2008-08-17T00:01:18 <dennda> hm my browser seems to fail that test
2008-08-17T00:23:35 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4829:0bcdeed16618 1.8-storage-cdenter/MoinMoin/storage/__init__.py: storage: docstrings: we want unicode for itemnames rather than strings...
2008-08-17T01:40:44 <dreimark> re
2008-08-17T01:57:06 <ThomasWaldmann> hi dreimark
2008-08-17T01:57:36 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4830:41d77f12f7bf 1.8-storage-cdenter/MoinMoin/storage/backends/memory.py: storage: memory backend bugfix in create_revision. start with 0.
2008-08-17T02:08:32 <dreimark> moin ThomasWaldmann
2008-08-17T02:17:00 <dreimark> someone enjoys techno music and I can watch the beat, look like sleepless night
2008-08-17T02:17:15 * dreimark tries to sleep again
2008-08-17T02:30:25 <dennda> ask him to turn the volume down. it's 2:30
2008-08-17T10:23:47 <xorAxAx> lol, dilbert
2008-08-17T10:25:13 <dreimark> moin
2008-08-17T12:04:31 <ThomasWaldmann> moin
2008-08-17T12:29:38 <dreimark> bbl lunch
2008-08-17T15:08:27 <dreimark> re
2008-08-17T15:43:44 <dennda> Aha
2008-08-17T15:44:46 <dennda> So the idea of list_revision() returning a list of revnos was, that the API should not deny any "real delete" operations on revisions from the beginning and allow such implementation when that need arises
2008-08-17T15:54:11 <dennda> lol? hg buggs
2008-08-17T15:54:27 <dennda> ah ok
2008-08-17T15:54:46 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4831:effe18fd8e62 1.8-storage-cdenter/MoinMoin/storage/__init__.py: storage: little docstring adjustment and removal of blank lines where not necessary
2008-08-17T16:16:25 <ThomasWaldmann> dennda: The list must be ordered, starting with
2008-08-17T16:16:25 <ThomasWaldmann> 200 the first revision-number.
2008-08-17T16:16:34 <dennda> yes
2008-08-17T16:17:05 <dennda> what if you decide later on, after deletion-functionality has been added, to delete the first revision of a page?
2008-08-17T16:17:12 <ThomasWaldmann> "first" is not telling much. Maybe "oldest"?
2008-08-17T16:17:17 <dennda> ah
2008-08-17T16:17:18 <dennda> ok
2008-08-17T16:17:48 <dennda> boy I need a lawyer or something next to me to make all those docstrings unambiguous
2008-08-17T16:17:59 <ThomasWaldmann> :P
2008-08-17T16:20:06 <ThomasWaldmann> and there is still nothing telling about subsequent revnos.
2008-08-17T16:20:24 <ThomasWaldmann> is that because they don't need to or because you forgot?
2008-08-17T16:20:30 <dennda> the former
2008-08-17T16:20:43 <dennda> if you delete revisions at one point you may well have [0, 1, 5, 6]
2008-08-17T16:21:04 <dennda> but the revisions must be created in subsequent order, which is why I mention that in create_revision()
2008-08-17T16:21:08 <ThomasWaldmann> ok, then please document that there/
2008-08-17T16:21:20 <dennda> i did
2008-08-17T16:21:32 <ThomasWaldmann> line number?
2008-08-17T16:21:40 <dennda> 212
2008-08-17T16:21:59 <ThomasWaldmann> i am talking about what list revisions returns
2008-08-17T16:22:32 <dennda> ah, you want me to document that the revnos must not necessarily be subsequent in list_revisions?
2008-08-17T16:22:47 <ThomasWaldmann> yes. and the reason for that.
2008-08-17T16:22:50 <dennda> ok
2008-08-17T16:25:37 <ThomasWaldmann> (because doing this is the only justification of using a list there)
2008-08-17T16:25:39 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4832:dfaf941ee187 1.8-storage-cdenter/MoinMoin/storage/__init__.py: storage: elaborating more on subsequency of the list-items returned by list_revisions
2008-08-17T16:25:43 <dennda> yes
2008-08-17T16:25:59 <PawelPacana> moin
2008-08-17T16:26:19 <ThomasWaldmann> hi pawel
2008-08-17T16:26:29 <dennda> but it doesn't break anything like this. for now it's just a bit nasty, but if you want to allow deletion some day and only have last_revno, you'd need to add that to the api again and alter all callers...
2008-08-17T16:26:44 <dennda> hi PawelPacana
2008-08-17T16:30:31 <dennda> gnarf
2008-08-17T16:30:54 <dennda> PawelPacana: you didn't change the ABC when you changed the API back then, did you?
2008-08-17T16:32:33 <PawelPacana> ABC? if that stands for abstract docstrings than no, i did not
2008-08-17T16:32:41 <dennda> abstract base class
2008-08-17T16:32:56 * dennda updates the docstring and abc to reflect that change
2008-08-17T16:33:00 <PawelPacana> only thing changed there:
2008-08-17T16:33:17 <PawelPacana> item to rev param on commit/rollback
2008-08-17T16:33:29 <PawelPacana> and i moved some type checks there
2008-08-17T16:33:56 <ThomasWaldmann> dennda: change/publish item metadata - one can guess it from the method name, but maybe also the docstring should talk about for what it is exactly and when it is called.
2008-08-17T16:34:47 <PawelPacana> ThomasWaldmann: do i have to conform to epydoc-style docstrings too?
2008-08-17T16:35:05 <dennda> ThomasWaldmann: ok. btw I was planning to elaborate the docstrings in the base class and refer to those in the corresponding item / revision methods docstrings
2008-08-17T16:35:17 <dennda> rather than duplicating the content and making it unmaintainable
2008-08-17T16:36:17 <ThomasWaldmann> dennda: also those two methods should refer to each other
2008-08-17T16:36:34 <ThomasWaldmann> PawelPacana: yes
2008-08-17T16:38:19 <ThomasWaldmann> PawelPacana: as long as you are just implementing the same interface as the ABC, you don't have to copy all docstrings from there
2008-08-17T16:38:38 <dennda> (but use what syntax to refer to the ABC methods docstrings?)
2008-08-17T16:39:52 <ThomasWaldmann> there could be one class level docstring telling to look it up in MoinMoin.storage.Backend ABC
2008-08-17T16:42:03 <dennda> PawelPacana: does this reflect the change you made properly? http://paste.pocoo.org/show/82468/
2008-08-17T16:42:24 <dennda> ok, I just was under the impression of having seen such special epydoc syntax once
2008-08-17T16:43:03 <ThomasWaldmann> dennda: _get_revision_timestamp is not telling what timestamp that is.
2008-08-17T16:43:34 <ThomasWaldmann> and as we are not supporting py 2.2 anymore, you can just tell the return value is int
2008-08-17T16:44:35 <ThomasWaldmann> but you should also tell that the meaning of that int is a UNIX timestamp (or seconds since the epoch)
2008-08-17T16:49:53 <ThomasWaldmann> dennda: _seek_revision_data - i guess that docstring is missing some apostrophes
2008-08-17T16:51:43 <ThomasWaldmann> dennda: btw, at some places you are talking about persistent storage. Although that is the usual usecase, your memory backend is not persistent, so maybe the base class shouldn't talk about persistency.
2008-08-17T16:54:10 <ThomasWaldmann> dennda: Item is missing docstrings. line 415 "access".
2008-08-17T16:54:26 <ThomasWaldmann> please proofread Item docstrings again, there are quite some problems.
2008-08-17T16:55:44 <PawelPacana> dennda: it's ok
2008-08-17T16:57:57 <ThomasWaldmann> dennda: use a list comprehension in Item.keys
2008-08-17T16:59:00 <ThomasWaldmann> dennda: you very often use foo-bar, but in English it usually is foo bar.
