T13:37:24 <ThomasWaldmann> oh, ok
2007-07-29T17:43:04 <ThomasWaldmann> johill: your id test is failing when no 4suite is installed
2007-07-29T17:43:32 <ThomasWaldmann> (or rather "crashing")
2007-07-29T17:48:11 <ThomasWaldmann> see the formatter tests about how to do it for now
2007-07-29T17:59:56 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2580:c3741ed41050 1.7/MoinMoin/formatter/_tests/test_formatter.py: don't test the docbook formatter if there is no 4suite installed
2007-07-29T18:02:09 <ThomasWaldmann> fixed
2007-07-29T18:08:16 <lanius_> there is another failing test: TestFilters().testTextGeneric
2007-07-29T18:10:21 <ThomasWaldmann> that's in your branch, right? :)
2007-07-29T18:10:29 <lanius_> no, i also tested 1.7
2007-07-29T18:12:02 <ThomasWaldmann> tests finished: 334 passed, 17 skipped in 17.01 second for moin/1.7
2007-07-29T18:12:13 <xorAxAx> ThomasWaldmann: wrong OS
2007-07-29T18:12:43 <lanius_> assert u'asdf\r\r\nghjk' == u'asdf\r\nghjk'
2007-07-29T18:13:26 <xorAxAx> hah!
2007-07-29T18:13:30 <xorAxAx> :-)
2007-07-29T18:13:44 <xorAxAx> the former is a buggy string
2007-07-29T18:13:46 <xorAxAx> mixed line endings
2007-07-29T18:14:07 <xorAxAx> lanius_: that weird line ending converter is disabled?
2007-07-29T18:14:38 <lanius_> ?
2007-07-29T18:14:41 <lanius_> ah, yes
2007-07-29T18:16:55 <xorAxAx> ok :)
2007-07-29T19:29:32 <lanius_> ThomasWaldmann: writing the file in binary mode in make_file fixes it
2007-07-29T19:36:27 <lanius_> ThomasWaldmann: what about the action stuff in edit-log, should that be kept?
2007-07-29T19:47:25 <ThomasWaldmann> lanius_: action stuff?
2007-07-29T19:54:41 <ThomasWaldmann> lanius_: # Emulate edilock
2007-07-29T19:56:39 <ThomasWaldmann> lanius_: what do you mean by make_file?
2007-07-29T20:08:20 <lanius_> action=SAVE,SAVENEW, etc.
2007-07-29T20:08:47 <lanius_> the make_file method in filter/_tests/test_filter.py
2007-07-29T20:08:51 <lanius_> ThomasWaldmann:
2007-07-29T20:10:49 <ThomasWaldmann> that action info is needed for rc (esp. for renames)
2007-07-29T20:11:00 <ThomasWaldmann> also the extra stuff
2007-07-29T20:11:15 <lanius_> extra stuff?
2007-07-29T20:11:51 <ThomasWaldmann> the extra field (used for recording the old name, name of attachment etc.). feel free to clean it up.
2007-07-29T20:12:02 <lanius_> comment?
2007-07-29T20:12:06 <lanius_> clean it up?
2007-07-29T20:13:08 <ThomasWaldmann> no, not the comment field, the field (iirc) one before that
2007-07-29T20:14:46 <ThomasWaldmann> did you fix the make_file wb thing in your branch or shall I do it in main?
2007-07-29T20:15:08 <ThomasWaldmann> (the last would make more sense)
2007-07-29T20:15:10 <lanius_> i didn't, i don't know if this is the right fix ;)
2007-07-29T20:15:22 <lanius_> ah this field, is it really needed?
2007-07-29T20:17:35 <ThomasWaldmann> fixed
2007-07-29T20:17:39 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2581:67aeed6f77ce 1.7/MoinMoin/filter/_tests/test_filter.py: fix filter tests for win32 by using binary mode
2007-07-29T20:18:32 <ThomasWaldmann> lanius_: there is no other way to track a rename currently (or the filename of an attachment)
2007-07-29T20:18:38 <ThomasWaldmann> just look where it is used
2007-07-29T20:20:29 <ThomasWaldmann> btw, i guess all codecs use in fs_storage is wrong
2007-07-29T20:20:58 <ThomasWaldmann> this should be either handled on a higher level or be optional
2007-07-29T20:22:50 <ThomasWaldmann> and, if you offer decoding files there, you should use some "coding" metadata (that should be 'utf-8' ever for moin16, but could be anything else and taken from metadata for future storage)
2007-07-29T20:25:37 <ThomasWaldmann> s/utf-8/config.charset/
2007-07-29T20:30:08 <ThomasWaldmann> hmm, that "optional" encoding/decoding could be driven by checking the mimetype field for text/*
2007-07-29T20:34:22 <lanius_> ThomasWaldmann: so we need all those fields in the new code as well
2007-07-29T20:35:45 <ThomasWaldmann> it should be cleaned up (not putting everything into same metadata key)
2007-07-29T20:36:21 <ThomasWaldmann> so if the action is a rename, there could be some oldname key
2007-07-29T20:36:49 <ThomasWaldmann> and the attachfile stuff will have to get changed anyway
2007-07-29T21:27:30 <lanius_> ThomasWaldmann: the codecs stuff is correct for metadata and userbackend, isn't it?
