MoinMoin: MoinMoinChat/Logs/moin-dev/2008-09-07

2008-09-07T00:05:53  <dennda> puh
2008-09-07T00:06:00  <dennda> blog is finally online
2008-09-07T00:06:04  <dennda> the-space-station.com @ dreimark
2008-09-07T00:06:12  <dennda> posting on moin will follow :-)
2008-09-07T00:07:26  * dreimark reads
2008-09-07T00:12:12  <dreimark> dennda: nice
2008-09-07T00:16:16  <dennda> :-)
2008-09-07T00:44:20  <dreimark> gn
2008-09-07T05:10:53  <dreimark> http://search.yahoo.com/mrss describes all meta-data for multimedia files which we should be able to provide too
2008-09-07T05:58:25  <dreimark> byeongweon_pc: please join #moin there is one with question about fckeditor improvment
2008-09-07T06:02:42  <dreimark> his nick is whyameye
2008-09-07T06:03:10  <dreimark> ThomasWaldmann: test18 is broken "sys.stdout access restricted by mod_wsgi"
2008-09-07T06:03:56  <dreimark> MoinMoin/converter/text_html_text_moin_wiki.py
2008-09-07T06:04:15  <dreimark> 1309 print attrs
2008-09-07T06:06:55  <dreimark> gn
2008-09-07T10:33:51  <dreimark> moin
2008-09-07T11:55:53  <ThomasWaldmann> hi dreimark
2008-09-07T12:01:42  <dreimark> that print bug I assume was a debugging feature by byeongweon_pc
2008-09-07T12:07:54  <ThomasWaldmann> see hg annotate
2008-09-07T12:34:45  <CIA-25> Reimar Bauer <rb.proj AT googlemail DOT com> default * 4058:92be0252883d 1.8/MoinMoin/converter/text_html_text_moin_wiki.py: text_html_text_moin_wiki: removed superfluously print statements
2008-09-07T14:16:49  <CIA-25> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3807:b1a690282f20 1.7/MoinMoin/macro/TableOfContents.py: TableOfContents: show correct TOC for editor preview
2008-09-07T14:24:32  <dreimark> ThomasWaldmann: in cache.key I have had a misunderstanding by reading If wikiname, itemname, attachname is given, we don't touch the content
2008-09-07T14:25:04  <dreimark> It is not meant that you can use conent only
2008-09-07T14:35:33  <dreimark> s/conent/content/
2008-09-07T14:40:17  <ThomasWaldmann> dreimark: i am not sure I understand the problem
2008-09-07T14:40:41  <ThomasWaldmann> hmm, looks like noone is using 1.7 with mailimport
2008-09-07T14:43:46  <dreimark> ThomasWaldmann: I was dhortly looking at it, some users wanted it, but then I recognized I have to enable xmlrpc too.
2008-09-07T14:44:17  <dreimark> and if xmlrpc is eanbled you can use all the xmlrpc methods.
2008-09-07T14:46:07  <CIA-25> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3808:65bf5767d8ad 1.7/MoinMoin/mail/mailimport.py: mailimport: fix attachment filename processing
2008-09-07T14:47:43  <dreimark> and for the mailimport I have to configure a local mail daemon.
2008-09-07T14:51:53  <ThomasWaldmann> dreimark: well, it doesn't need to be local, that's why we use xmlrpc
2008-09-07T14:52:32  <ThomasWaldmann> (local meaning on same machine as wiki server)
2008-09-07T14:52:50  <dreimark> ThomasWaldmann: yeah but when you have to use a mailrelay at a company this is bad if it has no xmlrpc support.
2008-09-07T14:53:48  <dennda> ugh need to send my diffs :-)
2008-09-07T14:54:23  <ThomasWaldmann> dreimark: yes, you need python and xmlrpclib (and maybe a moin installation, for convenience)
2008-09-07T14:55:38  <ThomasWaldmann> if somebody has time, please test my TOC modification (not only the preview behaviour, but also other TOC usage)
2008-09-07T15:10:08  * ThomasWaldmann looks at EmailActivation
2008-09-07T15:15:59  <dreimark> ThomasWaldmann: I found probably the reason for the 500er with Load
2008-09-07T15:16:15  <dreimark> MoinMoin/request/__init__.py", line 1326
2008-09-07T15:16:26  <dreimark> if self.cfg.log_timing:
2008-09-07T15:16:33  <dreimark> AttributeError: 'Request' object has no attribute 'cfg'
2008-09-07T15:27:55  <dennda> ThomasWaldmann: http://cia.vc/stats/project/Moin <-- the left hand side shows I had three nicks beginning with "Christopher Denter" when committing. does your command hg export `hg log -M --template "#rev# " -k "Christopher Denter" -r SOC2008-START:SOC2008-END` export them all?
