PANIC in `x_tv.e’: Memory has been corrupted
Stefano Zibetti wrote on Jan 19, 2006
Dear all,
I have the following problem ovarlaying a mask on a displayed image.
When I run a command like this:
display (image=image.fits[0], frame=1, bpmask=, bpdisplay=none, overlay=tmp$mask.pl, ocolors=green, zscale=no, zrange=no, z1=475., z2=712.)
the image displays correctly and the mask is also overlayed, but the process hangs and issues the error message:
PANIC in `[path]/x_tv.e': Memory has been corrupted.
The cl must be killed then.
Note that the image is a multi-extension fits image that correctly displays without any problem if no mask overlay is requested.
I am running IRAF V12.2.2 on a customized debian Linux distribution for pc, with kernel 2.6.8-2-686. ds9 is version 3.0.3
Any hint?
Thanks!
Stefano
I have the following problem ovarlaying a mask on a displayed image.
When I run a command like this:
display (image=image.fits[0], frame=1, bpmask=, bpdisplay=none, overlay=tmp$mask.pl, ocolors=green, zscale=no, zrange=no, z1=475., z2=712.)
the image displays correctly and the mask is also overlayed, but the process hangs and issues the error message:
PANIC in `[path]/x_tv.e': Memory has been corrupted.
The cl must be killed then.
Note that the image is a multi-extension fits image that correctly displays without any problem if no mask overlay is requested.
I am running IRAF V12.2.2 on a customized debian Linux distribution for pc, with kernel 2.6.8-2-686. ds9 is version 3.0.3
Any hint?
Thanks!
Stefano
Mike Fitzpatrick wrote on Jan 19, 2006
This has come up before in cases where the mask is not the same dimension as the image. While the pixel dimensions may be the same, magnification or the extraction of an image section can affect the logical dimension of the image.
If that's not it, try explicitly saying
bpmask="" instead of just bpmask=
especially if you have a BPM keyword in the header. Since the task is failing at the end (i.e. when pointers are being freed) it likely means memory was corrupted/overrun earlier someplace. If you've got ElectricFence installed you can use that to better trap the error at the point of the corruption.
-Mike
If that's not it, try explicitly saying
bpmask="" instead of just bpmask=
especially if you have a BPM keyword in the header. Since the task is failing at the end (i.e. when pointers are being freed) it likely means memory was corrupted/overrun earlier someplace. If you've got ElectricFence installed you can use that to better trap the error at the point of the corruption.
-Mike
Last post on Jan 19, 2006