2009-08-04T01:16:21 *** amartani has joined #moin-dev
2009-08-04T01:16:22 <xorAxAx> amartani: appended my comments to the projects page. can you clean up the latter? (This message has been postponed.)
2009-08-04T01:16:25 <xorAxAx> amartani: http://paste.pocoo.org/show/132335/ (This message has been postponed.)
2009-08-04T01:27:48 <xorAxAx> amartani: did you see my comments on the project page?
2009-08-04T01:29:20 <amartani> xorAxAx: yes, thanks, I'm looking them
2009-08-04T01:29:46 <xorAxAx> ok
2009-08-04T01:31:26 <amartani> can I commit directly to storage branch?
2009-08-04T01:49:45 *** grzywacz has quit IRC
2009-08-04T01:50:14 *** tpfennig has quit IRC
2009-08-04T01:53:11 <xorAxAx> amartani: you can commit to it, but you probably cannot push to the repo :)
2009-08-04T01:53:39 <xorAxAx> in any case, that should not be a reason to avoid committing to it because you can always send bundles
2009-08-04T01:53:45 <xorAxAx> to e.g. dennda
2009-08-04T02:02:31 <amartani> xorAxAx: ok, I'll send him, but I think only caching can be commited on main
2009-08-04T02:04:23 <xorAxAx> i think you dont need to do it retrospectively
2009-08-04T02:04:28 <xorAxAx> just remember it for the next time
2009-08-04T02:04:38 <xorAxAx> and work on the code to be good enough to be merged :)
2009-08-04T02:04:49 <xorAxAx> into the main 2.0 branch
2009-08-04T02:05:06 <xorAxAx> (or, more precisely, storage branch)
2009-08-04T02:05:46 <amartani> sure
2009-08-04T02:09:21 <amartani> about the reviewed files: they were redoing the acls (on storage branch), so I skipped that part
2009-08-04T02:09:52 <amartani> on ajax_modify, I did as it is done on modify, with most of code on items package
2009-08-04T02:11:52 <xorAxAx> so thats cleaned up nowadays?
2009-08-04T02:12:10 <amartani> I think so
2009-08-04T02:14:17 <amartani> I have put a config variable to disable mobwrite, and ThomasWaldmann thinks it is better to not display that message, and just document that on help
2009-08-04T02:25:24 <xorAxAx> is it already in the code?
2009-08-04T02:25:35 <xorAxAx> because i still saw unconditional code in the lock module
2009-08-04T02:25:42 <xorAxAx> amartani:
2009-08-04T03:50:31 <amartani> xorAxAx: not there yet, will be soon
2009-08-04T05:47:37 <CIA-43> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6369:4bc6e9dc82dd 2.0-storage/MoinMoin/storage/backends/hg.py:
2009-08-04T05:47:37 <CIA-43> hg storage backend: use explicit lock.release() to get rid of DeprecationWarning
2009-08-04T05:47:37 <CIA-43> Mercurial 1.3.1 emits a DeprecationWarning if you use "del lock",
2009-08-04T05:47:37 <CIA-43> it rather wants "lock.release()".
2009-08-04T05:47:37 <CIA-43> The release() method existed in the lock class since ever, so
2009-08-04T05:47:41 <CIA-43> calling lock.release() is no problem even for older mercurial versions.
2009-08-04T05:48:21 <ThomasWaldmann> PawelPac1na: please check ^^
2009-08-04T05:57:09 <ThomasWaldmann> dennda: http://hg.moinmo.in/moin/2.0-storage/rev/f96578042635#l3.16 there must not be a Config attr just called "default" (or "before" or "after"), because acl context is not given there.
2009-08-04T05:57:50 <ThomasWaldmann> (this is fine in the context of initializing and acl wrapper, but not without acl context)
2009-08-04T05:58:03 <ThomasWaldmann> s/and/an/
2009-08-04T06:05:07 <ThomasWaldmann> dennda: http://hg.moinmo.in/moin/2.0-storage/rev/59e6ce3b0e02#l1.12 se
2009-08-04T06:05:31 <ThomasWaldmann> "separate" (like in german)
2009-08-04T06:07:36 <ThomasWaldmann> dennda: http://hg.moinmo.in/moin/2.0-storage/rev/9ee14a4ecf33 you are doing unclean changesets with completely unrelated things
2009-08-04T06:17:39 <ThomasWaldmann> dennda: please refactor enduser.py -> _create_folders (note: os.makedirs exists)
2009-08-04T06:20:01 <ThomasWaldmann> be less verbose, must not contain more than 5 lines
2009-08-04T06:22:12 <ThomasWaldmann> please add support for Trash/ in separate backend
2009-08-04T06:24:52 <ThomasWaldmann> and fix the docstring afterwards
2009-08-04T06:32:16 <ThomasWaldmann> dennda: if you iterate over the namespace_mapping, do use sane names instead of line, line[0], line[1]...
2009-08-04T06:40:13 <CIA-43> Alexandre Martani <amartani AT gmail DOT com> default * 6478:3ce5b65aeb56 2.0-storage-editor-amartani/MoinMoin/items/lock.py: Real-time editor: on itemlock, change message to not offer synchronized editing. call update_editor even if lock is disabled.
2009-08-04T06:40:15 <CIA-43> Alexandre Martani <amartani AT gmail DOT com> default * 6479:a3d7228eb8b9 2.0-storage-editor-amartani/MoinMoin/ (items/__init__.py templates/modify_text.html): Real-time editor: display last save also without javascript
2009-08-04T06:40:17 <CIA-43> Alexandre Martani <amartani AT gmail DOT com> default * 6480:a75183b6e852 2.0-storage-editor-amartani/MoinMoin/support/ (6 files in 2 dirs): Add simplejson to MoinMoin/support
2009-08-04T06:40:22 <CIA-43> Alexandre Martani <amartani AT gmail DOT com> default * 6481:a852855803c1 2.0-storage-editor-amartani/MoinMoin/web/static/htdocs/common/js/jquery-1.3.2.js: Add jquery javascript library
2009-08-04T06:40:25 <CIA-43> Alexandre Martani <amartani AT gmail DOT com> default * 6482:503ecd5d53e3 2.0-storage-editor-amartani/MoinMoin/theme/__init__.py: theme: include jquery script
2009-08-04T06:40:28 <ThomasWaldmann> dennda: everything using cfg.storage still looks like needing to get fixed
2009-08-04T06:40:34 <CIA-43> Alexandre Martani <amartani AT gmail DOT com> default * 6483:ecf24d698319 2.0-storage-editor-amartani/MoinMoin/ (2 files in 2 dirs): Real-time editor: ajax_info action writes json
2009-08-04T06:43:26 <ThomasWaldmann> dennda: because if you provide the namespace_mapping and derive everything from it (request.storage and request.unprotected_storage), then cfg.storage is rather superfluous, right?
2009-08-04T06:44:36 <ThomasWaldmann> and your "enduser" helper maybe should just help to create the namespace_mapping
2009-08-04T06:45:20 <ThomasWaldmann> dennda: please update diary, daily
2009-08-04T07:22:20 <ThomasWaldmann> dennda: kill that threading stuff you added yesterday. after starting current code and trying to edit a (new) page, the wikiserver process completely hangs and not even ctrl-c works any more.
2009-08-04T07:24:02 <ThomasWaldmann> amartani: editing is rather borked right now
2009-08-04T07:24:32 <ThomasWaldmann> and it often saves stuff with comment "deleted" although size shows it is not.
2009-08-04T07:25:08 <ThomasWaldmann> did something go wrong with your last main repo merge?
