2008-08-13T00:06:53 <CIA-54> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4766:1f7ed6920cb5 1.8-storage-cdenter/MoinMoin/action/ (AttachFile.py info.py): merge storage repo
2008-08-13T00:06:53 <CIA-54> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4765:e9dff376282f 1.8-storage-cdenter/MoinMoin/storage/ (_tests/test_backends_fs17.py backends/fs17.py): fs17 storage backend: for deleted rev N, get metadata from rev N-1
2008-08-13T00:09:12 <ThomasWaldmann> oh, cia is really fast today :P
2008-08-13T00:10:43 <dreimark> oh a new day
2008-08-13T00:13:33 <dreimark> PawelPacana: btw. it is funny that the backend uses my hg configuration username=...
2008-08-13T00:16:27 <dreimark> looks like upuntu has an old werkzeug
2008-08-13T00:17:07 <dennda> it's a debian based distro, what do you expect? :p
2008-08-13T00:17:44 <ThomasWaldmann> hehe
2008-08-13T00:18:27 <ThomasWaldmann> zenhase: i guess you wanted to include some usable werkzeug in moin, right?
2008-08-13T00:18:46 <dreimark> arrgh, ge pro is only for win und mac available.
2008-08-13T00:19:52 <ThomasWaldmann> upuntu sounds pythonic :D
2008-08-13T00:23:37 <dreimark> time to sleep
2008-08-13T00:23:39 <dreimark> gn
2008-08-13T00:23:55 <ThomasWaldmann> gn dreimark
2008-08-13T00:36:10 * dennda tries to get an impression of conflict handling
2008-08-13T04:32:50 <byeongweon> dreimark: ok I will check that.
2008-08-13T05:54:03 <CIA-54> Byeongweon [tasyblue@gmail.com] default * 4000:42526183b82b 1.8-guieditor-mbyeongweon/MoinMoin/converter/text_html_text_moin_wiki.py: fix normal list notation didn't work problem
2008-08-13T08:30:03 <dreimark> moin
2008-08-13T08:31:03 <dreimark> byeongweon_pc: do you have seen that in the {{{{{ section in the first line afterwards the blank is removed before the list bullet ?
2008-08-13T08:31:54 <dreimark> Line 27:
2008-08-13T08:32:21 <byeongweon> dreimark: I checked with syntax page. I found several things to fix.
2008-08-13T08:32:48 <dreimark> ok, then I don't ask for those ;)
2008-08-13T08:32:52 <byeongweon> I also found that thing.
2008-08-13T08:33:41 <byeongweon> actually my laptop doesn't work well. I just buy desktop about 3 days ago.
2008-08-13T08:34:13 <byeongweon> after setting has done, I will continue fix such a bugs job.
2008-08-13T08:34:37 <dreimark> that's bad. what do you install? ubuntu?
2008-08-13T08:36:17 <byeongweon_t60> dreimark: I just install windows vista for my desktop(in korea windows is not a option). and I will install hackintoshi for my laptop.
2008-08-13T09:02:28 <dreimark> waldi: there is no quickhelp shown at action=edit
2008-08-13T09:06:48 <ThomasWaldmann> moin
2008-08-13T09:14:44 <ThomasWaldmann> btw, if someone is uses converted 1.7 pages, maybe he wants to convert again with the updated fs17 converter
2008-08-13T09:15:34 <ThomasWaldmann> it should handle deleted rev's metadata and attachment acls better than before
2008-08-13T09:16:18 <xorAxAx> fs17 converter?
2008-08-13T09:16:30 <ThomasWaldmann> aka r/o backend
2008-08-13T10:19:49 <zenhase> moin
2008-08-13T10:21:09 <zenhase> ThomasWaldmann: similar to flup-server
2008-08-13T10:22:53 <zenhase> i used the last 2 days to clear my head a little bit, i was getting quite unconcentrated
2008-08-13T10:23:32 <zenhase> the rest was also quite good in terms of my wrist
2008-08-13T10:23:39 <zenhase> i will catch up now
2008-08-13T11:18:39 <waldi> dreimark: known. they need a new home
2008-08-13T11:21:35 <waldi> but I don't know what is an appropriate place for them
2008-08-13T11:22:32 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4343:3efe76bf0b82 1.8-dom-bblank/MoinMoin/converter2/ (_tests/test_moinwiki_in.py moinwiki_in.py): Moin Wiki input converter - Support subscript and superscript
2008-08-13T11:26:38 <waldi> äh, Page uses a usersupplied string as regular expression without escaping? re is not considered safe and can be used for a DoS
2008-08-13T11:27:34 <xorAxAx> waldi: search does as well
2008-08-13T11:40:13 <CIA-54> Byeongweon [tasyblue@gmail.com] default * 4001:0c9ba3c96ac8 1.8-guieditor-mbyeongweon/MoinMoin/converter/text_html_text_moin_wiki.py: fix first line of pre tag strip problem
2008-08-13T11:40:14 <CIA-54> Byeongweon [tasyblue@gmail.com] default * 4002:a258a234af5d 1.8-guieditor-mbyeongweon/MoinMoin/converter/text_html_text_moin_wiki.py: fix MoinMoinBugs/GuiEditorExtendsAnchorsToThePageName bug
2008-08-13T11:46:29 <ThomasWaldmann> zenhase: as moin does not run without werkzeug, you can't handle it like flup
2008-08-13T11:48:16 <ThomasWaldmann> zenhase: we want to keep that "untar and run" style of installation as you see in 1.7
2008-08-13T11:49:14 <zenhase> ok
2008-08-13T11:49:56 <ThomasWaldmann> zenhase: MoinMoin.support has other "buitlin" 3rd party stuff
2008-08-13T11:50:11 <xorAxAx> oh, an austrian IP
2008-08-13T11:50:21 <xorAxAx> blackbird: vacations in austria? :)
2008-08-13T11:52:30 <zenhase> hi blackbird
2008-08-13T11:53:38 <zenhase> ThomasWaldmann: but i don't feel well patching all the imports in a 3rd party dependency to ship it with moin
2008-08-13T11:54:13 <zenhase> i guess you refer to putting it in the MoinMoin-tree
2008-08-13T11:56:10 <ThomasWaldmann> hmm, instead of patching, you could think about adding MoinMoin.support to sys.path
2008-08-13T11:57:18 <CIA-54> Byeongweon [tasyblue@gmail.com] default * 4003:31a903106d19 1.8-guieditor-mbyeongweon/MoinMoin/converter/text_html_text_moin_wiki.py: additional fix for MoinMoinBugs/GuiEditorExtendsAnchorsToThePageName bug
2008-08-13T11:57:38 <zenhase> hmm that would be an idea too :)
2008-08-13T11:58:32 <zenhase> i actually never thought about adding something to sys.path that is beneath another sys.path entry, but it makes sense too
2008-08-13T11:59:29 <ThomasWaldmann> it's a bit more fragile than importing from MoinMoin.support as it depends on sys.path order
2008-08-13T12:00:15 <blackbird> hi xorAxAx
2008-08-13T12:00:33 <blackbird> no, back home
2008-08-13T12:00:45 <blackbird> just at my friends place, arrived here at 5am in the morning
2008-08-13T12:01:04 <xorAxAx> oh
2008-08-13T12:01:56 <ThomasWaldmann> zenhase: for that, you likely need to update comments/code in all scripts modifying sys.path
2008-08-13T12:02:58 <xorAxAx> no
2008-08-13T12:03:13 <ThomasWaldmann> zenhase: and then you also can move the patched flup there
2008-08-13T12:03:17 <xorAxAx> depending on the interface
2008-08-13T12:03:39 <xorAxAx> he could e.g. expose a factory function that handles import hackery and werkzeug instantiation
2008-08-13T12:04:10 <blackbird> sys.path hackery is ugly ;)
2008-08-13T12:04:13 <blackbird> use virtualenv ^^
2008-08-13T12:05:29 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4344:483a6b068211 1.8-dom-bblank/MoinMoin/converter2/moinwiki_in.py: Moin Wiki input converter - Handle subscript and superscript as entity
2008-08-13T12:05:35 <zenhase> i was already looking into virtualenv :)
2008-08-13T12:05:36 <xorAxAx> or zc.buildout :)
2008-08-13T12:06:01 <zenhase> but that will need a bundled virtualenv ;)
2008-08-13T12:06:01 <blackbird> buildout has a strange idea of egg handling
2008-08-13T12:06:13 <blackbird> zenhase: you just need a virtualenv bootstraping script
2008-08-13T12:06:30 <blackbird> that handles that for you
2008-08-13T12:06:43 <xorAxAx> well, all virtualenv approaches would need to copy data
2008-08-13T12:06:59 <xorAxAx> umm, python modules
2008-08-13T12:07:03 <blackbird> yes, at least the python executable
2008-08-13T12:07:17 <blackbird> modules are mostly symlinked
2008-08-13T12:07:21 <xorAxAx> but moin wants to be able to run after unpacking
2008-08-13T12:07:30 <blackbird> wtf. heise.de has a new design?
