2008-08-06T00:04:22 <dreimark> data_backend = ?
2008-08-06T00:04:48 <PawelPacana> MercurialBackend("path_to_empty_dir")
2008-08-06T00:05:18 <TheSheep> dreimark: from MoinMoin.storage import hg data_backend = hg.MercurialStorage("some empty dir here")
2008-08-06T00:06:19 <PawelPacana> >> from MoinMoin.storage.backends
2008-08-06T00:06:41 <johill> should we move that up?
2008-08-06T00:06:50 <johill> I don't really see a reason for the sub-package
2008-08-06T00:07:15 <dennda> Doesn't matter too much, does it?
2008-08-06T00:07:27 <johill> no. just more typing this way
2008-08-06T00:07:30 <dennda> One may argue that it's a bit more structured, but I don't care
2008-08-06T00:08:00 <johill> if nobody cares let's just keep it, less work :)
2008-08-06T00:08:07 <dennda> haha
2008-08-06T00:08:11 <PawelPacana> ;)
2008-08-06T00:08:30 <TheSheep> you copy-paste it anyways ;)
2008-08-06T00:08:54 <johill> heh, clearly you didn't though ;)
2008-08-06T00:09:33 * johill braces for freenode service upgrade
2008-08-06T00:10:06 <johill> hmm. someone should port my sql backend ;)
2008-08-06T00:10:26 <dennda> someone should write an sqlalchemy one
2008-08-06T00:10:31 <dennda> I want to reserve that task for me
2008-08-06T00:10:42 <dennda> Want to get my hands dirty with sqlalchemy anyways :)
2008-08-06T00:10:51 <johill> :)
2008-08-06T00:11:32 <PawelPacana> zodb backend maybe? ;)
2008-08-06T00:12:11 <TheSheep> PawelPacana: not such a bad idea actually
2008-08-06T00:12:22 <dennda> Last time they updated the services the result was a complete mess
2008-08-06T00:12:55 <dennda> I hope they do better this time ;)
2008-08-06T00:13:21 <dreimark> probably my fault or ConfigurationError: AttributeError: 'module' object has no attribute 'MercurialStorage' ?
2008-08-06T00:13:53 * TheSheep copy-and-pastes this time
2008-08-06T00:14:12 <TheSheep> from MoinMoin.storage.backends import hg
2008-08-06T00:14:22 <TheSheep> data_backend = hg.MercurialBackend("/home/sheep/dev/moin/1.8-mercurialbackend-ppacana/wikidata")
2008-08-06T00:14:38 <TheSheep> dreimark: sorry :)
2008-08-06T00:15:37 <dreimark> no problem
2008-08-06T00:19:26 <dreimark> PawelPacana: TheSheep should I upload tracebacks ?
2008-08-06T00:19:54 <johill> not if they include 'timestamp' or ''
2008-08-06T00:20:02 <johill> 'size'
2008-08-06T00:20:28 <PawelPacana> (in fact zodb would be persistent version of not much changed memory backend, i think)
2008-08-06T00:20:54 <johill> except it's not thread-safe ;)
2008-08-06T00:21:05 <PawelPacana> timestamp/size is 'in progress' so wont work now
2008-08-06T00:21:12 <dreimark> ok
2008-08-06T00:22:19 <dennda> toi toi toi
2008-08-06T00:22:38 <johill> :)
2008-08-06T00:22:57 <johill> I'm only on two freenode channels any more, apparently
2008-08-06T00:23:20 <dennda> I have 48 Windows open
2008-08-06T00:23:25 <dennda> Most of them freenode channels
2008-08-06T00:23:46 <dennda> What's the other channel?
2008-08-06T00:23:51 <johill> #moin
2008-08-06T00:23:56 <dennda> ah ;)
2008-08-06T00:23:56 <dennda> sure
2008-08-06T00:24:48 <dreimark> good night
2008-08-06T00:25:14 <johill> ok, I have the rename with metadata stuff working
2008-08-06T00:25:49 <johill> but I'm now starting to think maybe we should have a rename log with each item or so?
2008-08-06T00:27:11 <johill> (and provide them as part of news() for optimisation)
2008-08-06T00:27:39 *** irc.freenode.net sets mode: +o ChanServ
2008-08-06T00:29:45 <johill> oww. http://arstechnica.com/news.ars/post/20080805-high-tech-peeping-tom-rigged-laptop-webcam-to-snap-nude-pics.html
2008-08-06T00:35:11 <TheSheep> johill: I guess webcam covers will be a fashionable laptop accessory soon :)
2008-08-06T00:35:47 <johill> closing the lid would work...
2008-08-06T00:36:08 <TheSheep> johill: not if you want to use it
2008-08-06T00:36:57 <johill> if you're naked with your boyfriend, how are you using your laptop?
2008-08-06T00:37:24 <TheSheep> johill: watch a porn movie with him? listening to music?
2008-08-06T00:37:36 <johill> heh
2008-08-06T00:37:46 <johill> oh well. I suggest duck tape
2008-08-06T00:37:57 <TheSheep> hamster tape
2008-08-06T00:38:17 *** irc.freenode.net sets mode: +o ChanServ
2008-08-06T00:45:23 <zenhase> ThomasWaldmann: they are already in removed state in my repo, but i want to remove them bit by bit as they have proper replacements
2008-08-06T01:05:15 <CIA-53> Johannes Berg <johannes AT sipsolutions DOT net> default * 4563:f62ae77f82a4 1.8-storage-cdenter/MoinMoin/storage/backends/ (sql1.py sql1.sql): storage: remove sql1 backend
2008-08-06T01:05:30 <johill> xorAxAx: http://paste.pocoo.org/show/81342/
2008-08-06T01:16:48 <dennda> gn8
2008-08-06T08:31:42 <ThomasWaldmann> moin
2008-08-06T08:51:36 <xorAxAx> johill: uhm
2008-08-06T08:52:07 <xorAxAx> johill: 'SAVE/RENAME' should probably be changed to something new
2008-08-06T08:52:36 <ThomasWaldmann> anybody opposing that we just drop mod_python support completely in the moin-wsgi repo?
2008-08-06T08:53:01 <ThomasWaldmann> (the more I read about mod_python, the more brokenness surfaces to me)
2008-08-06T08:54:48 <xorAxAx> johill: but besides that, it looks bearable ... hmm, EDIT_LOG_ACTION should probably be gone their completly, no?
2008-08-06T08:57:26 <xorAxAx> johill: that ars technica story probably wouldnt have happened with linux
2008-08-06T09:22:13 <johill> xorAxAx: because the webcam wouldn't work ;)
2008-08-06T09:22:25 <johill> or nobody would have noticed
2008-08-06T09:22:30 <johill> (indeed, macbook webcams do work)
2008-08-06T09:23:08 <johill> xorAxAx: yes, edit log action should probably be gone for renames, but this is so much easier, and it's not a low-level hack, it's the upper layer adding a workaround for itself
2008-08-06T09:23:13 <xorAxAx> johill: because you can install a driver that doesnt light up the lamp :)
2008-08-06T09:23:52 <xorAxAx> yes, but i consider it legacy
2008-08-06T09:26:19 <johill> yes, but I really don't have time to clean up all the junk right now
2008-08-06T09:27:20 <xorAxAx> yeah, of course
2008-08-06T09:27:27 <xorAxAx> it wasnt there before you anyway
2008-08-06T09:27:32 <xorAxAx> s/wasnt/was/
2008-08-06T09:56:46 <zenhase> moin
2008-08-06T09:56:58 <zenhase> ThomasWaldmann: i wouldn't oppose dropping it
2008-08-06T09:58:10 <zenhase> i think mod_python will cease in the long run and mod_wsgi will become the defacto standard
2008-08-06T09:59:10 <xorAxAx> i think its fine if the docs are cut down to a short description where to get a wsgi 2 modpy adapter
2008-08-06T10:02:05 <johill> xorAxAx: what do you think though of storing the renames per item and weaving them into news() as the optimisation?
2008-08-06T10:02:18 <johill> xorAxAx: that'd easily allow us to have them pop up on action=info too
2008-08-06T10:03:29 <xorAxAx> johill: hmm, an item has a history of renames and revisions? probably
2008-08-06T10:03:34 <xorAxAx> model-wise its sensible
2008-08-06T10:03:45 <xorAxAx> or better, sound
2008-08-06T10:03:57 <johill> nowadays it doesn't, but it'd be doable
2008-08-06T10:04:20 <xorAxAx> it would also make news() just an optimisation again
2008-08-06T10:04:34 <johill> yeah
2008-08-06T10:04:57 <johill> although it currently is too, folding revision news and renames from item_log()
2008-08-06T10:05:14 <johill> (so item_log() is the actual data)
2008-08-06T10:05:43 <johill> doing it per revision might again require some duplicate storage, but only inside the news file that only fs backends would need
2008-08-06T10:08:03 <johill> anyway, no time, let's discuss with more folks after the meeting or so
2008-08-06T10:08:36 <ThomasWaldmann> zenhase: if nobody else does, drop it today evening. if someone really wants to use that crap, he can find a wsgi adaptor on his own.
2008-08-06T10:10:24 <ThomasWaldmann> dennda: there is still some getPagePath/caching related code in Page
2008-08-06T10:13:02 <xorAxAx> johill: hmm, wikisync wise, i like per-item rename logs
2008-08-06T10:16:32 <zenhase> ThomasWaldmann: ok
2008-08-06T10:18:56 <xorAxAx> wtf, 8 eye doctors in PB, none of them wants to give an appointment before october (or in one case, is pretty bad friendlyness and craftsmen wise)
2008-08-06T10:19:15 <xorAxAx> just phoned the single doctor in the town nearby (bad lippspring, 7 km), got an appointment in 2 weeks
2008-08-06T10:20:08 <xorAxAx> (and she actually asked me for the insurance caste after giving me the appointment :))
2008-08-06T10:23:27 <zenhase> insurance caste :D
2008-08-06T10:24:09 <zenhase> nice way to explain our strange health insurance system to non-germans in 2 words ;)
2008-08-06T10:33:00 <xorAxAx> :)
2008-08-06T10:45:03 <CIA-53> Johannes Berg <johannes AT sipsolutions DOT net> default * 3934:be75172aaf9b 1.8/MoinMoin/security/_tests/test_security.py: security tests: convert to iterators
2008-08-06T10:47:48 <CIA-53> Johannes Berg <johannes AT sipsolutions DOT net> default * 4565:643796f59c5a 1.8-storage-cdenter/MoinMoin/security/_tests/test_security.py: merge 1.8
2008-08-06T11:36:12 <PawelPacana> moin
2008-08-06T11:36:22 <xorAxAx> moin PawelPacana
2008-08-06T11:37:46 <dreimark> moin
2008-08-06T11:38:47 <PawelPacana> xorAxAx: johill: storing renames per item, so you get rename history from item?
