Tuesday, 2009-07-28

*** tpb has joined #tp00:00
*** ChanServ sets mode: +o tpb00:00
Greywhindtansell-laptop, llnz, you have time to talk?00:04
tansell-laptopin about 20 minutes00:04
llnzgoing to be leaving in 20 minutes00:04
llnzi'll be back about 45 minutes later00:05
alanpgrrrr browser00:16
*** llnz has quit IRC00:34
tansell-laptopGreywhind, what did you want to talk about?00:38
tansell-laptopmedia stuff I'm assuming?00:38
Greywhindllnz isn't back yet :-/01:25
*** llnz has joined #tp01:32
* llnz is back01:34
Greywhindhey llnz01:34
Greywhindstill there tansell-laptop ?01:34
tansell-laptopnot really01:35
Greywhindwell, anyway, llnz - do you have any ideas?01:40
llnzwell... i guess i have two01:42
llnzone is locks on files being download/written, the other is to avoid that by having a local media server01:42
Greywhindwhat would the local media server entail?01:43
llnzbrb, swapping room, wireless might drop out01:43
llnzwell, it would act like a proxy server01:44
Greywhindi'm not terribly familiar with proxy servers01:48
Greywhindhow do they work?01:48
llnzbasically, it will provide the media straight away if it's avaliable locally, otherwise it will download it in the background and provide it when complete01:51
Greywhindmm. k.01:51
llnzit might not be a literal proxy server, it might talk a slightly different protocol01:51
Greywhindso each client on startup would have to check if it was running, if not, start it?01:51
llnzwhich i guess would still require a lock file01:52
tansell-laptopI would recommend writing the file as01:52
llnztansell-laptop: normally the pid gets written into the file01:52
tansell-laptopno I mean write the file to that name01:52
tansell-laptopthen do a rename at the last minute01:52
Greywhindwell, i don't think this kind of thing is within the scope of what i could do now01:53
Greywhindmaybe i could leave it for someone else to write01:53
llnzthe other option is to have a lock file every media file, as needed01:54
Greywhindthe question then, though, is what the client does when it sees a lock file01:54
Greywhinddoes it return None, wait until the lock disappears, or what?01:55
llnzit should wait, i think01:57
Greywhindbut then, how does it wait without busy looping to check whether the file is there?01:58
llnzcheck periodically01:58
Greywhindcheck, sleep 1 second, check?01:59
Greywhindthat kind of thing?01:59
llnzcheck, set timer for 2-5 seconds time, return to event loop, ... timer expire, check...01:59
llnzwhatever is decided, it should be documented on the Media.gz page on the wiki02:04
Greywhindtansell-laptop: what do you think is the better or more reasonable option?02:05
llnza proxy could act like a client and use the lock files as well02:06
*** tansell-laptop has quit IRC02:16
*** tansell-laptop has joined #tp02:18
llnzafk, dinner02:32
*** tansell-laptop has quit IRC03:20
*** gau_veldt has joined #tp03:26
*** gau_veldt has left #tp03:26
*** nash has quit IRC03:27
*** tansell has joined #tp03:29
*** mithro_ has joined #tp03:40
*** Vadtec_ has quit IRC04:02
*** Vadtec has quit IRC04:02
*** edison has quit IRC04:02
*** Vadtec_ has joined #tp04:02
*** Vadtec has joined #tp04:02
*** edison has joined #tp04:02
*** Greywhind has quit IRC04:26
*** Demitar has quit IRC04:39
epyonllnz, houston I have a problem05:42
epyonI guessed sow05:42
epyonllnz: ping?05:43
llnzhi epyon05:43
epyonI have a problem with sequence number working05:43
epyonA sequence generates a modified list05:44
epyonbut that list is stored on server05:44
epyonSo if two players request modlists at the same time, the server will send them both Invalid Sequence number05:44
epyonin the second frame of the sequence05:45
epyonI assume that is a known bug?05:45
llnzno, not a known bug05:45
llnzalmost all of the seq num for id list frames are per-user05:46
epyonAaaaaaah, because that's a PlayerView!05:46
llnzthose that aren't are generally fixed and only change when the underlying list changes05:46
epyonso it's per-player :D05:46
*** mithro_ has quit IRC05:46
epyonGreat, that clarifies things05:47
epyonThanks :)05:47
llnzeach player has their own playerview05:47
llnzno problem05:47
*** mithro_ has joined #tp06:08
*** mithro_ has quit IRC06:27
*** mithro_ has joined #tp06:28
*** Demitar has joined #tp06:38
*** tuna-fish has joined #tp06:40
*** jmtan has joined #tp06:47
*** peres has joined #tp06:49
CIA-26epyon tpserver-cpp-refactor * rf0b1fb445bc5 /tpserver/ (playerview.cpp playerview.h): GenerateModList added to playerview EntityList06:54
CIA-26epyon tpserver-cpp-refactor * r7604b99574b4 /tpserver/ (persistence.cpp persistence.h): Persistence::saveProtocolObject added06:54
CIA-26epyon tpserver-cpp-refactor * rcaddb95f1b9f /tpserver/ (component.cpp component.h design.cpp design.h): Design and Component made ProtocolObjects06:55