2008-09-07T15:28:16  <dreimark> dennda: did you tried ?
2008-09-07T15:29:18  <dennda> ah looks like
2008-09-07T15:29:20  <dennda> nevermind
2008-09-07T15:33:58  <dreimark> ThomasWaldmann: changeset 3794  560cdfdfffc3 removes cfg from self
2008-09-07T15:37:04  <dennda> """Please upload your code sample by 15 September. If you do not upload
2008-09-07T15:37:04  <dennda> your code sample by 15 September, then we will chase you down for it
2008-09-07T15:37:06  <dennda> and, if need be, rescind payment. """
2008-09-07T15:52:30  <dreimark> ThomasWaldmann: I think we should remove from various actions self.request.finish()
2008-09-07T15:53:03  <dreimark> Load, CopyPage, RenamePage
2008-09-07T15:53:57  <ThomasWaldmann> why?
2008-09-07T15:54:41  <dreimark> because fisnish removes cfg from the request object
2008-09-07T15:55:08  <dreimark> and this gives a failure in request.run
2008-09-07T15:55:31  <dreimark> at the bottom  "if self.cfg.log_timing:"
2008-09-07T15:55:47  <dreimark> then cfg is gone it these actions were used
2008-09-07T15:56:10  <ThomasWaldmann> we emit stuff after calling finish?
2008-09-07T15:56:11  <dreimark> return self.finish() will remove it afterwards
2008-09-07T15:56:51  <dreimark> we call finish in these three actions and then it continues in run
2008-09-07T15:57:18  <dreimark> and checks self.cfg.log_timing
2008-09-07T15:57:39  <ThomasWaldmann> i'll have a look in a few mins
2008-09-07T15:57:43  <dreimark> ok
2008-09-07T16:14:38  <CIA-25> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4059:1d6d0d9a7c9d 1.8/MoinMoin/action/ (CopyPage.py Load.py RenamePage.py): actions: request.run will call request.finish, do not call it in the action
2008-09-07T16:14:40  <CIA-25> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4060:c40f5e5560d4 1.8/ (4 files in 4 dirs): merge main
2008-09-07T16:14:46  <ThomasWaldmann> dreimark: you are right
2008-09-07T16:15:59  <dreimark> ThomasWaldmann: :) but it is the same in 1.7
2008-09-07T16:16:25  <ThomasWaldmann> ehrm, shit :|
2008-09-07T16:16:34  <ThomasWaldmann> I'll fix it.
2008-09-07T16:20:26  <CIA-25> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3809:e2107f4b6217 1.7/MoinMoin/action/ (CopyPage.py Load.py RenamePage.py): actions: request.run will call request.finish, do not call it in the action
2008-09-07T16:23:11  <CIA-25> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4072:8bbb028087f1 1.8/ (12 files in 12 dirs): merge changes from moin/1.7
2008-09-07T16:33:59  * dennda uploaded his diff
2008-09-07T16:35:13  <dreimark> 149K
2008-09-07T16:44:15  <dreimark> ThomasWaldmann: looks like this is fixed too http://moinmo.in/MoinMoinBugs/TableOfContentsBrokenForIncludedPages or what do you think?
2008-09-07T16:49:14  <dennda> oh yes
2008-09-07T16:49:17  <dennda> it's quite big
2008-09-07T16:49:29  <dennda> all I did is execute thomas' command
2008-09-07T16:49:35  <dennda> hmm
2008-09-07T16:49:43  <dennda> will check later
2008-09-07T16:54:27  <ThomasWaldmann> dreimark: maybe, but not related to my recent changeset
2008-09-07T17:00:37  * dennda writes his GSoC posting
2008-09-07T17:48:59  <ThomasWaldmann> dennda: btw, did you get an email from google about your GSOC outcome?
2008-09-07T17:49:04  <dennda> ThomasWaldmann: no
2008-09-07T17:49:11  <dennda> why?