2008-08-06T11:39:38 <xorAxAx> PawelPacana: yes, thats his newest idea
2008-08-06T11:39:43 <xorAxAx> PawelPacana: does it make sense to you? :)
2008-08-06T11:40:07 <xorAxAx> not that anybody expects you to implement that before the deadline
2008-08-06T11:40:21 <xorAxAx> but does that make sense design wise?
2008-08-06T11:40:24 <PawelPacana> sounds much better
2008-08-06T11:40:47 <xorAxAx> but note that it would be distinct from the revision history
2008-08-06T11:41:09 <xorAxAx> (but you can correlate/weave both histories and get the full story of an item :))
2008-08-06T11:41:53 <PawelPacana> i think it's ok
2008-08-06T11:47:14 <ThomasWaldmann> PawelPacana: did you try the backend converter already?
2008-08-06T11:52:11 <PawelPacana> ThomasWaldmann: didnt have much time to focus on converter yet (had some merging issues), i'll work on it after meeting
2008-08-06T11:52:39 <ThomasWaldmann> it would be nice to at least find out if the api is usable for it
2008-08-06T11:53:05 <ThomasWaldmann> I especially think of initialization of the backends.
2008-08-06T12:01:23 * dennda is away to pick up his new screen, be back in time I hope
2008-08-06T12:01:59 * dreimark has a meeting at the same time today (and no idea if there is internet)
2008-08-06T12:31:59 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4600:130c9324c278 1.8-mercurialbackend-ppacana/MoinMoin/storage/_tests/test_backends_hg.py: (hg tests) Merge test refactoring. Strange timing issue to investigate.
2008-08-06T12:32:00 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4601:75b057d1851b 1.8-mercurialbackend-ppacana/MoinMoin/storage/_tests/test_backends_hg.py: (hg) Removal of obsoleted hg file heads test.
2008-08-06T12:32:02 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4602:4fd3c50fa4c2 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Timestamp implementation.
2008-08-06T12:32:02 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4603:689974038364 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Added _get_revision_size.
2008-08-06T12:32:04 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4604:d1c750c5fc58 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Added news() implementation from fs backend.
2008-08-06T12:32:07 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4605:9fbe7bdfa9df 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Fixed _get_revision_metadata.
2008-08-06T12:32:10 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4606:0e35bd226dfc 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) _list_revisions change to include merged revisions.
2008-08-06T12:32:15 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4607:0aee1679d3cd 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py:
2008-08-06T12:32:18 <CIA-53> (hg) _commit_item: Get rid of file heads checking (unreliable on items with the
2008-08-06T12:32:20 <CIA-53> same data, but different meta). Fixed revision parents.
2008-08-06T12:32:22 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4608:296a56d65e22 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Modified get_revision to include merged revisions.
2008-08-06T12:39:52 <byeongweon> ThomasWaldmann: I have a question about unicode special character
2008-08-06T12:44:51 <xorAxAx> byeongweon: just ask :)
2008-08-06T12:45:57 <byeongweon> he suggest add additonal unicode character into special character selection
2008-08-06T12:46:16 <byeongweon> I was search about that few ways.
2008-08-06T12:46:33 <byeongweon> but I couldn't determine what character is required.
2008-08-06T12:46:56 <byeongweon> I think its base on language culture.
2008-08-06T12:47:27 <byeongweon> I don't use any other latin extension character. but european use that.
2008-08-06T12:47:45 <byeongweon> I want to know what character are required.
2008-08-06T12:47:51 <xorAxAx> to fix which bug?
2008-08-06T12:48:11 <byeongweon> no. it's just impove function.
2008-08-06T12:48:25 <xorAxAx> ah
2008-08-06T12:48:58 <byeongweon> second element of FeatureRequests on MoonByeongweon/SOC2008
2008-08-06T12:49:36 <xorAxAx> ah
2008-08-06T12:49:49 <xorAxAx> i think he was asking whether fckeditor probably ships a new plugin already
2008-08-06T12:49:51 <byeongweon> and I want to ask about http://moinmo.in/MoinMoinBugs/GuiEditorMixedListsCrash this bug after meeting
2008-08-06T12:49:55 <xorAxAx> that has more characters or whatever
2008-08-06T12:53:50 <xorAxAx> byeongweon: do you think that fckeditor has something like that?
2008-08-06T12:54:35 <byeongweon> I try to found something like that. but I couldn't
2008-08-06T12:54:41 <xorAxAx> ok
2008-08-06T12:54:44 <xorAxAx> then its fine
2008-08-06T12:56:02 * xorAxAx cleans up the soc page of byeongweon a bit
2008-08-06T12:56:20 <byeongweon> thanks :)
2008-08-06T13:00:32 <ThomasWaldmann> meeting in a minute
2008-08-06T13:00:48 <xorAxAx> but its 13:00:47 :)
2008-08-06T13:00:49 <ThomasWaldmann> no, now, welcome to our last coding-phase meeting
2008-08-06T13:00:58 <ThomasWaldmann> waldi: please begin
2008-08-06T13:01:06 <dennda> here now
2008-08-06T13:01:41 <CIA-53> Johannes Berg <johannes AT sipsolutions DOT net> default * 4566:ad8450d358de 1.8-storage-cdenter/MoinMoin/storage/ (__init__.py backends/fs.py):
2008-08-06T13:01:41 <CIA-53> storage: remove tuple special casing for get_revision_{metadata,timestamp}
2008-08-06T13:01:41 <CIA-53> implement it with a special attribute in the fs backend instead, much
2008-08-06T13:01:41 <CIA-53> cleaner.
2008-08-06T13:03:46 <ThomasWaldmann> waldi: are you there?
2008-08-06T13:03:46 <zenhase> . . ~#
2008-08-06T13:04:00 <xorAxAx> zenhase: you forgot enter :)
2008-08-06T13:04:18 <ThomasWaldmann> ok, waldi seems not to be here right now, thus zenhase begins
2008-08-06T13:04:19 <zenhase> xorAxAx: hum? :)
2008-08-06T13:04:24 <zenhase> ok
2008-08-06T13:05:22 <zenhase> i replaced request_cli with a context and started on the .cgi .fcg etc. scripts
2008-08-06T13:05:54 <zenhase> i contacted allan saddi for including flup and also asked if it's ok if its distributed patched
2008-08-06T13:06:29 <zenhase> i patched it with the singleserver-patch, which resolves those run-with-no-threading-issues
2008-08-06T13:06:49 <ThomasWaldmann> zenhase: did you care about its license terms?
2008-08-06T13:07:19 <TheSheep> zenhase: author's permission overrides license
2008-08-06T13:07:26 <ThomasWaldmann> did you document what exactly you used as source and as patch? versions, urls, date?
2008-08-06T13:07:33 <johill> it says "BSD style license" but I can find no license document
2008-08-06T13:07:48 <zenhase> well, it's something like 2-clause BSD
2008-08-06T13:07:57 <johill> debian's package has a two-clause bsd
2008-08-06T13:08:04 <johill> so I guess they did find it somewhere
2008-08-06T13:08:09 <zenhase> and i didn't remove any lines containing copyright, so we are ok :)
2008-08-06T13:08:16 <xorAxAx> yep
2008-08-06T13:08:22 <zenhase> (also he permitted it explicitely, so anyway ...)
2008-08-06T13:08:27 <johill> :)
2008-08-06T13:08:57 <zenhase> yesterday i wrote the combined cgi/fcgi-script on the base of flup
2008-08-06T13:08:58 <ThomasWaldmann> zenhase: please add some lines documenting the code sources you used and why you patched it
2008-08-06T13:09:06 <zenhase> ok
2008-08-06T13:09:39 <zenhase> i will also document fastcgi/cgi setup today and try to add the other methods from flup (scgi, ajp) right away
2008-08-06T13:10:15 <ThomasWaldmann> is your wrist ok for working?
2008-08-06T13:10:23 <zenhase> will also write a small fallback CGI adapter, for those not wanting to install flup and still run from apache
2008-08-06T13:10:46 <ThomasWaldmann> good idea.
2008-08-06T13:10:50 <zenhase> my wrist is better now, the bandage the doctor put on is quite tight and makes typing ok again
2008-08-06T13:11:51 <zenhase> i shouldn't strain it too much, but he did not insist on resting it all the time
2008-08-06T13:12:21 <ThomasWaldmann> mitsuhiko: are you with us?
2008-08-06T13:12:42 <zenhase> my guess is, he is somewhere between scotland and germany :)
2008-08-06T13:12:48 <ThomasWaldmann> zenhase: blockers?
2008-08-06T13:13:13 <zenhase> no blockers, right now it's quite obvious what to do :)
2008-08-06T13:13:31 <zenhase> finish the other deployment methods and document
2008-08-06T13:13:31 <ThomasWaldmann> plans until end-of-codings phase?
2008-08-06T13:13:35 <xorAxAx> zenhase: you mean you want to ship flup but require separate installation?
2008-08-06T13:13:54 <zenhase> xorAxAx: actually yes
2008-08-06T13:14:02 <zenhase> it's in contrib right now
2008-08-06T13:14:22 <ThomasWaldmann> if you patched it anyway, patching the imports would be an idea
2008-08-06T13:14:26 <zenhase> hmm
2008-08-06T13:14:32 <zenhase> yes, could think about that
2008-08-06T13:14:37 <ThomasWaldmann> (and including that patch and also the other one)
2008-08-06T13:14:42 <xorAxAx> well, trac manages to ship a single module and support fcgi out of the box
2008-08-06T13:14:53 <zenhase> they ship a precursor of flup
2008-08-06T13:14:57 <zenhase> afaik
2008-08-06T13:15:00 <TheSheep> xorAxAx: not to mention zope :)
2008-08-06T13:15:13 <zenhase> and they support ONLY fcgi
2008-08-06T13:15:35 <zenhase> flup is based on the code in fcgi.py (which trac ships)
2008-08-06T13:15:47 <xorAxAx> TheSheep: :-)
2008-08-06T13:16:00 <xorAxAx> zenhase: ah
2008-08-06T13:16:17 <ThomasWaldmann> zenhase: plans until next monday?