2008-08-17T17:00:34 <ThomasWaldmann> dennda: 500 Note: If you pass -1 as revno, this shall return the latest revision of the item.
2008-08-17T17:00:46 <ThomasWaldmann> I am not sure if "shall" is right here.
2008-08-17T17:01:13 <dennda> hochlichtalarm
2008-08-17T17:01:36 <ThomasWaldmann> hm?
2008-08-17T17:01:41 <dennda> yes, I only altered the "Backend"-class' docstrings
2008-08-17T17:02:23 <ThomasWaldmann> and I guess _list_revisions should point to Item.list_revisions and the docstrings should be moved there.
2008-08-17T17:02:30 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4833:ec16afcbd8ec 1.8-storage-cdenter/MoinMoin/storage/__init__.py: storage: adjusting docstring to reflect changed parameters in commit_item and rollback_item
2008-08-17T17:02:44 <dennda> ThomasWaldmann: that, I think, is a bad idea
2008-08-17T17:03:07 <ThomasWaldmann> that's public api :)
2008-08-17T17:03:16 <dennda> I'll rather do it the other way around and let Item.list_revisions point to the backends docstring
2008-08-17T17:03:53 <ThomasWaldmann> that sounds backwards :)
2008-08-17T17:03:57 <dennda> well, the one who implements the methods is the developer. I think it's much more handy to just copy the backend class and adjust it than looking up all docstrings in the corresponding class
2008-08-17T17:04:29 <ThomasWaldmann> the problem you have there is that you'll have 2 kinds of developers
2008-08-17T17:04:32 <dennda> what is the benefit of describing it on the method that actually doesn't do anything but call the backends method?
2008-08-17T17:04:41 <ThomasWaldmann> one wants to implement new backends
2008-08-17T17:04:52 <ThomasWaldmann> the other one wants to use storage api
2008-08-17T17:05:08 <ThomasWaldmann> counting them, i guess it is 1:100
2008-08-17T17:05:41 <dennda> well, but moving them just sounds wrong
2008-08-17T17:05:59 <dennda> in Backend._list_revision you can say that the list must be ordered
2008-08-17T17:06:46 <dennda> if you move that docstring to Item.list_revision, you are implying that the list must be ordered there, although that method does nothing but call the backend
2008-08-17T17:07:46 <ThomasWaldmann> well, do it how you like, but at least include pointers
2008-08-17T17:08:26 <xorAxAx> dennda: think about contracts
2008-08-17T17:08:38 <xorAxAx> dennda: you promise that list_revision returns an ordered list
2008-08-17T17:08:52 <xorAxAx> dennda: how that is done is completly irrelevant to the caller
2008-08-17T17:08:58 <xorAxAx> thats how interfaces work
2008-08-17T17:09:10 <xorAxAx> you only see the interface of the module, not the implementation
2008-08-17T17:09:31 <xorAxAx> if you have incorrect coupling in your system, you actually need to see beyond the interface. thats bad
2008-08-17T17:09:42 <ThomasWaldmann> dennda: there are some sentences like "By default this uses the foo method the backend specifies internally."
2008-08-17T17:10:02 <xorAxAx> ThomasWaldmann: thats fine, thats an implementation note :)
2008-08-17T17:10:10 <ThomasWaldmann> dennda: maybe "specifies" -> implements / offers
2008-08-17T17:10:21 <xorAxAx> yep
2008-08-17T17:10:34 * dennda hates writing docstrings
2008-08-17T17:10:47 <xorAxAx> dennda: well, there are good guides
2008-08-17T17:11:16 <dennda> xorAxAx: in fact, I only recently completed a lecture that covered that
2008-08-17T17:11:34 <dennda> but I already hated it back then :)
2008-08-17T17:11:54 <xorAxAx> we dont have such lectures
2008-08-17T17:12:12 <dennda> """Softwaretechnik"""
2008-08-17T17:12:30 <xorAxAx> all we were tought is that software architecture is something that is the result of the third step in the waterfall modell (which is the only one they taught us)
2008-08-17T17:12:56 <xorAxAx> also lines of code is the only way to measure the productivity of programmers
2008-08-17T17:13:19 * dennda reverts his "remove blank lines that are not necessary" changeset
2008-08-17T17:13:37 <xorAxAx> thats obviously unfair on those programmers who have to use a system that supports only very small copy&paste buffers
2008-08-17T17:13:53 <xorAxAx> dennda: hopefully using backout ;-)
2008-08-17T17:13:58 <dennda> that's just stupid
2008-08-17T17:15:06 <xorAxAx> dennda: did you read all of thomas' comments?
2008-08-17T17:15:32 <dennda> the ones he made the last hour?
2008-08-17T17:15:34 <dennda> yes
2008-08-17T17:15:48 <xorAxAx> dennda: http://java.sun.com/j2se/javadoc/writingapispecs/index.html is pretty interesting
2008-08-17T17:15:53 <xorAxAx> dennda: good :)
2008-08-17T17:16:14 <xorAxAx> of course you have to filter out the java crap and restrictions
2008-08-17T17:16:15 <dennda> I can hardly overlook them. that turned my screen yellow
2008-08-17T17:16:18 <xorAxAx> but it gives some general advice
2008-08-17T17:16:45 <ThomasWaldmann> waldi: i didn't see much checklist activity yet.
2008-08-17T17:16:50 <xorAxAx> dennda: i prefer to disagree here, actually that motivates people to rather ignore them
2008-08-17T17:16:54 <dennda> hm we were taught javadoc
2008-08-17T17:17:34 <dennda> I am not "people" :)
2008-08-17T17:18:00 <xorAxAx> dennda: i just remember your attempts to memorize certain things communicated here :)
2008-08-17T17:18:29 <xorAxAx> as a response to your questions (think of the decorators question :))
2008-08-17T17:18:31 <dennda> well yes, but that's not purposly ignoring but my brain malfunctioning
2008-08-17T17:18:42 <xorAxAx> its a matter of self-organization :)
2008-08-17T17:18:59 <dennda> yes, I should get one of those notebooks
2008-08-17T17:19:07 <dennda> the old school ones
2008-08-17T17:19:14 <xorAxAx> yes of course you didnt do it on purpose ... but the general pattern is that you defer the question and finally come back to it without any progress
2008-08-17T17:19:46 <dennda> heh ok I take that as constructive critizism
2008-08-17T17:20:13 <xorAxAx> yeah, note taking helps
2008-08-17T17:20:23 <xorAxAx> some people even write full logs of what they did
2008-08-17T17:21:20 <xorAxAx> effective learning can be learned :-)
2008-08-17T17:21:38 <dennda> i just watched a talk called "getting things done" the other day
2008-08-17T17:21:53 <xorAxAx> but its one of those things that are never taught
2008-08-17T17:22:09 <xorAxAx> except in expensive self management courses held for MBA graduates :)
2008-08-17T17:22:59 <ThomasWaldmann> waldi: the converter2 and macro2 package level docstrings should be much more verbose and telling what the package is about
2008-08-17T17:23:29 <xorAxAx> dennda: and which new things did you leanr?