2008-09-07T17:50:16  <ThomasWaldmann> well, I just wondered about whether they send some kind of "official" notification to students about whether they made it or not.
2008-09-07T17:50:37  <dennda> I didn't receive a notification "you passed"
2008-09-07T17:50:41  <dennda> should I worry? :-)
2008-09-07T17:51:27  <ThomasWaldmann> waldi: byeongweon_pc: zenhase: did you get some mail?
2008-09-07T17:52:03  <dennda> ThomasWaldmann: we all passed, didn't we?
2008-09-07T17:52:14  <waldi> ThomasWaldmann: nope
2008-09-07T17:52:37  <waldi> "PS - Don't worry about buying matching shoes, black goes with *everything.*"
2008-09-07T17:53:56  <ThomasWaldmann> waldi: hm? :)
2008-09-07T17:54:42  <waldi> the question was about the color of the gsoc shirt
2008-09-07T17:55:28  <dennda> ThomasWaldmann: afaict they only add students who passed to the page where we need to upload our diffs
2008-09-07T17:57:11  <ThomasWaldmann> waldi: byeongweon_pc: zenhase: dennda: pawel: you passed finals (in case somebody is still unsure)
2008-09-07T17:57:31  <dennda> congrats :-)
2008-09-07T17:57:41  <ThomasWaldmann> expect some $$$$, t-shit and certificate from google
2008-09-07T17:57:44  <dennda> ok then I don't need to adjust the style of writing in my article ^^
2008-09-07T17:57:51  <dennda> t-shit?
2008-09-07T17:57:57  <ThomasWaldmann> +r :)
2008-09-07T17:58:13  <waldi> "Unless you see "Disqualified." you passed."
2008-09-07T18:02:05  <ThomasWaldmann> (and you all entered MoinMoinAcknowledgements page "hall of fame" some time ago :)
2008-09-07T18:09:29  <dennda> uh
2008-09-07T18:12:11  <waldi> ThomasWaldmann: more than once ...
2008-09-07T18:14:17  <dreimark> yeah "BastianBlank" != "Bastian Blank"
2008-09-07T18:14:44  <dreimark> bbl
2008-09-07T18:35:00  <dennda> http://the-space-station.com/2008/09/07/my-work-on-moinmoins-new-storage-engine-gsoc-2008
2008-09-07T18:35:17  <dennda> my blog is a bit deadlocky at the moment, though
2008-09-07T18:35:46  <dennda> should be on planet ubuntu soon
2008-09-07T18:50:57  <dennda> yup there we go: planet.ubuntu.com is now increasing moins popularity :-)
2008-09-07T18:53:43  <zenhase> finally i have decent internet again!
2008-09-07T18:54:53  <zenhase> our universities computing centre (which is the ISP for our students dorm) has massive problems this weekend :/
2008-09-07T18:55:11  <zenhase> 20% packetloss, tcp-connections dropping every few minutes
2008-09-07T18:55:36  <zenhase> makes ssh-sessions and instant-messaging a huge pain
2008-09-07T18:57:03  <zenhase> i also discovered, that jabber does not seem to honor any integrity in message delivery ... the server happily blasts out received messages to the user and deletes them without considering that TCP-connections might die (or already have died somewhere in between)
2008-09-07T18:59:03  <zenhase> cool, i passed! :)
2008-09-07T18:59:25  * zenhase talks first and reads backlogs afterwards ;)
2008-09-07T19:00:00  <zenhase> also congrats to the other students
2008-09-07T19:00:13  <ThomasWaldmann> zenhase: screen rules :)
2008-09-07T19:01:19  <zenhase> ThomasWaldmann: yeah it does
2008-09-07T19:02:35  <ThomasWaldmann> all else you need is some internet server with better connectivity
2008-09-07T19:03:08  * ThomasWaldmann requested that moin gets into heise's software directory
2008-09-07T19:04:14  <zenhase> my server has great connectivity
2008-09-07T19:04:28  <zenhase> my university is the problem ;)
2008-09-07T19:05:01  <zenhase> (which is kind of sad, being a former corner stone of the german internet)
2008-09-07T19:06:08  <ThomasWaldmann> do you get static IPs?