2007-07-29T21:28:11 <ThomasWaldmann> yes
2007-07-29T21:28:29 <lanius_> but you would remove it from AbstractData?
2007-07-29T21:28:47 <ThomasWaldmann> well, it is not correct in general
2007-07-29T21:29:10 <ThomasWaldmann> you maybe didnt fall over it yet because you did not yet have binary files :)
2007-07-29T21:29:25 <lanius_> there won't be binary files for user stuff and metadata
2007-07-29T21:29:43 <ThomasWaldmann> no, but for items
2007-07-29T21:29:47 <xorAxAx> abstractdata?!?!?
2007-07-29T21:29:49 <lanius_> yes
2007-07-29T21:29:52 <xorAxAx> data is binary, no unicode involved
2007-07-29T21:30:11 <lanius_> so i will remove it from the data and move it up to the Page level?
2007-07-29T21:30:47 <xorAxAx> umm, yes
2007-07-29T21:30:54 <xorAxAx> or offer a wrapper
2007-07-29T21:31:08 <xorAxAx> that calls codecs open etc.
2007-07-29T21:31:21 <lanius_> wrapper, how?
2007-07-29T21:31:26 <xorAxAx> no diea
2007-07-29T21:31:28 <xorAxAx> idea
2007-07-29T21:31:35 <lanius_> ^^
2007-07-29T21:33:11 <ThomasWaldmann> it should be at a place where metadata is available
2007-07-29T21:33:26 <ThomasWaldmann> so it can do the right thing after looking into metadata
2007-07-29T21:33:46 <lanius_> but currently we don't have a mimetype metadata yet
2007-07-29T21:35:47 <xorAxAx> the key is called encoding, not mimetype :)
2007-07-29T21:35:52 <ThomasWaldmann> yeah, because you didnt need it for 1.6
2007-07-29T21:36:05 <ThomasWaldmann> xorAxAx: no, the key is called coding :P
2007-07-29T21:36:26 <ThomasWaldmann> but it will be only used if mimetype is text/*
2007-07-29T21:37:24 <lanius_> so i will move it to Page.py for now
2007-07-29T21:37:53 <xorAxAx> ThomasWaldmann: ok
2007-07-29T21:39:38 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2726:fe0440a46160 1.7-storage-hwendel/MoinMoin/formatter/_tests/test_formatter.py: merge main
2007-07-29T21:39:41 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2727:39119b2c9505 1.7-storage-hwendel/MoinMoin/ (Page.py PageEditor.py storage/fs_storage.py): move codecs stuff to the right place
2007-07-29T21:40:12 <xorAxAx> ThomasWaldmann: no, its called encoding :)
2007-07-29T21:40:50 <ThomasWaldmann> xorAxAx: it looks a bit strange if you do .decode(encoding)
2007-07-29T21:41:16 <ThomasWaldmann> and python also uses coding: btw :)
2007-07-29T21:41:26 <xorAxAx> ThomasWaldmann: welll
2007-07-29T21:41:32 <xorAxAx> thats emacs
2007-07-29T21:42:11 <xorAxAx> http://en.wikipedia.org/wiki/Character_encoding
2007-07-29T21:46:23 <ThomasWaldmann> emacs?
2007-07-29T21:46:49 <xorAxAx> yes, the encoding declaration was stolen from emacs
2007-07-29T21:48:46 <ThomasWaldmann> ok, anyway, doesnt matter that much. coding would be less to type also.
2007-07-29T21:48:58 <lanius_> ^^
2007-07-29T21:52:02 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2728:dfc265db6be9 1.7-storage-hwendel/MoinMoin/ (4 files in 3 dirs): add missing edit-log metadata
2007-07-29T21:55:02 <lanius_> ThomasWaldmann: anymore comments about the current code? :-)
2007-07-29T21:56:32 <ThomasWaldmann> values[1] and [3] is what?
2007-07-29T21:57:15 <lanius_> 1 is the revno and 3 is the pagename
2007-07-29T21:57:31 <xorAxAx> where is that?