2008-08-06T13:17:00 <zenhase> finish those different deployment-methods (implementation-wise) and some testing code
2008-08-06T13:17:13 <ThomasWaldmann> looks like melita is not here (and seems to make 5d vacactions although she has lots to do left 8( )
2008-08-06T13:17:20 <zenhase> from monday on i would concentrate fully on improving documentation
2008-08-06T13:17:36 <zenhase> i think that's all from me right now :)
2008-08-06T13:17:36 <ThomasWaldmann> zenhase: and cleanup the not used old code
2008-08-06T13:17:40 <zenhase> yes
2008-08-06T13:17:48 <ThomasWaldmann> ok, thanks zenhase
2008-08-06T13:17:51 <zenhase> that comes this week while implementing the methods
2008-08-06T13:18:11 <ThomasWaldmann> dreimark: if you are here, anything to add to what I've said a few lines above?
2008-08-06T13:18:57 <dennda> Seems he has no internet at the meeting table
2008-08-06T13:19:00 <ThomasWaldmann> looks like not. byeongweon: xorAxAx: your stage
2008-08-06T13:19:07 <byeongweon> yes
2008-08-06T13:19:19 <byeongweon> I was struggle with bugs next week.
2008-08-06T13:19:42 <byeongweon> and
2008-08-06T13:19:48 <byeongweon> recheck fixed bugs and
2008-08-06T13:19:59 <byeongweon> make clear that bugs.
2008-08-06T13:20:13 <byeongweon> blockers:
2008-08-06T13:20:50 <byeongweon> I concentration on specific bug.
2008-08-06T13:21:07 <byeongweon> I talk about that bug with xorAxAx after this meeting.
2008-08-06T13:21:24 <byeongweon> except that notthing's block.
2008-08-06T13:21:37 <xorAxAx> hmm
2008-08-06T13:21:56 <xorAxAx> we have a few bugs related the converter left, but i think most of them arent regressions
2008-08-06T13:22:06 <xorAxAx> there are 2 bullet points under unfixed currently
2008-08-06T13:22:11 <xorAxAx> are those client or server side issues?
2008-08-06T13:22:15 <xorAxAx> (e.g. image roundtrips)
2008-08-06T13:22:45 <waldi> ThomasWaldmann: *gna* there
2008-08-06T13:22:50 <byeongweon> yes.
2008-08-06T13:23:38 <xorAxAx> byeongweon: that was an or-question
2008-08-06T13:23:48 <byeongweon> until next meeting. I will reordering remained bugs and fix high priority.
2008-08-06T13:24:01 <xorAxAx> ok, so you didnt check them yet
2008-08-06T13:24:02 <xorAxAx> ok
2008-08-06T13:24:14 <byeongweon> xorAxAx: they are server side issue.
2008-08-06T13:24:15 <ThomasWaldmann> waldi: you're last after pawel
2008-08-06T13:24:28 <xorAxAx> i saw that you correctly changed the category of a lot of bugs but didnt add an appropriate comment regarding where they were fixed
2008-08-06T13:24:35 <xorAxAx> (in fact in 1.8 on your branch)
2008-08-06T13:24:47 <xorAxAx> not sure whether that is important currently
2008-08-06T13:25:00 <xorAxAx> just as an advise for the future - would be nice to add that just in case you close another bug
2008-08-06T13:25:07 <byeongweon> xorAxAx: it because most bugs are fixed with recent version of fckeditor
2008-08-06T13:25:26 <xorAxAx> yes, i know, but people want to know which moin version they need at least to get the bug fix from moin
2008-08-06T13:25:35 <ThomasWaldmann> it could be that some bugs just not happen with updates fckeditor. but IF moon did a changeset to fix stuff, the changeset url should be in the bug report.
2008-08-06T13:25:36 <byeongweon> ah..
2008-08-06T13:25:56 <byeongweon> ok I understand. I will notify about that on bug page.
2008-08-06T13:26:16 <xorAxAx> cool
2008-08-06T13:26:26 <byeongweon> I think I'm done.
2008-08-06T13:26:48 <xorAxAx> yeah
2008-08-06T13:26:52 <xorAxAx> PawelPacana, TheSheep
2008-08-06T13:27:01 <xorAxAx> or dennda? :)
2008-08-06T13:27:13 <xorAxAx> dennda, johill
2008-08-06T13:27:15 <dennda> ok
2008-08-06T13:27:43 <dennda> As of this last week, showing the right revision of a Page now works. There have been several fixes to Page.py, PageEditor.py and the diff action.
2008-08-06T13:27:54 <dennda> Furthermore, I started rewriting the AttachFile-action. Here is what works: Uploading an attachment, getting an attachment, deleting an attachment, listing all attachments with size and timestamp.
2008-08-06T13:28:01 <ThomasWaldmann> dennda is next
2008-08-06T13:28:09 <dennda> I already started :)
2008-08-06T13:28:30 <dennda> Additionally, johill committed a bunch of changes, so he may (or may not) want to say a few words about these.
2008-08-06T13:28:41 <johill> well I did a few api changes wrt. revision.timestamp/size
2008-08-06T13:29:04 <dennda> Those are already used in attachfile
2008-08-06T13:29:04 <johill> those we had discussed before, so I won't elaborate on them now, basically it's to have news() be a trivial optimisation _only_ and not have extra data
2008-08-06T13:29:22 <johill> wrt. attachments, I have a few questions though
2008-08-06T13:29:28 <dennda> Shoot
2008-08-06T13:29:30 <johill> right now, attachments are created as sub-items as we wanted
2008-08-06T13:29:54 <johill> do we want, instead of /page/action=attach;do=view;name=abc have
2008-08-06T13:30:01 <johill> /page/abc
2008-08-06T13:30:13 <johill> and instead of /page?actionn=attach;do=get;name=abc have /page/abc?raw
2008-08-06T13:30:17 <johill> action=raw
2008-08-06T13:30:32 <johill> action=info could work too, but many of the other actions need to be blocked, at lesat for the time being
2008-08-06T13:31:16 <ThomasWaldmann> johill: looking at the few days left, i would advise against changing that now
2008-08-06T13:31:29 <johill> we have to touch the code anyway, currently it just breaks
2008-08-06T13:31:36 <johill> ("unicode decode error" or so)
2008-08-06T13:31:45 <ThomasWaldmann> but rather cleanup Page/PageEditor/AttachFile and remove all the old crap and fix the TODOs
2008-08-06T13:31:55 <johill> but yeah, we can just block all actions on attachment items for now
2008-08-06T13:32:02 <xorAxAx> yep
2008-08-06T13:32:07 <johill> including show
2008-08-06T13:32:38 <dennda> Why do we need to block them?
2008-08-06T13:32:43 <johill> because they crash?
2008-08-06T13:33:04 <dennda> Havn't had a crash yet, but ok :)
2008-08-06T13:33:10 <johill> well just upload a jpeg, and go to the item
2008-08-06T13:33:14 <dennda> I wanted to look into the one you mentioned anyway
2008-08-06T13:33:17 <johill> i.e. /mypage/myimg.jpg
2008-08-06T13:33:44 <johill> the problem is that this is listed in RC
2008-08-06T13:33:57 <dennda> Ah I didn't know that kind of url already worked
2008-08-06T13:33:57 <johill> so that'll be a bit weird, at least I suppose it should redirect to the attachfile action
2008-08-06T13:34:38 <johill> ok another thing
2008-08-06T13:34:40 <johill> about testing
2008-08-06T13:34:44 <ThomasWaldmann> johill: in the end, the urls will be like your proposed (but action=raw for now as other page code also uses that)
2008-08-06T13:35:18 <johill> currently all testing code assumes that we have underlay present
2008-08-06T13:35:24 <johill> do we really want that? the tests are very fragile due to this
2008-08-06T13:35:43 <xorAxAx> well, the new storage branch doesnt handle backend setup correctly anyway
2008-08-06T13:35:45 <johill> this would be a good opportunity to just run them on a memory backend, and indeed we already do, but that means a lot of tests fail
2008-08-06T13:35:54 <xorAxAx> and it hasnt solved the plugin package problem yet, johill
2008-08-06T13:36:02 <johill> plugin package problem?
2008-08-06T13:36:16 <xorAxAx> the data directory is not only used to hold structured data but also code
2008-08-06T13:36:37 <xorAxAx> the plugin loading code needs to be fixed to have a notion of a user plugin directory instead of a data directory which would vanish
2008-08-06T13:37:03 <johill> eh, first of all, so far only data/pages and data/user vanishes
2008-08-06T13:37:17 <johill> secondly, I already committed arbitrary plugin dir paths support to 1.8
2008-08-06T13:38:31 <xorAxAx> when i had my last screen session with dennda, multiconfig didnt reflect that
2008-08-06T13:38:35 <xorAxAx> not sure when you did :)
2008-08-06T13:38:44 <johill> it might not have been merged yet
2008-08-06T13:38:48 <johill> but I merged up 1.8 now
2008-08-06T13:38:58 <ThomasWaldmann> I would like to see a working wiki, with underlay pages, out of the box of that repo
2008-08-06T13:39:32 <dennda> So do we :)
2008-08-06T13:39:34 <johill> we still need to implement a read-only 1.7-style backend
2008-08-06T13:39:50 <ThomasWaldmann> didn't we have that already?
2008-08-06T13:39:57 <johill> no
2008-08-06T13:40:15 <johill> the old API stuff had, but it's be pointless to port that horrible code
2008-08-06T13:40:35 <TheSheep> at least the bugs would be ported accurately
2008-08-06T13:40:35 <ThomasWaldmann> lots to do
2008-08-06T13:40:35 <johill> anyhow
2008-08-06T13:40:58 <johill> I'd like to slow convert to a model where tests are run on a clean slate
2008-08-06T13:41:00 <ThomasWaldmann> so the plan until monday is?
2008-08-06T13:41:03 <johill> does anyone disagree?
2008-08-06T13:41:15 <johill> (probably not going to happen soon though)
2008-08-06T13:41:47 <xorAxAx> johill: makes perfect sense
2008-08-06T13:42:09 <xorAxAx> johill: you could e.g. migrate from the underlay snapshot into the memory backend :)
2008-08-06T13:42:24 <xorAxAx> clean state != empty state :-)
2008-08-06T13:42:27 <johill> yeah
2008-08-06T13:42:40 <johill> pulling all of underlay into memory would probably take a long time though
2008-08-06T13:42:50 <johill> so it'd be smarter to just have some sort of "minimal testing state"
2008-08-06T13:42:53 <xorAxAx> yeah
2008-08-06T13:43:02 <xorAxAx> easy to do
2008-08-06T13:43:10 <xorAxAx> if migration/copying code is there
2008-08-06T13:44:05 <johill> right
2008-08-06T13:44:10 <johill> ok that's all from me, back to dennda
2008-08-06T13:45:04 <dennda> ok
2008-08-06T13:45:13 <dennda> What's next: Complete work on AttachFile, start work on migration scripts.