2008-08-13T12:07:37 <zenhase> yes
2008-08-13T12:07:41 <xorAxAx> yes, they have a news item about it
2008-08-13T12:07:45 <xorAxAx> 1500+ comments
2008-08-13T12:07:55 <zenhase> i find the new design disturbing
2008-08-13T12:08:16 <zenhase> i can't read any articles, i get to distracted by the fact everything looks different ;)
2008-08-13T12:08:55 <ThomasWaldmann> if you look into their css, you see most of them is about ad placement :X
2008-08-13T12:09:06 <blackbird> it's ugly on a eee
2008-08-13T12:09:12 <blackbird> half of the screen is the retarded sidebar
2008-08-13T12:09:24 * ThomasWaldmann currently has a big white block at the top and didn't find out yet how to remove it
2008-08-13T12:10:49 <zenhase> apropos eee ... the new 900a looks intriguing :)
2008-08-13T12:11:07 <ThomasWaldmann> except the fscked battery life
2008-08-13T12:11:37 <byeongweon_t60> xorAxAx: do you know why `` notation needed?
2008-08-13T12:12:40 <byeongweon_t60> maybe I should ask this to dreimark.
2008-08-13T12:12:51 <ThomasWaldmann> btw, meeting in 45mins
2008-08-13T12:12:58 <CIA-54> Bastian Blank <bblank@thinkmo.de> default * 4345:1b272904b556 1.8-dom-bblank/MoinMoin/converter2/ (creole_in.py moinwiki_in.py):
2008-08-13T12:12:58 <CIA-54> Creole, Moin Wiki input converter - Handle text outside of the regex
2008-08-13T12:12:58 <CIA-54> The regex was only able to find one character at a time, because this was a
2008-08-13T12:12:58 <CIA-54> "anything else"-rule.
2008-08-13T12:14:44 <zenhase> hmm
2008-08-13T12:14:48 <zenhase> 'if sys.version_info <= (2, 3):'
2008-08-13T12:15:09 <zenhase> virtualenv bootstrap-script does not like 2.3 very much it seems
2008-08-13T12:16:27 <xorAxAx> byeongweon_t60: needed when?
2008-08-13T12:16:39 <xorAxAx> byeongweon_t60: generally its shorter and nicer to read in a paragraph
2008-08-13T12:17:39 <byeongweon_t60> relate with this bug http://moinmo.in/MoinMoinBugs/GuiEditorRemovesEscapingCharForPlural
2008-08-13T12:18:20 <waldi> bah, "class" is a keyword but also a html attribute
2008-08-13T12:18:29 * waldi needs to hack the element generator ...
2008-08-13T12:18:43 <ThomasWaldmann> waldi: it is rather _ifnonempty than _iftrue, right?
2008-08-13T12:18:45 <xorAxAx> byeongweon_t60: ah, thats a trick to "escape" wikiwords
2008-08-13T12:18:52 <xorAxAx> byeongweon_t60: obviously, the link changes
2008-08-13T12:18:59 <byeongweon_t60> not only gui editor remove two ` . text editor also didn't show them.
2008-08-13T12:19:06 <xorAxAx> byeongweon_t60: ?
2008-08-13T12:19:15 <xorAxAx> byeongweon_t60: the text editor doesnt remove things :)
2008-08-13T12:19:57 <waldi> ThomasWaldmann: hmm, its true in boolean context
2008-08-13T12:20:04 <waldi> which is essentialy the same
2008-08-13T12:20:43 <ThomasWaldmann> sure, but on a semantic level, you are interested in nonempty stuff, not in True stuff.
2008-08-13T12:20:58 <waldi> yes
2008-08-13T12:21:45 <byeongweon_t60> xorAxAx: could you tell me what they escape?
2008-08-13T12:22:25 <xorAxAx> byeongweon_t60: well, the idea is that you want to write a link to the page FooBar but the inflection of the word is FooBari, so you write FooBar``i to get a link and read the correct word
2008-08-13T12:23:07 <xorAxAx> or course you dont need it in isolating languages :)
2008-08-13T12:23:35 <xorAxAx> hmm, korean isnt analytic, is it ... :)
2008-08-13T12:23:47 <byeongweon_t60> oh. I understand.
2008-08-13T12:24:23 <xorAxAx> of course its not completly correct in the sense that the suffix is linked
2008-08-13T12:24:34 <xorAxAx> but its kind of the wiki way - fast to write and easy to understand
2008-08-13T12:26:44 <byeongweon_t60> but we can use [[link|desc]] intead of that notation. isn't it?
2008-08-13T12:31:04 <waldi> <option value="application">all <strong class="highlight">application</strong> files</option>
2008-08-13T12:31:24 <waldi> what is wrong?
2008-08-13T12:31:47 <xorAxAx> byeongweon_t60: well, you could transform it, yes
2008-08-13T12:31:56 <xorAxAx> byeongweon_t60: but only if it is a link
2008-08-13T12:32:04 <xorAxAx> byeongweon_t60: it has different use cases as well
2008-08-13T12:32:18 <xorAxAx> think of Foo``Bar (equivalent to !FooBar)
2008-08-13T12:32:27 <xorAxAx> so its problably easiest just generating ``
2008-08-13T12:32:36 <xorAxAx> if there is no whitespace around it
2008-08-13T12:34:57 <dreimark> byeongweon_t60: you can use it for pre syntax inline e.g. `a = 1`
2008-08-13T12:37:05 <dennda> heya johill
2008-08-13T12:37:15 <johill> hi
2008-08-13T12:37:28 * dennda takes a fast shower
2008-08-13T12:40:27 <dreimark> hi johill
2008-08-13T12:40:30 <PawelPacana> hi johill
2008-08-13T12:40:55 * dreimark has some boot problems missing filesystem
2008-08-13T12:41:02 <dreimark> but the filesystem is there
2008-08-13T12:45:07 <TheSheep> dreimark: maybet he UUID changed for some reson?
2008-08-13T12:45:09 <TheSheep> reason
2008-08-13T12:46:57 <ThomasWaldmann> brb
2008-08-13T12:55:02 * byeongweon_t60 going somewhere about 10 min.
2008-08-13T12:57:56 * waldi turns the music a little louder
2008-08-13T12:58:13 * TheSheep kills some more kittens
2008-08-13T12:59:00 <waldi> ihh
2008-08-13T13:00:05 <waldi> good evening, its exactly 1100 UTC on our clocks. welcome to the last meeting
2008-08-13T13:00:34 <xorAxAx> last?
2008-08-13T13:00:50 <xorAxAx> we probably should another one next week as well
2008-08-13T13:00:58 <waldi> okay s/last //
2008-08-13T13:01:41 <ThomasWaldmann> waldi: please begin
2008-08-13T13:01:47 <waldi> what works
2008-08-13T13:01:47 <waldi> - uris, in a non-unicode representation for now
2008-08-13T13:01:47 <waldi> - pagelinks
2008-08-13T13:01:47 <waldi> - element generator, reduces boilerplate
2008-08-13T13:01:47 <waldi> what not work
2008-08-13T13:01:49 <waldi> - table and link parameters in moin wiki
2008-08-13T13:01:52 <waldi> - editor quick help. i think this belongs to the editor and not the converter
2008-08-13T13:01:52 <ThomasWaldmann> all other: welcome to our last meeting :)
2008-08-13T13:01:54 <waldi> - line anchors
2008-08-13T13:01:56 <xorAxAx> ThomasWaldmann: ?
2008-08-13T13:01:57 <waldi> - mark links to not existant pages
2008-08-13T13:01:59 <waldi> - per instance converters
2008-08-13T13:02:03 <xorAxAx> ThomasWaldmann: i strongly suggest to have another one next week
2008-08-13T13:02:13 * xorAxAx sighs
2008-08-13T13:02:37 <xorAxAx> waldi: hmm, but why broke the help?
2008-08-13T13:02:46 <xorAxAx> waldi: and why are non-existant pages not linked?
2008-08-13T13:03:17 <waldi> xorAxAx: the help broken during the removal of several old parsers
2008-08-13T13:03:17 <johill> waldi: I intentionally moved the quickhelp to the converter so it can be shipped with the converter plugin
2008-08-13T13:03:40 <xorAxAx> waldi: ?