2008-09-07T19:09:57  <dennda> I expect you to comment on that posting!!111
2008-09-07T19:09:59  <dennda> :-)
2008-09-07T19:23:29  <zenhase> ThomasWaldmann: no
2008-09-07T19:23:36  <zenhase> ThomasWaldmann: at least not at my dorm
2008-09-07T19:23:40  <zenhase> hmm
2008-09-07T19:23:47  <zenhase> then i think no dorm in karlsruhe does
2008-09-07T19:24:15  <zenhase> the computing centre monopolizes and directs all dorm-traffic through a single small NAT
2008-09-07T19:34:39  <ThomasWaldmann> dennda: commented :)
2008-09-07T19:35:28  <ThomasWaldmann> dennda: btw, there is stuff like NFS (yuck!), SANs, DRBD, ...
2008-09-07T19:35:47  <dennda> ThomasWaldmann: I know
2008-09-07T19:36:16  <dreimark> re
2008-09-07T19:36:17  <dennda> ThomasWaldmann: i just wanted to add that because the ubuntuusers webteam once complained about it. afair it was the main reason they ditched moin
2008-09-07T19:37:00  <dennda> ThomasWaldmann: approved your comment
2008-09-07T19:37:05  <ThomasWaldmann> complained about what?
2008-09-07T19:37:12  <dreimark> byeongweon_pc: or anyone else, try to answer the question with the gui editor on http://test18.moinmo.in/TestWiki18
2008-09-07T19:38:00  <dennda> ThomasWaldmann: that it is not really trivial to have moin use separate "database" servers
2008-09-07T19:38:30  <ThomasWaldmann> managing busy systems is in general not trivial :)
2008-09-07T19:38:41  <dennda> indeed
2008-09-07T19:39:15  <ThomasWaldmann> and as we currently don't support databases at all, there is not much point in asking for seperate database servers :)
2008-09-07T19:39:53  <dennda> well that's just their use case
2008-09-07T19:39:56  <ThomasWaldmann> also, a database (SQL or whatever) won't magically solve all your problems
2008-09-07T19:40:09  <dreimark> no you get new problems
2008-09-07T19:40:14  <dennda> yes, I should have made that more clear
2008-09-07T19:40:24  <dennda> too late now :-)
2008-09-07T19:40:49  <dreimark> and we will need probably some database related migration scripts
2008-09-07T19:40:57  <ThomasWaldmann> some stuff will be easier, some will be harder and some stuff will still suck (if the algorithm using the backend is just not optimized for large scale)
2008-09-07T19:42:58  <ThomasWaldmann> dreimark: except from backend init (create DB/tables) the clone script should do it
2008-09-07T19:43:31  * ThomasWaldmann for sure won't touch sql stuff :)
2008-09-07T19:43:46  <dennda> ok I will let you read the preview next time :-)
2008-09-07T19:46:51  <dreimark> ThomasWaldmann: I don't mean at the initialisation time. That step is easy. The question is if we need to take care on optimisation of DB tables from 1.9 to 2.0 and further or how we can test migration data on such backends
2008-09-07T19:50:09  <ThomasWaldmann> that will be a problem of the guy implementing the SQL stuff :)
2008-09-07T19:50:46  <ThomasWaldmann> what we need much more and rather soon is the ACL and router stuff
2008-09-07T19:50:56  <ThomasWaldmann> dennda: did you look at the stuff I committed?
2008-09-07T19:51:07  <dennda> back then? :-)
2008-09-07T19:51:49  <ThomasWaldmann> 2wk ago or so
2008-09-07T19:52:48  <dennda> we spoke about it, old man :D
2008-09-07T19:53:36  <ThomasWaldmann> well, maybe i used the wrong word. s/look at/work on/ :)
2008-09-07T19:54:34  <dennda> no, sorry. was too busy with that blog
2008-09-07T19:54:48  <dennda> I hope 0.1 is reached soon
2008-09-07T19:54:52  <dennda> :-)
2008-09-07T20:06:22  <ThomasWaldmann> TheSheep: btw, does modernized have open todo?
2008-09-07T20:07:12  <TheSheep> ThomasWaldmann: I rather try to keep a DONE list
2008-09-07T20:07:41  <ThomasWaldmann> ok, so what's not on that list? :)
2008-09-07T20:07:52  <ThomasWaldmann> (that reminds me of that goedel xkcd)
2008-09-07T20:08:01  <TheSheep> that was a good one
2008-09-07T20:09:06  <TheSheep> ThomasWaldmann: obviously, anything that doesn't look/work right with the current version
2008-09-07T20:10:48  <ThomasWaldmann> because after 1.7.2 is out, next thing should be some 1.8beta.