2008-08-06T13:45:20 <dennda> and what was just discussed
2008-08-06T13:45:24 <dennda> No Blockers
2008-08-06T13:46:22 <ThomasWaldmann> dennda: you coordinate the mig script stuff with PawelPacana
2008-08-06T13:46:38 <dennda> Yep
2008-08-06T13:46:56 <ThomasWaldmann> ok, thanks dennda / johill
2008-08-06T13:47:04 <ThomasWaldmann> PawelPacana: TheSheep: your stage
2008-08-06T13:47:08 <PawelPacana> ok
2008-08-06T13:47:38 <PawelPacana> i was messing with merging stuff: merging pages on concurrent edits which have same parent revision
2008-08-06T13:47:57 <johill> I thought we made it abundantly clear that you're not supposed to do that, why are you playing with it?
2008-08-06T13:48:00 <PawelPacana> currently merging is automated and implemented in backend (diff3)
2008-08-06T13:48:21 <PawelPacana> it will be left to user
2008-08-06T13:48:24 <PawelPacana> johill: when?
2008-08-06T13:49:22 <johill> uh, I don't know, probably half a dozen times? maybe TheSheep failed to pass you the memo
2008-08-06T13:49:52 <TheSheep> johill: who's the "we"?
2008-08-06T13:50:08 <johill> xorAxAx and me
2008-08-06T13:50:30 <TheSheep> johill: I recall you mentioned several times that our whole project doesn't make sense, but somehow I didn't think it was worth paying attention to
2008-08-06T13:51:01 <johill> I never said anything like it
2008-08-06T13:51:13 <PawelPacana> we mentioned merging even on last meeting
2008-08-06T13:51:19 <TheSheep> johill: lets let PawelPacana talk for now
2008-08-06T13:51:22 <johill> but we repeatedly stated that merging isn't supposed to be done in the backend
2008-08-06T13:51:29 <PawelPacana> it wont be
2008-08-06T13:51:29 <xorAxAx> well
2008-08-06T13:51:32 <PawelPacana> if you let mi finish
2008-08-06T13:51:34 <PawelPacana> ok?
2008-08-06T13:51:39 <xorAxAx> hg needs to merge always
2008-08-06T13:51:41 <xorAxAx> the manifest
2008-08-06T13:52:07 <PawelPacana> merging wont be done in backendd level
2008-08-06T13:52:22 <PawelPacana> it is now, until we have action to switch page heads
2008-08-06T13:52:30 <PawelPacana> and action to merge such heads
2008-08-06T13:52:40 <PawelPacana> which is in the upper-level
2008-08-06T13:52:55 <johill> why does it have to be done there now until you get that
2008-08-06T13:53:14 <johill> the backend API requires to just fail one of concurrent edits
2008-08-06T13:53:29 <PawelPacana> merging in backend is done there to work on tree info action
2008-08-06T13:54:58 <TheSheep> johill: it's temporary, because for now there is no user interface to invoke merge manually
2008-08-06T13:55:11 <TheSheep> johill: this action is the next point on the schedule
2008-08-06T13:55:19 <xorAxAx> johill: well, hg preserves exact history
2008-08-06T13:55:37 <xorAxAx> johill: and in order to come back to linear history, a merge changeset is needed
2008-08-06T13:55:55 <xorAxAx> as long as the api doesnt provide a merge op, this means that it is disjunct from the contents merge
2008-08-06T13:56:56 <johill> I don't understand, but I'm not going to bother if you'll just accuse me of sabotaging your project anyway
2008-08-06T13:57:12 <johill> so just go on
2008-08-06T13:57:35 <PawelPacana> for metadata merging:
2008-08-06T13:57:41 <PawelPacana> (revision)
2008-08-06T13:58:05 <PawelPacana> if there are any conflicts, newer values are taken
2008-08-06T13:58:30 <johill> ok can you explain one last thing?
2008-08-06T13:58:33 <PawelPacana> as i mentioned: merging goes to user merge and heads actions
2008-08-06T13:58:39 <johill> how can you possibly even get into a conflict situation?
2008-08-06T13:59:25 <PawelPacana> i two pages have same revision meta keys and different values for them
2008-08-06T13:59:49 <PawelPacana> how into conflict invoking merging?
2008-08-06T14:00:08 <PawelPacana> or what?
2008-08-06T14:00:15 <johill> describe a series of backend API operations that triggers this conflict/merging stuff
2008-08-06T14:00:27 <johill> but we can postpone to after the meeting for benefit of all the others
2008-08-06T14:00:49 <PawelPacana> ok, this will be better
2008-08-06T14:01:17 <PawelPacana> so
2008-08-06T14:01:29 <PawelPacana> the new stuff was added: news, timestamp, size
2008-08-06T14:01:43 <PawelPacana> on my two merge tests fail /115
2008-08-06T14:02:21 <PawelPacana> todo: info, merge, heads actions, migration script with dennda, merge fixes
2008-08-06T14:02:46 <PawelPacana> blockers none for now
2008-08-06T14:02:59 <ThomasWaldmann> waldi: you're next after PawelPacana has finished
2008-08-06T14:03:20 <PawelPacana> s/on/only/
2008-08-06T14:04:30 <ThomasWaldmann> PawelPacana: TheSheep: you are finished?
2008-08-06T14:04:37 <PawelPacana> if i disable merging, backend is purely API compatible [and has no special features]
2008-08-06T14:04:50 <PawelPacana> we're done for now
2008-08-06T14:05:09 <ThomasWaldmann> ok, thanks, waldi now please :)
2008-08-06T14:05:16 <waldi> what work: moinmoin wiki parser except of table and link parameters
2008-08-06T14:05:16 <waldi> what not work:
2008-08-06T14:05:16 <waldi> - query strings in wiki links vs. ? in link
2008-08-06T14:05:16 <waldi> - wiki formatted getText with percent hack
2008-08-06T14:05:16 <waldi> - editor quick help
2008-08-06T14:05:19 <waldi> - pagelinks: links with query strings/fragments
2008-08-06T14:05:21 <waldi> next:
2008-08-06T14:05:23 <ThomasWaldmann> :)
2008-08-06T14:05:24 <waldi> - factor out common code of creole and moin wiki parser
2008-08-06T14:06:23 <ThomasWaldmann> you wanted to look at lineno anchors, did you have a look already?
2008-08-06T14:07:16 <waldi> the line informations are available, but no real id how to write them into the output
2008-08-06T14:07:24 <waldi> s/id/idea/
2008-08-06T14:07:43 <ThomasWaldmann> the old stuff used spans
2008-08-06T14:08:03 <ThomasWaldmann> the new stuff could do the same, but just emit them once
2008-08-06T14:08:07 <ThomasWaldmann> per line
2008-08-06T14:08:48 <ThomasWaldmann> and where exactly is that line information available?
2008-08-06T14:09:45 <waldi> i'm not sure if i commit it already. its in the iterator, which operates on lines anyway
2008-08-06T14:10:29 <ThomasWaldmann> ok, just don't forget, a diff display jumping to the right place in full text is nice to have
2008-08-06T14:10:57 <ThomasWaldmann> the most important thing of that list is query string support I think
2008-08-06T14:11:05 <ThomasWaldmann> because much other code depends on that
2008-08-06T14:11:12 <TheSheep> .oO(it would be nice to only display the anchors that are actually used)
2008-08-06T14:11:20 <waldi> the old code does not even support query strings
2008-08-06T14:11:50 <waldi> TheSheep: maybe they should only be rendered if requested (by the link in the diffs)
2008-08-06T14:12:29 <ThomasWaldmann> TheSheep: that would be a nice additional feature (after SOC)
2008-08-06T14:12:59 <ThomasWaldmann> waldi: what's needed is an easy way to create elements including qs kv pairs
2008-08-06T14:13:08 <ThomasWaldmann> (in their unescaped form)
2008-08-06T14:13:24 <ThomasWaldmann> and then have the html converter escape and assemble the stuff
2008-08-06T14:14:19 <waldi> ThomasWaldmann: in the wiki links?
2008-08-06T14:14:33 <ThomasWaldmann> yes
2008-08-06T14:14:35 <CIA-53> Johannes Berg <johannes AT sipsolutions DOT net> default * 4567:b741018b7d45 1.8-storage-cdenter/MoinMoin/storage/ (__init__.py backends/fs.py): storage: remove useless timestamp parameter from news() iterator
2008-08-06T14:14:55 <ThomasWaldmann> [[foo||&rev=42]]
2008-08-06T14:15:16 <waldi> hmm
2008-08-06T14:15:17 <ThomasWaldmann> should make a link element to wiki page foo with rev=42
2008-08-06T14:15:58 <ThomasWaldmann> or &template=SomethingWithUmlautsandQuotes should also work and not need escaping at that place
2008-08-06T14:16:35 <waldi> this have to be done in the wiki parser, but should work
2008-08-06T14:16:38 <ThomasWaldmann> quite some of the macro I did not touch yet neet such stuff
2008-08-06T14:16:57 <ThomasWaldmann> s/t/d/ :)
2008-08-06T14:17:32 <ThomasWaldmann> before you use it in the parser, there needs to be a converter offering that feature
2008-08-06T14:17:50 <ThomasWaldmann> (and then the macros that are blocked currently can be done also)
2008-08-06T14:18:10 <waldi> the converter already does it, if the wiki parser provides a proper escaped local url
2008-08-06T14:18:42 <ThomasWaldmann> but the goal is not to have to assemble and escape early
2008-08-06T14:19:32 <ThomasWaldmann> btw, images/icons are also partly missing still, right?
2008-08-06T14:19:41 <waldi> right
2008-08-06T14:19:50 <waldi> icons are missing
2008-08-06T14:20:14 <ThomasWaldmann> ok, so plan until monday is?