2008-08-13T13:04:00 <johill> waldi: before, it was global (you could say it was configured for the editor code) but that obviously doesn't work properly when the admin installs more converter plugins
2008-08-13T13:04:04 <waldi> johill: i understand the intention. my problem with it is that there may be more then one
2008-08-13T13:04:15 <johill> more than one?
2008-08-13T13:04:20 <waldi> .. converter for the same input type
2008-08-13T13:04:28 <blackbird> ThomasWaldmann: can we move me and zenhase to the beginning?
2008-08-13T13:04:34 <johill> waldi: to different outputs?
2008-08-13T13:04:38 <blackbird> i have to leave in a few minutes i'm afraid
2008-08-13T13:04:38 <waldi> johill: yes
2008-08-13T13:04:49 <blackbird> after waldi i mean
2008-08-13T13:05:08 <johill> waldi: ok; dunno, I guess it shouldn't really matter which one you use ;)
2008-08-13T13:05:10 <waldi> johill: a * -> * converter will show up for raw data (images and anything else which is currently handled as attachment)
2008-08-13T13:05:15 <ThomasWaldmann> blackbird: you have to request such stuff before we have to interrupt all other people
2008-08-13T13:05:24 <dennda> (fine with me)
2008-08-13T13:05:43 <johill> waldi: don't you do a 'best match wins' policy?
2008-08-13T13:06:01 <ThomasWaldmann> blackbird: and you are after waldi anyway
2008-08-13T13:06:56 <waldi> johill: sure. but it needs the source and the destination. the help is only dependant on one, but it should be sufficient to set the second type fixed to the internal tree type
2008-08-13T13:07:20 <johill> mostly, yeah
2008-08-13T13:07:55 <ThomasWaldmann> i think we can easy sort out the quickhelp stuff later (after editor cleanup)
2008-08-13T13:07:57 <johill> is there any point in text/x-moin-wiki -> text/xhtml direct conversion?
2008-08-13T13:08:07 <waldi> johill: no
2008-08-13T13:08:36 <ThomasWaldmann> waldi: is the % stuff in i18n solved somehow?
2008-08-13T13:09:19 <johill> waldi: so maybe converters /either/ go from something that's not the DOM-tree, /or/ to something that isn't, but not both, in which case the quickhelp really isn't much of a problem since there won't be more than one
2008-08-13T13:09:44 <xorAxAx> johill: or they are tree -> tree probably? but probably not in his scheme
2008-08-13T13:09:57 <xorAxAx> (then they would be called transformation anyway)
2008-08-13T13:10:05 <ThomasWaldmann> waldi: blockers?
2008-08-13T13:10:06 <waldi> johill: another solution, add a "converter" input -> quickhelp
2008-08-13T13:10:36 <waldi> ThomasWaldmann: no, the %-stuff is not solved and I was not able to find a way without doing the expand before the parsing step
2008-08-13T13:10:50 <waldi> which means changing all callers
2008-08-13T13:11:04 <waldi> ThomasWaldmann: documentation as usual
2008-08-13T13:11:11 <johill> xorAxAx: I'm not convinced yet tree->tree makes much sense as a converter, since you'd want to explicitly select which transformations to do, whereas with the conversions you just want any that makes your output work
2008-08-13T13:11:24 <ThomasWaldmann> ok, so i suggest to fix the important stuff, but not begin new stuff (like per instance converters)
2008-08-13T13:11:42 <waldi> sure
2008-08-13T13:11:48 <johill> waldi: that makes little sense, and also again means that admins need to do more than just drop in a plugin
2008-08-13T13:11:49 <ThomasWaldmann> then work on the docs checklist
2008-08-13T13:12:08 <waldi> johill: no, there can be more converters in one file
2008-08-13T13:12:24 <johill> waldi: ok
2008-08-13T13:12:25 <ThomasWaldmann> everything not done has to be on the TODO, everything known broken has to be on the KNOWNBROKEN list
2008-08-13T13:12:34 <ThomasWaldmann> so we have a clear status at end of SOC
2008-08-13T13:13:18 <ThomasWaldmann> waldi: documentation is a blocker?
2008-08-13T13:13:54 <waldi> ThomasWaldmann: sort of, its always underdocumented
2008-08-13T13:14:22 <ThomasWaldmann> that's because you are mostly coding not documenting so much. not doing something is not a blocker.
2008-08-13T13:14:31 <waldi> okay, then not
2008-08-13T13:14:57 <ThomasWaldmann> ok, anything important left? zenhase and blackbird are next after waldi has finished.
2008-08-13T13:15:22 <waldi> nothing more then on the list
2008-08-13T13:15:32 <ThomasWaldmann> ah and please make a priority list for the todo
2008-08-13T13:15:40 <waldi> okay
2008-08-13T13:15:46 <blackbird> okay
2008-08-13T13:15:47 <ThomasWaldmann> ok, thanks waldi
2008-08-13T13:16:47 <ThomasWaldmann> zenhase: your stage
2008-08-13T13:18:05 <ThomasWaldmann> zenhase: ?
2008-08-13T13:18:41 <blackbird> zenhase: !
2008-08-13T13:18:48 <johill> haha
2008-08-13T13:18:54 <ThomasWaldmann> ok, zenhase seems to not be here
2008-08-13T13:18:56 <johill> we could postpone ;)
2008-08-13T13:19:08 <ThomasWaldmann> byeongweon_t60: xorAxAx: your stage
2008-08-13T13:19:15 <byeongweon_pc> ok
2008-08-13T13:19:21 <byeongweon_pc> what I do
2008-08-13T13:19:22 <byeongweon_pc> - fix several unfixed regressions
2008-08-13T13:19:22 <byeongweon_pc> - fix several unfixed old bugs
2008-08-13T13:19:22 <byeongweon_pc> - editing HelpOnMoinWikiSyntax almost work except escaping char (``) problem and colspan problem
2008-08-13T13:19:22 <byeongweon_pc> what I didn't do
2008-08-13T13:19:22 <byeongweon_pc> - image roundtrips and PI process exists in unfixed regressions
2008-08-13T13:19:24 <byeongweon_pc> - opened bugs
2008-08-13T13:19:38 <byeongweon_pc> blockers
2008-08-13T13:19:39 <byeongweon_pc> I still have lots of bugs that I didn't fixed.
2008-08-13T13:19:39 <byeongweon_pc> to solve this problem I will fix critical bugs first. but hopeful, there's not so many critical bugs remained.
2008-08-13T13:19:55 <xorAxAx> what about the others regressions, byeongweon_pc? (like GuiEditorCantNest or GuiEditorRemovesEscapingCharForPlural)?
2008-08-13T13:20:09 <byeongweon_pc> I think I will make editor useful until end of gsoc.
2008-08-13T13:20:11 <ThomasWaldmann> yes, great. please do a priority list.
2008-08-13T13:20:32 <xorAxAx> BRB
2008-08-13T13:20:39 <byeongweon_pc> xorAxAx: yes. nest escape problem
2008-08-13T13:21:07 <byeongweon_pc> to the next week :
2008-08-13T13:21:17 <byeongweon_pc> make editor useful and
2008-08-13T13:21:17 <ThomasWaldmann> byeongweon_t60: don't forget the docs checklist. in your case, it has to be handled a bit different as you were mostly upgrading and fixing bugs, but you have to process it.
2008-08-13T13:21:31 <byeongweon_pc> check ThomasWaldmann's checklsit
2008-08-13T13:22:21 <byeongweon_pc> ThomasWaldmann: I know. I already copy checklist to my page and start it.
2008-08-13T13:22:22 <xorAxAx> re
2008-08-13T13:22:26 <ThomasWaldmann> all users using gui editor (or trying to)
2008-08-13T13:22:37 <ThomasWaldmann> will be very interested in your CHANGES fragment
2008-08-13T13:23:13 <byeongweon_pc> hehe. ok. I will consider such a side.
2008-08-13T13:23:15 <ThomasWaldmann> and your project will be likely the first one that gets integrated
2008-08-13T13:23:33 <xorAxAx> yep
2008-08-13T13:23:36 <byeongweon_pc> wow.
2008-08-13T13:24:01 <byeongweon_pc> anyway, I will do these things.
2008-08-13T13:24:35 <ThomasWaldmann> byeongweon_pc: there is a note in CHANGES about gui editor brokenness and how to disable it
2008-08-13T13:24:35 <byeongweon_pc> I think it's end.
2008-08-13T13:24:57 <byeongweon_pc> ThomasWaldmann: where can I find that?
2008-08-13T13:25:01 <ThomasWaldmann> maybe look at it from your point of view and decide whether we can remove that after merging your stuff
2008-08-13T13:25:10 <ThomasWaldmann> it is at top of docs/CHANGES
2008-08-13T13:25:42 <byeongweon_pc> ThomasWaldmann: ok. I understand.