2008-09-07T20:11:23  <TheSheep> ThomasWaldmann: and when 1.7.2 is out?
2008-09-07T20:11:24  <dennda> ThomasWaldmann: what release do we plan to ship the storage things?
2008-09-07T20:11:26  <dennda> 2.0?
2008-09-07T20:11:52  <ThomasWaldmann> i wanted to do it this weekend, but some fscking hardware problems at a customer kept me busy otherwise
2008-09-07T20:12:12  <ThomasWaldmann> dennda: yes
2008-09-07T20:12:16  <dennda> ok
2008-09-07T20:12:26  <dennda> plenty of time XD
2008-09-07T20:12:28  * dennda runs
2008-09-07T20:12:36  <ThomasWaldmann> that doesn't imply there will be 1.9 :)
2008-09-07T20:12:45  <TheSheep> ThomasWaldmann: ok, I will try to... polish it this week a little :)
2008-09-07T20:13:21  <ThomasWaldmann> dennda: and we should have that stuff merged before next soc...
2008-09-07T20:13:54  <dennda> ThomasWaldmann: hmmm sweet fun
2008-09-07T20:14:24  <dennda> especially merging zenhase's work and mine :-)
2008-09-07T20:15:26  <ThomasWaldmann> ehrm. wsgi == 1.9
2008-09-07T20:15:37  <ThomasWaldmann> so we will have 1.9 :)
2008-09-07T20:16:31  <ThomasWaldmann> (it'll be first to give us more time for storage/dom stuff and because wsgi stuff is less likely to cause unnoticed breakage / lots of other work)
2008-09-07T20:22:24  <ThomasWaldmann> ok, so lets talk about repos and releases :)
2008-09-07T20:22:45  <ThomasWaldmann> any problem with feature freezing 1.8 asap?
2008-09-07T20:23:28  <dennda> yes but you still need to merge that
2008-09-07T20:23:46  <dennda> at least my branch won't have all the things zenhase did magically :-)
2008-09-07T20:23:54  <dennda> and I used request objects quite often
2008-09-07T20:24:39  <ThomasWaldmann> that will have to be resolved by merging moin/1.9 (with wsgi stuff) into moin/storage
2008-09-07T20:25:00  <dennda> which is what I was talking about :-)
2008-09-07T20:25:12  <ThomasWaldmann> but first we need 1.9 repo
2008-09-07T20:26:25  <ThomasWaldmann> we can have it, but that means: 1.7 == just major bug and security fixes, 1.8 == no new features, just bug fixes
2008-09-07T20:27:34  <dennda> what's the point of 1.8 then?
2008-09-07T20:27:44  <dennda> if you don't add new features
2008-09-07T20:28:09  <ThomasWaldmann> byeongweon_pc: dreimark: TheSheep: xorAxAx: any stuff you think that needs to get into 1.8, any major fixes?
2008-09-07T20:28:22  <ThomasWaldmann> dennda: 1.8 is to get stuff out that we have production ready
2008-09-07T20:29:17  <ThomasWaldmann> i.e. GUI editor work, modernized theme, some other stuff done already in 1.8 repo
2008-09-07T20:29:18  <dennda> like what?
2008-09-07T20:29:32  <dennda> isn't that "new features"
2008-09-07T20:29:44  <ThomasWaldmann> no, we already have that
2008-09-07T20:29:54  <ThomasWaldmann> new == not in 1.8 already
2008-09-07T20:30:04  <dennda> ah
2008-09-07T20:33:11  <ThomasWaldmann> http://moinmo.in/MoinMoinTodo/Release%201.8
2008-09-07T20:47:55  * ThomasWaldmann added release schedule there
2008-09-07T20:48:54  <ThomasWaldmann> (rather quick beta/rc/release cycle as the changes in 1.8 are not that big, so hopefully that'll work)
2008-09-07T21:41:34  <dreimark> ThomasWaldmann: I want to add arnica soon
2008-09-07T21:47:36  <ThomasWaldmann> not sure if it is in the wanted state already
2008-09-07T21:53:50  <CIA-25> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4073:83d99e08e931 1.8/MoinMoin/i18n/ (42 files): updated i18n
2008-09-07T22:03:33  * dreimark thinks it is in a better state than the backup and restore action was over years.
