[Insight-developers] ITK 4.4 release candidate cycles about to begin

[Sean McBride]

On Wed, 8 May 2013 11:58:57 -0400, Sean McBride said:

>But presumably the people that care enough to submit dashboards would be
>willing to help out.  My dashboards have a fair bit of red, and although
>I don't have a lot of time for ITK, I can at least provide backtraces of
>crashing tests, etc.

For example, itkQuaternionRigidTransformTest seems to crash only on Rogue7:

Here's a quick gdb session:

Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: 13 at address: 0x0000000000000000
0x00000001008c52f0 in vnl_sse<double>::matrix_x_vector (m=0x1022bb5c8, v=0x7fff5fbff158, r=0x1022bb7d0, rows=3, cols=3) at vnl_sse.h:482
482	        y = VNL_SSE_HEAP_LOAD(pd)(v+j);

(gdb) bt
#0  0x00000001008c52f0 in vnl_sse<double>::matrix_x_vector (m=0x1022bb5c8, v=0x7fff5fbff158, r=0x1022bb7d0, rows=3, cols=3) at vnl_sse.h:482
#1  0x00000001008c519e in vnl_vector<double>::vnl_vector (this=0x7fff5fbfd448, M=@0x7fff5fbfd398, v=@0x7fff5fbfd438) at vnl_vector.txx:246
#2  0x00000001008c5053 in vnl_vector<double>::vnl_vector (this=0x7fff5fbfd448, M=@0x7fff5fbfd398, v=@0x7fff5fbfd438) at vnl_vector.h:247
#3  0x00000001001b17e2 in operator*<double> (m=@0x7fff5fbfd398, v=@0x7fff5fbfd438) at vnl_vector.h:503
#4  0x00000001001c4efe in operator*<double, 3u, 3u> (a=@0x1022bb5c8, b=@0x7fff5fbfd438) at vnl_matrix_fixed.h:1013
#5  0x00000001001c4e22 in itk::Matrix<double, 3u, 3u>::operator* (this=0x1022bb5c8, vc=@0x7fff5fbfd438) at itkMatrix.hxx:220
#6  0x00000001001bf38c in itk::MatrixOffsetTransformBase<double, 3u, 3u>::TransformVector (this=0x1022bb4c0, vect=@0x7fff5fbff158) at itkMatrixOffsetTransformBase.hxx:209
#7  0x0000000100258c25 in itkQuaternionRigidTransformTest () at /Users/builder/external/ITK/Modules/Core/Transform/test/itkQuaternionRigidTransformTest.cxx:172
#8  0x0000000100008cdc in main (ac=1, av=0x7fff5fbffa48) at /Users/builder/external/ITK-clang-dbg-x86_64-static/Modules/Core/Transform/test/ITKTransformTestDriver.cxx:332
(gdb) p pd
No symbol "pd" in current context.
(gdb) p v
$1 = (const double *) 0x7fff5fbff158
(gdb) p j
$2 = 0

Maybe Rogue7 is one of the few dashboards with VNL_CONFIG_ENABLE_SSE2 ON?

Cheers,

-- 
____________________________________________________________
Sean McBride, B. Eng                 sean at rogue-research.com
Rogue Research                        www.rogue-research.com 
Mac Software Developer              Montréal, Québec, Canada