2008-08-17T17:23:34 * xorAxAx hasnt seen it
2008-08-17T17:24:35 <ThomasWaldmann> byeongweon_t60: i also didn't see much checklist activity from you either
2008-08-17T17:25:18 <xorAxAx> hmm, he just vanished after my last questions
2008-08-17T17:25:31 <CIA-54> Byeongweon [tasyblue@gmail.com] default * 4005:64ecf99e41f6 1.8-guieditor-mbyeongweon/wiki/htdocs/applets/moinFCKplugins/moinimage/fck_image.js: improve usability : make fckeditor select default text box everytime change protocol box on attachment dialog
2008-08-17T17:25:32 <CIA-54> Byeongweon [tasyblue@gmail.com] default * 4006:05f8344e6085 1.8-guieditor-mbyeongweon/wiki/htdocs/applets/moinFCKplugins/ (moinimage/fck_image.js moinlink/fck_link.js): Code reformatting : indent and insert new line to moinmoin fckeditor plugin
2008-08-17T17:25:33 <CIA-54> Byeongweon [tasyblue@gmail.com] default * 4007:6d7d40363f3c 1.8-guieditor-mbyeongweon/wiki/htdocs/applets/moinFCKplugins/ (9 files in 8 dirs): Code reformatting : indent and insert new line to most of moinmoin fckeditor plugin
2008-08-17T17:25:34 <xorAxAx> byeongweon_t60: still waiting for answers to the various questions i asked :)
2008-08-17T17:27:24 <byeongweon_t60> ThomasWaldmann: I just update that. I coulnd'
2008-08-17T17:27:32 <ThomasWaldmann> zenhase: package docstring of MoinMoin.web needs updating.
2008-08-17T17:27:37 <byeongweon_t60> I couldn't connect internet before
2008-08-17T17:28:23 <byeongweon_t60> xorAxAx: insert empty line problem is fixed
2008-08-17T17:28:29 <ThomasWaldmann> zenhase: _fallback_cgi needs docstring
2008-08-17T17:28:47 <xorAxAx> byeongweon_t60: should i repaste my questions?
2008-08-17T17:29:07 <byeongweon_t60> and non-exists image problem?
2008-08-17T17:29:39 <byeongweon_t60> xorAxAx: what I remembered is you asked about image alt="" problem
2008-08-17T17:29:56 <dennda> xorAxAx: I usually watch those talks while trying to fall asleep. Unfortunately I managed to do so with that talk. I guess I need to watch it again
2008-08-17T17:30:02 * dennda is away, cooking
2008-08-17T17:30:07 <xorAxAx> yes! and i said that the only converter issue that is related to "minimal diffs" currently is that the fckeditor generates empty lines
2008-08-17T17:30:23 <xorAxAx> so after a few saves, your page will have a lot of empty lines
2008-08-17T17:31:08 <byeongweon_t60> xorAxAx: as I said to you. I fixed that problem.
2008-08-17T17:31:16 <xorAxAx> ok
2008-08-17T17:31:34 <xorAxAx> then your explanation about alt="" remains to be written so i have a chance to see the problem :)
2008-08-17T17:31:56 <byeongweon_t60> xorAxAx: I thoutgh that. but I was wrong.
2008-08-17T17:33:06 <byeongweon_t60> when user insert non exists image using fckeditor. fckeditor insert broken image file instead of wrong image
2008-08-17T17:33:34 <ThomasWaldmann> zenhase: flup_frontend module docstring needs to be much more verbose
2008-08-17T17:33:38 <xorAxAx> "wrong image"?
2008-08-17T17:33:42 <xorAxAx> you mean nonexistant image?
2008-08-17T17:33:48 <byeongweon_t60> yes
2008-08-17T17:33:58 <xorAxAx> ok, so it might be enough to disable that logic
2008-08-17T17:34:01 <xorAxAx> somehow :)
2008-08-17T17:34:05 <xorAxAx> probably its new
2008-08-17T17:34:29 <ThomasWaldmann> zenhase: profile.py module docstring needs update
2008-08-17T17:35:05 <byeongweon_t60> xorAxAx: I think I didn't expain clearly about this problem.
2008-08-17T17:35:26 <xorAxAx> hehe
2008-08-17T17:36:04 <ThomasWaldmann> zenhase: session.py:48 typo
2008-08-17T17:36:51 <byeongweon_t60> written in unfixed regression is show "broken image" graphics when reopen gui editor after user saved nonexistant image using gui editor.
2008-08-17T17:38:02 <xorAxAx> byeongweon_t60: isnt the problem rather that after saving again, the reference to the image will be gone?
2008-08-17T17:38:44 <byeongweon_t60> no. reference is not changed.
2008-08-17T17:39:15 <xorAxAx> what should it show instead of the broken image?
2008-08-17T17:39:55 <byeongweon_t60> I think this is not a bug. it's more similar to improving user experience of gui editor
2008-08-17T17:40:27 <byeongweon_t60> what we show is broken image :)
2008-08-17T17:41:00 <byeongweon_t60> "broken image" graphic
2008-08-17T17:43:16 <xorAxAx> so it should show the broken image graphic and currenty its showing what?
2008-08-17T17:43:33 <xorAxAx> hmm
2008-08-17T17:43:36 <xorAxAx> "then it will just display the attachment name)"
2008-08-17T17:43:44 <xorAxAx> how was it in the old version?
2008-08-17T17:43:58 <xorAxAx> i cant believe that broken image is useful here - the attachment name is more useful
2008-08-17T17:44:32 <byeongweon_t60> xorAxAx: do you think so? I agree with that.
2008-08-17T17:44:53 <xorAxAx> yes. can you just check what the old editor did?
2008-08-17T17:45:02 <byeongweon_t60> in fact, old version also show attachment name.
2008-08-17T17:45:20 <xorAxAx> see :)
2008-08-17T17:45:26 <byeongweon_t60> hehe
2008-08-17T17:45:44 <ThomasWaldmann> the rendering of {{attachment:foo.png}} by the moin wiki parser is currently either foo.png image (if it exists) or a grey clip, if it doesn't exist
2008-08-17T17:46:23 <ThomasWaldmann> (that's paralleling the rendering of links to existing/non-existing pages)
2008-08-17T18:45:06 <PawelPacana> bbl
2008-08-17T19:11:10 <dreimark> byeongweon_t60: xorAxAx if I save a textpage of http://www.ooowiki.de/CalcFunktionenDatum/RechnenMitDatumUndZeit?action=raw with the gui editor
2008-08-17T19:11:28 <xorAxAx> dreimark: so? thats not a regression :)
2008-08-17T19:11:47 <dreimark> the size increases from 23188 to 41280
2008-08-17T19:12:23 <dreimark> it is no regression but it looks like the text editor is more optimised
2008-08-17T19:12:47 <dreimark> it does not look wrong there is only an optimisation issue (delayed for later)
2008-08-17T19:28:22 <CIA-54> Byeongweon [tasyblue@gmail.com] default * 4019:db5023512482 1.8-guieditor-mbyeongweon/ (18 files in 6 dirs): merged main
2008-08-17T19:49:40 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4834:388f563d335f 1.8-storage-cdenter/MoinMoin/storage/__init__.py: storage: adjust Item and Revision docstrings
2008-08-17T20:03:31 <dennda> how important is it to obey the 80-chars limit per line?
2008-08-17T20:06:19 <TheSheep> dennda: not critical
2008-08-17T20:06:41 <TheSheep> dennda: actually, it's not important, all developers except me use wider terminals
2008-08-17T20:10:57 <dennda> I like `some` consistency in width, but 80 chars is a bit small sometimes
2008-08-17T20:11:52 <TheSheep> I think 137 is the next 'satndard' width
2008-08-17T20:11:58 <TheSheep> but nobody uses it either
2008-08-17T20:12:03 <dennda> not going to be that wid
2008-08-17T20:12:03 <TheSheep> standard
2008-08-17T20:12:04 <dennda> e
2008-08-17T20:15:04 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4835:2b69259107aa 1.8-storage-cdenter/MoinMoin/storage/__init__.py: storage: remove superfluos newlines and shorten some lines in length (split them)
2008-08-17T20:18:23 <dennda> thanks TheSheep
2008-08-17T20:25:16 <dennda> < ThomasWaldmann> dennda: Item is missing docstrings. line 415 "access". <-- can you please elaborate?