2009-08-04T07:29:26 <ThomasWaldmann> and it sets the deleted: True metadata
2009-08-04T07:32:34 <ThomasWaldmann> and "enable sharing" brings back the living dead
2009-08-04T07:37:46 <ThomasWaldmann> amartani: please note that deletion works different now in 2.0-storage (changed some days ago):
2009-08-04T07:38:25 <ThomasWaldmann> instead of setting the deleted meta to True and saving an empty revision, we have the trashbin approach now
2009-08-04T07:39:06 <ThomasWaldmann> to delete an item, it is just renamed to "Trash/itemname (deletion timestamp)"
2009-08-04T07:40:32 <ThomasWaldmann> thus, saving empty revision data now is legal, does not delete the item and does not do anything special any more. deleted meta flag is gone.
2009-08-04T08:31:37 <ThomasWaldmann> dennda: wake up :)
2009-08-04T08:44:20 <dreimark> dimazest: ping
2009-08-04T08:45:20 <CIA-43> Reimar Bauer <rb.proj AT googlemail DOT com> default * 4355:efd8087a3062 1.8/MoinMoin/action/diff.py: action diff: enable previous / next button only in the range of given revisions
2009-08-04T08:50:43 <dennda> ThomasWaldmann: I am awake
2009-08-04T08:51:35 <ThomasWaldmann> dennda: moin :) just read the backlog for some feedback.
2009-08-04T08:53:36 <dennda> will do
2009-08-04T08:53:53 <dennda> need to take care of plumber, just a sec
2009-08-04T09:19:36 *** tpfennig has joined #moin-dev
2009-08-04T09:22:36 <dennda> ThomasWaldmann: Thanks. About that default thing: I wasn't sure how to do it, since namespace must be defined in the config which itself needs to provide something as acl protection. (If I pass "", tests won't be happy because they cannot operate any longer)
2009-08-04T09:25:37 <dennda> .oO(suggestions for a better term than 'line'?)
2009-08-04T09:28:17 * dreimark vreakfast
2009-08-04T09:28:23 <dreimark> s/v/b/
2009-08-04T09:41:01 <dennda> Why exactly are we not using the werkzeug debugger for development, btw?
2009-08-04T09:41:20 <dennda> http://werkzeug.pocoo.org/documentation/dev/debug.html#using-the-debugger
2009-08-04T09:41:32 <dennda> may I enable it?
2009-08-04T09:45:56 *** devilsadvocate has quit IRC
2009-08-04T09:47:46 <dreimark> dennda: see the server config there is a var
2009-08-04T09:48:46 <dreimark> wikiserverconfig.py
2009-08-04T09:48:52 <dreimark> # How to debug? Your options:
2009-08-04T09:49:21 <dennda> oh
2009-08-04T09:49:23 <dennda> havn't seen that
2009-08-04T09:49:32 <dreimark> you may want to do it in your wikiserverconfig_local
2009-08-04T09:54:19 *** devilsadvocate has joined #moin-dev
2009-08-04T09:56:03 <dennda> ah great
2009-08-04T09:56:14 <dennda> thanks
2009-08-04T09:59:22 *** devilsadvocate has joined #moin-dev
2009-08-04T10:22:35 *** devilsadvocate has quit IRC
2009-08-04T10:22:42 <dennda> http://mercurial.selenic.com/wiki/ColorExtension
2009-08-04T10:22:43 <dennda> nice
2009-08-04T10:26:21 *** devilsadvocate has joined #moin-dev
2009-08-04T10:32:14 <ThomasWaldmann> dennda: not line, but ns, backend, acls
2009-08-04T10:33:05 * dennda fscked his repo
2009-08-04T10:33:11 <dennda> s/repo/branch/
2009-08-04T10:37:57 <dennda> ThomasWaldmann: how would you provide the acl parm in the namespace mapping of the test config if that's not yet present?
2009-08-04T10:38:49 *** tpfennig has quit IRC
2009-08-04T10:39:36 *** devilsadvocate has quit IRC
2009-08-04T10:54:39 <CIA-43> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6370:1528921def1b 2.0-storage/MoinMoin/wsgiapp.py: storage: Throw out threading for cfg.preloaded_xml again. It caused problems. The first request to your freshly started wikiserver will be potentially slow if this option is enabled.
2009-08-04T10:54:41 <CIA-43> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6371:c74c346b1f5c 2.0-storage/MoinMoin/wsgiapp.py: storage: line -> mapping + comments&docstrings
2009-08-04T10:54:43 <CIA-43> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6372:e3d28aa4684c 2.0-storage/wikiconfig.py: storage: use separate backend for trash
2009-08-04T10:54:45 <CIA-43> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6373:a15ce6cedd90 2.0-storage/wikiconfig.py: storage: fix namespace_mapping comments
2009-08-04T10:54:46 <CIA-43> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6374:3a299e3050e6 2.0-storage/MoinMoin/storage/backends/hg.py: storage: merge remote 2.0-storage
2009-08-04T10:55:07 <dennda> NOTE: mkdir instance/trash
2009-08-04T10:55:18 * dennda adjusts enduser
2009-08-04T10:56:48 <dennda> ThomasWaldmann: or was it just the name "default" that annoyed you?
2009-08-04T10:58:27 *** devilsadvocate has joined #moin-dev
2009-08-04T11:03:36 <CIA-43> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6375:5ab2bef8cc86 2.0-storage/MoinMoin/wsgiapp.py: storage: make processing of namespace_mapping more readable
2009-08-04T11:05:36 *** devilsadvocate has quit IRC
2009-08-04T11:09:22 * dennda thinks about how enduser should be like now
2009-08-04T11:22:47 *** devilsadvocate has joined #moin-dev
2009-08-04T11:29:05 * dreimark wonders if I have ever seen a different diff as by "if charobj.ratio() < 0.5:" from diff_html
2009-08-04T11:29:32 <dreimark> that ratio seems always close to 0 for me
2009-08-04T11:29:44 <dreimark> what is the usecase of that else?
2009-08-04T11:30:09 *** devilsadvocate has joined #moin-dev
2009-08-04T11:31:26 <dreimark> and if that is wanted/needed why don't we call quick_ratio()
2009-08-04T11:32:17 <dreimark> or real_quick_ratio()
2009-08-04T11:37:57 *** devilsadvocate has quit IRC
2009-08-04T11:41:39 *** devilsadvocate has joined #moin-dev
2009-08-04T11:46:43 <ThomasWaldmann> dennda: as I wrote, it is just about the bad naming
2009-08-04T11:49:17 <dennda> ok, I got that wrong then, sorry
2009-08-04T11:51:15 <ThomasWaldmann> 1.8 + unprotected_mapping = [(ns, backend) for ns, backend, acls in [mapping for mapping in ns_mapping]]
2009-08-04T11:51:22 <ThomasWaldmann> WTF?
2009-08-04T11:52:08 *** devilsadvocate has quit IRC
2009-08-04T11:53:43 <dennda> urghs
2009-08-04T11:54:21 <dennda> well yeah that was my brother, he's 11. forgive him... /me runs
2009-08-04T11:55:08 *** devilsadvocate has joined #moin-dev
2009-08-04T12:02:24 *** devilsadvocate has quit IRC
2009-08-04T12:02:49 *** devilsadvocate has joined #moin-dev
2009-08-04T12:03:32 *** devilsadvocate has joined #moin-dev
2009-08-04T12:06:40 *** devilsadvocate has quit IRC
2009-08-04T12:09:22 *** devilsadvocate has joined #moin-dev
2009-08-04T12:12:30 *** devilsadvocate has quit IRC
2009-08-04T12:12:48 <ThomasWaldmann> dennda: did you read 05:48+?
2009-08-04T12:14:16 <dennda> yes sure
2009-08-04T12:16:22 *** devilsadvocate has joined #moin-dev
2009-08-04T12:17:14 <dreimark> anyone with a windows box around?