2008-09-07T22:04:33  <dreimark> (but that comparision is not a good one)
2008-09-07T22:07:32  <dreimark> it is one of the tools which do use the new action cache code
2008-09-07T22:08:17  <dreimark> both should get in use
2008-09-07T22:10:29  <CIA-25> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4074:d712b7ef6043 1.8/docs/CHANGES: CHANGES: refactored header a bit, 1.8: first mention new features
2008-09-07T22:11:02  <ThomasWaldmann> dreimark: there are 2 problems in arnica:
2008-09-07T22:11:09  <dreimark> tell
2008-09-07T22:12:04  <ThomasWaldmann> a) the code uses lots of lowlevel AttachFile stuff and will have to get major changes when storage is merged.
2008-09-07T22:12:44  <dreimark> ok, this is known
2008-09-07T22:13:05  <ThomasWaldmann> (there is not much alternative currently to that, but still, if it is merged into moin, it will put an additional burden because we'll have to fix it before getting storage stuff out)
2008-09-07T22:13:36  <ThomasWaldmann> [this is not the case if it lives as an option on *Market]
2008-09-07T22:14:33  <ThomasWaldmann> b) the code seems a bit overcomplicated and uses multiple "parallel" arrays a lot
2008-09-07T22:15:06  <ThomasWaldmann> (not one list of tuples or dicts - that could maybe simplify things a lot)
2008-09-07T22:15:18  <dreimark> I've choosen the arrays because dict's don't keep the order
2008-09-07T22:15:42  <ThomasWaldmann> i mean a list of dicts or tuples
2008-09-07T22:15:57  <ThomasWaldmann> with the dict or tuple holding the information for one image
2008-09-07T22:18:24  <dreimark> that can work too and indead it will be easier. but this point is about refactoring which does not change any of the features it has now
2008-09-07T22:19:13  <dreimark> it sound like make the code simpler which I can totally agree
2008-09-07T22:19:48  <dreimark> so if this is done a) is still there
2008-09-07T22:20:49  <ThomasWaldmann> yes, b) is about refactoring
2008-09-07T22:22:15  <ThomasWaldmann> btw, there might be also some code duplication to reduce (not sure, didn't look into it the last days)
2008-09-07T22:27:24  <dreimark> I think that will happen during this change. but lets better talk on a) because I will work on b) anyway but I am not sure if it is worth to make it in haste (not so much haste because of my vacation)
2008-09-07T22:28:13  <ThomasWaldmann> refresh.py has some strange code
2008-09-07T22:28:37  <dreimark> If I haven't missed one getAttachDir, _get_files, .getAttachUrl, getFilename, move_file are used fromn AttachFile
2008-09-07T22:30:23  <dreimark> if file.endswith(key) ?
2008-09-07T22:32:35  <ThomasWaldmann> a) it seems you use internal knowledge about how the caching works
2008-09-07T22:33:04  <ThomasWaldmann> b) what's the sense of removing those files?
2008-09-07T22:33:22  <dreimark> yes we talked in vilnius about that they have always the same length
2008-09-07T22:33:47  <dreimark> b) for recreating with a differrent width and height
2008-09-07T22:34:51  <dreimark> a) arnica uses some prefix's and all should be removed
2008-09-07T22:36:07  <dreimark> +cachefiles
2008-09-07T22:37:06  <ThomasWaldmann> right now it is not clear to me why you need 2 nested loops to do that
2008-09-07T22:44:15  <dreimark> hmm, seems I wanted to be double safe
2008-09-07T22:44:35  <dreimark> the second loop is not needed
2008-09-07T22:44:53  <ThomasWaldmann> you could simply iterate over all attachments, calculate the cache keys (and add all prefixes/suffixes you use) and remove those
2008-09-07T22:45:50  <ThomasWaldmann> also, you could add w/h to the cache key, so it will never access a wrong cache
2008-09-07T22:47:09  <dreimark> the last johill suggested too.
2008-09-07T22:50:15  <ThomasWaldmann>             "thumbnail": "%s%s" % (AttachFile.getAttachUrl(self.pagename, '', self.request), self.thumb_key[idx]),
2008-09-07T22:50:24  <dreimark> how much files can the fs moin is used on store?
2008-09-07T22:50:42  <ThomasWaldmann> didn't you change the code to not store thumbnails as attachments (but in cache)?