2008-08-17T20:30:24 <dennda> hm now I b0rked something
2008-08-17T20:32:00 <ThomasWaldmann> dennda: typo
2008-08-17T20:34:13 <dennda> ah, thx
2008-08-17T20:36:43 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4378:02463fa1a22b 1.8-dom-bblank/MoinMoin/util/ (_tests/test_tree.py tree.py): Tree generator - Make namespace callable to produce names with programmable values easier
2008-08-17T20:36:44 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4379:8e366bb6df45 1.8-dom-bblank/MoinMoin/converter2/compatibility_in.py: Compatibility formatter - The parsers don't expect None as arguments, use an empty string
2008-08-17T20:36:45 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4380:a191322f51ff 1.8-dom-bblank/MoinMoin/formatter/compatibility.py: Compatibility formatter - Use util.tree instead of util.namespaces
2008-08-17T20:36:46 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4381:a58c77897f8a 1.8-dom-bblank/MoinMoin/formatter/compatibility.py: Compatibility formatter - Use more util.tree
2008-08-17T20:36:49 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4382:009043b612c9 1.8-dom-bblank/MoinMoin/converter2/text_in.py: Text input converter - Add comments, cleanup
2008-08-17T20:36:52 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4383:f932060de7fa 1.8-dom-bblank/MoinMoin/converter2/moinwiki_in.py: Moin Wiki input converter - Cleanup
2008-08-17T20:36:55 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4384:0b2cb7978486 1.8-dom-bblank/MoinMoin/converter2/compatibility_in.py: Compatibility input converter - Comment
2008-08-17T20:39:46 <dennda> ImportError: cannot import name Backend
2008-08-17T20:39:47 <dennda> wtf?
2008-08-17T20:40:24 <TheSheep> dennda: case?
2008-08-17T20:40:38 <dennda> no, everything correct afaict
2008-08-17T20:40:42 <TheSheep> dennda: or do you have a file called Backend in your sys.path?
2008-08-17T20:40:47 <dennda> after restarting ipython it works magically
2008-08-17T20:40:49 <dennda> strange
2008-08-17T20:41:00 <dennda> I've done this dozens of times before
2008-08-17T20:41:05 <TheSheep> magic should be banned
2008-08-17T20:41:09 <dennda> yes
2008-08-17T20:44:04 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4836:32d79900fd51 1.8-storage-cdenter/MoinMoin/storage/__init__.py: storage: Item.keys(): Use LC instead of filter/lambda expression
2008-08-17T20:53:54 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4385:9bc6f1daf6b2 1.8-dom-bblank/MoinMoin/util/tree.py: Tree generator - Don't produce recursion in Namespace __repr__ method
2008-08-17T20:53:55 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4386:7982139dbfec 1.8-dom-bblank/MoinMoin/i18n/__init__.py: i18n - Return unicode representation of tree fragment
2008-08-17T20:53:56 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4387:ecae3b7b030f 1.8-dom-bblank/MoinMoin/macro2/__init__.py: Macro package - Add comment
2008-08-17T20:53:56 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4388:4526281c84b1 1.8-dom-bblank/MoinMoin/converter2/__init__.py: Converter package - Add comment
2008-08-17T20:56:53 <gizmach> moin, dreimark I fixed tests failing and code not working
2008-08-17T20:59:46 <gizmach> dreimark: ofcourse I need to fix my tests (other tests are working fine now)
2008-08-17T21:10:37 <dennda> less than 24 hours left :)
2008-08-17T21:14:40 <PawelPacana> dennda: what is the official UTC deadline?
2008-08-17T21:14:46 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4389:1adf69d286d3 1.8-dom-bblank/docs/TODO: Add TODO
2008-08-17T21:14:47 <waldi> 1900
2008-08-17T21:18:16 <PawelPacana> i guess it time for push...
2008-08-17T21:18:22 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4809:de49d6d724e0 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) 'tip' getter sugar.
2008-08-17T21:18:23 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4810:7f36e1bc0c2b 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Cache upgrade: recreation, storing revno:ctxrevno pairs. _iterate_changesets simplification and corresponding adjustments.
2008-08-17T21:18:24 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4811:b1406e04ff9e 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) PEP8 fix.
2008-08-17T21:18:25 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4812:64b64c2a62fa 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Dropped cdb. Name-mapping slower but clone-safe
2008-08-17T21:18:27 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4813:c062f346e9e6 1.8-mercurialbackend-ppacana/ (MoinMoin/action/GraphInfo.py wiki/htdocs/graph/graph.css): (graphinfo) Added missing stuff: diff, size. Adjusted formatting. Cleaned up.
2008-08-17T21:18:31 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4814:7b9352dec111 1.8-mercurialbackend-ppacana/MoinMoin/action/GraphInfo.py: (graphinfo) Bugfix: appending same stylesheet more than once.
2008-08-17T21:18:39 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4842:e154ecb7aa1a 1.8-mercurialbackend-ppacana/ (28 files in 13 dirs): Merge with 1.8-storage-cdenter.
2008-08-17T21:18:40 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4843:b2662161f9fd 1.8-mercurialbackend-ppacana/MoinMoin/action/GraphInfo.py: (graphinfo) Added some notes: wrt attachments, and exceptions handling. Docstring update.
2008-08-17T21:18:43 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4844:4a6a70b3da12 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Made extended API use _iterate_changesets.
2008-08-17T21:18:48 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4845:edc7e49c239b 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Reverting name-mapping versioning. Now doing so only when needed: pull/push repository hooks.
2008-08-17T21:18:52 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4846:2756999262f1 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) 'hg backup': wrapper for 'hg clone' to backup backend data.
2008-08-17T21:18:55 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4847:2bda267ad0af 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Preventing hg use console encoding.
2008-08-17T21:18:58 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4848:bcc69c00fc1b 1.8-mercurialbackend-ppacana/MoinMoin/action/GraphInfo.py: (graphinfo) Fixed stylesheet not being loaded at first request.
2008-08-17T21:19:01 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4849:d2cea32ec9dc 1.8-mercurialbackend-ppacana/MoinMoin/storage/_tests/test_backends_hg.py: (hg tests) Killing outdated part of test_backend_init_non_empty_datadir.
2008-08-17T21:19:08 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4856:f36d608df7dd 1.8-mercurialbackend-ppacana/MoinMoin/storage/__init__.py: Merge with 1.8-storage-cdenter.
2008-08-17T21:19:34 * dennda wonders when CIA gets KLined for spamming
2008-08-17T21:38:51 <dreimark> hi gizmach
2008-08-17T21:46:37 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 50:d3767204a325 EmeraldTree/emeraldtree/ElementTree.py: ElementTree - Bail out if attrib argument is no dict, force uri argument to be unicode
2008-08-17T21:46:46 <dreimark> gizmach: your new line for 312 group_manager = request.group_manager(request)
2008-08-17T21:46:57 <dreimark> can you show it please ?