2007-07-29T21:57:40 <lanius_> nowhere
2007-07-29T21:57:51 <xorAxAx> ThomasWaldmann: i cant find any usage of coding on the net
2007-07-29T21:58:53 <ThomasWaldmann> revno and itemname. for an item's metadata, this is redundant, but not for global edit-log
2007-07-29T21:59:59 <ThomasWaldmann> xorAxAx: http://en.wikipedia.org/wiki/Coding
2007-07-29T22:00:15 <lanius_> ThomasWaldmann: right
2007-07-29T22:00:17 <xorAxAx> lanius_: + newrev.data.write(codecs.encode(text, config.charset))
2007-07-29T22:00:24 <xorAxAx> lanius_: please use the str method
2007-07-29T22:00:49 <xorAxAx> umm, unicode method
2007-07-29T22:00:53 <xorAxAx> unicode.encode
2007-07-29T22:01:09 <xorAxAx> ThomasWaldmann: thats something different
2007-07-29T22:01:51 <xorAxAx> ThomasWaldmann: thats on a different level - a coding is used on the wires and an encoding is used for characters
2007-07-29T22:01:52 <lanius_> xorAxAx: unicode? i just changed it to text.encode(config.charset)
2007-07-29T22:02:06 <xorAxAx> lanius_: yes, text is an unicode object (hopefully)
2007-07-29T22:02:27 <lanius_> xorAxAx: well, i do text.decode(config.charset) before
2007-07-29T22:02:32 <xorAxAx> (as the wikipedia page says)
2007-07-29T22:02:38 <xorAxAx> lanius_: hmm?
2007-07-29T22:02:42 <xorAxAx> lanius_: the same object?
2007-07-29T22:02:48 <lanius_> to parse it
2007-07-29T22:02:48 <xorAxAx> lanius_: that doesnt make sense
2007-07-29T22:02:53 <lanius_> in Page.py get_body
2007-07-29T22:03:14 <xorAxAx> then why is it called text there?
2007-07-29T22:03:23 <xorAxAx> in get_body
2007-07-29T22:03:33 <xorAxAx> data would be more appropriate :)
2007-07-29T22:04:06 <lanius_> yeah
2007-07-29T22:04:18 <xorAxAx> but thats nitpicking :)
2007-07-29T22:04:25 <ThomasWaldmann> zenhase: how's cleanup going?
2007-07-29T22:04:27 <lanius_> riiiight :D
2007-07-29T22:04:49 <xorAxAx> lanius_: i was just confused because you jumped to a totally different place in the code base :)
2007-07-29T22:05:07 <lanius_> moinmoin codebase is not too small ;)
2007-07-29T22:06:23 <ThomasWaldmann> your mission is to make the non-modular code part smaller :D
2007-07-29T22:06:44 <ThomasWaldmann> (and cleaner)
2007-07-29T22:08:28 <xorAxAx> lanius_: well, i am not sure if you should pass extra as EXTRA
2007-07-29T22:08:36 <xorAxAx> lanius_: maybe you can abstract it
2007-07-29T22:09:00 <xorAxAx> its ill-defined currently
2007-07-29T22:09:31 <lanius_> maybe at the layer above
2007-07-29T22:10:13 <xorAxAx> not sure how you would roundtrip it ... you have to check where it is used how
2007-07-29T22:10:21 <ThomasWaldmann> maybe keep extra as is until after attachment refactoring
2007-07-29T22:10:21 <xorAxAx> and introduce fields that are filled depending on the action
2007-07-29T22:10:23 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2728:fd3fb23438ab 1.7-maninfo-vpv/MoinMoin/action/SisterDiff.py: Fix SisterDiff action to work with the new namespace hierarchy after info importing support got introduced
2007-07-29T22:10:28 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2729:25d5898eb29d 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: Simplify filename handling a bit
2007-07-29T22:10:34 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2732:135b97002685 1.7-maninfo-vpv/MoinMoin/ (2 files in 2 dirs): merge from main
2007-07-29T22:29:29 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2729:a3ffe7a6622d 1.7-storage-hwendel/MoinMoin/ (Page.py PageEditor.py): nitpicking
2007-07-29T22:33:39 <xorAxAx> funny but nitpickingly viewn, incomplete comment
2007-07-29T22:39:03 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2121:6509d58b3f18 1.6/wiki/server/ (moin.wsgi moin_flup_wsgi.py moinwsgi.py): add mod_wsgi driver script, rename flup/wsgi based driver script
2007-07-29T22:39:06 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2122:1303ea4f1ee0 1.6/ (MoinMoin/macro/Action.py wiki/server/moinmodpy.htaccess): merge main - and finally fix moinmodpy.htaccess
2007-07-29T23:45:32 * neagulm is bored of calibrating the whole the ANN
2007-07-29T23:45:36 <neagulm> good night
2007-07-29T23:50:48 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2733:ea6a4e2d394f 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: Save information on packages whose info files are handled
MoinMoin: MoinMoinChat/Logs/moin-dev/2007-07-29 (last edited 2007-10-29 19:20:39 by localhost)