2008-09-07T22:50:54  <dreimark> it stores in cache
2008-09-07T22:51:21  <ThomasWaldmann> what does that line do?
2008-09-07T22:53:48  <ThomasWaldmann> looks like unused stuff
2008-09-07T22:53:56  <dreimark> shiiit
2008-09-07T22:54:03  <dreimark> missed to kill that line
2008-09-07T22:55:43  <ThomasWaldmann> that other thumbnail entry seems also to be unused
2008-09-07T22:55:47  <ThomasWaldmann>             except IOError:
2008-09-07T22:55:47  <ThomasWaldmann>                 msg = _("attachment: %(filename)s not exists") % {"filename": os.path.basename(imagef)}
2008-09-07T22:56:17  <ThomasWaldmann> that is evil. you just tell that it does not exist for any kind of IOError.
2008-09-07T22:58:38  <dreimark> yeah, but mostly the file has 0 bytes
2008-09-07T22:59:11  <dreimark> the text can be changed
2008-09-07T23:00:02  <dreimark> _("Can't access attachment: %(filename)s")
2008-09-07T23:02:02  <ThomasWaldmann> that sort_by_*=boolean - maybe just use sort_by=what
2008-09-07T23:05:03  <ThomasWaldmann> some methods have an empty line after the docstrings, some not. maybe remove the empty lines.
2008-09-07T23:06:25  <dreimark> both ok :)
2008-09-07T23:10:57  <ThomasWaldmann> another source of storage problems will be places where you used os.path (or other places working with pathes instead of item/page/attachment names)
2008-09-07T23:20:05  * dreimark wonders about how Image.open can work then 
2008-09-07T23:21:51  <ThomasWaldmann> of course you need to create pathes to files where needed, but the arnica methods and functions should not deal with pathes as parameters or in data structures
2008-09-07T23:22:44  <ThomasWaldmann> maybe everything could a bit more OO, too :)
2008-09-07T23:26:44  <dreimark> a lot of other code has this os.path problem
2008-09-07T23:27:18  <ThomasWaldmann> maybe first think about how it could work later, with items and make your api for that
2008-09-07T23:27:44  <ThomasWaldmann> just dealing with attachments hidden somewhere inside some objects/methods
2008-09-07T23:28:17  <ThomasWaldmann> yes, I know and it will be lots of work changing all that
2008-09-07T23:32:34  <ThomasWaldmann> (and we only have about 3 months for it)
2008-09-07T23:41:23  <dreimark> how do you get to this 3month count ?
2008-09-07T23:42:23  <dreimark> until 2008-11-01 release 1.8.0 is 2 months
2008-09-07T23:42:50  <ThomasWaldmann> i mean integrating storage code
2008-09-07T23:43:26  <dreimark> you want to release 1.9 this year ?
2008-09-07T23:43:35  <ThomasWaldmann> if we get 1.8 out until november, maybe we get 1.9 out late december or early january
2008-09-07T23:43:42  <ThomasWaldmann> (with wsgi stuff)
2008-09-07T23:44:08  <ThomasWaldmann> and then we have until march or so until we maybe can participate in next SOC
2008-09-07T23:44:19  * dreimark things that is to fast. people will think we always do 4-5 major releases in a year
2008-09-07T23:44:30  <dreimark> they will hate us for this
2008-09-07T23:45:06  <ThomasWaldmann> if they don't like it, they don't need to install every new version
2008-09-07T23:45:37  <ThomasWaldmann> but if we don't have storage included until SOC, we will block further work dependant on it
2008-09-07T23:47:33  <ThomasWaldmann> and I really don't want to have another "make storage code production ready" project
2008-09-07T23:49:00  <dreimark> we will have to care for many releases over the next two years
2008-09-07T23:50:28  <dreimark> I agree with the storage project but we should slow down a bit afterwards to 2 releases a year
2008-09-07T23:51:14  <ThomasWaldmann> it doesn't make a big difference if we do 2 big releases or 4 smaller ones
2008-09-07T23:51:15  <dreimark> or lets see how much code makes a new release
2008-09-07T23:53:33  <dreimark> well we should work on documentation too not only on new code and there is still more time within 2 big releases
2008-09-07T23:56:55  <ThomasWaldmann> in 1.9, we'll get rid of a lot of outdated docs
2008-09-07T23:59:00  <dreimark> is the interview online ?