2008-08-13T13:25:48 <ThomasWaldmann> zenhase: you just will tell us when you are here
2008-08-13T13:26:11 <ThomasWaldmann> any questions for moon?
2008-08-13T13:26:48 <ThomasWaldmann> ok, then thanks moon and xorAxAx. dennda and johill is next.
2008-08-13T13:26:51 <dreimark> byeongweon_pc: do you have enough pages to check the gui editor?
2008-08-13T13:27:11 <dennda> zenhase still not around?
2008-08-13T13:27:26 <dennda> Ok ;)
2008-08-13T13:27:42 <dennda> What works: AttachFile finished, fixes, bla foo. Easier to tell what still needs to be done, i.e.:
2008-08-13T13:27:52 <dennda> What needs to be done: Gracefully handle editor conflicts. That is just missing up to now.
2008-08-13T13:27:52 <byeongweon_pc> dreimark: I'm using underconstruct page that you provide me and helponmoinsyntax page
2008-08-13T13:28:06 <dreimark> ok
2008-08-13T13:28:19 <dreimark> we talk later, let continue dennda
2008-08-13T13:28:19 <dennda> Refactor mimetype-handling a bit.
2008-08-13T13:28:29 <dennda> Squash remaining bugs and start working on the TODO list.
2008-08-13T13:28:40 <dennda> If there is much time left I will also start working on an underlay proxy/wrapper/whatever you call it, but that is optional.
2008-08-13T13:28:50 <dennda> Blockers: None
2008-08-13T13:29:00 <johill> wrt. mimetype handling, I pointed out yesterday to dennda that it's not possible to completely sanitise it without breaking the plugin API, since for #format mysomething the wiki will not know what mimetype that is
2008-08-13T13:29:19 <johill> jspwiki does something like 'text/x-wiki.%s' % format
2008-08-13T13:29:24 <ThomasWaldmann> #format has to die
2008-08-13T13:29:34 <johill> in a longer-term refactoring project, yes
2008-08-13T13:29:35 <dreimark> why
2008-08-13T13:29:43 <johill> not in a short-term storage refactoring that should keep things working
2008-08-13T13:29:49 <zenhase> re
2008-08-13T13:29:55 <ThomasWaldmann> (because it is replace by mimetype metadata)
2008-08-13T13:29:56 <zenhase> uh
2008-08-13T13:30:12 <zenhase> i thought the last weeks meeting was the last one *scratches head*
2008-08-13T13:30:30 <dreimark> zenhase: it is dennda currently talking
2008-08-13T13:31:24 <ThomasWaldmann> dennda: i would like if you first finish the docs stuff as it looks like your todo list is so long that docs won't be done otherwise
2008-08-13T13:31:33 <dreimark> johill: how do you think it can in short term be solved
2008-08-13T13:32:30 <ThomasWaldmann> (docs stuff meaning the checklist, that includes src cleanup etc.)
2008-08-13T13:32:31 <johill> dreimark: well one solution is the one we have, with using mimetype=='text/x-unidentified-wiki-format' and format to specify it
2008-08-13T13:32:46 <johill> dreimark: the other one would be jspwiki's, to use 'text/x-wiki.%s' % format or similar
2008-08-13T13:33:26 <dennda> Is it that long? But ok, I will prioritize that higher.
2008-08-13T13:33:43 <dennda> (sorry, my connection lags a lot at the moment. someone is ddosing me or something)
2008-08-13T13:33:55 <ThomasWaldmann> the checklist will take you at least a full day, if not 2 or 3
2008-08-13T13:34:43 <dreimark> johill: if it is removed long term then we should not introduce it into a different way and remove it afterwards.
2008-08-13T13:35:19 <dreimark> if long term is very long we can do this, but not if we are continuing in the current speed
2008-08-13T13:35:22 <xorAxAx> dreimark: ??
2008-08-13T13:35:32 <xorAxAx> dreimark: if you dont introduce anything new, it wont work
2008-08-13T13:35:49 <xorAxAx> sometimes it makes sense to write a bit of code to remove it afterwards
2008-08-13T13:35:50 <dreimark> I mean to different for the user
2008-08-13T13:36:37 <xorAxAx> why would it be different to the user?
2008-08-13T13:37:02 <dreimark> well if current #format parser does not work, then this needs changes or me got it wrong
2008-08-13T13:37:18 <xorAxAx> ?
2008-08-13T13:37:34 <dreimark> 13:29 < ThomasWaldmann> #format has to die
2008-08-13T13:37:39 <xorAxAx> johill made suggestions how to make format work again
2008-08-13T13:37:46 <johill> it currently works
2008-08-13T13:37:54 <johill> so that the storage code can be used without refactoring everything else
2008-08-13T13:38:10 <johill> which, despite being less "clean" in some abstract way, means it can actually _work_
2008-08-13T13:38:15 <ThomasWaldmann> if it currently really works, I suggest not to touch it
2008-08-13T13:38:23 <ThomasWaldmann> just because there is about no time left
2008-08-13T13:39:22 <dreimark> ok, but it can be added to the ToDo
2008-08-13T13:39:43 <ThomasWaldmann> (we can continue discussiong the #format vs. mimetype stuff after the meeting, btw)
2008-08-13T13:40:22 <ThomasWaldmann> (and there needs to be done a lot after soc, because it can't be used for production as it is anyway)
2008-08-13T13:41:07 <ThomasWaldmann> ok, dennda / johill, you are finished?
2008-08-13T13:41:15 <johill> nothing else from me
2008-08-13T13:41:20 <dennda> I have nothing more to say either
2008-08-13T13:41:28 <ThomasWaldmann> ok, PawelPacana and TheSheep: your stage :)
2008-08-13T13:41:36 <PawelPacana> ok
2008-08-13T13:41:45 <PawelPacana> what was done:
2008-08-13T13:41:45 <PawelPacana> - backend converter
2008-08-13T13:41:45 <PawelPacana> - graphinfo using mercurial's graph.js
2008-08-13T13:41:45 <PawelPacana> - controversial concurrent edits work (uncommited, patch with disclamer here: http://moinmo.in/PawelPacana/MercurialBackend/ApiHassles)
2008-08-13T13:41:46 <ThomasWaldmann> btw, melita seems to prefer not to be with us anymore
2008-08-13T13:41:56 <ThomasWaldmann> (just in case someone is wondering)
2008-08-13T13:42:01 <PawelPacana> what is still missing/problems:
2008-08-13T13:42:01 <PawelPacana> - hg producing tracebacks -> find source and fix
2008-08-13T13:42:01 <PawelPacana> - graphinfo missing diff action, size, actions for attachments
2008-08-13T13:42:01 <PawelPacana> - merging heads (mostly putting this into action and removing from backend)
2008-08-13T13:42:01 <PawelPacana> - history not recreated on pull (thanks johill for pointing this out), can be sorted out quickly
2008-08-13T13:42:04 <PawelPacana> (3 last problems doable today, tracebacks - unknown)
2008-08-13T13:42:42 <PawelPacana> i assume you need some time to read this link posted, can be discussed after meeting
2008-08-13T13:43:03 <PawelPacana> as it has contoversial nature
2008-08-13T13:43:10 <ThomasWaldmann> PawelPacana: do you have completely cloned a underlay for testing?
2008-08-13T13:43:29 <PawelPacana> yes i have, and then this tracebacks came
2008-08-13T13:43:47 <PawelPacana> something is wrong and i need to find out what
2008-08-13T13:44:11 <PawelPacana> i have some assumptions with extra dict
2008-08-13T13:44:12 <ThomasWaldmann> ah, ok. i think clone underlay fs17 -> hg, and then hg -> fs is good for testing
2008-08-13T13:44:40 <johill> fs17 -> fs is fast, why is ->hg slow?
2008-08-13T13:44:42 <ThomasWaldmann> (and also, of course, just using a hg backed wiki)
2008-08-13T13:44:52 <PawelPacana> johill: tracebacks
2008-08-13T13:45:07 <PawelPacana> when you use converted data on hg backend
2008-08-13T13:45:11 <johill> oh so it just doesn't work?
2008-08-13T13:45:18 <PawelPacana> they sometimes show up...
