MoinMoin: MoinMoinChat/Logs/moin-dev/2006-07-31

2006-07-31T10:05:08  <ffesti> moin
2006-07-31T10:12:46  <ThomasWaldmann> moin ffesti
2006-07-31T10:13:41  <ThomasWaldmann> ffesti: how is the state of your form stuff? we keep getting requests about forms...
2006-07-31T10:14:21  <ffesti> I'll put what I have on my Wiki Hp today
2006-07-31T10:14:31  <ffesti> as soon as I got though my mail and stuff
2006-07-31T10:15:19  <ffesti> +r
2006-07-31T10:15:30  <ffesti> and had my first cup of coffee ;)=
2006-07-31T10:20:48  <Kepplar> hi
2006-07-31T10:25:43  <ThomasWaldmann> ffesti: thanks
2006-07-31T10:41:05  <fpletz> moin. I might be a bit late for the meeting today but I'll be here in any case.. need to make some phone calls..
2006-07-31T10:43:50  * Kepplar props head up against wall
2006-07-31T10:44:01  <Kepplar> Warning: I may make even less sense than usual
2006-07-31T10:46:33  <ThomasWaldmann> then better go to bed |-)
2006-07-31T10:46:59  <Kepplar> nah i need to stay awake for a few more hours if im to regain normal pattern by tomrorow
2006-07-31T10:48:30  <xorAxAx> Kepplar: what happened? :)
2006-07-31T10:49:39  <ffesti> Someone set us up the bomb!
2006-07-31T10:49:53  <xorAxAx> All your base are belong to us!
2006-07-31T10:50:40  <Kepplar> xorAxAx: its called re-structuring ones Industrial Placement report the night before the deadline
2006-07-31T10:51:00  <xorAxAx> Kepplar: who wants to see the report?
2006-07-31T10:51:22  <Kepplar> uni
2006-07-31T10:51:32  <Kepplar> currently stands at 8000 words
2006-07-31T10:51:37  <Kepplar> more or less done
2006-07-31T10:51:40  <Kepplar> not entirely happy
2006-07-31T10:51:48  <Kepplar> should be sufficiant to keep a first for this year
2006-07-31T11:01:05  <xorAxAx> 11:01 here :)
2006-07-31T11:01:24  <xorAxAx> i think we can start to paste our 3 lines:
2006-07-31T11:01:37  <xorAxAx>  LAST: Implemented IWID support, added function to generate random strings. Added support for editing the InterWikiMap in the wiki, added automatic reloading of intermap files. Added locking to the PickleTagStore and the MetaDict classes. Added handling of various options and detection of anonymous wikis to the SyncPages action.
2006-07-31T11:01:42  <xorAxAx>  NEXT: Get actual synchronisation working (SyncPages action), experiment with different merge algorithms.
2006-07-31T11:01:45  <xorAxAx>  BLOCKERS: Storage system doesn't allow r-locks per page. Where is the new one?
2006-07-31T11:01:57  <Kepplar> um
2006-07-31T11:02:06  <Kepplar> you want me to comment on blockers, or let tw handle the first two?
2006-07-31T11:02:30  <xorAxAx> maybe you paste your lines first :) then we can talk about blockers
2006-07-31T11:02:39  <xorAxAx> its very sensible to resolve blockers together
2006-07-31T11:02:42  <Kepplar> kk
2006-07-31T11:02:47  <Kepplar> well I dont prewrite lines so...
2006-07-31T11:03:26  <Kepplar> LAST: Implimented the Rubik \o/. Bug fixes through testing. Removed hard coded engine set up and generally removed hardcoded temp code. Began integration of the new Storage Engine into Page
2006-07-31T11:03:42  <Kepplar> NEXT: Integration with Page, PE, PGE, user, acl etc
2006-07-31T11:04:22  <Kepplar> BLOCKERS: None, although there might be some issues regarding to legacy file format wierdness and I havent implimented edit-log and locks
2006-07-31T11:04:39  <Kepplar> (well i have sort implimented locks, just not completly)
2006-07-31T11:04:51  <Kepplar> Oh I've also done some more UML
2006-07-31T11:07:54  <ThomasWaldmann> ok, I am here now, too (had customer phone calls, sorry).
2006-07-31T11:08:43  <ThomasWaldmann> xorAxAx: interwikimap in the wiki, how is it performance-wise and how to support farms with that?