2009-08-04T12:17:40 *** devilsadvocate has quit IRC
2009-08-04T12:17:41 <dreimark> http://moinmo.in/4ct10n/diff/ReimarBauer/HelpOnArnica close to the bottom cache needs to be debugged
2009-08-04T12:18:44 *** devilsadvocate has joined #moin-dev
2009-08-04T12:19:55 <ThomasWaldmann> <silence>
2009-08-04T12:27:53 *** devilsadvocate has quit IRC
2009-08-04T12:30:04 *** devilsadvocate has joined #moin-dev
2009-08-04T12:33:03 <dennda> <wuestenbuschvorbeiflieg>
2009-08-04T12:35:02 <ThomasWaldmann> hehe
2009-08-04T12:35:28 *** devilsadvocate has quit IRC
2009-08-04T12:38:27 <dennda> ThomasWaldmann: How would I reference test_items.xml from a test case so it works independently of my pwd?
2009-08-04T12:42:28 *** devilsadvocate has joined #moin-dev
2009-08-04T12:42:54 *** devilsadvocate has quit IRC
2009-08-04T12:43:34 *** devilsadvocate has joined #moin-dev
2009-08-04T12:44:26 <ThomasWaldmann> i already told you the way
2009-08-04T12:44:47 *** devilsadvocate has quit IRC
2009-08-04T12:45:48 <dennda> i.e. it needs to be relative to the file?
2009-08-04T12:45:53 <dennda> i'll try
2009-08-04T12:46:22 <ThomasWaldmann> if you just have 1 or 2 xml files, one for filling the dev playing wiki and one to run the tests,
2009-08-04T12:46:58 <ThomasWaldmann> you just compute the path to them once and assign it to some attr (e.g. in cfg)
2009-08-04T12:47:20 <ThomasWaldmann> using __file__ (grep for it)
2009-08-04T12:48:34 *** devilsadvocate has joined #moin-dev
2009-08-04T12:49:10 <ThomasWaldmann> devilsadvocate: can you make your irc client not connect/disconnect all 5 mins?
2009-08-04T12:50:17 *** devilsadvocate has quit IRC
2009-08-04T12:53:54 *** devilsadvocate has joined #moin-dev
2009-08-04T12:54:52 *** devilsadvocate has quit IRC
2009-08-04T13:06:53 *** devilsadvocate has joined #moin-dev
2009-08-04T13:07:52 *** devilsadvocate has left #moin-dev
2009-08-04T13:08:04 *** dreimark has quit IRC
2009-08-04T13:08:04 *** PawelPac1na has quit IRC
2009-08-04T13:08:04 *** franklin has quit IRC
2009-08-04T13:08:04 *** dennda has quit IRC
2009-08-04T13:08:04 *** amartani has quit IRC
2009-08-04T13:08:04 *** ThomasWaldmann has quit IRC
2009-08-04T13:10:18 *** amartani has joined #moin-dev
2009-08-04T13:10:18 *** dreimark has joined #moin-dev
2009-08-04T13:10:18 *** PawelPac1na has joined #moin-dev
2009-08-04T13:10:18 *** franklin has joined #moin-dev
2009-08-04T13:10:18 *** dennda has joined #moin-dev
2009-08-04T13:10:18 *** ThomasWaldmann has joined #moin-dev
2009-08-04T13:15:38 *** ChanServ has quit IRC
2009-08-04T13:15:38 *** TheSheep has quit IRC
2009-08-04T13:16:01 *** ChanServ has joined #moin-dev
2009-08-04T13:16:01 *** irc.freenode.net sets mode: +o ChanServ
2009-08-04T13:17:26 *** TheSheep has joined #moin-dev
2009-08-04T13:22:49 <dennda> hm I did the setup in wsgiapp wrong
2009-08-04T13:23:08 *** amartani has quit IRC
2009-08-04T13:24:09 *** ChanServ has quit IRC
2009-08-04T13:24:09 *** y84 has quit IRC
2009-08-04T13:24:45 *** ChanServ has joined #moin-dev
2009-08-04T13:24:45 *** y84 has joined #moin-dev
2009-08-04T13:24:45 *** irc.freenode.net sets mode: +o ChanServ
2009-08-04T13:25:24 *** dreimark has quit IRC
2009-08-04T13:25:24 *** TheSheep has quit IRC
2009-08-04T13:25:24 *** franklin has quit IRC
2009-08-04T13:25:24 *** PawelPac1na has quit IRC
2009-08-04T13:25:24 *** dennda has quit IRC
2009-08-04T13:25:24 *** ThomasWaldmann has quit IRC
2009-08-04T13:26:26 *** TheSheep has joined #moin-dev
2009-08-04T13:26:26 *** dreimark has joined #moin-dev
2009-08-04T13:26:26 *** PawelPac1na has joined #moin-dev
2009-08-04T13:26:26 *** franklin has joined #moin-dev
2009-08-04T13:26:26 *** ThomasWaldmann has joined #moin-dev
2009-08-04T13:26:26 *** dennda has joined #moin-dev
2009-08-04T13:27:06 <ThomasWaldmann> dennda: > ConfigurationError: OSError: [Errno 2] No such file or directory: 'instance/trash/name-mapping.lock'
2009-08-04T13:27:25 <ThomasWaldmann> please use absolute pathes for instance dir, made the same way as for the xml file
2009-08-04T13:27:58 <dennda> 10:55 < dennda> NOTE: mkdir instance/trash
2009-08-04T13:28:07 <dennda> yes, that will work as soon as enduser is redone
2009-08-04T13:28:24 <dennda> however i've a more serious problem at the moment
2009-08-04T13:30:02 <dennda> I put init_unprotected_storage in init. but as already the old docstring of init_unprotected_storage suggests, both should be separate so that conftest can create fresh backends after init. i am not sure how to do that properly.
2009-08-04T13:32:02 <ThomasWaldmann> i suggest you first fix the easy bugs, so the test (failure) output gets shorter
2009-08-04T13:32:40 <ThomasWaldmann> and if a change makes trouble, try not doing it :)
2009-08-04T13:33:32 <dennda> well, ok, but a few tests won't work due to this
2009-08-04T13:34:24 <ThomasWaldmann> is that needed for xml preload in tests?
2009-08-04T13:35:23 <dennda> yes
2009-08-04T13:35:38 <ThomasWaldmann> why?
2009-08-04T13:36:07 <ThomasWaldmann> i mean, if xml preload works for the normal wikiserver dev wiki, why doesn't it work for tests?
2009-08-04T13:37:02 <ThomasWaldmann> btw, is main repo current or do you have unpushed commits?
2009-08-04T13:37:38 <CIA-43> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6376:77a04608f802 2.0-storage/MoinMoin/_tests/ (data/README wiki/data/plugin/__init__.py): storage: Add empty folders for tests
2009-08-04T13:37:39 <CIA-43> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6377:bce5a4bbbe01 2.0-storage/MoinMoin/ (3 files in 3 dirs): storage: Start cleanup of test's wikiconfig
2009-08-04T13:37:41 <CIA-43> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6378:7242c10c8eac 2.0-storage/MoinMoin/wsgiapp.py: storage: don't drink and code. remove redundant LC
2009-08-04T13:37:43 <CIA-43> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6379:8d9646d53c73 2.0-storage/MoinMoin/_tests/wikiconfig.py: storage: oops. forgot 'admin'. adding it.