2008-08-13T13:45:30 <PawelPacana> has problems on underlay
2008-08-13T13:46:06 <PawelPacana> didnt observe such behaviour and slowness on fresh backend and creating some random stuff
2008-08-13T13:46:15 <PawelPacana> however the latter is not measurable
2008-08-13T13:46:40 <dreimark> I did some tests yesterday and have had some tracebacks
2008-08-13T13:47:07 <dreimark> some weren'T easy to reproduce
2008-08-13T13:47:28 <PawelPacana> i updated also my wiki pages and put some current design problems with came after playing with merging
2008-08-13T13:47:52 <PawelPacana> s/with/which/
2008-08-13T13:48:18 <johill> I still disagree with the automatic merging, but maybe we should modify the API in a way that allows to commit a "here a merge of revisions 4 and 5 into 6 happened"
2008-08-13T13:48:56 <johill> the fs backend could even store that info somehow and we could display a graph for that in case of edit conflicts
2008-08-13T13:49:24 <PawelPacana> johill: merge will be involved when you explicitly select heads from UI (from graphinfo as example)
2008-08-13T13:49:36 * ThomasWaldmann asks himself whether fs backend will run out of file handles even if we have removed all evil code willingly opening lots of item revs at once. A wiki with many normal users just looking around could also have lots of revs open at once.
2008-08-13T13:50:21 <PawelPacana> johill: so that i agree with you on disagreement on automated merging
2008-08-13T13:51:09 <xorAxAx> johill: have you seen the XXXs by thomas?
2008-08-13T13:51:09 <PawelPacana> johill: there is a screenshot/mockup on TodoNotes showing interface to select heads to merge
2008-08-13T13:51:21 <johill> xorAxAx: haven't really looked
2008-08-13T13:51:59 <xorAxAx> in a lot of scenarios, the rev objects keep their file alive
2008-08-13T13:52:00 <johill> PawelPacana: the thing is, I don't see how you can get into a situation with multiple heads in the first place
2008-08-13T13:52:18 <xorAxAx> johill: thats not the normal api :)
2008-08-13T13:52:22 <johill> xorAxAx: yeah, I know, I didn't think it was a problem since all things are iterators and revs should be gc'ed soon
2008-08-13T13:52:41 <xorAxAx> johill: surely it is, a lot of places keep a lot of items alive
2008-08-13T13:52:54 <TheSheep> johill: there are two ways, one requires edit locks to be disabled in the wiki, the other requires explicit edit of an old revision
2008-08-13T13:52:59 <ThomasWaldmann> xorAxAx: i recently removed some of them
2008-08-13T13:53:17 <johill> it would be simple to add a .close() to the API
2008-08-13T13:53:18 <xorAxAx> ThomasWaldmann: in the backend?
2008-08-13T13:53:22 <xorAxAx> or in the frontend?
2008-08-13T13:53:26 <PawelPacana> TheSheep: besides obvious one, after pull
2008-08-13T13:53:26 <ThomasWaldmann> but still, lots of parallel requests -> lots of open files
2008-08-13T13:53:38 <xorAxAx> yeah, it doesnt make sense to open all rev files anyway
2008-08-13T13:53:48 <johill> xorAxAx: hm?
2008-08-13T13:54:03 <ThomasWaldmann> most stuff is opened by some metadata access i think
2008-08-13T13:54:28 <johill> the thing is if you load metadata and then want data, you'd end up opening the same file twice
2008-08-13T13:54:32 <xorAxAx> johill: the file is always opened for a rev object, even if its not needed
2008-08-13T13:54:34 <ThomasWaldmann> deleted, acl, timestamp...
2008-08-13T13:54:42 <johill> I suppose it's not really a problem since syscalls are fast on linux :P
2008-08-13T13:55:00 <xorAxAx> opening a file twice is much better than not closing it
2008-08-13T13:55:20 <PawelPacana> johill: we want to neutralize edit-lock and substitute merging conflicted edits -> we create new head
2008-08-13T13:55:24 <xorAxAx> and you dont have any explicit closing semantics, so it should probably be more conservative
2008-08-13T13:56:16 <ThomasWaldmann> btw, how many files can be open on linux / win32 typically, per process and system-wide?
2008-08-13T13:56:27 <johill> default is about 1k on linux
2008-08-13T13:56:32 <xorAxAx> per process
2008-08-13T13:57:01 <PawelPacana> johill: and if we're on XXX an API, what is the status of XXX: can be moved to abstract (from FSBackend)
2008-08-13T13:57:30 <johill> PawelPacana: can be done, but I probably won't do it, it's just five lines of code or so anyway
2008-08-13T13:57:44 <johill> PawelPacana: it would be cleaner though because then the backends wouldn't have to access any item._ or rev._ variables
2008-08-13T13:57:59 <PawelPacana> ok, i need just a green light, its your stuff ;)
2008-08-13T13:58:11 <johill> if you want to do it go ahead
2008-08-13T13:58:33 <johill> just needs a slight API change to pass the rev instead of the item to commit/rollback
2008-08-13T13:59:16 <PawelPacana> ok, can we move on?
2008-08-13T13:59:28 * dreimark phonecall
2008-08-13T13:59:32 <johill> sure
2008-08-13T13:59:38 <PawelPacana> blockers:
2008-08-13T13:59:48 <PawelPacana> - one can say 'concurrent edits' are blocked by current API state, i would just say it is unapplicable right _now_ and needs discussion of future API (i would insist on a near future, though)
2008-08-13T14:00:20 <PawelPacana> and from me thats all
2008-08-13T14:01:09 <ThomasWaldmann> PawelPacana: do you think this will also affect dennda's todo of edit conflict handling?
2008-08-13T14:01:27 <johill> we'd have to first decide whether concurrent edits are appropriate at all
2008-08-13T14:01:43 <johill> I don't think any wiki has before ventured into that, as it creates two 'head' versions of a page
2008-08-13T14:01:51 <PawelPacana> ThomasWaldmann: it totally messes up what dennda does right now
2008-08-13T14:02:11 <PawelPacana> so that i propose to leave it for a while
2008-08-13T14:02:23 <PawelPacana> and discuss how to do it best
2008-08-13T14:02:39 <johill> maybe we should start from scratch
2008-08-13T14:02:46 <johill> what's the rationale for allowing this in the first place?
2008-08-13T14:02:47 <PawelPacana> and for concurrent edits:
2008-08-13T14:02:56 <johill> "we want to use hg to its full potential" doesn't cut it :)
2008-08-13T14:02:57 <PawelPacana> http://www.gossamer-threads.com/lists/trac/users/38155 some discussion pop up
2008-08-13T14:04:10 <johill> that thread seems mostly unrelated
2008-08-13T14:04:10 <PawelPacana> when conflict pops up you have to resolve it instantly
2008-08-13T14:04:13 <TheSheep> http://www.communitywiki.org/en/AlexandreDulaunoy <-- some more discussion and diagrams :)
2008-08-13T14:04:24 <ThomasWaldmann> PawelPacana: dennda: if there is no quick idea how to easily solve, I think this qualifies as "don't begin new stuff now" and should be on the after-SOC todo
2008-08-13T14:04:46 <johill> PawelPacana: the wiki could also allow you to save your edit somewhere as a diff for later processing, but that's unrelated
2008-08-13T14:05:16 <TheSheep> ThomasWaldmann: I fully agree here
2008-08-13T14:05:22 <johill> TheSheep: that's unrelated again, there the conflicts come from outside of the wiki scope
2008-08-13T14:06:36 <johill> I can't come up with a valid in-wiki use case of forking a page
2008-08-13T14:06:46 <johill> if you can, write it down somewhere please
2008-08-13T14:06:47 <TheSheep> johill: are you trying to say that because the conflicts come from the outside, there is no need to care about them?
2008-08-13T14:06:49 <xorAxAx> can we talk about zenhase now?
2008-08-13T14:06:55 <xorAxAx> and defer that topic
2008-08-13T14:06:56 <zenhase> about me?
2008-08-13T14:07:04 <johill> sure
2008-08-13T14:07:06 <xorAxAx> well, you havent had your turn
2008-08-13T14:07:07 <dreimark> yep
2008-08-13T14:07:14 <ThomasWaldmann> mitsuhiko: are you there?
2008-08-13T14:07:17 <TheSheep> yes
2008-08-13T14:07:20 <ThomasWaldmann> (he told me he is moving)
2008-08-13T14:07:29 <dreimark> zenhase: and your mentor is gone
2008-08-13T14:07:42 <zenhase> yeah, i am sorry i missed the start of the meeting :/
2008-08-13T14:07:58 <ThomasWaldmann> ok, then let's discuss without mentor (sigh)
2008-08-13T14:08:11 <ThomasWaldmann> zenhase: please update us :)
2008-08-13T14:08:26 <ThomasWaldmann> (and you will have some talk with mitsuhiko later, additionally)
2008-08-13T14:09:24 <zenhase> well, after some adjustments here and there with the flup-adapters and other stuff i finally removed request and server
2008-08-13T14:09:25 <PawelPacana> johill: i think we could post on wiki with TheSheep our needs and motivations, and you can comment them and raise objections (if any) then, ok?