2008-08-17T21:47:07 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4390:5ba3990b06e1 1.8-dom-bblank/MoinMoin/Page.py: Page - Provide URI object to converters
2008-08-17T21:47:08 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4391:9bf0ae606575 1.8-dom-bblank/MoinMoin/util/ (_tests/test_tree.py tree.py): Revert 4363:f777fb1e18d5: Tree generator - Derive the namespace object from unicode
2008-08-17T21:47:08 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4392:a5f54274cd89 1.8-dom-bblank/MoinMoin/util/tree.py: Tree generator - Add __str__ and __unicode__ to Namespace object
2008-08-17T21:51:27 <gizmach> dreimark: it is the same (explained in mail)
2008-08-17T21:53:41 <dreimark> gizmach: everything of wikiconfig* is in request.cfg
2008-08-17T21:54:12 <gizmach> dreimark: I need to call somehow cfg
2008-08-17T21:54:23 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4393:32fb0c076c15 1.8-dom-bblank/MoinMoin/converter2/compatibility_in.py: Compatibility input formatter - Use page url object
2008-08-17T21:54:24 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4394:f5666e2fe165 1.8-dom-bblank/MoinMoin/util/uri.py: URI - Use additional parameters als override of parsed informations
2008-08-17T21:54:35 <dreimark> gizmach: it is request.cfg
2008-08-17T21:54:56 <dreimark> look into your wikisync changeset
2008-08-17T21:55:11 <gizmach> dreimark: I saw that but that didn't work
2008-08-17T21:55:29 <dreimark> because you don't have self
2008-08-17T21:55:43 <gizmach> dreimark: not because of self
2008-08-17T21:56:02 <gizmach> because I didn't call fuction from wikiconfig_locale
2008-08-17T21:56:05 <dreimark> what does not work?
2008-08-17T21:56:32 <gizmach> it does not make an object it only passes te finction if I don't make it like in 312 line
2008-08-17T22:00:49 <dreimark> I don't understand what you want me to tell
2008-08-17T22:00:53 <dreimark> from MoinMoin.request import request_cli
2008-08-17T22:00:58 <dreimark> request = request_cli.Request()
2008-08-17T22:01:06 <dreimark> group_manager = request.cfg.group_manager
2008-08-17T22:01:11 <dreimark> help (group_manager)
2008-08-17T22:01:15 <dreimark> is for me an object
2008-08-17T22:02:19 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4395:e6a1cc6afc98 1.8-dom-bblank/docs/TODO: TODO - Update
2008-08-17T22:02:22 <gizmach> dreimark: how does the wikiconfig_local looks to you
2008-08-17T22:02:29 <dreimark> i have none
2008-08-17T22:02:44 <dreimark> I do use the predefined values from multiconfig
2008-08-17T22:02:51 <dreimark> as I told in the mail
2008-08-17T22:03:01 <gizmach> dreimark: ah maybe it's the wrong in multiconfig
2008-08-17T22:03:08 <dreimark> if you want a wiki comarable to the tests you don't need any own config
2008-08-17T22:03:12 <gizmach> because I had to change things in wikiconfig_local
2008-08-17T22:03:20 <dreimark> I have your last tip
2008-08-17T22:03:55 <dreimark> s/comarable/comparable/
2008-08-17T22:04:21 <dreimark> 4066:8bdb2fa2fbf7
2008-08-17T22:05:01 <dreimark> the tests don't read your own wikiconfig_local
2008-08-17T22:05:02 <gizmach> dreimark: multiconfig there doesn't work properly that needs to be changed
2008-08-17T22:05:19 <gizmach> dreimark: ok :)
2008-08-17T22:11:55 <dreimark> PawelPacana: wow - speed increased
2008-08-17T22:12:38 <dennda> dreimark: he said it's a little faster than fs backend now, though I don't know how he managed to do so
2008-08-17T22:14:15 <xorAxAx> maybe he doesnt keep all files open like johill :)
2008-08-17T22:14:20 <PawelPacana> dreimark: conversion can go under 2:40, on my machine (goes around 8:00 on underlay)
2008-08-17T22:14:39 <dreimark> PawelPacana: dennda SystemInfo looks broken
2008-08-17T22:14:51 <dennda> SystemInfo?
2008-08-17T22:14:54 <PawelPacana> it just a matter of disabling working copy updates after memory commits and doing so on clone end
2008-08-17T22:15:05 <dreimark> dennda: it is a wiki page
2008-08-17T22:15:12 <dennda> ah
2008-08-17T22:15:30 <PawelPacana> dennda: its not faster on all apects
2008-08-17T22:15:35 <dreimark> TypeError
2008-08-17T22:15:37 <dreimark> object of type 'generator' has no len()
2008-08-17T22:16:08 <PawelPacana> and surely no if you delete backend-cache
2008-08-17T22:16:22 <PawelPacana> until its recreated
2008-08-17T22:17:36 <dennda> dreimark: Looks fine here. What is displayed for you?
2008-08-17T22:17:50 <dennda> (At least I get the same result as with 1.8 main)
2008-08-17T22:18:44 <TheSheep> dennda: I think the problem might be that some backends (tm) might return an iterator instead of a list in some cases...
2008-08-17T22:19:07 <dennda> TheSheep: Talking of the SystemInfo "problem"?
2008-08-17T22:19:16 <TheSheep> aargh, scre quotes
2008-08-17T22:19:30 <TheSheep> please, have mercy :)
2008-08-17T22:20:09 <dreimark> http://moinmo.in/PawelPacana/MercurialBackend/BugReports?action=AttachFile&do=view&target=traceback6.html
2008-08-17T22:20:16 <dreimark> dennda: PawelPacana ^^
2008-08-17T22:20:41 <TheSheep> yup
2008-08-17T22:21:06 <TheSheep> PawelPacana: maybe you could monkey-patch a 'len' on it? :)
2008-08-17T22:21:41 <dennda> dreimark: I don't have that problem with my backend
2008-08-17T22:21:56 <TheSheep> I mean __len__
2008-08-17T22:22:31 <dreimark> dennda: ok, just wanted to inform, I didn't know at which level the bug occured
2008-08-17T22:22:47 <dennda> I don't know yet either. Thanks dreimark
2008-08-17T22:23:05 <dennda> PawelPacana: I don't have the appropriate mercurial version installed to test that with your backend. Can you reproduce that bug?
2008-08-17T22:23:20 <dreimark> what is the name of this new action info with that graph?
2008-08-17T22:23:47 <PawelPacana> dennda: just a sec
2008-08-17T22:23:52 <PawelPacana> dreimark: GraphInfo
2008-08-17T22:24:02 <PawelPacana> last one on list
2008-08-17T22:24:26 <dreimark> PawelPacana: thanks
2008-08-17T22:24:34 <dreimark> PawelPacana: should attachments work?
2008-08-17T22:24:51 <PawelPacana> i dont think so
2008-08-17T22:24:56 <dennda> why not?
2008-08-17T22:25:22 <PawelPacana> they dodont get listed in my info action
2008-08-17T22:25:36 <PawelPacana> s/dodont/dont/
2008-08-17T22:25:59 <dennda> click on "attachments" and you should see them
2008-08-17T22:26:02 <PawelPacana> so they are not visible in GrapInfo either
2008-08-17T22:26:24 <dreimark> I was asking because of another bug
2008-08-17T22:26:30 <PawelPacana> ah, that way they work
2008-08-17T22:26:45 <PawelPacana> but they screw up graph tree
2008-08-17T22:27:01 <dreimark> 'NoneType' object has no attribute '_backend'
2008-08-17T22:27:57 <PawelPacana> (so that subsequent revisions have no edge to parent drawn)
2008-08-17T22:28:07 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4837:e1ce71f0ab77 1.8-storage-cdenter/MoinMoin/storage/__init__.py: storage: use LC in Revision.keys as well
2008-08-17T22:28:09 <PawelPacana> if there is attachment beetwen them
2008-08-17T22:28:26 <dennda> dreimark: traceback please
2008-08-17T22:30:02 <dreimark> http://moinmo.in/PawelPacana/MercurialBackend/BugReports?action=AttachFile&do=view&target=traceback7.html
2008-08-17T22:31:14 <dreimark> on an other page it works
2008-08-17T22:31:42 <dennda> I cannot click "Show debugging information"
2008-08-17T22:32:41 <dennda> ah "get"
2008-08-17T22:33:08 <dreimark> or download it and call it with browser
2008-08-17T22:33:21 <dennda> yep, will look into that
2008-08-17T22:33:43 <dreimark> btw. the conversion comment is a strange 0
2008-08-17T22:34:23 <dreimark> for the underlay pages. it should be localhost or the name of the converter
2008-08-17T22:35:16 <dennda> eh
2008-08-17T22:35:30 <dennda> now I get that. the attachment is that html page and it is shown in the browser
2008-08-17T22:35:47 <dennda> is that correct?