2008-08-06T14:20:43 <waldi> urls, table parameters
2008-08-06T14:21:12 <waldi> links to not available wiki pages
2008-08-06T14:22:40 <ThomasWaldmann> btw, maybe that percent hack solves itself it the other stuff is done right
2008-08-06T14:23:19 <ThomasWaldmann> (we can't use % anyway if we work with ETs)
2008-08-06T14:23:35 <waldi> the percents needs to be expanded before the wiki parser. what does currently happen if someone uses "[[%s]]"?
2008-08-06T14:24:14 <ThomasWaldmann> it depends on whether percent=True or False
2008-08-06T14:24:15 <waldi> we could use the zope page templates style translations. they act on a tree
2008-08-06T14:24:21 <waldi> with percent=True
2008-08-06T14:24:51 <waldi> i think it will do something like <a href="%%s">%s</a>
2008-08-06T14:24:58 <ThomasWaldmann> well, i think it will create a link to a page called %s X)
2008-08-06T14:25:15 <waldi> exactly
2008-08-06T14:25:28 <ThomasWaldmann> (but we have no use case yet, afaik)
2008-08-06T14:26:04 <ThomasWaldmann> waldi: maybe you should also try to fix some of the other points until monday
2008-08-06T14:26:13 <waldi> will try
2008-08-06T14:26:15 <ThomasWaldmann> monday is suggested "stop coding" date
2008-08-06T14:26:29 <waldi> at least the paggelinks
2008-08-06T14:26:35 <ThomasWaldmann> and you will also need some time to clean up and document afterwards
2008-08-06T14:27:16 <ThomasWaldmann> ok, looks like we are finished. thanks to everyone being here.
2008-08-06T14:27:36 <zenhase> :)
2008-08-06T14:28:05 <zenhase> btw ... when reading 'blockers' i somehow always have think of beta blockers :O
2008-08-06T14:28:06 * waldi found a poster about tdd while he tries to get a date for its exam
2008-08-06T14:28:37 <ThomasWaldmann> waldi: maybe you can look at the qs stuff next, so we can talk later in tamm
2008-08-06T14:29:25 <PawelPacana> johill: this is API behaviour you mentioned http://hg.moinmo.in/moin/1.8-mercurialbackend-ppacana/rev/d510df9a945c , anyway your fs backend is not passing this test
2008-08-06T14:30:18 <PawelPacana> here is what hg-back does: http://hg.moinmo.in/moin/1.8-mercurialbackend-ppacana/rev/41bd415980e5
2008-08-06T14:31:00 <johill> odd, I'd have thought it does pass
2008-08-06T14:31:27 <johill> yeah that part is bogus
2008-08-06T14:31:36 <johill> moin's upper layer should be able to expect that behaviour
2008-08-06T14:32:34 <PawelPacana> johill: maybe you/dennda didnt merge this test?
2008-08-06T14:33:20 <johill> I don't think dennda has merged in a while unfortunately
2008-08-06T14:33:31 <dennda> umm
2008-08-06T14:33:39 <PawelPacana> this is the only one failing for fs
2008-08-06T14:33:47 <dennda> actually I did, maybe I didn't commit
2008-08-06T14:34:02 * dennda checks
2008-08-06T14:34:38 <dennda> I'll just merge again
2008-08-06T14:36:27 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4568:04058f3528af 1.8-storage-cdenter/MoinMoin/action/AttachFile.py: storage: action/AttachFile.py: Use recently added rev.timestamp rather than manually saving EDIT_LOG_MTIME in a revisions metadata
2008-08-06T14:36:37 <dennda> zzzzZZzzzZzzZ
2008-08-06T14:37:02 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4569:69d13294dae0 1.8-storage-cdenter/MoinMoin/ (6 files in 4 dirs): Merge with my own branch, 1.8-storage-cdenter
2008-08-06T14:37:28 <TheSheep> cildren of the night :)
2008-08-06T14:37:35 <johill> yeah, there's really a stupid thinko in my code
2008-08-06T14:37:36 <ThomasWaldmann> dennda: how long do commits take / irc postings of CIA take?
2008-08-06T14:38:29 <dennda> ThomasWaldmann: hm, this commit roughly took about 1-2 minutes until hg ended, and CIA reported ca. 20-50 secs after the changeset was pushed
2008-08-06T14:39:52 <ThomasWaldmann> ok, then it seems to be slightly below the timeout now at least...
2008-08-06T14:39:59 <dennda> yes
2008-08-06T14:40:14 <dennda> It's still annoying, but one can live with it for the moment
2008-08-06T14:40:52 <ThomasWaldmann> (you don't need to wait for it :)
2008-08-06T14:41:24 <dennda> for it to finish? I know
2008-08-06T14:41:46 <dennda> .
2008-08-06T14:41:50 <dennda> here we go again
2008-08-06T14:42:01 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4617:50694191da29 1.8-storage-cdenter/ (6 files in 4 dirs): Merge with 1.8-mercurialbackend-ppacana
2008-08-06T14:42:17 <johill> ok, hmm
2008-08-06T14:42:28 <johill> this makes it really unix specific, using os.link()
2008-08-06T14:43:14 <dennda> don't ask me, I keep away from windows as far as possible... (just had to install it the other day. god was that a pain) :)
2008-08-06T14:43:39 <ThomasWaldmann> byeongweon: + self.text.extend([("{" * (len(longest_inner_formater) + 1)) + bang_args, \
2008-08-06T14:43:57 <ThomasWaldmann> there should be at least a comment that this is KNOWN BROKEN
2008-08-06T14:44:15 <ThomasWaldmann> (we already talked about it)
2008-08-06T14:44:36 <dennda> johill: Why are those empty lines annoying? I actually think it makes the whole thing more readable
2008-08-06T14:46:13 <CIA-53> Johannes Berg <johannes AT sipsolutions DOT net> default * 4618:43ab81146e22 1.8-storage-cdenter/MoinMoin/storage/backends/fs.py:
2008-08-06T14:46:13 <CIA-53> fs storage: fix revision creation bug
2008-08-06T14:46:13 <CIA-53> Unfortunately this makes the backend Unix specific. We'll have to
2008-08-06T14:46:13 <CIA-53> see if we can come up with a Windows one, probably adding a lot of
2008-08-06T14:46:13 <CIA-53> filesystem locks.
2008-08-06T14:46:31 <johill> dennda: I think it tears apart the if/else block and thus makes it visually less apparent that it's one block, so I tend to get confused
2008-08-06T14:46:55 <dennda> Ok, will change it then
2008-08-06T14:47:08 * dennda checks whether PEP8 mentions how to handle it
2008-08-06T14:47:50 <ThomasWaldmann> ehrm
2008-08-06T14:48:24 <johill> well it has a section on blank lines
2008-08-06T14:48:41 <johill> "Use blank lines in functions, sparingly, to indicate logical sections.
2008-08-06T14:48:42 <johill> "
2008-08-06T14:48:51 <ThomasWaldmann> johill: what bug did you fix in that cs?
2008-08-06T14:48:57 <dennda> alright
2008-08-06T14:49:08 <johill> ThomasWaldmann: os.rename() will overwrite, that's not what we want here
2008-08-06T14:50:27 <ThomasWaldmann> i think there is a comment in the src missing (not just in the cs comment)
2008-08-06T14:51:06 <johill> double standard
2008-08-06T14:51:18 <ThomasWaldmann> ?
2008-08-06T14:51:46 <johill> besides, I don't see how
2008-08-06T14:51:58 <johill> if you're not familiar with unix fs semantics, you're lost in that backend anyway
2008-08-06T14:52:48 <ThomasWaldmann> lik # TODO XXX this is a regression as it incompatible to non-posix platform, find a way how to rename (and not overwrite) that works for posix and win32
2008-08-06T14:53:28 <ThomasWaldmann> that's not the point, the point is that the problem is more visible
2008-08-06T14:53:47 <ThomasWaldmann> and not hidden in the 3rd line of a commit comment
2008-08-06T14:54:08 <johill> the whole thing doesn't work on windows anyway
2008-08-06T14:55:00 <ThomasWaldmann> then there is another comment in module level docstring missing
2008-08-06T14:56:18 <johill> feel free
2008-08-06T14:56:24 <ThomasWaldmann> 462 # XXXX: doesn't work on windows
2008-08-06T14:56:34 <ThomasWaldmann> there are multiple of those comments
2008-08-06T14:56:41 <ThomasWaldmann> at other places
2008-08-06T14:58:32 <johill> those are non-obvious
2008-08-06T14:58:43 <johill> because it might work under no-races conditions
2008-08-06T14:58:51 <ThomasWaldmann> everthing not automatically found is non-obvious
2008-08-06T14:58:59 <johill> this is found automatically
2008-08-06T14:59:04 <johill> there's a large test suite
2008-08-06T14:59:10 <ThomasWaldmann> some people use tools that extract XXX (3 X btw, not 4) and TODO automatically
2008-08-06T14:59:27 <johill> do I look like I care/
2008-08-06T14:59:27 <ThomasWaldmann> but if there aren't any, you can only find it by manual, line by line review
2008-08-06T14:59:28 <johill> ?
2008-08-06T14:59:38 <ThomasWaldmann> or some windows user submitting an angry bug report
2008-08-06T14:59:45 <johill> look
2008-08-06T14:59:48 <johill> if you want to fix it, go ahead
2008-08-06T15:00:09 <johill> I can spend my time in better ways than to document things that can obviously not work on windows and will never work because the function simply doesn't exist
2008-08-06T15:00:12 <ThomasWaldmann> and I guess your tests ill NOT find it if they are executed on linux
2008-08-06T15:00:20 <johill> obviously
2008-08-06T15:00:26 <johill> this is why I cannot actually implement a windows backend
2008-08-06T15:00:32 <johill> I don't happen to have a windows machine to test on
2008-08-06T15:00:42 <ThomasWaldmann> so why are we discussing that, this is just sloppyness and hiding stuff
2008-08-06T15:00:56 <ThomasWaldmann> and you should know better
2008-08-06T15:01:01 <johill> it's well known that this thing cannot work on windows
2008-08-06T15:01:07 <johill> so anyone wanting to make it work needs to test it
2008-08-06T15:01:15 <ThomasWaldmann> then document it!
2008-08-06T15:01:29 <ThomasWaldmann> line 3: XXX does not work on win32
2008-08-06T15:02:15 <johill> I don't give a fuck though
2008-08-06T15:02:17 <ThomasWaldmann> and for the case that some win32 dev wants to fix it, it would be really nice to also have that XXX mark at all known places, not just at some and some others not
2008-08-06T15:02:40 <ThomasWaldmann> pfff
2008-08-06T15:08:20 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4619:708f4de213e1 1.8-storage-cdenter/MoinMoin/storage/backends/fs.py: adding a clear comment about win32 incompatibility is not that hard
2008-08-06T15:11:42 <ThomasWaldmann> ok, that basically means that we have not a single working backend on win32, because the project mentor doesn't give a fuck (quoting him).