2008-08-13T14:09:28 <zenhase> the packages that ist
2008-08-13T14:09:29 <zenhase> -t
2008-08-13T14:10:10 <johill> PawelPacana: sure
2008-08-13T14:10:13 <zenhase> i still don't know how to provide twisted support :o
2008-08-13T14:10:23 <johill> does anyone really care about twisted?
2008-08-13T14:10:39 <zenhase> if noone does i won't do either
2008-08-13T14:10:48 <ThomasWaldmann> zenhase: i think we already talked about twisted and to drop it if it is not easily done
2008-08-13T14:11:08 <ThomasWaldmann> if some twisted folk cares, they will provide a new adaptor
2008-08-13T14:11:24 <zenhase> ok, when that is the consensus
2008-08-13T14:11:49 <zenhase> now everything revolves around deployment with the new dependencies
2008-08-13T14:11:53 <ThomasWaldmann> (you could get into contact with that folk, just asking if someone is interested in providing some code)
2008-08-13T14:12:40 <zenhase> i already packaged flup in the contrib directory, but for core moin there is werkzeug as a dependency that has to be met anyway
2008-08-13T14:13:07 <dreimark> zenhase: there should be a version check in the code
2008-08-13T14:13:17 <zenhase> question is which way to follow: thomas suggested support-package
2008-08-13T14:13:20 <dreimark> because of the wrong versions of debian++
2008-08-13T14:13:32 <zenhase> and my mentor suggest using virtualenv bootstrap scripts
2008-08-13T14:13:41 <zenhase> dreimark: wrong versions?
2008-08-13T14:14:03 <dreimark> python-werzeug from ubuntu seems not to work
2008-08-13T14:14:08 <dreimark> +k
2008-08-13T14:14:27 <ThomasWaldmann> zenhase: looking at the time left and the length of your todo, i suggest doing a simple solution
2008-08-13T14:14:35 <zenhase> ThomasWaldmann: yes
2008-08-13T14:14:58 <ThomasWaldmann> that is, just add another sys.path entry and add some more comment above it that is understandable to non-pythonistas
2008-08-13T14:15:03 <johill> zenhase: I have a question when you're mostly done
2008-08-13T14:15:28 <zenhase> tho i find the virtualenv idea quite intriguing
2008-08-13T14:15:39 <ThomasWaldmann> zenhase: and if you have solved it that way for werkzeug, just move flup package besides it and it also will work without requiring installation
2008-08-13T14:15:44 <zenhase> it could be made into a wiki-instance-bootstrapper
2008-08-13T14:16:12 <ThomasWaldmann> zenhase: that is totally new stuff and also qualifies under "do not touch completely new stuff now"
2008-08-13T14:16:23 <blackbird> hi zenhase
2008-08-13T14:16:25 <blackbird> back
2008-08-13T14:16:28 <zenhase> ok, then move this idea to after-soc?
2008-08-13T14:16:38 <ThomasWaldmann> the virtualenv stuff for sure
2008-08-13T14:16:42 <zenhase> and go with MoinMoin.support for the time?
2008-08-13T14:16:44 <ThomasWaldmann> sys.path for sure not :)
2008-08-13T14:17:20 <zenhase> blackbird: hi there
2008-08-13T14:17:34 <ThomasWaldmann> do you have other major todos aside from the checklist?
2008-08-13T14:17:49 <zenhase> nothing i would touch in the remaining time
2008-08-13T14:18:06 <ThomasWaldmann> major bugs?
2008-08-13T14:18:18 <ThomasWaldmann> how about the slow tests?
2008-08-13T14:18:37 <zenhase> those aren't slow anymore here
2008-08-13T14:19:11 <zenhase> it was related to the logging-calls for attribute access i had there
2008-08-13T14:19:33 <ThomasWaldmann> ok, I'll retry if the are not 8x slower any more :)
2008-08-13T14:19:54 <ThomasWaldmann> (but they also should not be more than 2x slower than before)
2008-08-13T14:20:13 <johill> zenhase: how does all the context object stuff work? are you envisioning passing only some specialised contexts to some pieces of moin?
2008-08-13T14:20:14 * zenhase crosses his fingers
2008-08-13T14:20:24 <ThomasWaldmann> otherwise devs will run them less often than they to anyway
2008-08-13T14:20:46 <zenhase> johill: actually this is one of the things that did not work out the way i imagined
2008-08-13T14:21:14 <johill> hm, why?
2008-08-13T14:21:29 <johill> I mean, I have a hard time imagining this context stuff anyway
2008-08-13T14:21:40 <ThomasWaldmann> zenhase: you know that checklist will take you some full days? your task is requiring quite some docs.
2008-08-13T14:21:41 <johill> why split it up and stuff like that?
2008-08-13T14:22:17 <zenhase> johill: right now i pass something that behaves more like the traditional request object anyway
2008-08-13T14:22:24 <johill> yeah I know
2008-08-13T14:22:27 <johill> but why did you split it up?
2008-08-13T14:23:09 <zenhase> i thought i could split up the functionality in different contexts but actually you would have to switch context several times while it's passed through different parts of moin
2008-08-13T14:23:19 <zenhase> i mean switching back and forth
2008-08-13T14:24:00 <zenhase> ThomasWaldmann: yeah i know, there is still lot to document
2008-08-13T14:24:11 <johill> so that, for example, parsers won't see the request environment?
2008-08-13T14:24:42 <zenhase> yeah, that they get exactly what they need and not more
2008-08-13T14:24:46 <xorAxAx> any blockers, zenhase?
2008-08-13T14:24:48 <johill> ok
2008-08-13T14:25:11 <zenhase> the old way of passing request everywhere has lead to quite some strange dependencies
2008-08-13T14:25:59 <johill> yeah I know
2008-08-13T14:26:00 <xorAxAx> if not, we can close the weekly meeting for now - but we should first clarify the question whether to meet again next week, i would strongly suggest so because its only 1-2 days after pencil down date
2008-08-13T14:26:02 <zenhase> xorAxAx: since most of the stuff i will do now until end of soc is documentation i would be only subject to 'writers block'
2008-08-13T14:26:24 <xorAxAx> does anybody disagree? ThomasWaldmann?
2008-08-13T14:27:12 <ThomasWaldmann> I am fine with having a meeting. as it is after pencils down, it won't have meaning for SOC, but the projects need to continue after SOC anyway, so why not :)
2008-08-13T14:27:24 <xorAxAx> sure it has a meaning
2008-08-13T14:27:27 <johill> I'd prefer an internal meeting about evaluations
2008-08-13T14:27:33 <xorAxAx> johill: that as well, yes
2008-08-13T14:27:49 <xorAxAx> we update everybody about the stuff that was done, like always
2008-08-13T14:28:06 <ThomasWaldmann> s/won't have meaning/students won't be able to change SOC outcome any more/
2008-08-13T14:28:07 <xorAxAx> also we should request a short list of things that werent done
2008-08-13T14:28:18 <xorAxAx> so that everybody can judge how successful the project of the other student was
2008-08-13T14:28:23 <ThomasWaldmann> that's part of the checklist
2008-08-13T14:28:35 <xorAxAx> yes, but not every student reads every other student's checklist
2008-08-13T14:29:39 <xorAxAx> ok, then i suggest to continue with the eval meeting after the general meeting next week
2008-08-13T14:30:19 <johill> ok
2008-08-13T14:30:25 <johill> well
2008-08-13T14:30:35 <johill> blackbird might be the one to ask ;)
2008-08-13T14:30:46 <xorAxAx> why?
2008-08-13T14:30:51 <xorAxAx> he is in austria again
2008-08-13T14:30:57 <johill> ok, haven't followed
2008-08-13T14:32:14 <xorAxAx> blackbird: dont forget the meeting next week :)
2008-08-13T14:32:30 * zenhase puts it into his calendar again too
2008-08-13T14:32:52 <zenhase> i really misinterpreted last weeks 'last meeting before soft pencils down'
2008-08-13T14:32:58 <ThomasWaldmann> ok, so we are finished for today? thanks for everyone attending.
