[Bf-committers] Interesting cycle

[Ton Roosendaal]

Hi,

This bugreport had a forcefield, vertexparented to a softbody.
Caused a very interesting eternal loop, but only when the forcefield  
had a frame offset. For those who like to study Blender internals, a  
nice cycle :)

Daniel; the fix I made was to prevent the give_parvert() to build a  
mesh. Actually, no code in Blender should create mesh data, apart from  
depsgraph... not sure how you intented this?

-Ton-


#113 0x002ae998 in where_is_object_time (ob=0x0, ctime=1) at  
object.c:1348
#114 0x002c1ae0 in pdDoEffectors (lb=0x3000000, opco=0x0,  
force=0x2bddacc, speed=0x0, cur_time=2, loc_time=1000, flags=0) at  
effect.c:472
#115 0x002d1600 in softbody_calc_forces (ob=0x3162a1c, forcetime=1) at  
softbody.c:590
#116 0x002d316c in sbObjectStep (ob=0x0, framenr=2,  
vertexCos=0x3162a1c, numVerts=3) at softbody.c:1363
#117 0x002bc5b0 in softbodyModifier_deformVerts (md=0x3000000, ob=0x54,  
derivedData=0x1, vertexCos=0x3, numVerts=3) at modifier.c:1163
#118 0x002aa400 in mesh_calc_modifiers (ob=0x0, inputVertexCos=0x0,  
deform_r=0xbf8050a0, final_r=0x3162d0c, useRenderParams=45996748,  
useDeform=0) at DerivedMesh.c:1490
#119 0x002aafb0 in mesh_build_data (ob=0x0) at DerivedMesh.c:1833
#120 0x002ab23c in mesh_get_derived_deform (ob=0x3162a1c,  
needsFree_r=0xbf805850) at DerivedMesh.c:1898
#121 0x002ae4b0 in give_parvert (par=0x3000000, nr=0, vec=0xbf8058e0)  
at object.c:1189
#122 0x002ae698 in ob_parvert3 (ob=0x0, par=0x3162a1c, mat=0xbf8058e0)  
at object.c:1247
#123 0x002aeca4 in solve_parenting (ob=0x0, par=0x0, slowmat=0x3162a1c,  
simul=-1082107664) at object.c:1437
#124 0x002ae998 in where_is_object_time (ob=0x0, ctime=1) at  
object.c:1348



------------------------------------------------------------------------ 
--
Ton Roosendaal  Blender Foundation ton at blender.org  
http://www.blender.org