2006-07-31T11:09:20  <xorAxAx> ThomasWaldmann: in the old version, it loaded the files on the first request (for every request with cgi)
2006-07-31T11:09:35  <xorAxAx> ThomasWaldmann: now it caches all files and the Page contents for 3 minutes
2006-07-31T11:10:17  <xorAxAx> so it reloads them every 3 minutes
2006-07-31T11:10:25  <xorAxAx> farms can still use shared intermaps
2006-07-31T11:10:39  <xorAxAx> or sync the interwikimap page via wikisync
2006-07-31T11:11:33  <ThomasWaldmann> it checks page and files for mtime against cache?
2006-07-31T11:11:40  <xorAxAx> no
2006-07-31T11:11:45  <xorAxAx> it simply reloads
2006-07-31T11:11:52  <xorAxAx> could be implemented
2006-07-31T11:12:35  <xorAxAx> but getting the mtime of the page is expensive
2006-07-31T11:12:48  <xorAxAx> (opening the log etc.)
2006-07-31T11:13:09  <xorAxAx> so i thought that this is the fastest amortimised solution
2006-07-31T11:13:09  <ThomasWaldmann> only get it when the 5mins are over
2006-07-31T11:13:20  <xorAxAx> ok, that would be possible as well
2006-07-31T11:13:56  <ThomasWaldmann> did you fix cache_dir usage in MetaDict already?
2006-07-31T11:14:06  <xorAxAx> currently doing so
2006-07-31T11:14:17  <xorAxAx> i have to fix the interface because it doesnt have a request object
2006-07-31T11:15:03  <ThomasWaldmann> what does "detection of anon wikis" mean and what is it used for?
2006-07-31T11:15:29  <xorAxAx> anon == anonymous.
2006-07-31T11:15:45  <Kepplar> even i guessed that :P
2006-07-31T11:15:46  <xorAxAx> the problem i had talked about earlier: not every admin sets an interwikiname
2006-07-31T11:16:02  <Kepplar> xorAxAx: use hostname?
2006-07-31T11:16:03  <Kepplar> er
2006-07-31T11:16:06  <Kepplar> i mean dns name
2006-07-31T11:16:07  <xorAxAx> so i consider these wikis anonymous and have to show the URL in those cases
2006-07-31T11:16:18  <Kepplar> yep
2006-07-31T11:16:22  <xorAxAx> Kepplar: yes, but a wiki doesnt know its hostname nor URL
2006-07-31T11:16:30  <xorAxAx> so it doesnt know its name
2006-07-31T11:16:40  <xorAxAx> and thats where the IWID comes into play
2006-07-31T11:17:00  <Kepplar> surely you can forward on the http request location coming in?
2006-07-31T11:17:08  <Kepplar> (so badly written i know)
2006-07-31T11:17:28  <ThomasWaldmann> xorAxAx: explain why you have those blockers
2006-07-31T11:17:41  <xorAxAx> ThomasWaldmann: which blockers? i had listed one above
2006-07-31T11:18:00  <xorAxAx> or do you mean this anon wiki thing?
2006-07-31T11:18:10  <ThomasWaldmann> blockers mean that you have to stop working because of that (r-lock)
2006-07-31T11:18:29  <xorAxAx> right! i dont have to stop directly working, but i cannot guarantee data safety
2006-07-31T11:18:44  <xorAxAx> because i cannot correctly lock pages locally
2006-07-31T11:19:09  <xorAxAx> lupy introduced a special page lock, but the PageEditor doesnt respect it, does it?
2006-07-31T11:19:15  <ThomasWaldmann> this is rather a TODO comment than a blocker.
2006-07-31T11:19:25  <xorAxAx> ok, right. but nevertheless it is annoying
2006-07-31T11:19:45  <ThomasWaldmann> well, this is why kepplar is doing his job :)
2006-07-31T11:19:47  <xorAxAx> no, i wont stop working there just because of that :)
2006-07-31T11:20:09  <Kepplar> souldnt blockers be handled transparently for you xorAxAx?
2006-07-31T11:20:16  <xorAxAx> Kepplar: ?
2006-07-31T11:20:21  <Kepplar> I'm probably just going to have an exception raised if a page is in use
2006-07-31T11:20:49  <xorAxAx> Kepplar: you need to use locks (and the locks code is already there)
2006-07-31T11:20:50  <ThomasWaldmann> ok, kepplar's time (before he falls asleep <g>)
2006-07-31T11:21:07  <Kepplar> xorAxAx: it will, but thats not presented to you
2006-07-31T11:21:09  * xorAxAx just committed the metadict fix to the sync branch
2006-07-31T11:21:12  <Kepplar> xorAxAx: at least i dont think its a good idea
2006-07-31T11:21:27  <xorAxAx> Kepplar: we can discuss about after this meeting
2006-07-31T11:21:32  <Kepplar> xorAxAx: storage should hide the inner working and just present you with a raised exception ItemRevisionInUse or something
2006-07-31T11:21:42  <Kepplar> xorAxAx: sure
2006-07-31T11:21:42  <xorAxAx> Kepplar: exactly
2006-07-31T11:22:01  <xorAxAx> Kepplar: but i need to lock directly as well
2006-07-31T11:22:02  <Kepplar> ThomasWaldmann: pasted my lines already :)
2006-07-31T11:22:20  <ThomasWaldmann> Kepplar: yes, how far is Page working right now with your code?
2006-07-31T11:22:49  <Kepplar> xorAxAx: yea, at the moment I'm thinking have a lock function to indicate the page is being modified (possibly setting it up automatically when clone() is run on an item revision or something else
2006-07-31T11:23:02  <Kepplar> ThomasWaldmann: not that far, im just doing the inital configurations
2006-07-31T11:23:26  <Kepplar> of the Item object
2006-07-31T11:23:32  <Kepplar> oh actually
2006-07-31T11:23:32  <Kepplar> yes
2006-07-31T11:23:35  <Kepplar> i have one blocker
2006-07-31T11:23:45  <Kepplar> its in regards with creating the engines
2006-07-31T11:24:05  <Kepplar> at the moment each time Page is created it "creates" a StorageEngineManager
2006-07-31T11:24:20  <Kepplar> in reality the object is singleton-eque as it uses the Borg design pattern
2006-07-31T11:24:43  <Kepplar> however, since hardcoded engine creation has come out (so i can support underlay as discussed) i have to create it somewhere
2006-07-31T11:24:46  <Kepplar> if I put it in page
2006-07-31T11:24:54  <Kepplar> then engines will be created over and over again when page is called
2006-07-31T11:25:13  <ThomasWaldmann> cant you create the S.E.M. in the cfg?
2006-07-31T11:25:25  <Kepplar> ThomasWaldmann: sure, but I'm not familar on that yet :)
2006-07-31T11:26:07  <Kepplar> it was either that or quickly hack up if statements to test if the data and underlay engines have been created
2006-07-31T11:26:09  <ThomasWaldmann> you can access it via request.cfg and you just put the code into wikiconfig.py
2006-07-31T11:26:27  <Kepplar> and its just an object I can write anything to?
2006-07-31T11:26:30  <ThomasWaldmann> what params does your SEM get?
2006-07-31T11:26:38  <Kepplar> it takes none
2006-07-31T11:26:57  <Kepplar> creat engine takes engine "instance" name and engine "class" type
2006-07-31T11:27:14  <Kepplar> and create item requires an engine to create the item (newly added) and the wikiname
2006-07-31T11:27:23  <ThomasWaldmann> example?
2006-07-31T11:27:30  <Kepplar> sem = StorageEngineManager()
2006-07-31T11:27:49  <Kepplar> sem.create_engine("mystandardflatfile",  "core/flatfile")
2006-07-31T11:28:17  <ThomasWaldmann> where does data_dir get in it?
2006-07-31T11:28:21  <Kepplar> item = sem.reteieve_item("MyWikiPage", "mystandardlfatfile")
2006-07-31T11:28:35  <Kepplar> ah
2006-07-31T11:28:37  <Kepplar> thats
2006-07-31T11:29:00  <Kepplar> sem.engines["mystandardflatfile"].open("my/uri/to/wiki/data")
2006-07-31T11:29:07  <Kepplar> that would come from config i'd imagine
2006-07-31T11:30:02  <ThomasWaldmann> ok, until we have refactored that, put that into wikiconfig
2006-07-31T11:30:21  <Kepplar> refactored what?
2006-07-31T11:30:38  <ThomasWaldmann> that it is easier and less code :)
2006-07-31T11:31:11  <ThomasWaldmann> (at least for the user doing the configuration)
2006-07-31T11:32:56  <ThomasWaldmann> can we have multiple "open" to one engine?