2009-08-04T13:37:51 <dennda> it's up to date :-)
2009-08-04T13:37:56 <ThomasWaldmann> ok
2009-08-04T13:40:50 <ThomasWaldmann> why does tests/wikiconfig still have this?:
2009-08-04T13:40:51 <ThomasWaldmann> self.test_num_pages = len(clone(flatfile.FlatFileBackend(self.flat_dir), self.storage)[0])
2009-08-04T13:41:18 <ThomasWaldmann> that directory is empty, btw
2009-08-04T13:42:30 <dennda> Isn't that related to the thing I'm working on? (using test_items.xml in the tests in need of it)
2009-08-04T13:44:36 <ThomasWaldmann> well, if it does do nothing, why is it there? :)
2009-08-04T13:46:41 <ThomasWaldmann> hmm, somehow my workdir didn't update, ignore that
2009-08-04T13:47:25 <ThomasWaldmann> you forgot to remove flat_dir
2009-08-04T13:47:52 <dennda> thanks
2009-08-04T13:53:26 <ThomasWaldmann> i am wondering a bit about AccessDeniedError for READ
2009-08-04T13:54:38 <dennda> hierarchic by chance?
2009-08-04T13:55:49 <dennda> hm doesn't look like
2009-08-04T13:58:10 <ThomasWaldmann> ah, ok, test_security needs Config fixing, it uses the old settings
2009-08-04T13:59:01 <dennda> I'm at it
2009-08-04T13:59:03 <ThomasWaldmann> btw, AccessDeniedErrors could tell for which username they reject access
2009-08-04T13:59:48 <dennda> they do?
2009-08-04T13:59:56 <dennda> > AccessDeniedError: WikiAdmin may not read 'AclTestMainItem'.
2009-08-04T14:01:33 <ThomasWaldmann> ehrm %) ok, ignore that. :)
2009-08-04T14:01:41 *** starshine_away has quit IRC
2009-08-04T14:01:41 *** vpv has quit IRC
2009-08-04T14:01:41 *** xorAxAx has quit IRC
2009-08-04T14:01:41 *** CIA-43 has quit IRC
2009-08-04T14:01:41 *** Submundo_____ has quit IRC
2009-08-04T14:02:07 <dennda> actually I don't think the old-style acl declaration is what's causing the problem
2009-08-04T14:02:07 <dennda> hm
2009-08-04T14:02:07 <ThomasWaldmann> i was rather thinking whether that request.user patchery in test_security can work that way
2009-08-04T14:02:08 <dennda> i'm thinking the same
2009-08-04T14:03:02 <ThomasWaldmann> but as it tells it is for WikiAdmin, there is some different problem
2009-08-04T14:03:06 *** Submundo_____ has joined #moin-dev
2009-08-04T14:03:06 *** starshine_away has joined #moin-dev
2009-08-04T14:03:06 *** CIA-43 has joined #moin-dev
2009-08-04T14:03:06 *** xorAxAx has joined #moin-dev
2009-08-04T14:03:06 *** vpv has joined #moin-dev
2009-08-04T14:03:09 <dennda> at the point where that user is patched in, the backends have already been protected (and thus use the old user)
2009-08-04T14:03:09 <ThomasWaldmann> esp. denial of read is rather strange
2009-08-04T14:03:25 <ThomasWaldmann> ah, so the acls are ""?
2009-08-04T14:03:25 <dennda> that's just the first acl to be checked since create_item tries to get the item first before creating it I assume
2009-08-04T14:03:31 <ThomasWaldmann> that would explain it :)
2009-08-04T14:03:55 <dennda> or Item.create, for that matter
2009-08-04T14:04:51 <dennda> so I guess it's the user patchery
2009-08-04T14:06:07 <dennda> bah segfaults for readline are SO annoying in the middle of debugging
2009-08-04T14:06:39 <dennda> hm
2009-08-04T14:06:56 <dennda> (Pdb) self.request.storage.get_backend("/").request.user.name
2009-08-04T14:06:56 <dennda> u'WikiAdmin'
2009-08-04T14:09:03 <ThomasWaldmann> i still guess it is rather some empty acl ("" means: do not give any right to anybody)
2009-08-04T14:09:20 <ThomasWaldmann> dennda: did you already fix test_caching?
2009-08-04T14:09:28 <dennda> no
2009-08-04T14:09:35 *** starshine_away has quit IRC
2009-08-04T14:09:35 *** vpv has quit IRC
2009-08-04T14:09:35 *** xorAxAx has quit IRC
2009-08-04T14:09:35 *** CIA-43 has quit IRC
2009-08-04T14:09:35 *** Submundo_____ has quit IRC
2009-08-04T14:09:35 <dennda> yes it's an empty acl
2009-08-04T14:09:36 <ThomasWaldmann> ok, I am trying that now
2009-08-04T14:10:25 *** Submundo_____ has joined #moin-dev
2009-08-04T14:10:25 *** starshine_away has joined #moin-dev
2009-08-04T14:10:25 *** CIA-43 has joined #moin-dev
2009-08-04T14:10:25 *** xorAxAx has joined #moin-dev
2009-08-04T14:10:25 *** vpv has joined #moin-dev
2009-08-04T14:11:44 <dennda> the question is where that empty acl comes from
2009-08-04T14:13:38 <dennda> investigating...
2009-08-04T14:14:12 *** starshine_away has quit IRC
2009-08-04T14:14:12 *** vpv has quit IRC
2009-08-04T14:14:12 *** xorAxAx has quit IRC
2009-08-04T14:14:12 *** CIA-43 has quit IRC
2009-08-04T14:14:12 *** Submundo_____ has quit IRC
2009-08-04T14:15:06 *** Submundo_____ has joined #moin-dev
2009-08-04T14:15:06 *** starshine_away has joined #moin-dev
2009-08-04T14:15:06 *** CIA-43 has joined #moin-dev
2009-08-04T14:15:06 *** xorAxAx has joined #moin-dev
2009-08-04T14:15:06 *** vpv has joined #moin-dev
2009-08-04T14:16:04 <dennda> hm
2009-08-04T14:16:31 <dennda> perhaps it's MoinMoin.conftest.Module's fault
2009-08-04T14:16:44 *** starshine_away has quit IRC
2009-08-04T14:16:44 *** vpv has quit IRC
2009-08-04T14:16:44 *** xorAxAx has quit IRC
2009-08-04T14:16:44 *** CIA-43 has quit IRC
2009-08-04T14:16:44 *** Submundo_____ has quit IRC
2009-08-04T14:16:46 *** xorAxAx has joined #moin-dev
2009-08-04T14:16:55 *** SuBmUnDo_ has joined #moin-dev
2009-08-04T14:16:58 <dennda> at least it has self.request = init_test_request(given_config=wikiconfig.Config) in it's __init__ and does hence not honor the config defined in the test
2009-08-04T14:17:24 <dennda> (as opposed to MoinClassCollector from the same module. no idea what the difference is)
2009-08-04T14:17:31 <dennda> that's too py.testish
2009-08-04T14:17:33 *** vpv has joined #moin-dev
2009-08-04T14:17:33 *** starshine_away has joined #moin-dev
2009-08-04T14:18:47 *** CIA-38 has joined #moin-dev
2009-08-04T14:19:12 <dennda> the incoming changesets contain NO fix for this problem
2009-08-04T14:20:48 * dennda kicks CIA-38
2009-08-04T14:20:48 <CIA-38> ow
2009-08-04T14:22:58 <dennda> great
2009-08-04T14:23:12 <dennda> it has self.obj but accessing that results in an exception %)
2009-08-04T14:25:55 <dennda> hypothesis proven by adding before="WikiAdmin:read,write,create,destroy,admin" to the standard config
2009-08-04T14:26:13 <dennda> no idea how to fix that properly with that py.test weirdness
2009-08-04T14:26:32 <dennda> if there's no other WikiAdmin in the tests we could use that as a workaround...
