Log In
New Account
Home My Page Projects Blender 2.x BF release
Summary Activity Tracker SCM Files

Blender 2.6 Bug Tracker: Browse

[#25797] Hair partcle crash when vertexgroup length enabled

Date:
2011-01-25 19:50
Priority:
3
State:
Closed
Submitted by:
Miika Hamalainen (miikah)
Assigned to:
Janne Karhu (jhk)
Category:
Physics
Status:
Fixed / Closed
Relates to:
Duplicates:
Patches:
 
Summary:
Hair partcle crash when vertexgroup length enabled
Detailed description
If you set hair particle length by a vertexgroup, it crashes randomly when painting weights.

Attachments:

hair_crash.blend
* Start painting those weights. For me it usually crashes within 10 seconds, sometimes takes longer.

corrupted_hair.png
* Just before it crashes, some hairs points get corrupted, and will be drawn as random lines to infinity.


Tested on SVN rev 34488, using both 32 and 64-bit Windows builds.


Crash backtrace:
Program received signal SIGSEGV, Segmentation fault.
0x008cbd29 in do_particle_interpolation (psys=0xf4c7b10, p=306, pa=0xf4d5400,
t=100.000015, frs_sec=24, pind=0x28fb20, result=0x28fb70)
at source/blender/blenkernel/intern/particle.c:1266
1266 while(pind->hkey[1]->time < real_t) {

#0 0x008cbd29 in do_particle_interpolation (psys=0xf4c7b10, p=306,
pa=0xf4d5400, t=100.000015, frs_sec=24, pind=0x28fb20, result=0x28fb70)
at source/blender/blenkernel/intern/particle.c:1266
#1 0x008d22da in psys_cache_paths (sim=0x28fcb0, cfra=1)
at source/blender/blenkernel/intern/particle.c:2956
#2 0x008efd25 in psys_update_path_cache (sim=0x28fcb0, cfra=1)
at source/blender/blenkernel/intern/particle_system.c:3189
#3 0x008f0450 in hair_step (sim=0x28fcb0, cfra=1)
at source/blender/blenkernel/intern/particle_system.c:3346
#4 0x008f2c31 in particle_system_update (scene=0xf4c5908, ob=0xf4c7688,
psys=0xf4c7b10) at source/blender/blenkernel/intern/particle_system.c:4129
#5 0x009b2d43 in deformVerts (md=0xa586b60, ob=0xf4c7688, derivedData=0x0,
vertexCos=0xa654a30, UNUSED_numVerts=289, UNUSED_useRenderParams=0,
UNUSED_isFinalCalc=1)
at source/blender/modifiers/intern/MOD_particlesystem.c:195
#6 0x008dd923 in mesh_calc_modifiers (scene=0xf4c5908, ob=0xf4c7688,
inputVertexCos=0x0, deform_r=0xf4c7a38, final_r=0xf4c7a3c,
useRenderParams=0, useDeform=1, needMapping=1, dataMask=1048601, index=-1,
useCache=1) at source/blender/blenkernel/intern/DerivedMesh.c:1720
#7 0x008decbf in mesh_build_data (scene=0xf4c5908, ob=0xf4c7688,
dataMask=1048601) at source/blender/blenkernel/intern/DerivedMesh.c:2245
#8 0x008dee47 in makeDerivedMesh (scene=0xf4c5908, ob=0xf4c7688, em=0x0,
dataMask=1048601) at source/blender/blenkernel/intern/DerivedMesh.c:2286
#9 0x0083838d in object_handle_update (scene=0xf4c5908, ob=0xf4c7688)
at source/blender/blenkernel/intern/object.c:2568
#10 0x00831e00 in scene_update_tagged_recursive (bmain=0xf494f08,
scene=0xf4c5908, scene_parent=0xf4c5908)
at source/blender/blenkernel/intern/scene.c:960
#11 0x00831e90 in scene_update_tagged (bmain=0xf494f08, scene=0xf4c5908)
at source/blender/blenkernel/intern/scene.c:980
#12 0x0040a1b4 in wm_event_do_notifiers (C=0x57b3b98)
at source/blender/windowmanager/intern/wm_event_system.c:305
#13 0x00403f5a in WM_main (C=0x57b3b98)
at source/blender/windowmanager/intern/wm.c:345
#14 0x00403796 in main (argc=1, argv=0x57b3480)
at source/creator/creator.c:1238

Followup

Message
  • Date: 2011-01-25 19:52
  • Sender: Janne Karhu
  • Assigned to self for investigation.
  • Date: 2011-01-26 01:54
  • Sender: Gabriel
  • no problems in ubuntu 10.10 64b r34490 and windows 7 64b blender 64b r34390
  • Date: 2011-01-26 07:41
  • Sender: Miika Hamalainen
  • I just noticed this is much easier to reproduce on a debug build. For a regular build you should increase hair count to +50.000, then it usually crashes within a minute. :p
  • Date: 2011-01-26 10:25
  • Sender: Gabriel
  • Nothing in 10.10 64b r34490 with 60000 particles and 10 minutes of play.

    I was thinking ... could be a hw/sf problem? Which 3d card , memory, cpu you have ? And, you could test memory with memtest or another utility ?
  • Date: 2011-01-26 13:22
  • Sender: Miika Hamalainen
  • @Gabriel

    I don't think it's a hardware problem, everything else works fine. And the crash backtrace always leads to that same code line, so I think jahka's new hair interpolation simply has a small bug.
  • Date: 2011-01-26 14:03
  • Sender: Janne Karhu
  • Fix committed, thanks for the report!
  • Date: 2011-01-26 15:20
  • Sender: Gabriel
  • Good work miika ,it dont fail in my comp until I tested it in windows and then after much time of testing (six minutes) at the end fail. This was a hard one!

    Thanks jahka!
 

Attached Files:

Name Date Download
hair_crash.blend 2011-01-25 19:50 Download
corrupted_hair.png 2011-01-25 19:50 Download

Changes:

Field Old Value Date By
status_idOpen2011-01-26 14:03jhk
close_dateNone2011-01-26 14:03jhk
StatusInvestigate2011-01-26 14:03jhk
assigned_tonone2011-01-25 19:52jhk
StatusNew2011-01-25 19:52jhk
summaryHair partcle crash when when using vertexgroup length2011-01-25 19:51miikah
File Added14709: corrupted_hair.png2011-01-25 19:50miikah
File Added14710: hair_crash.blend2011-01-25 19:50miikah