2008-08-17T22:36:35 <PawelPacana> dreimark: all comments come from what your clone source backend gives
2008-08-17T22:37:12 <dreimark> dennda: the attachment problem happens if one tries to add an attachment to a nonexting page
2008-08-17T22:37:31 <dennda> hm that shouldn't happen
2008-08-17T22:37:34 <dennda> the nonetype error?
2008-08-17T22:37:41 <dreimark> dennda: yes
2008-08-17T22:38:04 <dreimark> PawelPacana: sorry not the comment the username/localhost place
2008-08-17T22:38:11 <dreimark> was meant
2008-08-17T22:38:36 <xorAxAx> dreimark: that sounds like a bug in the backend, please ask thomas (about the 0)
2008-08-17T22:39:26 <dreimark> ThomasWaldmann: currently the converterbackend writes a 0 instead of localhost.
2008-08-17T22:40:38 <dreimark> PawelPacana: RC needs a booster too
2008-08-17T22:41:09 <ThomasWaldmann> dreimark: for what?
2008-08-17T22:41:31 <PawelPacana> dreimark: measure the time and compare with FSBackend on same dataset
2008-08-17T22:41:35 <PawelPacana> please :)
2008-08-17T22:42:32 <dreimark> ThomasWaldmann: if I did a clone(data_backend_source, data_backend, verbose=True)
2008-08-17T22:42:53 <dreimark> then RC don't tell localhost instead it writes a 0 on that place
2008-08-17T22:43:11 <PawelPacana> (it is slow on underlay mostly because nearly all of 1983 revisions have same timestamps and get listed)
2008-08-17T22:43:22 <PawelPacana> i mean, RC
2008-08-17T22:44:13 <dreimark> that's true, but why isn't that page not cached if there is no change
2008-08-17T22:44:33 <PawelPacana> we need a benchmark for backends, though
2008-08-17T22:45:27 <xorAxAx> PawelPacana: why does that slow it down? :)
2008-08-17T22:45:29 <ThomasWaldmann> dreimark: underlay pages have no edit-log. thus, there is no metadata about editor, creation time, editor host etc.
2008-08-17T22:46:02 <ThomasWaldmann> for those cases, the fs17 backend just uses something empty, 0.0.0.0, etc.
2008-08-17T22:46:04 <PawelPacana> xorAxAx: because it has go get those 1983 revisions first before displaying
2008-08-17T22:46:12 <xorAxAx> PawelPacana: ah
2008-08-17T22:47:14 <PawelPacana> xorAxAx: and only optimization an be done to get_revision (if it is not higher level page caching)
2008-08-17T22:47:24 <PawelPacana> s/an/can
2008-08-17T22:47:27 <ThomasWaldmann> zenhase: i started some docs on MoinMoin:NewDocs
2008-08-17T22:47:50 <dreimark> ThomasWaldmann: could it use the name of the backend instead of the 0
2008-08-17T22:48:41 <ThomasWaldmann> dreimark: why should it?
2008-08-17T22:49:28 <dreimark> because you know the source of this pages then
2008-08-17T22:49:37 <ThomasWaldmann> why is that interesting?
2008-08-17T22:50:31 <dreimark> anyway wrong place
2008-08-17T22:50:42 <ThomasWaldmann> (the point is we don't have any real metadata for those pages, so IMHO the most sane thing we can do is to use some obviously fake things)
2008-08-17T22:50:57 <dreimark> but you can now move different wiki instances pages to one new backend
2008-08-17T22:51:34 <ThomasWaldmann> I am not sure that is really supported
2008-08-17T22:51:53 <ThomasWaldmann> (see history stuff)
2008-08-17T22:52:16 <xorAxAx> surely that is supported
2008-08-17T22:52:33 * xorAxAx sighs
2008-08-17T22:52:38 <dreimark> hmm
2008-08-17T22:52:42 <ThomasWaldmann> history is required to be in order
2008-08-17T22:52:52 <xorAxAx> not yet in the code (nobody extended clone for multiple backends yet) but by the api easily
2008-08-17T22:52:55 <xorAxAx> ThomasWaldmann: so?
2008-08-17T22:53:08 <xorAxAx> you just cannot call clone() multiple times
2008-08-17T22:53:21 <ThomasWaldmann> the code doesn't support it, as you told right now
2008-08-17T22:53:29 <xorAxAx> its easy to support
2008-08-17T22:53:56 <ThomasWaldmann> and still, who cares about what backend was used before
2008-08-17T22:54:06 <dennda> dreimark: How did you add an attachment to a nonexisting page? I only tried moving an attachment to a nonexisting page once, which worked. I don't know the correct url syntax. I think I fixed that bug, but I want to make sure...
2008-08-17T22:54:16 <dreimark> dennda: by the url
2008-08-17T22:54:25 <ThomasWaldmann> i also don't care whether it was stored on scsi disk or usb stick :P
2008-08-17T22:54:29 <dennda> dreimark: Sure. How must the url look like?
2008-08-17T22:54:32 <dreimark> got to attachments of an existing page and chage the page name
2008-08-17T22:54:38 <dreimark> +n
2008-08-17T22:54:56 <xorAxAx> the correct idea is of course to default at least the author field
2008-08-17T22:55:07 <xorAxAx> with "moinmaster import" or something like that
2008-08-17T22:55:34 <dreimark> right
2008-08-17T22:55:56 <xorAxAx> but who wants to import it anyway
2008-08-17T22:56:18 <dreimark> the one who knows how easy it is :)
2008-08-17T22:57:12 <ThomasWaldmann> dreimark: you want to mix in system/help pages into same backend as your own content?
2008-08-17T22:57:44 <dreimark> dennda: PawelPacana the invoking of the right parser for the attachment is not working
2008-08-17T22:57:51 <dreimark> ThomasWaldmann: no
2008-08-17T22:58:28 <dreimark> but I want to mix e.g. a data wiki and an experiment wiki
2008-08-17T22:58:45 <dreimark> and prefix them
2008-08-17T22:58:59 <ThomasWaldmann> prefix what?
2008-08-17T22:59:05 <dreimark> every pagename
2008-08-17T22:59:29 <ThomasWaldmann> how is that related to SOC?
2008-08-17T22:59:41 <dreimark> it is not realted, it is an other use case
2008-08-17T22:59:51 <dreimark> s/realted/related/
2008-08-17T23:00:00 <ThomasWaldmann> ok, then i suggest we discuss that after soc
2008-08-17T23:00:05 <dreimark> :)
2008-08-17T23:00:59 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4838:5beca6c878e0 1.8-storage-cdenter/MoinMoin/Page.py: storage: Page: rewrite isUnderlayPage and isStandardPage for now. Remove them post-SoC.