2006-07-31T11:34:59  <ThomasWaldmann> or could we do mystandardflatfile = sem.create_engine("data/", "core/flatfile")
2006-07-31T11:36:18  <ThomasWaldmann> and maybe later layers = [mystandardflatfile, myunderlayflatfile]
2006-07-31T11:36:39  <Kepplar> er
2006-07-31T11:37:05  <Kepplar> sem.create_data argument changes sound good
2006-07-31T11:37:31  <Kepplar> multiple open doesnt work
2006-07-31T11:37:35  <Kepplar> incase someone closes it
2006-07-31T11:37:38  <Kepplar> and tries to write
2006-07-31T11:37:59  <Kepplar> hopefully im going to migrate all storae address generation on open
2006-07-31T11:38:12  <Kepplar> so they arent run every single query in the hope it may speed things up
2006-07-31T11:38:14  <ThomasWaldmann> if there is no multiple open, but one open has to happen anyway, you maybe could use __init__ for that
2006-07-31T11:38:49  <Kepplar> yea
2006-07-31T11:38:55  <Kepplar> if a open url is provided
2006-07-31T11:39:02  <Kepplar> then __init__ will just call .open()?
2006-07-31T11:39:14  <Kepplar> same way Page calls reset()
2006-07-31T11:39:18  <ThomasWaldmann> for example
2006-07-31T11:39:22  <Kepplar> sounds good
2006-07-31T11:40:07  <ThomasWaldmann> the question is rather if there is an application for creating an engine with no URL and open it later?
2006-07-31T11:40:26  <ThomasWaldmann> if there is no such application, you don't need a separate open() call
2006-07-31T11:40:53  <Kepplar> except might be nicer to keep the functionality of __init__ and this seperated *shrugs*?
2006-07-31T11:41:04  <Kepplar> maybe we have a remote file system
2006-07-31T11:41:06  <Kepplar> with various config
2006-07-31T11:41:10  <ThomasWaldmann> that's _open then :)
2006-07-31T11:41:12  <Kepplar> and it has round-robit access
2006-07-31T11:41:17  <Kepplar> er
2006-07-31T11:41:18  <Kepplar> robin*
2006-07-31T11:41:23  <Kepplar> or failover
2006-07-31T11:41:32  <Kepplar> but one doesnt want to change internal identifier
2006-07-31T11:41:53  <Kepplar> plus
2006-07-31T11:42:00  <ThomasWaldmann> ok, we can continue on this later
2006-07-31T11:42:01  <Kepplar> closing the engine means that the wiki can be shut down
2006-07-31T11:42:05  <Kepplar> ok
2006-07-31T11:42:05  <ThomasWaldmann> fpletz: your time
2006-07-31T11:42:19  <Kepplar> nn
2006-07-31T11:42:20  <Kepplar> :P
2006-07-31T11:44:31  <ThomasWaldmann> fpletz: you need to raise SOC priority
2006-07-31T11:46:10  <xorAxAx> he needs a preemptive scheduler :)
2006-07-31T11:46:43  <fpletz> ok, now
2006-07-31T11:46:51  <fpletz> well, problem is.. it's preemtive ;)
2006-07-31T11:46:56  <ThomasWaldmann> definitely. the current one has some days latency.
2006-07-31T11:47:35  <fpletz> LAST: I couldn't do any work last week because I finally got my formal admission to study at the TU Munich which I needed to register with the accomodation search service of the student union. So I ended up phoning countless people and inspecting apartments from thursday to saturday.
2006-07-31T11:47:40  <fpletz> I'm sorry, but this had top priority as the market for cheap accomodation is sparse. Furthermore, I surely weren't the only student receiving his admission that week so I had to take action.
2006-07-31T11:47:44  <fpletz> NEXT: finally the UI stuff
2006-07-31T11:47:46  <fpletz> BLOCKERS: -
2006-07-31T11:49:11  <ThomasWaldmann> how is your plan to assure FULL time work on SOC?
2006-07-31T11:49:24  <ThomasWaldmann> I guess this is your biggest problem currently.
2006-07-31T11:50:36  <fpletz> yes, I know.. however, sometimes there are more important things in my life to be done
2006-07-31T11:50:53  <fpletz> I know that there are about 20 days left, but this should be enough to complete all my tasks
2006-07-31T11:51:51  <ThomasWaldmann> the problem is that there will be some question like "should the student get the final $2000" in the final survey.
2006-07-31T11:52:46  <ThomasWaldmann> and in contrary to the mid-term survey, there will be no place for a "yes" just based on optimism.
2006-07-31T11:53:25  <ThomasWaldmann> btw, this comment of course applies to all SOC students, not only fpletz
2006-07-31T11:53:37  <fpletz> yeah, that's the thing.. if I don't succeed I honestly don't deserve it, it's that easy.. but I'll keep trying!