2009-08-04T14:35:25 <dimazest> dreimark: hi
2009-08-04T14:36:46 <dimazest> i have German course in the morning
2009-08-04T14:38:20 <dennda> ThomasWaldmann: *naturally* that workaround won't fix all the tests since there's two configs with different acls in test_security
2009-08-04T14:40:19 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6382:894fd6d803fc 2.0-storage/MoinMoin/ (conftest.py security/_tests/test_security.py): storage: Do the same for the other config defining acls in test_security
2009-08-04T14:54:34 * ThomasWaldmann hugs CIA-38
2009-08-04T14:54:35 * CIA-38 hugs ThomasWaldmann
2009-08-04T14:56:23 <dennda> it skipped 2 CS
2009-08-04T14:58:04 <ThomasWaldmann> is that cia behaviour new? :) did they read their log files? :)
2009-08-04T14:59:10 <dennda> hm? doesn't it freak out every now and then?
2009-08-04T15:00:53 <ThomasWaldmann> no, I mean that it reacts when addressed
2009-08-04T15:01:19 <dimazest> dreimark: http://hg.moinmo.in/moin/1.9-xapian-dmilajevs/file/eb5644419456/MoinMoin/search/queryparser.py#l702 is even more intersting, is continues is necessary there?
2009-08-04T15:01:38 <dennda> ah
2009-08-04T15:01:43 <ThomasWaldmann> btw, i am changing the caching api a bit
2009-08-04T15:01:45 <dennda> yes that's old hat
2009-08-04T15:09:26 *** dreimark has quit IRC
2009-08-04T15:09:26 *** PawelPac1na has quit IRC
2009-08-04T15:09:26 *** franklin has quit IRC
2009-08-04T15:09:26 *** dennda has quit IRC
2009-08-04T15:09:26 *** TheSheep has quit IRC
2009-08-04T15:09:26 *** ThomasWaldmann has quit IRC
2009-08-04T15:22:41 *** ThomasWaldmann has joined #moin-dev
2009-08-04T15:27:00 <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6383:d3228027385f 2.0-storage/MoinMoin/ (6 files in 5 dirs): (log message trimmed)
2009-08-04T15:27:00 <CIA-38> caching: cleaned up API a bit (see below), fixed failing test / all callers
2009-08-04T15:27:00 <CIA-38> CacheEntry(request, arena, ..., scope='item', ...)
2009-08-04T15:27:00 <CIA-38> Old: arena could be a Page or Item instance (it was used to determine the name)
2009-08-04T15:27:01 <CIA-38> New: just directly give the page / item name as arena [unicode]
2009-08-04T15:27:03 <CIA-38> cacheentry.needsUpdate(...)
2009-08-04T15:27:05 <CIA-38> Old: you had to give a single param with a list of Page or Item instances (they
2009-08-04T15:27:07 <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6384:416c61d1c944 2.0-storage/MoinMoin/ (4 files in 4 dirs): merged main
2009-08-04T15:30:16 *** dreimark has joined #moin-dev
2009-08-04T15:30:19 *** TheSheep has joined #moin-dev
2009-08-04T15:30:19 *** PawelPac1na has joined #moin-dev
2009-08-04T15:30:19 *** franklin has joined #moin-dev
2009-08-04T15:30:19 *** dennda has joined #moin-dev
2009-08-04T15:31:27 <ThomasWaldmann> dennda: there are new failures
2009-08-04T15:31:40 <ThomasWaldmann> assert py.test.raises(AccessDeniedError, backend.create_item, "I will never exist")
2009-08-04T15:31:44 <ThomasWaldmann> > ExceptionFailure: 'DID NOT RAISE'
2009-08-04T15:33:06 <ThomasWaldmann> and my xml test says "backend is full" again
2009-08-04T15:34:59 <ThomasWaldmann> dennda: still there?
2009-08-04T15:35:02 <dennda> Yes
2009-08-04T15:35:06 <dennda> I have to look into that
2009-08-04T15:35:41 <dennda> there's no automatic provide_fresh_backend call anymore when initializing the request
2009-08-04T15:35:48 <ThomasWaldmann> before you leave, push all stuff that is ready and maybe publish a diff somewhere of the stuff you rather not want to commit yet
2009-08-04T15:36:28 <dennda> I'm just using my mittagspause in order to tidy up here a bit before departure and bought a present :-)
2009-08-04T15:36:38 <dennda> hence no outstanding changes on my disk
2009-08-04T15:37:12 <dennda> es haben sich ein paar kaffeebecher angesammelt
2009-08-04T15:37:40 <ThomasWaldmann> ah, i thought you almost have some preload working maybe
2009-08-04T15:38:11 <dennda> oh didn't I commit that?
2009-08-04T15:38:17 <dennda> %-)
2009-08-04T15:38:30 <ThomasWaldmann> i just looked at the tests
2009-08-04T15:38:40 <dennda> well the test_items.xml thing must be redone anyways since it doesn't contain the new syspage metadata
2009-08-04T15:39:07 <dennda> and there are isSystemPage checks
2009-08-04T15:42:26 <ThomasWaldmann> # TODO Always uses the default wiki config? Is that correct?
2009-08-04T15:42:26 <ThomasWaldmann> self.request = init_test_request(given_config=wikiconfig.Config)
2009-08-04T15:42:40 <ThomasWaldmann> from MoinMoin._tests import maketestwiki, wikiconfig
2009-08-04T15:42:58 <ThomasWaldmann> do you mean "_tests.wikiconfig" by "default wiki config"?
2009-08-04T15:43:16 <dennda> yes
2009-08-04T15:43:39 <dennda> that's what I believe is causing the issue with the security tests not working
2009-08-04T15:44:34 <dennda> at least it broke there when I made before a mandatory parameter to AclWrapperBackend.__init__ and granting WikiAdmin access in the default config fixed some of the security errors
2009-08-04T15:44:39 <dennda> hence my conclusion
2009-08-04T15:45:09 <ThomasWaldmann> that is on module level only
2009-08-04T15:45:26 <dennda> aha?
2009-08-04T15:45:41 <dennda> i guess the ClassCollector is for classes :-)
2009-08-04T15:45:45 <ThomasWaldmann> class level is different, see conftest
2009-08-04T15:46:28 <dennda> i will later. need to hurry. relatives approaching and this place looks like vietnam
2009-08-04T15:46:43 <ThomasWaldmann> hehe
2009-08-04T15:47:52 <ThomasWaldmann> 1.12 + hierarchic = False,
2009-08-04T15:47:52 <ThomasWaldmann> 1.13 + before=u"WikiAdmin:admin,read,write,create,destroy",
2009-08-04T15:47:52 <ThomasWaldmann> 1.14 + default = u"All:read,write",
2009-08-04T15:47:52 <ThomasWaldmann> 1.15 + after = u"All:read",
2009-08-04T15:48:08 <dennda> style fuckup?
2009-08-04T15:48:19 <ThomasWaldmann> decide whether you want spaces or not. iirc pep8 says "no" in that case.
2009-08-04T15:48:32 <dennda> jup, must have overlooked that
2009-08-04T16:13:49 <ThomasWaldmann> dimazest: how is it going with xappy? you are so silent recently.
2009-08-04T16:17:45 <ThomasWaldmann> 1.12 + hierarchic = False,
2009-08-04T16:17:45 <ThomasWaldmann> 1.13 + before=u"WikiAdmin:admin,read,write,create,destroy",
2009-08-04T16:17:45 <ThomasWaldmann> 1.14 + default = u"All:read,write",
2009-08-04T16:17:45 <ThomasWaldmann> 1.15 + after = u"All:read",
2009-08-04T16:18:20 <ThomasWaldmann> oops
2009-08-04T16:19:25 * ThomasWaldmann .oO(now you can guess the "mouse button" count of my thinkpad :)
2009-08-04T16:28:12 <dreimark> dimazest: strange may be the one who did that wanted to have a placeholder instead of adding a #ToDo
2009-08-04T16:28:25 <dreimark> The continue statement, also borrowed from C, continues with the next iteration of the loop.
