[gnuastro-devel] [bug #49588] Segfault in mesh interpolation

[Mohammad Akhlaghi]

Update of bug #49588 (project gnuastro):

                  Status:             In Progress => Fixed                  
             Open/Closed:                    Open => Closed                 

    _______________________________________________________

Follow-up Comment #1:

The issue was fixed and has been pushed
<http://git.savannah.gnu.org/cgit/gnuastro.git/commit/?id=0c69f027f53> to the
main repo.

Just to complete the explanation, the sample image was 30x30 pixels wide while
the small mesh size was 32x32 pixels (the default value). So ultimately, the
mesh library (through NoiseChisel in this case) will abort, but there will be
an explanation to explain why it didn't complete.


astnoisechisel img.fits
NoiseChisel started on Fri Nov 11 21:30:56 2016
  - Using 8 CPU threads.
  - Input: test.fits (hdu: 1)
  - Kernel: FWHM=2 pixel Gaussian.
  - Convolved with kernel.                        0.000470 seconds
  - Starting to find initial detections.         
astnoisechisel: Interpolating quantile threshold: only 0 mesh(s) are filled
for interpolation in channel 0. Either set less restrictive requirements to
get more interpolation points or decrease the number of nearest points to use
for interpolation. Problem encountered on thread 0, for pixel 0. When running
on multiple threads, This message might be repeated for different threads


In such cases, it might be useful to allow no interpolation when the mesh and
image size are roughly similar (only one mesh) and the mesh could be given a
value (task #14218).

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?49588>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/