Hi all,

I'm pretty new to IRAF. That said, I've been trying to use the noao->twodspec->longslit package to flux calibrate some images. I can get all the way through til the final calibrate, but every time I do this, the memory goes corrupt:

Input spectra to calibrate (ccd0101_sub_stitch_qsosub_s_1.fits):
Output calibrated spectra (test.fits):
Warning: setting CD2_2 to 1.
Dispersion axis (1=along lines, 2=along columns, 3=along z) (1:3) (1):
test.fits: CTQ408

PANIC in `/pciraf/iraf/noao/bin.redhat/x_onedspec.e': Memory has been corrupted

This memory corruption has happened other tasks and resolved with a smaller image (stecf->specres->specinholucy). I did get a usable output once where it gave me the PANIC, but still produced output. But that has not happened twice. I've been trolling the forums for a while and found a number of posts:

http://iraf.net/phpBB2/viewtopic.php?p=124039&sid=27af5be67ada24628e68e26a79d4bdf1
http://iraf.net/phpBB2/viewtopic.php?t=86285&highlight=memory+corrupted
http://iraf.net/phpBB2/viewtopic.php?t=85994&highlight=memory+corrupted
-not sure what the last one there is actually suggesting, I don't know if I can or how to turn cache off

Anyways, here are some pars:
twodspec> dpar longslit
longslit.dispaxis = 1
longslit.nsum = "2032"
longslit.observatory = "observatory"
longslit.extinction = "onedstds$kpnoextinct.dat"
longslit.caldir = "onedstds$spec50cal/"
longslit.interp = "poly5"
longslit.records = ""
longslit.version = "February 1993"
longslit.mode = "ql"
# EOF

longslit> dpar calib
calibrate.input = "ccd0101_sub_stitch_qsosub_s_1.fits"
calibrate.output = "test.fits"
calibrate.extinct = yes
calibrate.flux = yes
calibrate.extinction = "onedstds$kpnoextinct.dat"
calibrate.observatory = )_.observatory
calibrate.ignoreaps = yes
calibrate.sensitivity = "novsens"
calibrate.fnu = no
calibrate.mode = "ql"
# EOF

I'm not sure what the "nsum" parameter means, but I've tried it on 1, 2032, and 6064 (the dimensions of my images). The one time it produced output, I think it was set to 2032. I'm quite certain my stack size is unlimited. I haven't tried it with smaller images, but I was hoping to do it on the full size. Unlearn and flpr do not seem to change anything. I don't see why my dispersion axis would be off (example, above).

I believe I'm using the latest fedora on a 64 bit machine. Iraf is site maintained so I'm sure it fairly up-to-date.
Any ideas?

Thanks,
Dan

I just retooled some of my images and sent through a 2032x2032 frame + standard in the longslit flux calibration routines. I got the exact same behavior. All steps up to 'calibrate' worked fine. When I tried to calibrate, the first frame went through and the correct output was written before the
PANIC in `/pciraf/iraf/noao/bin.redhat/x_onedspec.e': Memory has been corrupted.
The output looks fine, but now when I try to do another, it hangs before output even when processing the same image. Flpr and restarts make no difference.

Dan

Oh, also the correct setting for longslit parameter is one, which it is now set to.

Hi all,
This problem was resolved. The culprit was a custom setting in my login.cl. It was the cmbuflen or min_lenuserara settings that was causing the corruption.
Dan