2009-08-04T16:29:11 <dreimark> it is not necessary to add a continue at the end ofa loop to get it continued
2009-08-04T16:29:53 <dreimark> looks like we haven't good reviewed at the time this code was done
2009-08-04T16:45:10 <dimazest> ThomasWaldmann: it is going, but quite difficult
2009-08-04T16:46:17 <ThomasWaldmann> the search code is a bit ... "twisted"
2009-08-04T16:46:19 <dimazest> it is more reading than writting
2009-08-04T16:46:39 <dimazest> with repetitions
2009-08-04T16:46:56 <ThomasWaldmann> how much xapwrap is left?
2009-08-04T16:48:50 <dimazest> ThomasWaldmann: not that much, now i'm working on queryparser
2009-08-04T16:49:34 <dimazest> xapwrap is not used for indexing and searching, i will delete it soon
2009-08-04T16:50:01 <dimazest> dreimark: i also want to write more tests
2009-08-04T16:50:39 <dimazest> at least test examples which are on the http://moinmo.in/HelpOnSearching
2009-08-04T16:52:03 <ThomasWaldmann> good idea
2009-08-04T16:52:18 <ThomasWaldmann> btw, please test for both slow and xapian search
2009-08-04T16:53:05 <ThomasWaldmann> PawelPac1na: if you have time, could you look at the item/revision destroy storage api call for hg backend?
2009-08-04T16:54:47 <dimazest> ThomasWaldmann: have a look at http://hg.moinmo.in/moin/1.9-xapian-dmilajevs/file/eb5644419456/MoinMoin/search/_tests/test_search.py#l122
2009-08-04T16:55:19 <dimazest> then all test are added to the TestSearch and tests are performed for both
2009-08-04T16:56:24 <ThomasWaldmann> ok
2009-08-04T16:57:53 <ThomasWaldmann> PawelPac1na: if destroying can't be done with hg, it maybe could be emulated somehow. e.g. destroy item could just rename the item to some "nirvana" folder :)
2009-08-04T17:00:55 <ThomasWaldmann> dimazest: btw, if you could also have a look at that locking stuff the current 1.9 search code does for xapian, that would be great.
2009-08-04T17:01:06 <ThomasWaldmann> I sometimes asked myself whether that is really needed.
2009-08-04T17:01:53 <ThomasWaldmann> btw, did you see the xappy hints about thread safety and follow them?
2009-08-04T17:04:05 <dimazest> searching is thread safe
2009-08-04T17:04:49 <dimazest> and there must be only one indexing connection (searching connections may exist)
2009-08-04T17:05:50 <dimazest> ThomasWaldmann: now i try to do as little changes as possible
2009-08-04T17:06:34 <ThomasWaldmann> one reason I am ask for that is because that locking stuff itself quite often made trouble :)
2009-08-04T17:06:52 <dimazest> :) ok
2009-08-04T17:06:58 <ThomasWaldmann> e.g. it was a problem to build an index while the wiki was running
2009-08-04T17:07:23 <ThomasWaldmann> so my usual procedure was to kill the lockfile and run the indexing then :)
2009-08-04T17:07:58 <ThomasWaldmann> so, IF it is completely superfluous anyway, please get rid of it as soon as it makes sense
2009-08-04T17:09:00 <ThomasWaldmann> and make sure that there really can only be one indexing connection
2009-08-04T17:09:28 <ThomasWaldmann> (i guess for that, you need locking)
2009-08-04T17:13:18 <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6385:40752a9cf3a4 2.0-storage/MoinMoin/ (events/_tests/test_events.py parser/text_moin_wiki.py): skip events testing (needs Item refactoring and other work), fix other test
2009-08-04T17:31:35 <dreimark> dimazest: some of tests should use some chinese signs (or whatever unicode sign you can use)
2009-08-04T17:34:53 <ThomasWaldmann> for many purposes, even german umlauts or accented european chars are enough
2009-08-04T17:36:40 <dreimark> yeah
2009-08-04T17:46:34 <ThomasWaldmann> cloning to fs backend is fscking slow
2009-08-04T17:55:42 <dimazest> ok, i'll add some unicode
2009-08-04T18:00:28 <dennda> indeed
2009-08-04T18:00:54 <dennda> we need profilers :-)
2009-08-04T18:07:36 <dennda> and then we make a contest who has the fastest backend
2009-08-04T18:08:01 <dennda> the price being this gsoc's org money
2009-08-04T18:08:32 <dennda> i kinda like that idea %-)
2009-08-04T18:11:06 <dimazest> dennda: it is not fair, i do not write a backend
2009-08-04T18:21:48 <ThomasWaldmann> dennda: I'll make a /dev/null backend that is even fast than your memory one :P
2009-08-04T18:23:45 * ThomasWaldmann cleans wikiconfig
2009-08-04T18:28:47 <dennda> ThomasWaldmann: that won't pass the tests :-)
2009-08-04T18:29:21 <dennda> dimazest: it wouldn't be fair anyway. if test passing would be a criterion, i'd have already won :D
2009-08-04T18:32:13 <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6386:62a16e4a7266 2.0-storage/ (4 files in 4 dirs): wikiconfig: use absolute pathes, cleanup. create some new dirs under wiki/data/.
2009-08-04T18:32:15 <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6387:c4157f8a9b41 2.0-storage/wiki/data/ (3 files in 3 dirs): remove data/pages and data/user (not used any more)
2009-08-04T18:36:01 * dreimark suggests to add to the contest the length of doc/CHANGES and docs on master wiki
2009-08-04T18:42:00 <ThomasWaldmann> hmm, fs backend has n^2 backend filling complexity
2009-08-04T18:46:39 <dennda> doesn't sound optimal
2009-08-04T18:51:17 *** grzywacz has joined #moin-dev
2009-08-04T19:21:41 <dreimark> ThomasWaldmann: http://hg.moinmo.in/moin/1.8/file/efd8087a3062/MoinMoin/util/filesys.py#l148
2009-08-04T19:22:41 <dreimark> that looks like the problem, because when the file upload was before st.st_mtime + max_staleness you always get a new fuid
2009-08-04T19:23:26 <ThomasWaldmann> http://en.wikipedia.org/wiki/Cdb_(software) shit, that must be done that way...
2009-08-04T19:25:36 <dennda> hm why do you think it's n*n?
2009-08-04T19:25:59 <dennda> i've not looked at the code, but hashtables usually have constant complexity
2009-08-04T19:26:53 <ThomasWaldmann> dreimark: that just means that the fuid changes after the staleness interval, triggering a rebuild of the cache
2009-08-04T19:27:13 <dreimark> that happens on windows with arnica
2009-08-04T19:27:16 <ThomasWaldmann> after the cache is rebuilt, it should be ok for one staleness interval
2009-08-04T19:27:57 <ThomasWaldmann> that happens on windows with everything using that code :)
2009-08-04T19:28:31 <dreimark> hmm, the uploaded file does not change its time.
2009-08-04T19:29:14 <dreimark> you get new cache files the old won't be removed
2009-08-04T19:30:52 <ThomasWaldmann> can you point me to the place where you use fuid?
