[Bf-committers] physics concerns

[Michael Fox]

on this subject of physics a great Idea came up on IRC about this so
instead of explaining here is the log of the convo

[14:40:34] <joeedh> someone just needs to appropriately hook the cache
system into the DAG :/
[14:40:59] <mfoxdogg> yeah
[14:41:38] * Aligorith (n=chatzill at 219-88-199-206.jetstream.xtra.co.nz)
has joined #blendercoders
[14:42:43] * joeedh gives mfoxdogg a significant look.
[14:42:48] <joeedh> you know. . .I bet Aligorith could do it. . . :)
[14:43:35] <broken_work> it's more than just the depgraph though - it's
also things like tweaking settings in deflectors etc, not changing
geometry or object locations
[14:43:53] <mfoxdogg> joeedh, you want me to code it 0_o
[14:43:56] <broken_work> and you wouldn't want to calculate the whole
depgraph just by changing collision friction or something
[14:44:07] <broken_work> because of *
[14:44:20] <joeedh> mfoxdogg: no it's just Aligorith joined right after
we agreed someone needed to do it :)
[14:44:33] <mfoxdogg> hehehe
[14:45:13] <mfoxdogg> broken_work, im thinking of not rebuilding the
depsgraph just recieve signals from the depsgraph
[14:45:15] <joeedh> broken_work: well I was thinking making the particle
cache work better.
[14:45:27] <broken_work> as it should :)
[14:45:30] <Aligorith> I've got enough bugs and small features I want to
finish already today :)
[14:45:31] <joeedh> broken_work: the depsgraph relations would be
different then SB
[14:45:37] <joeedh> Aligorith: :) ok
[14:45:56] <broken_work> mfoxdogg: sure, but tweaking physics settings
generally dont' cause a depgraph update
[14:46:00] <Aligorith> depsgraph wrangling can be done another day...
[14:46:34] <mfoxdogg> broken_work, it should, or aleast the physics
depsgraph
[14:47:25] <mfoxdogg> now thats an idea, have a seperate depsgraph for
physics systems
[14:47:38] <joeedh> broken_work: well editing particle settings should
update the cache.  as should moving the object.  also moving effectors,
and deflector objects whose bounding box intersects the bounding box of
all the particles combined.
[14:48:05] <broken_work> joeedh: yes, all of those things for sure
[14:48:18] <broken_work> Im just saying it's more than making it respond
to the depgraph
[14:48:35] <joeedh> well duh
[14:48:57] <broken_work> :P
[14:49:28] <joeedh> you could also have SB detect which objects it hits,
and if you change them it clears the cache.
[14:49:35] * joeedh isn't sure this would be good, though
[14:49:40] * nicholasbishop (n=nicholas at nicholasbishop.student.umd.edu)
has joined #blendercoders
[14:50:40] <joeedh> hi nicholasbishop 
[14:51:11] <nicholasbishop> hi joeedh
[14:51:23] * venomgfx (i=venomgfx at 190.2.37.237) has joined
#blendercoders
[14:51:35] <mfoxdogg> joeedh, that could be done with a physics
depsgraph, along with and option to say delect this system but not this
and so on
[14:52:02] * leafw (n=Any at mbi136-221.mbi.ucla.edu) has joined
#blendercoders
[14:52:13] <joeedh> a physics subgraph would be interesting. . .
[14:52:33] * MrNightmare (n=sevalso at 70.44.209.159.res-cmts.wbo.ptd.net)
has joined #blendercoders
[14:52:41] <joeedh> that would allow specific signals just for physics.
[14:52:49] <joeedh> without hijacking the existing system
[14:53:12] <mfoxdogg> yeah
-- 
Michael Fox <mfoxdogg at gmail.com>