2008-08-13T14:33:00 <zenhase> into 'last meeting'
2008-08-13T14:33:35 <ThomasWaldmann> zenhase: that was just trying to wake some awareness that the coding end is near
2008-08-13T14:34:29 <ThomasWaldmann> btw, no student is required to stop coding in his repo, not even after the deadline
2008-08-13T14:34:37 <ThomasWaldmann> it just won't matter for SOC any more
2008-08-13T14:35:08 <xorAxAx> zenhase: i pinged you as well because i knew it was ambiguous
2008-08-13T14:35:08 <zenhase> ok :)
2008-08-13T14:35:21 * ThomasWaldmann will put tags into all repos marking the official end of SOC
2008-08-13T14:35:35 <xorAxAx> zenhase: so you must have overlooked that
2008-08-13T14:35:52 <ThomasWaldmann> btw, after that tagging, I suggested we:
2008-08-13T14:36:20 <ThomasWaldmann> * merge byeongweon_pc's stuff into moin/1.8
2008-08-13T14:36:57 <ThomasWaldmann> * merge PawelPacana's and dennda's stuff into a common storage repo
2008-08-13T14:37:52 <dennda> which shouldn't be to hard since we merge often
2008-08-13T14:38:13 <ThomasWaldmann> yes :)
2008-08-13T14:39:02 <ThomasWaldmann> and a bit later, we'll discuss about more merges and more main repos :)
2008-08-13T14:39:17 <dennda> heh
2008-08-13T14:39:31 <johill> wsgi + storage will be horrible
2008-08-13T14:39:43 <dennda> don't spoil!
2008-08-13T14:39:45 <zenhase> johill: why is that?
2008-08-13T14:39:51 <zenhase> what makes you think that?
2008-08-13T14:39:53 <johill> attachfile, for example
2008-08-13T14:41:11 <johill> anyway, I haven't tried :)
2008-08-13T14:42:19 <zenhase> will be fun to find out :)
2008-08-13T14:42:24 <ThomasWaldmann> maybe we can try to kill attachfile first
2008-08-13T14:42:48 <johill> next SoC ;)
2008-08-13T14:42:57 <ThomasWaldmann> and wsgi should be merged first
2008-08-13T14:43:11 <ThomasWaldmann> like for moin 1.9 or so
2008-08-13T14:43:22 <xorAxAx> johill: storage refactoring pass 5? :)
2008-08-13T14:43:30 <zenhase> yeah, i think so too
2008-08-13T14:43:32 <johill> xorAxAx: more like "UI refactoring"
2008-08-13T14:43:46 <xorAxAx> johill: but storage isnt finished neither :-P
2008-08-13T14:43:50 <zenhase> wsgi still needs some cosmetics after SoC
2008-08-13T14:44:07 <zenhase> but in the near future it will be ready for mainline i suppose
2008-08-13T14:45:02 <johill> xorAxAx: I don't think there's much missing
2008-08-13T14:45:24 <xorAxAx> johill: e.g. rename history et al.
2008-08-13T14:45:28 <xorAxAx> correct history handling
2008-08-13T14:45:36 <xorAxAx> clock dejittering in the fs backends
2008-08-13T14:45:59 <xorAxAx> because the global timestamp is not monotonic enough in concurrent usages
2008-08-13T14:46:04 <xorAxAx> (but that has to be tested)
2008-08-13T14:47:46 * ThomasWaldmann wants a flux compensator
2008-08-13T14:48:02 <dennda> a thermodynamic one?
2008-08-13T14:48:25 <xorAxAx> johill: or how would you handle that?
2008-08-13T14:48:26 <johill> xorAxAx: why does the timestamp matter?
2008-08-13T14:48:34 <johill> because of caching?
2008-08-13T14:48:46 <xorAxAx> johill: the problem is that the append is atomic
2008-08-13T14:48:53 <xorAxAx> johill: but timestamp generation is another event
2008-08-13T14:49:01 <xorAxAx> so you can get timestamps that run backwards
2008-08-13T14:49:05 <xorAxAx> and you likely will
2008-08-13T14:49:15 <johill> ok, so the history can be very slightly off
2008-08-13T14:49:21 <xorAxAx> but how bad this is depends on the full codepath which is of course backend dependant
2008-08-13T14:49:33 <xorAxAx> johill: yes, we dont want to handle that in the upper layer
2008-08-13T14:49:48 <xorAxAx> "please be robust if time runs backwards" is a bad api spec
2008-08-13T14:49:56 <johill> heh
2008-08-13T14:50:01 <johill> true
2008-08-13T14:50:53 <xorAxAx> johill: do you think that rename history is hard to implement?
2008-08-13T14:50:59 <johill> yeah but it's trivial to do really, just pre-load a few log entries
2008-08-13T14:51:04 <johill> or run with nfs=True
2008-08-13T14:51:20 <xorAxAx> what does nfs=True do?
2008-08-13T14:51:36 <johill> (and move around timestamp generation)
2008-08-13T14:51:47 <johill> lock around _addhistory, nfs has no O_APPEND
2008-08-13T14:52:20 <johill> rename history is a matter of the API of showing it in history()
2008-08-13T14:54:23 <xorAxAx> 14:50:53 < xorAxAx> johill: do you think that rename history is hard to implement?
2008-08-13T14:54:29 <johill> and adding the api to allow passing metadata to rename()
2008-08-13T14:54:37 <johill> well, judge for yourself
2008-08-13T14:54:53 <xorAxAx> i havent written a backend yet
2008-08-13T14:55:16 <johill> oh actually storing the data should be trivial
2008-08-13T15:35:46 <xorAxAx> PawelPacana: didnt you defer a discussion? :)
2008-08-13T15:35:55 <xorAxAx> that you wanted to hold now
2008-08-13T16:26:17 <byeongweon_t60> xorAxAx: I have something to tell you.
2008-08-13T16:26:38 <xorAxAx> i am here :)
2008-08-13T16:27:08 <byeongweon_t60> I should going my home tommorow and came back this sunday.
2008-08-13T16:27:24 <byeongweon_t60> I will do my works there.
2008-08-13T16:28:23 <byeongweon_t60> but I can't connect internet every hour there.
2008-08-13T16:29:32 <byeongweon_t60> so, until sunday, I usually work my job offline, and push them once a day or 2 days.
2008-08-13T16:31:47 <xorAxAx> yeah, thats fine
2008-08-13T16:31:57 <xorAxAx> you only need to be online anyway if you have a question
2008-08-13T16:32:07 <xorAxAx> we can always use mail if necessary
2008-08-13T16:32:07 <byeongweon_t60> thank, xorAxAx.
2008-08-13T16:33:01 <byeongweon_t60> ok. in fact, if I go to there I should going internet cafe to using internet. it's about 30-40 min. by car :)
2008-08-13T16:33:38 <xorAxAx> hehe
2008-08-13T16:34:04 <byeongweon_t60> so I think once a day will be maxinum during 3 days.
2008-08-13T16:34:16 <xorAxAx> i read that korea is well equipped wrt internet connectivity. but i figure that this only is true for urban areas
2008-08-13T16:36:13 <byeongweon_t60> actually, korea has very well internet enviroment. my case is special. after leave home me and younger brother for university
2008-08-13T16:36:51 <byeongweon_t60> there's no more neccesasy using internet. so my parent decide not to use.
2008-08-13T16:37:29 * ThomasWaldmann heard that working with ssh private keys in inet cafes is very secure :P
2008-08-13T16:39:53 <byeongweon_t60> hehe. that's true. I know. but I should visit grand parent with my parent during that period. that place where my grand parent lives is more worse situation for internet.
2008-08-13T16:40:17 <byeongweon_t60> anyway, I will try to connect as possible as I can :)
2008-08-13T16:49:32 <ThomasWaldmann> byeongweon_t60: if you go to an internet cafe, you take your machine with you and just use their lan?
2008-08-13T16:49:55 <byeongweon_t60> I will do that.
2008-08-13T16:50:16 <ThomasWaldmann> ok, that's safe then :)
2008-08-13T16:50:32 <byeongweon_t60> I don't want to make mercurial enviroment to there :)
2008-08-13T16:54:53 <byeongweon_t60> xorAxAx: did you read first element of Unfixed Regressions?
2008-08-13T16:55:03 <byeongweon_t60> it's about image thing.
2008-08-13T16:56:45 <xorAxAx> yes, but i didnt write it. can you reproduce it?
2008-08-13T17:04:48 <xorAxAx> byeongweon_t60:
2008-08-13T17:05:39 <byeongweon_t60> yes. I can reproduce.
2008-08-13T17:05:46 <byeongweon_t60> but what I want to know is
2008-08-13T17:06:04 <byeongweon_t60> should I remove alt attribute from image tag?
2008-08-13T17:06:37 <xorAxAx> in which case/step?
2008-08-13T17:06:48 <byeongweon_t60> hmm..
2008-08-13T17:06:50 <xorAxAx> i think you can set it using wiki syntax, in that case it shouldnt be removed
2008-08-13T17:07:24 <byeongweon_t60> that bugs cause formatter insert alt attribute to image tag.
2008-08-13T17:07:40 <xorAxAx> huh
2008-08-13T17:07:48 <xorAxAx> you mean the alt tag causes it? why?