2008-08-06T15:12:36 <ThomasWaldmann> (except hg maybe(?), that is requiring a development version of hg)
2008-08-06T15:13:09 <ThomasWaldmann> and there also is no moin17 backend
2008-08-06T15:13:29 <xorAxAx> who is using win32 anyway ... moin doesnt need 90% of its users :)
2008-08-06T15:13:45 <xorAxAx> (or more realistically, 50%)
2008-08-06T15:14:10 <dennda> are there that many?
2008-08-06T15:14:50 <xorAxAx> yep
2008-08-06T15:14:53 <ThomasWaldmann> moin 1.7 runs without problems on win32 and quite some people use that, yes
2008-08-06T15:15:03 <xorAxAx> when mmde was distributed separately, the download ratio was about 1:1
2008-08-06T15:15:04 <ThomasWaldmann> some even run it on windows servers
2008-08-06T15:15:11 <xorAxAx> and i suppose that nearly all of those were win32 users
2008-08-06T15:15:27 <xorAxAx> hmm, indeed, the .exe numbers were high and that download only made sense on windows
2008-08-06T15:16:36 <ThomasWaldmann> dennda: if I have some free time for this, I maybe try working on the 1.7 backend. do you know if there is uncommitted stuff somewhere for it?
2008-08-06T15:16:53 <dennda> not onmy disk
2008-08-06T15:17:03 <ThomasWaldmann> (because doing a migration script is rather pointless if you have no source data to do a practical test)
2008-08-06T15:18:17 <ThomasWaldmann> dennda: ok, find out if johill has. I guess I won't discuss with him until he is behaving reasonable again.
2008-08-06T15:18:50 <dennda> ok, will try
2008-08-06T15:19:11 <dennda> but for lunch now
2008-08-06T15:32:19 <ThomasWaldmann> dennda: ok, find out if johill has. I guess I won't discuss with him until he is behaving reasonable again.
2008-08-06T15:32:22 <ThomasWaldmann> oops
2008-08-06T15:39:40 <dennda> ThomasWaldmann: just wondering while cooking: what do you want to do with the 1.7 backend?
2008-08-06T15:41:35 <ThomasWaldmann> we need something that can read existing data (call it 1.6 or 1.7, it doesn't make a difference)
2008-08-06T15:42:06 <dennda> ah you mean that
2008-08-06T15:42:15 <dennda> ok, I understood it differently. that is great. thanks
2008-08-06T15:52:17 <xorAxAx> dennda: lanius already started some code
2008-08-06T15:52:28 <xorAxAx> but he also tried to make it support modifications
2008-08-06T15:52:34 <xorAxAx> which is not really necessary
2008-08-06T15:53:00 <ThomasWaldmann> dennda: please work on some not moin1.7 backend related field
2008-08-06T15:53:14 <xorAxAx> as there is no spec, i cant judge how much the new backend api is different
2008-08-06T15:53:27 <xorAxAx> but i dont think that not looking at the old code is a good idea :)
2008-08-06T15:54:56 <xorAxAx> ThomasWaldmann: hmm, how do you mean that?
2008-08-06T15:55:25 <ThomasWaldmann> there is lots of other stuff todo and I currently look at moin17 backend, so it is not useful to do conflicting work
2008-08-06T15:55:31 <xorAxAx> ah
2008-08-06T15:55:50 <xorAxAx> yeah, you could try to get it working with the existing test code
2008-08-06T15:56:03 <dennda> ThomasWaldmann: sure
2008-08-06T15:56:24 <xorAxAx> (theoretically its a matter of making all tests pass :))
2008-08-06T15:56:42 <xorAxAx> and practically one of adding new tests in case of issues
2008-08-06T15:57:11 <dennda> yeah we all know TDD, don't we? :)
2008-08-06T16:02:47 <xorAxAx> dennda: yeah, but because people like me forget the advantages of it too quickly, i like to repeat the spirit as a mantra :)
2008-08-06T16:05:12 <dennda> i see
2008-08-06T16:29:19 <dreimark> re
2008-08-06T17:00:21 <TheSheep> xorAxAx: the pain of not having tests is much more memorable
2008-08-06T17:13:28 <dreimark> bbl
2008-08-06T17:17:41 <xorAxAx> TheSheep: yeah
2008-08-06T18:15:02 <ThomasWaldmann> dennda: reading?
2008-08-06T18:16:41 <dennda> reading what?
2008-08-06T18:17:16 <xorAxAx> the question whether you are reading :)
2008-08-06T18:17:32 <ThomasWaldmann> yes
2008-08-06T18:18:16 <dennda> The screen in my peripheral view produces a red bar if someone hilights me, so yes :)
2008-08-06T18:18:26 <ThomasWaldmann> dennda: am I seeing it correctly that a deleted rev is a revision on its own with no content but deleted marker in metadata?
2008-08-06T18:18:37 <dennda> ThomasWaldmann: Yes
2008-08-06T18:18:55 <dennda> That's how it is implemented right now
2008-08-06T18:19:12 <ThomasWaldmann> what is that marker?
2008-08-06T18:20:04 <dennda> ThomasWaldmann: MoinMoin.storage.DELETED
2008-08-06T18:20:17 * ThomasWaldmann just sees some strange code in _write_file
2008-08-06T18:20:33 <ThomasWaldmann> ### newrev.deleted = True # XXX Does it still work when this is commented out completely? Maybe store the info in metadata
2008-08-06T18:21:48 <ThomasWaldmann> also, _list_revisions leaves you unclear about what to report for deleted revisions
2008-08-06T18:21:48 <dennda> Ah yes, that comes from lanius
2008-08-06T18:22:09 <dennda> I'll look into it
2008-08-06T18:22:17 <dennda> I should make a TODO-List
2008-08-06T18:23:51 <ThomasWaldmann> you have TODO markers in the code, just add some / look at them :)
2008-08-06T18:24:16 <dennda> "make a TODO-List" == grep -r TODO *
2008-08-06T18:24:24 <xorAxAx> hmm, not good :)
2008-08-06T18:24:34 <dennda> well, limit the scope of *
2008-08-06T18:24:38 <xorAxAx> if the other people also use TODO
2008-08-06T18:24:52 <xorAxAx> (todo lists are most useful if they support the 4 fields mehtod :))
2008-08-06T18:24:58 <dennda> I didn't want to put all the files on that line just to make the intention clear
2008-08-06T18:26:47 <xorAxAx> and todo tags dont support the eisenhower principle
2008-08-06T18:26:48 <xorAxAx> http://seminare4you.de/weblog/article/44/das-eisenhower-prinzip-und
2008-08-06T18:28:05 <ThomasWaldmann> dennda: I'll have to leave for LUG meeting soon. Just commit some better docstring in the ABC, so I will see it.
2008-08-06T18:28:16 <dennda> ThomasWaldmann: See what?
2008-08-06T18:28:33 <ThomasWaldmann> your improved documentation, leaving no questions open :)
2008-08-06T18:28:57 <dennda> I will not put information outlining higher level metadata usage into the ABC or johill is going to kill me
2008-08-06T18:29:09 <dennda> :)
2008-08-06T18:29:25 <ThomasWaldmann> documenting revision list contents is not high level
2008-08-06T18:29:35 <dennda> Ah, that you mean
2008-08-06T18:29:55 <dennda> Ok, I will give it some thought
2008-08-06T18:31:15 <ThomasWaldmann> btw, you maybe remember that this questions come up again and again and as long as youdon't put answers into good docstrings, the questions will follow you...
2008-08-06T18:32:06 <xorAxAx> umm
2008-08-06T18:32:11 <xorAxAx> how about a "wiki page"?
2008-08-06T18:32:26 <xorAxAx> overviews should probably go into the wiki, concrete info into a docstring
2008-08-06T18:32:28 <ThomasWaldmann> that's step 2, first docstrings
2008-08-06T18:32:28 <dennda> Why not a docstring?
2008-08-06T18:33:01 <ThomasWaldmann> missing docstrings block me :P
2008-08-06T18:33:04 <xorAxAx> because a docstring is located at a specific place, overview over multiple elements is not - so you need a neutral differently organized place
2008-08-06T18:33:12 <xorAxAx> but docstrings are useful as well
2008-08-06T18:33:18 <xorAxAx> they just serve a different purpose
2008-08-06T18:33:29 <xorAxAx> maybe moin will migrate to sphinx in some distant parallel future
2008-08-06T18:33:32 <dennda> If you copy that ABC in order to implement it, docstrings are much more handy
2008-08-06T18:33:41 <dennda> I don't argue against the rest
2008-08-06T18:33:45 <xorAxAx> in that case docstrings combine well with additional textual documentation as well
2008-08-06T18:33:49 <dennda> Yes
2008-08-06T18:33:51 <dennda> so we agree
2008-08-06T18:34:15 <xorAxAx> dennda: well, but if you dont know where the particular ABC is useful in the whole system, you probably end up doing wrong decisions while implementing hte methods
2008-08-06T18:36:21 <dennda> ThomasWaldmann: I wonder what's unclear about that docstring. "For a given Item, list all Revisions. ...". Deleted Revisions are still there, it's just some MetaData inserted by some higher level
2008-08-06T18:37:28 <dennda> Mentioning the "also deleted revisions are included" thing there is somehow implicitly saying that you can a) really delete revisions, which you cannot, at the moment b) the backend knows about the metadata, which is not the case either and a bit layer violating
2008-08-06T18:37:38 <dennda> But if you think it's more clear I can add it. No problem
2008-08-06T18:38:02 <xorAxAx> see, probably thomas is needing more meta-information what a revision and an item is. he could look at the class docstrings but if that is still too spotty, he might prefer some design wiki page
2008-08-06T18:38:02 <ThomasWaldmann> dennda: for the fs17 backend i write, deleted revisions are not there (in the filesystem)
2008-08-06T18:38:19 <xorAxAx> ThomasWaldmann: no, their data file is not htere
2008-08-06T18:38:24 <dennda> ThomasWaldmann: I don't know about 1.7
2008-08-06T18:38:26 <xorAxAx> a revision consists of data and metadata
2008-08-06T18:38:29 <ThomasWaldmann> and if you want to report me deleted revs also, i will just return range(1, current+1)
2008-08-06T18:38:56 <xorAxAx> but deleted revs have an editlog entry, right?