2006-07-31T11:54:38  <ThomasWaldmann> so better get work done EACH day, not just hope for you can do it all at the end somehow
2006-07-31T11:55:51  <fpletz> yes, will do.. I can work this afternoon but will probably be in munich tomorrow (the phone calls a few moments ago..)
2006-07-31T11:57:19  <fpletz> ok, so the search UI
2006-07-31T11:57:19  <fpletz> btw.
2006-07-31T11:57:42  <ThomasWaldmann> and it is not only based on "success" (meaning that something works), but also on quantity and quality of work.
2006-07-31T11:57:46  <fpletz> xorAxAx: what did you mean yesterday or the day before with erros in titlesearch or something?
2006-07-31T11:57:58  <ThomasWaldmann> quality is no problem for you, fpletz. quantity is.
2006-07-31T11:58:09  <xorAxAx> fpletz: i got a blank screen ... i didnt track it down
2006-07-31T11:58:12  <fpletz> ok, I see..
2006-07-31T11:58:22  <xorAxAx> i.e. the connection was canceled early
2006-07-31T11:58:23  <fpletz> xorAxAx: not using xapian?
2006-07-31T11:58:25  <ThomasWaldmann> it maybe was based on header trouble
2006-07-31T11:58:28  <xorAxAx> not using xapian, yes
2006-07-31T11:58:45  <fpletz> ok, I'll take a look at it, thanks
2006-07-31T11:59:12  <ThomasWaldmann> did you contact starshine for graphics?
2006-07-31T11:59:32  <ThomasWaldmann> I already warned her :)
2006-07-31T11:59:55  <fpletz> yeah, last week, but haven't heard back from her
2006-07-31T11:59:59  <ThomasWaldmann> So maybe make up some concrete requirements for the gfx and tell her.
2006-07-31T12:00:13  <fpletz> well, I'll ping her when she's here ;)
2006-07-31T12:00:22  <fpletz> ok
2006-07-31T12:00:39  <ThomasWaldmann> if she has no time, try thesheep, he is also very good at gfx
2006-07-31T12:01:08  <ThomasWaldmann> until you have the real stuff, you could even borrow gfx from google :)
2006-07-31T12:01:28  <fpletz> hehe, yeah..
2006-07-31T12:01:38  <fpletz> but I gues not commit them to the repo :)
2006-07-31T12:02:17  <ThomasWaldmann> maybe not :)
2006-07-31T12:02:36  <ThomasWaldmann> but to use the time to work in parallel
2006-07-31T12:02:47  <fpletz> btw. have you heard anything from richard?
2006-07-31T12:03:03  <ThomasWaldmann> not yet, but I guess he should return soon
2006-07-31T12:03:09  <fpletz> ok
2006-07-31T12:03:31  <ThomasWaldmann> but just ask here if you have questions
2006-07-31T12:04:41  <fpletz> yup, will do
2006-07-31T12:05:11  <ThomasWaldmann> do you have a laptop?
2006-07-31T12:06:34  <fpletz> yes, but it's pretty old, heavy and the battery is weak.. already ordered a new one for my studies which should arrive this week
2006-07-31T12:07:11  <ThomasWaldmann> it is nice for travelling by train, so travel time is not wasted time
2006-07-31T12:07:29  <xorAxAx> esp. using a DVCS :)
2006-07-31T12:07:36  <ThomasWaldmann> right
2006-07-31T12:07:39  <fpletz> yeah ;)
2006-07-31T12:08:16  <fpletz> oh
2006-07-31T12:08:23  <fpletz> great
2006-07-31T12:08:24  <fpletz> Sendung Befindet Sich In Der Zustellung
2006-07-31T12:08:28  <fpletz> perhaps today ;)
2006-07-31T12:09:02  <ThomasWaldmann> type of laptop?
2006-07-31T12:09:10  <xorAxAx> hopefully an IBM
2006-07-31T12:09:20  <fpletz> ibm was too expensive
2006-07-31T12:09:22  <fpletz> macbook ;)
2006-07-31T12:09:41  <xorAxAx> my ibm wasnt expensive
2006-07-31T12:09:43  <xorAxAx> 1100 EUR
2006-07-31T12:10:09  <fpletz> hmm, what model?
2006-07-31T12:11:08  * ThomasWaldmann increased rss_cache to 300 for the farm and changed the code to only do 5 full diffs in rss
2006-07-31T12:11:32  <xorAxAx> fpletz: bought it 1 year ago, a T41
2006-07-31T12:12:02  * ThomasWaldmann gtg & bbl
2006-07-31T12:14:32  <fpletz> ah, ok.. I was primarily looking at the X series.