2009-08-04T19:32:51 <dreimark> it is in the action cache called from Image
2009-08-04T19:33:34 <dreimark> http://hg.moinmo.in/moin/1.8/file/efd8087a3062/MoinMoin/action/cache.py#l94
2009-08-04T19:35:33 <ThomasWaldmann> you'ld need to set max_staleness to None or to a large value to avoid this
2009-08-04T19:36:17 <ThomasWaldmann> if you set it to None, it won't detect fast changes with same filesize due to mtime granularity
2009-08-04T19:36:44 <ThomasWaldmann> if you set it to 24*3600, it will rebuild the cache daily to avoid that effect
2009-08-04T19:37:32 <ThomasWaldmann> or someone finds out how to get the inode on windows :D
2009-08-04T19:39:30 <ThomasWaldmann> hash_name = request.cfg.hash_algorithm
2009-08-04T19:39:30 <ThomasWaldmann> hash_hexdigest = self.rev[hash_name]
2009-08-04T19:39:30 <ThomasWaldmann> cache_meta = [ # we use a list to have order stability
2009-08-04T19:39:30 <ThomasWaldmann> (hash_name, hash_hexdigest),
2009-08-04T19:39:30 <ThomasWaldmann> ('width', width),
2009-08-04T19:39:33 <ThomasWaldmann> ('height', height),
2009-08-04T19:39:35 <ThomasWaldmann> ('transpose', transpose),
2009-08-04T19:39:38 <ThomasWaldmann> ]
2009-08-04T19:39:40 <ThomasWaldmann> cache = SendCache.from_meta(request, cache_meta)
2009-08-04T19:39:43 <ThomasWaldmann> btw, that is how 2.0 solves that problem
2009-08-04T19:40:27 <ThomasWaldmann> (even on crap OSes)
2009-08-04T19:45:46 <dreimark> ok
2009-08-04T19:47:01 <ThomasWaldmann> dreimark: to solve that win32 problem (if someone wants another staleness value), one either needs to make that configurable or just monkey patch it
2009-08-04T19:51:36 <dreimark> None as default for windows may be better. otherwise you get after that staleness time on each request new cache files
2009-08-04T19:53:07 <ThomasWaldmann> yes, that's what it should do :)
2009-08-04T19:53:25 <dreimark> no
2009-08-04T19:53:31 <ThomasWaldmann> depends on whether you rather want to have cache rebuilds or rarely some wrong cache content
2009-08-04T19:53:35 <dreimark> it should then last for the staleness time
2009-08-04T19:53:40 <dreimark> but it didn't
2009-08-04T19:54:04 <ThomasWaldmann> hmm
2009-08-04T19:54:15 <dreimark> now >= st.st_mtime + max_staleness
2009-08-04T19:54:43 <dreimark> after max_staleness this is always true
2009-08-04T20:00:27 <ThomasWaldmann> hmm, maybe a modulo staleness or something is missing there
2009-08-04T20:01:27 <dreimark> the problem is that we get st.st_mtime from the attachment file and this is old then
2009-08-04T20:02:20 <dreimark> s/old/in the past/
2009-08-04T20:05:31 <ThomasWaldmann> only the attachment is relevant
2009-08-04T20:06:18 <dreimark> yeah, windows does not follow the inode concept
2009-08-04T20:06:58 <ThomasWaldmann> it doesn't need to be the inode btw, we just need something that likely changes when the file changes
2009-08-04T20:16:56 *** amartani has joined #moin-dev
2009-08-04T20:25:53 * dennda adjusted test_items.xml
2009-08-04T20:31:51 <dreimark> http://dandar3.blogspot.com/2008/06/filemon-for-windows-v704.html
2009-08-04T20:32:11 <dreimark> Filemon's timestamping feature will show you precisely when every open, read, write or delete, happens, and its status column tells you the outcome.
2009-08-04T20:36:20 <CIA-38> Dmitrijs Milajevs <dimazest@gmail.com> default * 4933:fc330376d50b 1.9-xapian-dmilajevs/MoinMoin/search/queryparser.py: Xapian2009: xapian_called attribute was removed. It is only set but not used.
2009-08-04T20:36:22 <CIA-38> Dmitrijs Milajevs <dimazest@gmail.com> default * 4934:8df5d749cf2d 1.9-xapian-dmilajevs/MoinMoin/search/ (Xapian.py queryparser.py): Xapian2009: xapian_term() was refactored. Code repetition was reduced by introducing BaseFieldSearch class. Field action definitions was updated.
2009-08-04T20:36:43 <dennda> 18 failed
2009-08-04T20:36:46 <dennda> new record :-)
2009-08-04T20:37:29 <dimazest> dreimark: tomorrow i want to work on tests, hope some failing tests will be added :))
2009-08-04T20:39:48 <dreimark> :)
2009-08-04T20:40:08 * dreimark sees a quite shortened xapian_term
2009-08-04T20:40:48 <dimazest> dreimark: there is a mess with self.pattern and self._pattern
2009-08-04T20:41:14 <dimazest> and sometimes it is lowercased (in CategorySearch)
2009-08-04T20:41:41 <dimazest> but i want to make changes on code which has tests
2009-08-04T20:41:49 <dennda> now I just need to get the reference to that file right...
2009-08-04T20:43:10 <PawelPac1na> moin
2009-08-04T20:43:18 * PawelPac1na is back from holiday
2009-08-04T20:43:24 *** PawelPac1na is now known as PawelPacana
2009-08-04T20:43:37 <dreimark> hi PawelPac1na please see backlog
2009-08-04T20:43:42 <PawelPacana> a lot of changes, a lot of reading
2009-08-04T20:43:54 <TheSheep> PawelPacana: hi there!
2009-08-04T20:44:09 <dimazest> Jetz muss ich mene deutche Hausafgabe machen. Tschüß!
2009-08-04T20:44:21 <dreimark> dimazest: viel spass
2009-08-04T20:44:25 <dimazest> :)
2009-08-04T20:44:46 <TheSheep> ja vohl
2009-08-04T20:48:11 <dennda> wohl
2009-08-04T20:50:18 <ThomasWaldmann> wb PawelPacana
2009-08-04T20:52:20 <dimazest> m*ei*ne
2009-08-04T20:52:28 <dimazest> anyway
2009-08-04T20:53:01 <dreimark> bbl
2009-08-04T21:00:33 * ThomasWaldmann boots Vista (oerks)
2009-08-04T21:01:52 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6388:2060396006e9 2.0-storage/MoinMoin/ (_tests/test_wikiutil.py wikiutil.py _tests/testitems.xml): storage: Fix wikiutil.isSystemPage, fix testcase, add testitems.xml
2009-08-04T21:02:02 <dennda> ThomasWaldmann: why that?
2009-08-04T21:02:31 <ThomasWaldmann> to try something for filesys.fuid on win32
2009-08-04T21:13:04 <ThomasWaldmann> we could use ntfs streams to store some extra stuff :)
2009-08-04T21:13:34 <ThomasWaldmann> e.g. some unique id or a more precise timestamp
2009-08-04T21:14:27 *** amartani has quit IRC
2009-08-04T21:30:37 * ThomasWaldmann sees a rather high res st_mtime
2009-08-04T21:40:45 <CIA-38> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6389:28a6b70fd1a1 2.0-storage/MoinMoin/storage/_tests/test_backends_hg.py: (tests) mark tests expecting to fail, skip tests on import error
2009-08-04T21:50:49 <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4356:a6b0b1fcd463 1.8/MoinMoin/util/filesys.py: fuid: keep same fake_mtime for intervals of max_staleness duration
2009-08-04T21:52:17 <ThomasWaldmann> dreimark: ^^
2009-08-04T21:53:05 <ThomasWaldmann> btw, on win32/ntfs that max_staleness stuff might do more harm than good
2009-08-04T21:53:32 <ThomasWaldmann> because st_mtime seems to be a float with about 1/100s resolution
2009-08-04T21:53:59 <ThomasWaldmann> (at least for recent python)
2009-08-04T21:55:21 <ThomasWaldmann> so maybe do some experiments with it max_staleness=None. I guess we can do another changeset then with setting it to None as default.