2008-08-17T23:01:04 <dennda> dreimark: ^^^
2008-08-17T23:01:07 <dennda> try again please
2008-08-17T23:01:23 <dreimark> dennda: can you post your w*_local.py
2008-08-17T23:01:53 <dreimark> or is there an example
2008-08-17T23:01:57 <dennda> http://paste.pocoo.org/show/82509/
2008-08-17T23:01:59 <ThomasWaldmann> zenhase: I hope you'll get up early tomorrow to work on the checklist and all the missing docs
2008-08-17T23:02:17 <PawelPacana> dreimark: add this prefixing to clone wishlist, along with xorAxAx 'weaving'
2008-08-17T23:03:13 * ThomasWaldmann sees some featureritis before the basic stuff is even stable
2008-08-17T23:05:32 <dennda> hehe
2008-08-17T23:05:41 <dennda> there was a specific term for that other than featuritis
2008-08-17T23:05:44 <dennda> cannot recall though
2008-08-17T23:05:56 <ThomasWaldmann> waldi: TODO.dom
2008-08-17T23:06:40 <xorAxAx> PawelPacana: :)
2008-08-17T23:07:21 <zenhase> ThomasWaldmann: actually i am working concentrated right now
2008-08-17T23:07:25 <xorAxAx> dennda: what happens if you try create_revision on a new item with a number != 0?
2008-08-17T23:07:35 <zenhase> (re btw.)
2008-08-17T23:08:22 <dennda> xorAxAx: RevisionNumberMismatchError
2008-08-17T23:08:33 <dennda> didn't I mention that in the fscking docstring? hmm...
2008-08-17T23:08:35 <xorAxAx> dennda: is that required by the api?
2008-08-17T23:08:51 <xorAxAx> if so, how would you ever implement missing/deleted revisions?
2008-08-17T23:08:58 <xorAxAx> in a sound way
2008-08-17T23:09:13 <xorAxAx> if you cannot even recreate the item state
2008-08-17T23:09:21 <ThomasWaldmann> zenhase: ah, good. just was not seeing anything from you recently.
2008-08-17T23:09:27 <dennda> xorAxAx: I don't see the problem
2008-08-17T23:09:31 <PawelPacana> deleted revisions?
2008-08-17T23:10:01 <dennda> xorAxAx: You yourself once proposed that "deleting" a revision could mean "tagging" it somehow and copying one backend to another, skipping all "tagged" revisions (renumbering them, of course)
2008-08-17T23:10:12 <xorAxAx> PawelPacana: dennda said that the api will allow deletion of revisions some day
2008-08-17T23:10:23 <xorAxAx> dennda: then list_revisions is nuts
2008-08-17T23:10:23 <ThomasWaldmann> waldi: nice TODO :)
2008-08-17T23:10:32 <ThomasWaldmann> maybe reorder it a bit
2008-08-17T23:10:52 <xorAxAx> dennda: hmm
2008-08-17T23:11:00 <xorAxAx> yes, it is.
2008-08-17T23:11:17 <xorAxAx> except! except if this flagging should happen on both layers
2008-08-17T23:11:17 <dennda> xorAxAx: Didn't we already agree on that? :)
2008-08-17T23:11:22 <xorAxAx> dennda: i cant remember
2008-08-17T23:11:31 <dennda> I, for myself, did so at least
2008-08-17T23:11:32 <xorAxAx> but that would mean mutable revisions
2008-08-17T23:11:40 <dennda> Maybe I didn't express it joyfully enough
2008-08-17T23:11:47 <dennda> why is that?
2008-08-17T23:11:53 <xorAxAx> also my flagging suggestion was about pruning items, and not revisions
2008-08-17T23:12:09 <dennda> you can do the same thing with revisions, can't you?
2008-08-17T23:12:13 <dennda> you just need to renumber them
2008-08-17T23:12:16 <xorAxAx> revisions are mutable
2008-08-17T23:12:19 <xorAxAx> immutable
2008-08-17T23:12:25 <xorAxAx> you cannot flag them
2008-08-17T23:12:39 <dennda> why not use metadata?
2008-08-17T23:12:41 <xorAxAx> ?
2008-08-17T23:12:54 <dennda> umm
2008-08-17T23:12:56 <dennda> yes
2008-08-17T23:12:58 <dennda> -_-
2008-08-17T23:13:19 <dennda> true
2008-08-17T23:13:25 <xorAxAx> well, then can you introduce a new method and mark the existing one (list_rev*) as deprecated?
2008-08-17T23:13:40 <xorAxAx> should be a trivial commit
2008-08-17T23:13:44 <dreimark> dennda: the attchment bug is fixed
2008-08-17T23:13:49 <xorAxAx> in the Backend class
2008-08-17T23:13:53 <xorAxAx> PawelPacana: what do you think?
2008-08-17T23:14:29 <dennda> xorAxAx: why not create a new "deleted" item and in each revision, store the info which other items / revisions to delete. the "purge"-script could then use that item
2008-08-17T23:14:50 * dreimark waits for RC ...
2008-08-17T23:15:01 <dennda> xorAxAx: ok, but you don't want me to change all callers now, do you?
2008-08-17T23:15:02 <xorAxAx> dennda: yes, but that is no reason to keep list_revs alive
2008-08-17T23:15:06 <xorAxAx> dennda: did i say that?
2008-08-17T23:15:11 <dennda> no, of course not
2008-08-17T23:15:16 <xorAxAx> deprecated method = you dont need to fix the callers :)
2008-08-17T23:15:19 <dennda> (answering the former one)
2008-08-17T23:15:21 <dennda> ok
2008-08-17T23:15:31 <xorAxAx> the students in the next and the year after that one will do so
2008-08-17T23:15:32 <dennda> just wanted to make sure before all hell breaks lose during evaluation
2008-08-17T23:15:37 <dennda> heh
2008-08-17T23:15:41 <dennda> shouldn't be too hard
2008-08-17T23:16:02 <ThomasWaldmann> waldi: http://hg.moinmo.in/moin/1.8-dom-bblank/rev/f5666e2fe165 i don't think that works as the comment tells
2008-08-17T23:16:12 <dennda> dreimark: ok thanks
2008-08-17T23:17:08 <PawelPacana> xorAxAx: another one list_revs, but accepting only tagged revisions?
2008-08-17T23:17:17 <xorAxAx> PawelPacana: no
2008-08-17T23:17:26 <xorAxAx> PawelPacana: we cannot tag on the backend layer because revs are immutable
2008-08-17T23:18:06 <xorAxAx> or do you think it makes sense to add that deleted revs handling?
2008-08-17T23:18:17 <xorAxAx> and make it manage some internal data structure
2008-08-17T23:18:20 <xorAxAx> i dont think so
2008-08-17T23:18:27 <xorAxAx> (and it was never planned to do so)
2008-08-17T23:18:36 <xorAxAx> rather it makes sense to handle it on the layers above
2008-08-17T23:18:57 * dennda hasn't read a comment on his brainstormed idea
2008-08-17T23:20:56 <xorAxAx> hmm?
2008-08-17T23:21:08 <dreimark> dennda: does Rc every finish?
2008-08-17T23:21:15 <dennda> dreimark: hm?
2008-08-17T23:21:28 <xorAxAx> dennda: a single deleted item?
2008-08-17T23:21:32 <dennda> yes
2008-08-17T23:21:52 <xorAxAx> dennda: so you have to traverse all revisions of that single item on every single operation of the frontend code that deals with e.g. wiki pages?
2008-08-17T23:22:06 <PawelPacana> xorAxAx: didnt follow the first discussion, what will be introduced after list_revs being deprecated?
2008-08-17T23:22:06 <xorAxAx> to know whether the revision was deleted or not
2008-08-17T23:22:21 <xorAxAx> PawelPacana: a function that returns the latest revision number
2008-08-17T23:22:25 <dreimark> dennda: oh, why is it empty (me was thinking it needs time to load)
2008-08-17T23:22:42 <xorAxAx> hmm
2008-08-17T23:22:45 <PawelPacana> xorAxAx: ah
2008-08-17T23:22:47 <xorAxAx> dennda: but wait
2008-08-17T23:23:12 <xorAxAx> doesnt it probably rather make sense to allow rev numbers starting at a laterrevision and retain the list?