2008-08-06T18:39:02 <ThomasWaldmann> hmm, or rather range(current) as you want it zero based
2008-08-06T18:39:25 <dennda> Do you want a wiki Page with UML-Diagrams? :)
2008-08-06T18:39:40 <ThomasWaldmann> xorAxAx: you know that edit-log is not extremely reliable
2008-08-06T18:40:10 <ThomasWaldmann> ok, you get range(current) :)
2008-08-06T18:40:11 <xorAxAx> ThomasWaldmann: yeah, i do. but how is that related?
2008-08-06T18:40:21 <xorAxAx> you still need the metadata from the editlog
2008-08-06T18:40:28 <xorAxAx> to build the revision
2008-08-06T18:40:35 * dennda creates a class-diagram
2008-08-06T18:40:38 <ThomasWaldmann> yes, but I can also create a rev if there is no edit-log entry
2008-08-06T18:40:43 <ThomasWaldmann> but just a file
2008-08-06T18:40:46 <ThomasWaldmann> data file
2008-08-06T18:40:57 <ThomasWaldmann> dennda: i want docstrings, not pics
2008-08-06T18:41:09 <xorAxAx> not sure a class diagram is the best solution
2008-08-06T18:41:13 <ThomasWaldmann> don't waste time you don't have
2008-08-06T18:41:14 <xorAxAx> think about a more textual one
2008-08-06T18:41:16 <xorAxAx> :-)
2008-08-06T18:41:23 <xorAxAx> (think about it, but dont do it :))
2008-08-06T18:41:30 <xorAxAx> (now)
2008-08-06T18:42:05 * dennda updates docstrings then
2008-08-06T18:42:11 <xorAxAx> ThomasWaldmann: when can you create it? i dont understand. the old wiki only created both at the same time, except when there was a bug in which case the old code nevertheless recognized a rev and showed some "-" in the info list
2008-08-06T18:42:12 <ThomasWaldmann> dennda: ok, so i return [0,1,2,3]?
2008-08-06T18:42:29 <dennda> ThomasWaldmann: in _list_revisions? If your Item has 4 Revisions, yes
2008-08-06T18:42:38 <xorAxAx> dennda: are those integers or objects?
2008-08-06T18:42:47 <xorAxAx> i meean, an intger is not a revision
2008-08-06T18:42:47 <ThomasWaldmann> ints
2008-08-06T18:42:47 <dennda> umm, in python: both
2008-08-06T18:42:52 <xorAxAx> so the docstring was wrong
2008-08-06T18:43:06 <xorAxAx> an integer is a revision identifier
2008-08-06T18:43:12 <xorAxAx> or revision number
2008-08-06T18:43:13 <ThomasWaldmann> if I have 10000 revs, I return range(10000)
2008-08-06T18:43:26 <dennda> xorAxAx: No, that's in the second sentence
2008-08-06T18:43:43 <dennda> For a given Item, list all Revisions. Returns a list of ints representing the Revision numbers.
2008-08-06T18:43:51 <dennda> that is what it says
2008-08-06T18:44:46 <xorAxAx> thats confusing
2008-08-06T18:44:53 <xorAxAx> see the javadoc guidelines
2008-08-06T18:45:01 <xorAxAx> please merge both sentences
2008-08-06T18:45:03 <dennda> No it's not, but OK :)
2008-08-06T18:45:05 <ThomasWaldmann> the fs17 backend shrinks :)
2008-08-06T18:45:08 <xorAxAx> it is, dennda
2008-08-06T18:45:15 <xorAxAx> it does /not/ list all revisions
2008-08-06T18:45:25 <dennda> Why not?
2008-08-06T18:45:26 <xorAxAx> it does neither return a list of all revisions
2008-08-06T18:45:46 <dennda> represented by their revision number as int
2008-08-06T18:45:52 <xorAxAx> exactly
2008-08-06T18:45:56 <xorAxAx> but numbers are not revisions
2008-08-06T18:46:02 <dennda> Which is clearly what the second sentence says.
2008-08-06T18:46:09 <xorAxAx> but the first one does not
2008-08-06T18:46:30 <dennda> Well, I do expect a dev to be able to read more than just the first sentence
2008-08-06T18:46:40 <dennda> but I can rephrase it, using a comma instead :)
2008-08-06T18:47:55 <xorAxAx> dennda: well, ambiguity at such places is bad
2008-08-06T18:48:10 <dennda> """For a given Item, return a list containing all revision numbers (as ints) of the Revisions the Item has."""
2008-08-06T18:48:14 <dennda> better?
2008-08-06T18:50:25 <xorAxAx> "revision numbers (as ints) of the Revisions" is a bit ambiguous
2008-08-06T18:50:34 <xorAxAx> because revs dont have revision numbers
2008-08-06T18:50:39 <xorAxAx> in the model
2008-08-06T18:50:58 <xorAxAx> but i think thats enough nit picking for today
2008-08-06T18:51:09 <dennda> You are not a lawyer, are you? :)
2008-08-06T18:51:20 <xorAxAx> i am, a language lawyer.
2008-08-06T18:51:25 <xorAxAx> well, i prefer well understandable tex
2008-08-06T18:51:26 <xorAxAx> t
2008-08-06T18:51:59 <xorAxAx> its an art to retain brevity, correctness and style in such api docs
2008-08-06T18:52:16 <xorAxAx> thats why nobody writes them :)
2008-08-06T18:52:42 <dennda> I always had good marks in my German and English classes, maybe that makes you feel a little more comfortable :)
2008-08-06T18:53:36 <xorAxAx> hmm, probably they didnt have any technical writing tasks in the exams :)
2008-08-06T18:54:08 <TheSheep> writing documentation is *hard*
2008-08-06T18:54:10 <dennda> You don't need a technical subject to produce precise text
2008-08-06T18:54:33 <xorAxAx> TheSheep: yep, and if its just a docstring
2008-08-06T18:55:01 <xorAxAx> dennda: is precise writing a goal in german courses?
2008-08-06T18:55:12 <dennda> In mine it was, yes
2008-08-06T18:55:47 <dennda> Or take philosophy. There you are probably even more concerned about your texts being precise
2008-08-06T18:55:58 <dennda> (I had good marks there as well, so don't worry)
2008-08-06T18:55:59 <xorAxAx> i had the impression that i didnt learn much there, the feedback lacked quality. it was completly different in english, though (also wrt the contents/style, not only the grammar)
2008-08-06T18:56:29 * TheSheep waves his copy of Strunk & White ;)
2008-08-06T18:56:42 <xorAxAx> dennda: brevity wasnt really a goal in my philosophy classes. people who formed the longest sentences with logical deductions from incorrect assumptions won
2008-08-06T18:57:04 <dennda> Brevity is not so important, that's true
2008-08-06T18:57:58 <TheSheep> it's not so much about brevity as about to-the-pointness
2008-08-06T18:58:55 <xorAxAx> depends on the reader
2008-08-06T19:23:22 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4620:169f00419458 1.8-storage-cdenter/MoinMoin/storage/backends/fs17.py: storage backend for moin 1.7 (currently rather pseudo code, needs more work)
2008-08-06T19:23:47 <dennda> \o/
2008-08-06T19:25:59 <ThomasWaldmann> don't try to run that, but please read it and add comments in case I have misunderstood something
2008-08-06T19:26:43 * ThomasWaldmann will add more (and more working) code later
2008-08-06T19:26:45 <ThomasWaldmann> bbl
2008-08-06T19:31:02 <dennda> pleasure
2008-08-06T19:31:56 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4609:b79ddf508944 1.8-mercurialbackend-ppacana/MoinMoin/storage/_tests/test_backends_hg.py: (hg tests) Adjusted merge tests. Documented weird behaviour.
2008-08-06T19:31:56 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4610:fababdad0113 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Moin's diff3 data merging (yet still in backend-level)
2008-08-06T19:33:01 <xorAxAx> PawelPacana: hmm, didnt johill ask for an operation sequence where diff3 is necessary? (just to allow us to understand)
2008-08-06T19:33:31 <dennda> ThomasWaldmann: No, johill has no uncommitted 1.7-readonly-backend changes of interest
2008-08-06T19:33:43 <PawelPacana> xorAxAx: yes, and i showed him link
2008-08-06T19:33:55 <xorAxAx> PawelPacana: ah, let me grep my logs, must have missed it
2008-08-06T19:33:55 <PawelPacana> xorAxAx: do yo want to see it?
2008-08-06T19:34:24 <xorAxAx> test_concurrent_create_revision?
2008-08-06T19:34:34 <PawelPacana> yes
2008-08-06T19:34:38 <xorAxAx> ok
2008-08-06T19:34:57 <PawelPacana> an additionaly merge tests for hg
2008-08-06T19:35:20 <xorAxAx> looks sensible - i guess thats what the revno argument was made for
2008-08-06T19:35:21 <PawelPacana> bbl
2008-08-06T19:36:22 <xorAxAx> PawelPacana: can you replace the pass in the overriding method by a skip?
2008-08-06T19:36:33 <xorAxAx> that contains a short version of the docstring
2008-08-06T19:37:13 <PawelPacana> ok, will change it
2008-08-06T19:37:26 <xorAxAx> otherwise it might look like the backend is passing the test even though it does not
2008-08-06T19:38:19 <xorAxAx> but nevertheless i think that the additional tests violate the API
2008-08-06T19:38:40 <TheSheep> xorAxAx: you mean the features that they test
2008-08-06T19:38:51 <xorAxAx> no, the tests itself
2008-08-06T19:39:02 <xorAxAx> the test is assuming that a concurrent commit is passing without exception
2008-08-06T19:39:21 <xorAxAx> while the api says that there should be an exception and thats what the upper layers in moin depend on
2008-08-06T19:39:23 <TheSheep> xorAxAx: you think it would be better to conform to the API completely, and have extended functions for the "nonstandard" behavior?
