[00:04:23] <thebolt> hm, sorry for becomming so quiet.. i am designing a mechanical project and got a bit caught up.. anyhow, i will try to look into how well boost::python+stackless+native threads work..
[00:06:14] <PK> no problem (about the becomming quiet).... anyway, good night
[00:06:22] PK has quit:
[00:08:58] sueastside has quit: "Peragro Tempus: Now in color!"
[00:59:53] thebolt changed nick to: thebolt|zzz
[02:34:42] iceeey_ changed nick to: iceeey
[02:34:45] ChanServ sets mode: +o iceeey
[07:08:24] |sueasts| has joined #peragro-dev
[07:40:18] |sueasts| changed nick to: sueastside
[07:40:37] ChanServ sets mode: +o sueastside
[08:24:28] thebolt|zzz changed nick to: thebolt
[09:37:56] sueastside has quit: Read error: 104 (Connection reset by peer)
[09:38:13] sueastside has joined #peragro-dev
[09:38:14] ChanServ sets mode: +o sueastside
[09:48:49] PK has joined #peragro-dev
[09:48:49] ChanServ sets mode: +o PK
[12:12:47] thebolt|away changed nick to: thebolt
[13:28:28] PK has quit: Remote closed the connection
[13:54:09] caedes has joined #peragro-dev
[13:54:09] ChanServ sets mode: +v caedes
[15:00:48] iceeey has quit: Read error: 110 (Connection timed out)
[16:50:02] Lightwave has quit: Read error: 110 (Connection timed out)
[17:13:03] Lightwave has joined #Peragro-Dev
[17:13:03] ChanServ sets mode: +v Lightwave
[17:25:18] iceeey has joined #peragro-dev
[17:25:19] ChanServ sets mode: +o iceeey
[17:26:40] <iceeey> hi
[18:08:53] iceeey has quit: Read error: 110 (Connection timed out)
[19:10:23] <sueastside> hi thebolt
[20:32:41] caedes has quit: "nightmare reset by peer"
[20:42:43] PK has joined #peragro-dev
[20:42:43] ChanServ sets mode: +o PK
[20:54:12] |sueasts| has joined #peragro-dev
[21:00:55] sueastside has quit: Nick collision from services.
[21:01:27] ChanServ sets mode: +o sueastside
[21:12:13] iceeey has joined #peragro-dev
[21:12:13] ChanServ sets mode: +o iceeey
[21:25:47] <thebolt> iceeey: hm, i'm looking at the stackless source to find out how one can do scheduling etc
[21:25:54] <thebolt> some conclusions of todays research
[21:26:01] <thebolt> 1. Stackless works with native c++ threads
[21:26:27] <thebolt> 2. you need to do a c++->python->c++ round to get the base (main) stack frame setup before you can start scheduling stuff
[21:26:53] <thebolt> (ie adding new tasklets for example)
[21:28:45] <thebolt> however, once created it isn't possible to have a tasklet migrate over threads
[21:29:46] <sueastside> not implicitly, but i think you can save their data, migrate, and start it again...
[21:29:59] <iceeey> smart
[21:30:27] <sueastside> i know i am, but what are you?
[21:30:31] <thebolt> sueastside: yes
[21:30:34] <iceeey> smarter..
[21:30:40] <sueastside> i know i am, but what are you?
[21:30:44] <sueastside> :P
[21:30:45] <thebolt> however, i think it is smarter to try to get the initial scheduling somewhat evenly
[21:31:15] <thebolt> (for example by keeping track of approximatly how many "long lived" tasklets you have per thread..)
[21:32:19] <iceeey> okay
[21:33:51] <iceeey> what kind of scheduling functions do you think we will provide?
[21:34:27] <thebolt> well, more or less just one.. "RunAsTasklet(...)" ;)
[21:34:42] <iceeey> well no control things like sleep()?
[21:34:51] <thebolt> within the tasklet code, of course
[21:34:55] <iceeey> yes
[21:34:58] <thebolt> a sleep/yield is needed :-)
[21:35:15] <thebolt> one question is, say every behaviour and every action is their own tasklets
[21:35:34] <thebolt> what if a behaviour call an action.. should it be run "in-line" or spawn a new one?
[21:36:51] <iceeey> doesn't call just execute c++ code to send a message over network?
[21:37:35] <thebolt> well, a server side behaviour calling an action results (probably) in a server side action, no?
[21:37:58] <iceeey> then it would be a new thread
[21:38:03] <iceeey> tasklet
[21:39:14] <iceeey> actions consist of validation and process methods
[21:39:44] <thebolt> hm, okay?
[21:40:00] <iceeey> so, you're going to have two methods being called
[21:40:45] <iceeey> i would say, the validation is inline, and process is a tasklet
[21:40:51] <thebolt> hm, then i would say call the verification inline then process in tasklet
[21:40:55] <iceeey> :)
[21:53:56] sueastside has quit: "A lapdance is soo much better when the stripper is your own wife."
[22:08:47] iceeey has quit: Remote closed the connection
[22:11:19] iceeey has joined #peragro-dev
[22:11:20] ChanServ sets mode: +o iceeey
[22:14:06] thebolt2 has joined #peragro-dev
[22:14:21] thebolt2 changed nick to: thebolt
[22:14:31] ChanServ sets mode: +o thebolt
[23:14:41] PK has quit: