Setback (possible) Bug

678
3
11-26-2011 06:47 AM
AndréCardoso
New Contributor III
I don't know if I should post these findings here... please correct me, or move this to a more appropriate section...

While moving a project to the 2011 version, I had multiple CE crashes while trying to generate geometry. The crash was related to a 'setback' operation. I had something like this:
HouseSetBack -->
 setback(gardenSetback){
  remainder : innerRect House |
  all: NIL
 }


The mere opening of the CGA rule file containing this, would crash CE, because of the default preview bar of the rule (I reckon it evaluates the code, thus crashing)....

If started through a terminal, CE outputs the following:
#
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x00007f6112ff80a3, pid=13790, tid=140054875539200
#
# JRE version: 6.0_26-b03
# Java VM: Java HotSpot(TM) 64-Bit Server VM (20.1-b02 mixed mode linux-amd64 compressed oops)
# Problematic frame:
# C  [libcom_procedural_grammarcore.so+0x29f0a3]  SetbackOperation::FaceSetback::extractSetbackFaces(std::vector<unsigned, std::allocator<unsigned> > const&, Mesh const*)+0x671


How to solve? Well, I merely changed the code to the following, reversing the selector's order:
HouseSetBack -->
 setback(gardenSetback){
                all: NIL |
  remainder : innerRect House
 }


Which, probably, is the way it is meant to be used... I'm a little confuse now... anyway I think it's not supposed to crash CE !

Can anyone confirm this?

Using Ubuntu Linux 64 bits, CityEngine 2011 64 bit.

Hope this helps,
André
Tags (2)
0 Kudos
3 Replies
MatthiasBuehler1
Frequent Contributor II
hi Andre.

can you check if the issue is that the scope is not yUp ?

before the setback operation, try alignScopeToAxes(y).

is it this ? please let me know.

matt
0 Kudos
MatthiasBuehler1
Frequent Contributor II
by the way, setback and the shapeLOU operations need a yUp scope.

as you certainly know, after a comp(f), the scope is zUp.
0 Kudos
AndréCardoso
New Contributor III
I've just tried what you said. But no luck. Also, the setback was one of the first rules on the footprint, so the scope must've been yUp already (correct me please) !

The test done was (now with the first rule of the footprint):
Lot -->
 90%:
  HouseSetBack
  Garden
 else:
  innerRect
  House

HouseSetBack -->
 alignScopeToAxes(y)
 setback(gardenSetback){
  remainder: innerRect House |
  all: NIL
 }


And again on the terminal:
# JRE version: 6.0_26-b03
# Java VM: Java HotSpot(TM) 64-Bit Server VM (20.1-b02 mixed mode linux-amd64 compressed oops)
# Problematic frame:
# C  [libcom_procedural_grammarcore.so+0x29f0a3]  SetbackOperation::FaceSetback::extractSetbackFaces(std::vector<unsigned, std::allocator<unsigned> > const&, Mesh const*)+0x671



Don't know if it helps (did not open it), but in the attachments follows the error log generated.

André


hi Andre.

can you check if the issue is that the scope is not yUp ?

before the setback operation, try alignScopeToAxes(y).

is it this ? please let me know.

matt
0 Kudos