2008-08-06T19:39:40 <xorAxAx> also deisgn wise i am not sure whether data-merging in the storage backend makes any sense
2008-08-06T19:39:49 <xorAxAx> TheSheep: the latter of course and also the former :)
2008-08-06T19:40:01 <xorAxAx> well, i sketched an idea today how one would solve this issue
2008-08-06T19:40:11 <PawelPacana> xorAxAx: data merging i backend level is temporary, just to remind you
2008-08-06T19:40:16 <PawelPacana> *in
2008-08-06T19:40:26 <xorAxAx> imagine a merge operation that can be used to commit the diff3 result generated by the upper layer
2008-08-06T19:40:41 <xorAxAx> or maybe just a flag in the commit parameters
2008-08-06T19:40:43 <TheSheep> xorAxAx: that's how it will work
2008-08-06T19:40:48 <xorAxAx> cool
2008-08-06T19:40:48 <TheSheep> xorAxAx: real soon now
2008-08-06T19:41:03 <xorAxAx> i wonder why one would implement it on that layer first but well :-)
2008-08-06T19:41:28 <PawelPacana> to have nice info action first
2008-08-06T19:41:33 <xorAxAx> hehe
2008-08-06T19:41:46 <xorAxAx> its hg, you can use it on the command line ;-)
2008-08-06T19:41:50 <TheSheep> xorAxAx: so that you can have worse user interface matched to the internals, not the other way around, of course :)
2008-08-06T19:50:35 <dennda> .
2008-08-06T19:50:37 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4621:73fe30e06ec4 1.8-storage-cdenter/MoinMoin/storage/__init__.py: storage: ABC docstring improvement
2008-08-06T19:50:37 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4622:6b080357abb0 1.8-storage-cdenter/MoinMoin/storage/backends/fs17.py: Merge with my own branch, 1.8-storage-cdenter
2008-08-06T19:50:44 <dennda> Ok, fast again
2008-08-06T19:50:52 <dennda> . being the moment where I pressed enter
2008-08-06T19:51:07 <xorAxAx> and i thought you wanted to repeat your last stroke :)
2008-08-06T19:51:13 <xorAxAx> or movement :)
2008-08-06T19:51:25 <dennda> Nope, time measurement for poor men
2008-08-06T19:51:44 * xorAxAx sometimes uses irc to do time measurement for his jogging runs
2008-08-06T19:52:19 <dennda> does your client display miliseconds?
2008-08-06T19:53:40 <TheSheep> dennda: you run so fast?
2008-08-06T19:55:55 <xorAxAx> no, milli seconds are completly useless on 10 km :)
2008-08-06T19:56:34 <xorAxAx> esp. considering that i have to manually subtract the time spent talking to my neighbours or tieing the knots of my shoes
2008-08-06T19:57:09 <xorAxAx> byeongweon: still awake? i wonder when the aforementioned hours will have passed :)
2008-08-06T20:39:52 <blackbird> hi guys
2008-08-06T20:40:01 <blackbird> sorry, wasn't able to come online today :(
2008-08-06T20:52:34 <xorAxAx> hi blackbird
2008-08-06T20:52:41 <xorAxAx> current gps position? :)
2008-08-06T20:54:27 <xorAxAx> or country+city
2008-08-06T20:55:29 <blackbird> xorAxAx: no gps, sorry :)
2008-08-06T20:56:28 <blackbird> sitting in sterling right now
2008-08-06T20:58:18 <ThomasWaldmann> dennda: you didn't document that _list_revisions returns a sorted list
2008-08-06T20:58:30 <dennda> Oh, sorry
2008-08-06T20:58:33 <ThomasWaldmann> is that because it doesn't need or because you forgot?
2008-08-06T20:59:08 <ThomasWaldmann> if you forgot, also document sort order.
2008-08-06T20:59:27 <xorAxAx> blackbird: http://deutsche-bahn.ebay.de/
2008-08-06T20:59:49 <byeongweon> ThomasWaldmann: thanks for notify me. I will add comment to that code.
2008-08-06T21:00:00 <blackbird> xorAxAx: what's that?
2008-08-06T21:00:05 <byeongweon> xorAxAx: sorry. I just came back.
2008-08-06T21:00:30 <xorAxAx> blackbird: DB selling cheap europe tickets
2008-08-06T21:00:41 <blackbird> xorAxAx: well, that doesn't beat interrail :)
2008-08-06T21:00:56 <xorAxAx> yes, except if you have only one target
2008-08-06T21:01:41 <xorAxAx> byeongweon: thats fine, as long as we dont disturb your sleeping patterns :)
2008-08-06T21:01:43 <blackbird> yep
2008-08-06T21:02:06 <byeongweon> xorAxAx: in fact, I didn't sleep yet.
2008-08-06T21:02:31 <xorAxAx> byeongweon: i assumed so, i think getting up at 4 o'clock wouldnt really work for me either :)
2008-08-06T21:02:39 <byeongweon> a friend of mine, want to talk about his girl friend problem
2008-08-06T21:02:42 <dennda> ThomasWaldmann: Only 99% sure how I used it in other places. I will check that and adjust the docstring accordingly. You are on the safe side however, if you sort it.
2008-08-06T21:03:04 <byeongweon> I talk with him about few hours.
2008-08-06T21:04:00 <xorAxAx> byeongweon: oh
2008-08-06T21:04:38 <ThomasWaldmann> dennda: did you notice anything special in fs17?
2008-08-06T21:04:50 <xorAxAx> byeongweon: did they break up already or is it just a crisis?
2008-08-06T21:05:03 <dennda> ThomasWaldmann: Sorry I got distracted until now, I will look into it in a minute
2008-08-06T21:05:19 <byeongweon> I'm not a good counsellor but I try to hear his worry.
2008-08-06T21:05:50 <byeongweon> It seems like they are in trouble.
2008-08-06T21:06:52 <byeongweon> actually, I don't know about girl well. I don't know why my friend ask me such a problem :)
2008-08-06T21:06:54 <xorAxAx> byeongweon: had to do something similar today, even though on a smaller scale. it was just a fellow student talking about his relationship to coworkers of the working group and how they didnt pay their respect for his work
2008-08-06T21:07:14 <xorAxAx> hehe
2008-08-06T21:08:00 <byeongweon> hehe.
2008-08-06T21:08:11 <byeongweon> anyway. what I want to ask is
2008-08-06T21:08:21 <byeongweon> MoinMoinBugs/GuiEditorMixedListsCrash this bug.
2008-08-06T21:08:40 <byeongweon> to tell the truth. I try to fix this bug few hours.
2008-08-06T21:09:04 <byeongweon> what I found is two
2008-08-06T21:09:20 <byeongweon> two way of problem solve.
2008-08-06T21:09:45 <byeongweon> firs, parser problem.
2008-08-06T21:10:25 <byeongweon> current moin->html parser didn't make dl element such a case
2008-08-06T21:10:38 <byeongweon> it makes only dt and dd element.
2008-08-06T21:11:22 <xorAxAx> huh
2008-08-06T21:11:30 <xorAxAx> a bug :)
2008-08-06T21:11:30 <byeongweon> bug html converter didn't accept such a way of dt using.
2008-08-06T21:11:43 <xorAxAx> assume that this was fixed in the parser
2008-08-06T21:11:47 <byeongweon> dt and dd element must encapsulation with dl
2008-08-06T21:11:47 <xorAxAx> would it work then?
2008-08-06T21:12:09 <byeongweon> I also it should fixed with parser.
2008-08-06T21:12:49 <byeongweon> second way of problem solve is fixing conveter make accept dt, dd element without dl.
2008-08-06T21:13:03 <byeongweon> but I think first is better.
2008-08-06T21:13:08 <xorAxAx> hmm
2008-08-06T21:13:22 <xorAxAx> this is one of these bugs that are caused by a severely broken parser
2008-08-06T21:13:44 <xorAxAx> and because waldi is writing a new one, the bug could be solved with it
2008-08-06T21:13:55 <byeongweon> do you think this bug caused by parser?
2008-08-06T21:14:03 <xorAxAx> well, if it doesnt generate a dl, it is
2008-08-06T21:14:10 <xorAxAx> garbage in, garbage out :)
2008-08-06T21:14:11 <byeongweon> oh. really?
2008-08-06T21:14:29 <xorAxAx> well, how does the html version of the wiki page look like?
2008-08-06T21:14:33 <xorAxAx> does it have the dl?
2008-08-06T21:14:40 <xorAxAx> or is it just missing in the gui formatter output
2008-08-06T21:14:44 <byeongweon> that also same code
2008-08-06T21:14:48 <byeongweon> without dl
2008-08-06T21:15:38 <xorAxAx> well, then its the parser and i think you should continue with another bug :)
2008-08-06T21:15:45 <xorAxAx> of course the converter could be made more robust
2008-08-06T21:15:50 <byeongweon> hmm, if waldi's parser could solve this problem. I should try another bug.
2008-08-06T21:15:51 <byeongweon> hehe
2008-08-06T21:15:53 <byeongweon> right.
2008-08-06T21:15:54 <xorAxAx> but i guess thats a more central issue
2008-08-06T21:16:58 <byeongweon> thanks xorAxAx, I try to fix parser few ways.
2008-08-06T21:17:07 <xorAxAx> hmm?
2008-08-06T21:17:42 <waldi> byeongweon: you could convert the problems into tests for the new parser ...
2008-08-06T21:17:50 <byeongweon> but it has lots of dependent.
2008-08-06T21:17:57 <xorAxAx> waldi: good idea!
2008-08-06T21:18:08 <xorAxAx> byeongweon: no, you shouldnt touch the parser, take another bug report :)
2008-08-06T21:18:15 <xorAxAx> byeongweon: and document in the old one that the parser is evil
2008-08-06T21:19:50 <byeongweon> waldi: ok I wil make testcase for your parser :)
2008-08-06T21:26:08 * byeongweon feel tired. he will going to bed
2008-08-06T21:27:08 * byeongweon say gn.
2008-08-06T21:29:09 <xorAxAx> gn byeongweon
2008-08-06T21:29:19 <xorAxAx> how unexpected at 4:29 am
2008-08-06T22:50:52 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4623:914ce552a9e2 1.8-storage-cdenter/MoinMoin/storage/backends/fs17.py: storage: Trivial Fix: Typo in exception name
2008-08-06T23:16:31 <dennda> ThomasWaldmann: Why FsRevision._fs_metadata?
2008-08-06T23:20:01 <dennda> huh, I am a bit confused
2008-08-06T23:38:49 <dennda> ThomasWaldmann: I know little about 1.7 storage, but seems fine to me
2008-08-06T23:39:03 <dennda> At least there was nothing terribly wrong I was able to find
2008-08-06T23:47:36 <ThomasWaldmann> re
2008-08-06T23:48:23 <ThomasWaldmann> dennda: that impl will get rather simple, because it is only used once per user and only for reading MoinMoin: MoinMoinChat/Logs/moin-dev/2008-08-06 (last edited 2008-08-05 22:15:01 by IrcLogImporter)