2008-08-17T23:23:26 <xorAxAx> ThomasWaldmann wants to remove old revsiions in some cases
2008-08-17T23:23:37 <dreimark> dennda: my RC is always empty
2008-08-17T23:23:37 <xorAxAx> and it would be bad if the revnumbers became invalid
2008-08-17T23:24:08 <ThomasWaldmann> xorAxAx: or even "thin out" revisions
2008-08-17T23:24:32 <xorAxAx> dennda: i think we keep it like that
2008-08-17T23:25:04 <xorAxAx> dennda: please rather fix that rev number invalid doc string saying that the exception means "the backend doesnt support non-contiguous (please check dict) revnumber sequences"
2008-08-17T23:25:20 <xorAxAx> and then we can add that stuff later on without violating any spec
2008-08-17T23:25:28 <xorAxAx> i think thats the sanest way :)
2008-08-17T23:25:40 <dennda> let me summarize:
2008-08-17T23:25:59 <dreimark> dennda: and it does not show the footer
2008-08-17T23:26:29 <dennda> a) keep list_revisions b) add no new method to get the latest revision c) ..? the docstring of create_revision already mentions that
2008-08-17T23:26:32 <ThomasWaldmann> dreimark: that likely means it crashed and hides exceptions
2008-08-17T23:26:52 <dennda> dreimark: what backend? last time I tried, RC worked
2008-08-17T23:27:02 <PawelPacana> xorAxAx: you want to prune only items? (in a future)
2008-08-17T23:27:06 <dreimark> dennda: I used your config
2008-08-17T23:27:09 <xorAxAx> PawelPacana: no
2008-08-17T23:27:31 <xorAxAx> PawelPacana: pruning could be done on the upper layers by cloning with a filter that looks for a item metadata setting
2008-08-17T23:27:38 <TheSheep> deleting old history entries (aka ForgiveAndForget) could be nice for some uses of wiki
2008-08-17T23:27:39 <xorAxAx> dennda: mentions what?
2008-08-17T23:27:57 <PawelPacana> xorAxAx: exactly
2008-08-17T23:28:04 <dennda> xorAxAx: that you have to create revisions with subsequent revision numbers
2008-08-17T23:28:20 <ThomasWaldmann> waldi: _re = re.compile(rules, re.X) # s/_re/rules_re/g
2008-08-17T23:28:22 <xorAxAx> TheSheep: we are thinking about technical reasons. forgetting old stuff can always be done by the upper layers through the hiding of old revs :)
2008-08-17T23:28:28 <xorAxAx> dennda: well, thats the problem
2008-08-17T23:28:51 <dreimark> dennda: your SystemInfo page shows the same brokeness
2008-08-17T23:28:52 <xorAxAx> dennda: you need to be able to specify numbers that are monotonic but not subsequent neither starting from 0
2008-08-17T23:29:10 <dreimark> *object of type 'generator' has no len()
2008-08-17T23:29:10 <dreimark> --> -->object of type 'generator' has no len()
2008-08-17T23:29:11 <dreimark> TypeError
2008-08-17T23:29:11 <xorAxAx> dennda: except if the backend doesnt implement it in which case the exception will be raised
2008-08-17T23:29:16 <TheSheep> xorAxAx: but that would invalidate all links to a specific revision
2008-08-17T23:29:19 <PawelPacana> but in order to prune revisions (same trick with cloning) you need separate place to store those pointeris to revs (because of immutable revs)
2008-08-17T23:29:20 <xorAxAx> TheSheep: ?
2008-08-17T23:29:50 <dennda> xorAxAx: why don't we just renumber the revisions while cloning?
2008-08-17T23:29:54 <xorAxAx> TheSheep: maybe you understood me incorrectly
2008-08-17T23:30:02 <xorAxAx> dennda: because that will invalidate all revision numbers
2008-08-17T23:30:07 <xorAxAx> but revision numbers may never change
2008-08-17T23:30:09 <TheSheep> xorAxAx: suppose you have some spec or something, and so you refer to it with a link to a specific page revision, to avoid confusion
2008-08-17T23:30:13 <xorAxAx> and they have to be monotnic
2008-08-17T23:30:22 <xorAxAx> because they are often used by external systems
2008-08-17T23:30:23 <dennda> indeed, hmm
2008-08-17T23:30:42 <xorAxAx> TheSheep: yes, works perfectly fine but didnt work before this spec amendment :)
2008-08-17T23:32:38 <dennda> xorAxAx: so it boils down to just changing the docstrings explanation for why RevisionNumberMismatchError may be raised to "not implemented yet"?
2008-08-17T23:33:01 <xorAxAx> dennda: yes, and create* may get any other rev number if it is monotonic
2008-08-17T23:33:11 <xorAxAx> increasing
2008-08-17T23:33:38 <dennda> that, as well, needs to be *mentioned* in the docstring but not changed yet, correct? :)
2008-08-17T23:34:12 <dennda> dreimark: hm, works here
2008-08-17T23:34:35 <xorAxAx> dennda: "changed"?
2008-08-17T23:34:42 <dennda> dreimark: both, SystemInfo and RecentChanges
2008-08-17T23:34:50 <dennda> dreimark: do you have a traceback for that?
2008-08-17T23:35:26 <dennda> xorAxAx: "changed" == "the code does not need to be changed, reflecting the ideas we just discussed"
2008-08-17T23:35:38 <xorAxAx> dennda: no code change, only 2 docstring changes
2008-08-17T23:35:42 <dennda> ok
2008-08-17T23:35:45 <dennda> sounds doable
2008-08-17T23:35:46 <dennda> ;)
2008-08-17T23:36:01 <dennda> but I guess I will mention that in KNOWN_BROKEN or something
2008-08-17T23:36:06 <xorAxAx> yes
2008-08-17T23:40:01 <xorAxAx> TheSheep: does that make sense?
2008-08-17T23:44:07 <TheSheep> xorAxAx: yeah
2008-08-17T23:45:33 <xorAxAx> of course weaving will interfere if you have conflicting item names
2008-08-17T23:45:39 <xorAxAx> but nobody will be able to help you then :)
2008-08-17T23:50:57 <dreimark> dennda: http://moinmo.in/ChristopherDenter?action=AttachFile&do=view&target=traceback1.html
2008-08-17T23:51:02 <dreimark> RC gives no traceback
2008-08-17T23:52:25 <dennda> dreimark: ? that link leads to a traceback of the bug I just fixed a few minutes ago (the one you confirmed is fixed) :)
2008-08-17T23:52:42 <dennda> that will be a bit hard to debug then
2008-08-17T23:54:15 <dreimark> dennda: oops sorry
2008-08-17T23:57:25 <dreimark> http://moinmo.in/ChristopherDenter?action=AttachFile&do=view&target=traceback1.html
2008-08-17T23:57:33 <dennda> dreimark: yes, that is a known bug of the FSBackend I am not able to fix
2008-08-17T23:57:46 <dreimark> http://moinmo.in/ChristopherDenter?action=AttachFile&do=view&target=RecentChanges.html
2008-08-17T23:57:53 <dreimark> and the one for RC
2008-08-17T23:58:31 <dennda> hm strange, I wonder why that error doesn't occurr here
2008-08-17T23:59:02 <dennda> but yes, johill or someone familiar with all the things the FSBackend uses needs to fix that
2008-08-17T23:59:25 <dennda> By the way, isn't it a bit dangerous to render the html attachment on that page?
2008-08-17T23:59:37 <dennda> (not only very misleading, but also prone to XSS attacks imho) MoinMoin: MoinMoinChat/Logs/moin-dev/2008-08-17 (last edited 2008-08-16 22:15:01 by IrcLogImporter)
