iraf.net

I have been unable to use fixpix to correct an image file. Below is some history of the problem. I have already checked all of the situations listed in the text below. The actual error message is at the bottom. How can I resolve this problem? Or how do I begin to resolve the problem? (I am also having a similar problem with epix.) I don't think it is an installation problem, but it could turn out to be so.
-Amber Sierra
University of Arkansas at Little Rock
adsierra@ualr.edu
________________________________________________
Found this while googling on iraf fixpix ERROR: segmentation violationPlease let me know if this applies to this case.http://www.physics.hmc.edu/Astronomy/Ireduce.html"""
A Note on Errors
"ERROR:segmentation violation" is an all-purpose error. It is possible
that you have accidentally entered a filename that doesn't exist.
Check your path, directories, and spelling to be sure that you have
entered a valid filename.Another cause for this error may be due to Unix's case-sensitive
nature. If the images you are trying to combine are titled *.FIT, you
will have to change them to *.fit. IRAF will not process them otherwise.If you are using lists, check what is inside. If there is an object
which is not an image file, there is your problem. (Be careful when
using * to select images for a list. You might accidentally include
something you did not intend to.)If you get, "* is not an image or a number", you probably have not
entered a valid filename. Check to be sure you did not mispell or
point to an incorrect directory.If you get a "no images to combine" error, it may be due to the
IMAGETYP parameter in the fits files. If you are trying to run
zerocombine or ccdproc, check that the IMAGETYP is set correctly or
set ccdtype equal to "none" in the parameters for zerocombine.This error will also occur when you have not set the default image
type to "fits,noinherit". Go to the cl prompt and enter show imtype to
be sure that you have fits files set as default.If you get, "Warning: Cannot open image", you have entered an
incorrect filename or path. Check for spelling and correct path."""On Feb 24, 2010, at 10:10 AM, Marcus Seigar wrote:> Hi Albert,
>
> We seem to have an IRAF problem on ganymede. We're trying to run
> the task 'fixpix' and we've been getting the following error:
>
> --> fixpix
> List of images to be fixed ('irx_00026_c3_001.fits'):
> List of bad pixel masks ('mask.fits'):
> Killing IRAF task `fixpix'
> Traceback (innermost last):
> File "<CL script CL1>", line 1, in <module>
> iraf.fixpix(_save=1)
> IrafError: Error running IRAF task fixpix
> IRAF task terminated abnormally
> ERROR (501, "segmentation violation")
>
> Looks like a segmentation violation. We were running this is
> PyRAF. If we just use the 'old fashioned' version of IRAF to run
> 'fixpix' we get:
>
> ecl> fixpix
> List of images to be fixed (irx_00026_c3_001.fits):
> List of bad pixel masks (mask.fits):
> ERROR: segmentation violation
>
> Can you help? Hopefully we can get this fixed soon. Amber would
> like to continue with her data reduction. Ok... I would like Amber
> to continue as well.
>
> Cheers,
> Marc
>
>
> ---
> Dr. Marc S. Seigar (mxseigar@ualr.edu)
> Assistant Professor of Astrophysics
> University of Arkansas at Little Rock
> Department of Physics & Astronomy
> 2801 S. University Avenue
> Little Rock, AR 72204
> Tel: (501) 569-8964
> Fax: (501) 569-3314
> http://www.ualr.edu/mxseigar

By the way I am running the following version of IRAF on RHEL5.NOAO/IRAFNET PC-IRAF Revision 2.14 Fri Nov 30 15:27:05 MST 2007
This is the RELEASED version of IRAF V2.14 supporting PC systems.-Amber Sierra

A segfault usually implies a coding bug of some kind, but perhaps one that is only triggered by a certain combination of parameters or something specific about the data. To debug the problem we'll need to be able to reproduce it, could you upload the image and the mask file to the anonftp on ftp://iraf.noao.edu/pub? It would also help if you could post all the parameters you use (e.g. the output of the "dpar fixpix" command). Otherwise, there are no know issues with the task that could explain it.-Mike

Thanks, Mike. I uploaded the file files2.tar. It contains the image, mask file, and the dpar fixpix output. (Initially, there was an error as I was uploading the file, so file.tar is also mine. It should be removed from the site but I was unable to delete it.)
-Amber

Hi Amber,There are two problems: Less fatal is the fact that the FITS file is not properly zero-padded to an integral number of 2880 byte blocks. The cause of the segfault however is a bug in the FITS kernel where the BZERO value is causing the image to be seen as a real and the kernel is trying to read the pixels as real values rather than as shorts with a scaling. This means it's reading beyond the physical length of the file and returning an invalid pixel pointer leading to the error.The fix is trivial, simply copy the images in place and allow the system to rewrite the images as real, i.e.[code:1:9ab6309b33]
cl> imcopy *.fits *.fits
[/code:1:9ab6309b33]The resulting images will be properly padded and written with 32-bit pixels. Doing this to your sample image allowed FIXPIX to work properly, you might want to backup your image first just to be safe.-Mike

Mike,
Fixpix is now working. I copied the images in place like you recommended. Thank you so much for your help.
-Amber