CIA-26epyon tpserver-cpp-refactor * r3780d2ddfd43 /tpserver/ (persistence.cpp persistence.h):06:55
CIA-26Persistence ProtocolObject support updated06:55
CIA-26* added retrieveProtocolObject06:55
CIA-26* added getMaxProtocolObjectId06:55
CIA-26* added getProtocolObjectIds06:55
CIA-26epyon tpserver-cpp-refactor * rb0e59195f269 /tpserver/ (object.cpp object.h persistence.cpp):06:55
CIA-26Object as ProtocolObject06:55
CIA-26* In game Object is now a ProtocolObject06:55
CIA-26* Removed assignment operator06:55
CIA-26* added Persistence ProtocolObject support for Object06:55
mithro_Landon: ping?07:20
*** verhoevenv has joined #tp08:05
llnzepyon: mysqlpersistence, tpmzscheme and avahi are broken currently08:30
* llnz wanders off08:33
llnzlater all08:33
*** llnz has quit IRC08:33
*** alanp_ has joined #tp08:41
*** verhoevenv has quit IRC09:22
*** verhoevenv has joined #tp09:33
*** alanp_ is now known as alanp-09:44
*** alanp_ has joined #tp09:44
*** alanp has quit IRC09:59
totewhen i apply a order to the cache, how can i send it to the server? cache.update doesnt seem to work10:26
*** bddebian has joined #tp10:33
*** peres has quit IRC10:41
totenvm, solved it11:01
*** DTRemenak has quit IRC11:59
*** DTRemenak has joined #tp12:15
*** alanp- has quit IRC12:36
*** Greywhind has joined #tp13:05
*** Erroneous has joined #tp13:36
*** alanp has joined #tp14:27
Landonmithro_: pong15:04
*** alanp has quit IRC15:13
*** llnz has joined #tp16:04
*** alanp_ is now known as alanp16:38
*** verhoevenv has quit IRC16:57
*** verhoevenv has joined #tp17:06
*** bddebian has quit IRC17:18
llnzmorning all18:01
cherezMust... wake... up....18:55
*** kinglion has joined #tp19:04
*** kinglion has left #tp19:05
*** nash has joined #tp19:12
*** mithro_ has quit IRC19:37
*** mithro_ has joined #tp19:58
*** verhoevenv has quit IRC20:32
* llnz sighs20:40
cherezDo you need encouragement, llnz?20:45
llnzthankfully it's lunch time, so no20:46
tansellcherez, pong?20:56
tansellLandon, pong?20:56
tansellLandon, so have you had a chance to catch up with jmtan?20:56
Landonnot recently, no20:57
chereztansell: I... don't remember pinging you?20:57
cherezAlmost have my blog post up.20:57
tansellcherez, well I guess it should be a ping :)20:57
cherezMy code is worthless today. :(20:58
tansellcherez, oh?21:12
cherezI think I'm sick; I've had periods staring at code for half an hour not sure how to continue.21:13
tansellcherez, sometimes it's just an idea to jump on IRC and chat with people about it21:13
cherezThat is probably a good idea.21:14
tansellcherez, so it seems that you have been doing pretty will with the daily blog posts21:25
tansellso where are you at?21:26
*** Erroneous has quit IRC21:26
Greywhindtansell: i've updated issue 9001 and 10002 on codereview21:27
tanselllink me?21:27
tpbTitle: Issue 9001: Improved media support by centralizing media calls and adding a download manager. - Code Review (at codereview.mithis.com)21:28
Greywhindand the same with 10002 on it instead of 900121:28
tansellcherez, poke?21:30
tansellcherez, so21:31
tansellwhere are you at?21:31
cherezSorry, keyboard went out on me.21:31
tanselldo you have a recently coverage report?21:31
tansellif you recall, I asked for you to start uploading coverage reports regularly21:32
tansellpreferably in a way which means that I can compare the history21:33
chereztansell: I'm uploading.21:33
cherezAlright, I'll put together a script to put them up with timestamps.21:34
tansellthat would be great21:34
tansellit might be worth to go back and do it for older commits too21:34
tansellgreat way to show how you have progressed over GSoC21:35
chereztansell: I'm a little lost on how to unit test the SSL wrapper.21:35
tansellcherez, oh?21:35
tansellcherez, might be worth doing something closer to an integration test21:36
cherezIt supports I think 3 different SSL libraries as fallbacks, and I'm not too sure how I could set up a test for falling back correctly.21:36
tansellIE start up a test https server21:36
tansells/https/ssl server/21:36
cherezI'm also a bit uncertain about testing the socket wrapper. I just always feel that I might be overlooking some special case where they'll stop working.21:38
tansellwell upload it to the code review and go "mithro can you see anything I'm missing" :)21:39
tanselland if I say no - then its fine right? :)21:39
tansellcherez, oh something else I wanted to mention21:40
cherezI guess.21:40
tanselldoes your parser validate against the dtd?21:40
cherezNo; I'm not really sure how to do that.21:40
cherezShould I be looking it up?21:40
tansellnormally it's part of the xml parsing stuff21:41