2008-08-13T17:08:22 <byeongweon_t60> no. what I said is that's not a bug behavior.
2008-08-13T17:09:45 <byeongweon_t60> if image must containe alt attribute, that should fix with insert alt attribute when non-existing image inserted.
2008-08-13T17:10:32 <dreimark> byeongweon_t60: if you store a page (Syntaxreference) twice or more times by the gui editor it does always do some changes on the page content. I would assume ideally that it has nothing to change at the second or further savings.
2008-08-13T17:11:54 <byeongweon_t60> dreimark: yes. when I save helponmoinsyntax page with gui editor. it's something chaged.
2008-08-13T17:14:25 <byeongweon_t60> I try to reduce that change as possible as I can. but I think some change is inevitable. to make perfect html<->moin converting need more time and effort I think.
2008-08-13T17:17:24 * byeongweon_t60 going outside to take a break for a while.
2008-08-13T17:19:50 <xorAxAx> byeongweon_t60: i dont understand the image problem
2008-08-13T17:22:50 * dennda sometimes loves the debugger
2008-08-13T17:32:50 <dreimark> bbl
2008-08-13T17:38:11 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4768:cdfc5b1bd0a5 1.8-storage-cdenter/MoinMoin/PageEditor.py: storage: PageEditor: Handle edit conflicts gracefully
2008-08-13T17:38:49 <dennda> Someone please tell me whether that message in the Exception is OK
2008-08-13T17:40:43 <dennda> err
2008-08-13T17:43:11 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4769:691a16566821 1.8-storage-cdenter/MoinMoin/PageEditor.py: storage: PageEditor: Quickfix: Remove trailing whitespace eclipse left and make _ known
2008-08-13T18:42:43 <ThomasWaldmann> uhoh
2008-08-13T18:43:23 <ThomasWaldmann> page_category_regex = ur'(?P<all>(Category(?P<key>(?!Template)\S+))|((?P<key>\S+)Kategorie))'
2008-08-13T18:43:55 <ThomasWaldmann> intent was to have CategoryFoo recognized as well as FooKategorie
2008-08-13T18:44:12 <ThomasWaldmann> but: redefinition of group name u'key' as group 5; was group 3
2008-08-13T20:48:43 <ThomasWaldmann> re
2008-08-13T20:52:31 <xorAxAx> byeongweon_t60: hmm, when saving syntaxreference i see that the list ident is not handled correctly
2008-08-13T20:52:34 <xorAxAx> this looks like a regression
2008-08-13T20:53:56 <xorAxAx> byeongweon_t60: also it generates empty lines on every edit
2008-08-13T20:54:14 <xorAxAx> byeongweon_t60: after a few saves, your page will consist of a lot of empty lines
2008-08-13T20:54:38 <xorAxAx> byeongweon_t60: also i am still waiting for an explanation of the alt="" problem
2008-08-13T21:10:01 <CIA-54> Florian Krupicka <florian.krupicka@googlemail.com> default * 4073:c689dfa55de1 1.8-wsgi-fkrupicka/MoinMoin/support/werkzeug/ (50 files in 5 dirs): Import werkzeug 0.3.1 package into MoinMoin.support (still need to insert CREDITS)
2008-08-13T21:10:01 <CIA-54> Florian Krupicka <florian.krupicka@googlemail.com> default * 4074:775e8f24c1b1 1.8-wsgi-fkrupicka/MoinMoin/web/__init__.py: Fixup sys.path to accomodate for complete libraries under MoinMoin.support
2008-08-13T21:38:31 <ThomasWaldmann> zenhase: 2x slower still (fixing this is lower prio than checklist)
2008-08-13T21:42:10 <zenhase> ThomasWaldmann: ok, but 2x is a big gain over 8x
2008-08-13T21:42:34 <zenhase> i will try to test moin-1.8 and moin-1.8-wsgi on different machines of mine
2008-08-13T21:42:51 <zenhase> my laptop isn't perhaps a good reference
2008-08-13T21:43:02 <zenhase> it may be easily the fastest machine i have cpu wise
2008-08-13T21:43:52 <zenhase> i will test it but don't put a lot of prio on it for now
2008-08-13T21:51:20 <ThomasWaldmann> yeah
2008-08-13T22:23:19 <ThomasWaldmann> zenhase: why sys.path.append?
2008-08-13T22:30:04 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4787:6e05a886697b 1.8-mercurialbackend-ppacana/MoinMoin/ (6 files in 4 dirs): Merge with 1.8-storage-cdenter.
2008-08-13T22:30:05 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4788:259b86d2101c 1.8-mercurialbackend-ppacana/tests/wikiconfig.py: (tests) Fix for running tests from dirs other than repo root.
2008-08-13T22:30:06 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4789:60448707fd78 1.8-mercurialbackend-ppacana/MoinMoin/ (action/GraphInfo.py storage/_tests/test_backends.py): PEP8 fixes.
2008-08-13T22:30:06 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4790:53e3b694c878 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/fs.py: (fs) Unreferenced variable fix.
2008-08-13T22:30:09 <CIA-54> Pawel Pacana <pawel.pacana@gmail.com> default * 4791:f21cd562fa2f 1.8-mercurialbackend-ppacana/MoinMoin/storage/ (5 files in 2 dirs): (storage) API refresh: moved repeated fragments to abstract, changed _commit_item, _rollback_item arguments.
2008-08-13T22:30:54 <PawelPacana> dennda: i advise you to merge :)
2008-08-13T22:31:11 <PawelPacana> there is some failing test which is not in your repo
2008-08-13T22:41:18 <ThomasWaldmann> zenhase: you move flup also to support/?
2008-08-13T22:41:53 <dennda> PawelPacana: ... so I am not the one to fix it? Or what does that sentence imply?
2008-08-13T22:43:07 <PawelPacana> dennda: you cant see what to fix, because this test is not in your repo
2008-08-13T22:43:20 <dennda> ahh
2008-08-13T22:43:30 <dennda> you added a failing test to point me to the flaw in the code? :)
2008-08-13T22:44:11 <PawelPacana> it is a bit old test ;P
2008-08-13T22:44:23 <PawelPacana> few days i would say ;)
2008-08-13T22:45:17 <dennda> but you want *me* to fix it? that is the part I am most interested in ;)
2008-08-13T22:46:34 <PawelPacana> it's up to you, happens to memory backend only and is quite easy
2008-08-13T22:46:44 <dennda> ok thanks, finally understood :)
2008-08-13T22:49:34 <dennda> will do that then, thanks
2008-08-13T23:05:04 <dreimark> re
2008-08-13T23:16:12 <dreimark> gn
2008-08-13T23:39:31 <dennda> lo and behold
2008-08-13T23:39:41 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4770:efe9f1c4d91a 1.8-storage-cdenter/MoinMoin/Page.py: storage: Page.py: Beginning code cleanup. Removing all code I commented out previously. Leaving all other comments and other peoples TODOs in place
2008-08-13T23:39:42 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4771:ee96c3fe85f8 1.8-storage-cdenter/MoinMoin/PageEditor.py: storage: PageEditor.py: Removing my comments. Leaving other peoples comments in place.
2008-08-13T23:39:43 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4772:b25e7d7c4ef3 1.8-storage-cdenter/MoinMoin/action/AttachFile.py: storage: AttachFile: Removing obsolete code I commented out. Leaving other peoples comments and TODOs in place.
2008-08-13T23:39:46 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4773:0fb0d8da11f1 1.8-storage-cdenter/MoinMoin/action/diff.py: storage: diff: Removing obsolete code I commented out. Leaving other peoples comments and TODOs in place.
2008-08-13T23:39:50 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4774:0d0d92fe0827 1.8-storage-cdenter/MoinMoin/user.py: storage: user: Removing obsolete code I commented out. Leaving other peoples comments and TODOs in place.
2008-08-13T23:39:54 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4775:e7217360a276 1.8-storage-cdenter/MoinMoin/wikiutil.py: storage: wikiutil: cleanup: Remove old code I commented
2008-08-13T23:40:00 <CIA-54> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4776:08046d79cbdc 1.8-storage-cdenter/MoinMoin/storage/__init__.py: storage: ABCs: cleanup: Remove old code I commented
2008-08-13T23:44:19 <xorAxAx> dennda: you didnt delete your TODOs, did you? :)
2008-08-13T23:44:27 <dennda> no :)
2008-08-13T23:45:02 <dennda> though that would be the cheapest way to get rid of them ;)
2008-08-13T23:52:23 <dennda> PawelPacana: After you changed the API, did all tests pass that used to pass? MoinMoin: MoinMoinChat/Logs/moin-dev/2008-08-13 (last edited 2008-08-12 22:15:01 by IrcLogImporter)