2009-08-04T22:11:23 *** xorAxAx has quit IRC
2009-08-04T22:14:09 *** xorAxAx has joined #moin-dev
2009-08-04T22:36:12 <dennda> Is the package system from MoinMoin.packages accessible for users? (in other words, should ACLs be applied?)
2009-08-04T22:41:01 <ThomasWaldmann> E py.test.importorskip('mercurial')
2009-08-04T22:41:02 <ThomasWaldmann> > AttributeError: importorskip
2009-08-04T22:41:23 <ThomasWaldmann> PawelPacana: ^^
2009-08-04T22:41:27 <PawelPacana> probably from py 1.0
2009-08-04T22:42:11 <ThomasWaldmann> maybe rather use try/except
2009-08-04T22:42:27 <PawelPacana> i can do backward compatbile skip, or you can upgrade py.test
2009-08-04T22:42:41 <PawelPacana> the same for mark.xfail
2009-08-04T22:43:57 <ThomasWaldmann> well, it is not only me running the tests. even some users who download moin do that...
2009-08-04T22:45:25 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6390:5f0e212f13e0 2.0-storage/MoinMoin/security/_tests/test_security.py: storage: style consistency fix
2009-08-04T22:45:26 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6391:ba05e66794cd 2.0-storage/MoinMoin/ (3 files in 3 dirs): storage: Fix some cfg.storage leftovers. There's still two remaining for which it's unclear whether they should use protected or unprotected storage.
2009-08-04T22:49:13 <ThomasWaldmann> dennda: i have a look at the interwiki test failure (MoinMoin:...)
2009-08-04T22:52:29 <dennda> i'm working on the amw failure
2009-08-04T22:57:01 <dennda> ARGH
2009-08-04T22:57:02 <dennda> !!
2009-08-04T22:57:15 <dennda> request != self.request
2009-08-04T22:57:23 <dennda> bloody thing
2009-08-04T22:57:59 <dennda> 16 failed tests now
2009-08-04T22:58:24 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6392:c8a8603ec62b 2.0-storage/MoinMoin/storage/_tests/test_middleware_acl.py: storage: Fix AMW test. (Bug in test)
2009-08-04T23:03:19 * dennda looks at test_security
2009-08-04T23:04:47 <CIA-38> Pawel Pacana <pawel.pacana (at) gmail.com> default * 6393:201b516e7e4b 2.0-storage/MoinMoin/storage/_tests/test_backends_hg.py: (tests) backward compatibility for py.test < 1.0
2009-08-04T23:05:27 <PawelPacana> ThomasWaldmann: well py.test is separate package, they have to install it anyway, although i understand that major distros will ship codespeak-lib 1.0 packages at least after half a year
2009-08-04T23:08:51 <dennda> 14 fails %-)
2009-08-04T23:11:08 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6394:50ff2cb8ee5b 2.0-storage/MoinMoin/security/_tests/test_security.py: storage: Fix security tests
2009-08-04T23:11:10 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6395:c3265bd3ea42 2.0-storage/MoinMoin/storage/_tests/test_backends_hg.py: storage: merge remote 2.0-storage
2009-08-04T23:25:20 <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6396:01fa8d35e8b9 2.0-storage/MoinMoin/_tests/ (test_wikiutil.py wikiconfig.py): add shared_intermap_files to test wikiconfig, add test for resolve_interwiki
2009-08-04T23:26:39 <ThomasWaldmann> 10
2009-08-04T23:28:18 <dennda> i got a few more, just a sec :-)
2009-08-04T23:34:59 <dennda> Why is johills search stuff not used anywhere btw?
2009-08-04T23:35:05 <dennda> hm why do I get 11 now?
2009-08-04T23:35:08 <dennda> (after merge)
2009-08-04T23:35:13 <dennda> 13 or so before merge
2009-08-04T23:38:00 <ThomasWaldmann> iirc it is used, but not for the UI search
2009-08-04T23:38:12 <ThomasWaldmann> and maybe you have a newer py.test :)
2009-08-04T23:38:24 * ThomasWaldmann fixes 200 vs 404 now
2009-08-04T23:39:05 <dennda> py.test has no option to show the version 8)
2009-08-04T23:39:50 <dennda> codespeak lib version 0.9.1
2009-08-04T23:39:59 <dennda> cloning your stuff into a fresh branch to verify
2009-08-04T23:40:21 <dennda> btw: the source got quite big. takes me ages to clone a repo with dsl 20.000
2009-08-04T23:40:32 <dennda> where "ages" == some minutes
2009-08-04T23:41:21 <dennda> and why exactly would that be dependant on my py.test version?
2009-08-04T23:42:20 <PawelPacana> 2009-07-28T23:40:26 <ronny> (given that hg has an own nice way to deal with per file metadata since years
2009-08-04T23:42:23 <dennda> there was some other error that depended on the system you use to run the tests. not sure if that's still there. johill once complained
2009-08-04T23:42:56 <dennda> oh pawel is catching up :-)
2009-08-04T23:42:57 <PawelPacana> ronny: sure it has, but you can't write it unless patched (see http://moinmo.in/PawelPacana/MercurialBackend/PreviousIdeas, second paragraph)
2009-08-04T23:43:27 <PawelPacana> ronny: since years
2009-08-04T23:43:47 <PawelPacana> filelog provides only _readmeta
2009-08-04T23:48:07 <PawelPacana> if you wonder why metadata is pickled, see mercurial.changelog.encodeextra: http://selenic.com/repo/hg-stable/file/2de7d96593db/mercurial/changelog.py#l34
2009-08-04T23:50:07 <ronny> PawelPacana: hmm, i see, i can fix that :P
2009-08-04T23:50:52 <PawelPacana> finally, locking is for writing working directory files, you won't be doing this if single user/process has his own working directory
2009-08-04T23:51:25 <ronny> commitctz handles locking
2009-08-04T23:51:30 <PawelPacana> but on multithreaded server working directory is shared
2009-08-04T23:51:41 <PawelPacana> yeah, but not writing to file
2009-08-04T23:52:09 <PawelPacana> you re not commiting from memoryctx
2009-08-04T23:52:27 <ronny> well, i do commit from memoryctx
2009-08-04T23:52:30 <ronny> works fine
2009-08-04T23:52:36 <PawelPacana> there was a branch which had memory commits
2009-08-04T23:52:43 <PawelPacana> and no locks
2009-08-04T23:52:56 <PawelPacana> i ran into serious trouble with merging
2009-08-04T23:53:26 <ronny> there isnt yet a sane way to merge in memory
2009-08-04T23:53:35 <ronny> i will implement that in anyvc
2009-08-04T23:55:03 <PawelPacana> what do you mean by saying 'i can fix that'?
2009-08-04T23:56:18 <dennda> ThomasWaldmann: hm strange, got 10 with your stuff as well.
2009-08-04T23:56:31 <ronny> PawelPacana: i think i can fix merging given enough time
2009-08-04T23:56:53 <ronny> (i want to get it working for hg, bzr svn and git)
2009-08-04T23:57:36 <dennda> wtf is wrong with py.test
2009-08-04T23:57:48 <dennda> 2835235 failed tests with test_sourcecode??
2009-08-04T23:57:58 <dennda> (or some similar amount)
2009-08-04T23:59:47 <dennda> ah.
MoinMoin: MoinMoinChat/Logs/moin-dev/2009-08-04 (last edited 2009-08-03 23:30:02 by IrcLogImporter)