tansellI would put it at a low priority thanks to your other testing21:41
*** bddebian has joined #tp21:41
cherezAs long as I can't competently write any complex code today, that might be a good use of my time.21:42
tansellcherez, so your currently using figleaf to generate the coverage report right?21:43
cherezYeah, it's proven more reliable than coverage.21:43
tansellbut your test.sh in libtpproto2-py still uses coverage21:44
cherezOh, I haven't uploaded the  test.sh I'm using for it.21:44
cherezI'll put that in my next CL.21:44
tansellcherez, I would recommend putting in a separate CL21:45
tansellso are we on-target to reach our goal?21:45
cherezI believe so.21:47
tansellGreywhind, do you know about mixin classes?21:47
Greywhindtansell: not really.21:47
tansellGreywhind, as well you didn't look into the for/else loop in python21:47
tansellGreywhind, they let you put common functionality into a class which you then inherit from21:48
Greywhindtansell: is there a difference from just a base class?21:49
tansellexcept mixin's are limited to a very specific set to functionality21:49
tansellcherez, what is there still to do and how long have you set for each part?21:50
tansellGreywhind, the waiting stuff is a classic candidate for a mixin class21:50
Greywhindtansell: well, it's a bit different for each one though21:50
tansellGreywhind, is it?21:50
chereztansell: Well, there's Client, Server, and integrating the library with libtpclient-py.21:51
cherezClient is pretty much written, so I expect it will be pretty methodical.21:51
Greywhindtansell: because some of them want everything for an object, some want everything for multiple objects, the info panel wants to track by URL21:52
cherezI'm expecting that to only be a few days.21:52
tansellcherez, so you have common done then?21:52
tansellGreywhind, why does the info panel want to track by URL?21:53
Greywhindtansell: because it takes images by URL rather than by object21:53
cherezFunctionally I just have a few bits to finish up for Common. It would probably be done if I I hadn't lost half my IQ the last few days.21:54
tansellGreywhind, that sounds wrong - everything should be looking at the media the same way21:55
Greywhindtansell: but it creates separate panels for each URL, so it separates it out into URLs first.21:55
tansellGreywhind, splitting on URL doesn't really make sense21:56
Greywhinddef AddPanelForParam(self, group, parent):21:57
Greywhindelif isinstance(group, parameters.ObjectParamMedia):21:57
Greywhindit has to do it by URL, because it's creating panels by parameter21:57
Greywhindotherwise, it would need a separate construction to add the media panels21:58
Greywhindwhich is unnecessary21:58
Greywhindit still uses the same media code in threads.py21:58
tansellwhy is panelPicture still doing averaging?21:59
Greywhindtansell: it has a separate progress bar from the download window's progress bar. i should probably remove it now22:00
tansellGreywhind, yeah I think so22:00
tansellGreywhind, but the waiting on image always waits on a URL right?22:01
Greywhindtansell: which waiting on image?22:02
Greywhindtansell: oh, you mean the list of downloading images?22:02
Greywhindtansell: yeah, it always contains URLs, just gets them from different places22:02
tansellGreywhind, so I think it still belongs in a mixin class22:03
Greywhindtansell: i'll look into it22:03
*** mithro_ has quit IRC22:03
tansellcomments added too 900122:03
Greywhindi'll be back22:03
tansellcherez, I recommend uploading the CL now22:04
tansellcherez, upload early and often :)22:04
*** mithro_ has joined #tp22:04
shenkisome projects who have inferoir version control use rietveld for snapshotting their changes22:29
*** jmtan has quit IRC22:43
Landonho-boy, eyerest time oO23:13
*** jmtan has joined #tp23:23
tansellLandon, you still here?23:24
Greywhindtansell: http://codereview.mithis.com/8003/show23:34
tpbTitle: Issue 8003: Made a popup for all SystemLevelOverlays containing panelPicture. - Code Review (at codereview.mithis.com)23:34
tansellGreywhind, is this the image popup?23:34
Greywhinddecided to fix them up real quick before going back to media stuff23:35
tansellgot a new screenshot?23:40
Greywhindtansell: well, it's just the same with the panel either in the lower right or upper left corner23:42
Greywhindbut i'll take two23:42
tansellGreywhind, oh?23:42
tpb<http://ln-s.net/3ipD> (at img169.imageshack.us)23:44
tpb<http://ln-s.net/3ipJ> (at img259.imageshack.us)23:45
tansellGreywhind, apparently jeffery signed on :)23:48
Greywhindtansell: heh23:48
tansellGreywhind, that looks okay23:49
Greywhindtansell: k23:50
tansellGreywhind, have you thought about adding tests for the media stuff?23:57

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!