Scombine consistently gives a Segmentation Violation
sakenjm wrote on May 04, 2011
Since upgrading to IRAF V2.15.1.a I'm consistently get a segmentation violation error in scombine. This has happened on 2 different machines running Debian Linux.
I've tried to unlearn scombine and reran mkiraf. No luck.
On this particular machine I am running kernel version 2.6.26-2-amd64 and
glibc 2.11.
Here is the command and the results from d_trace:
************************************************
ecl> scombine(input="sg0001.ms.fits,sg0002.ms.fits,sg0003.ms.fits",output="test.fits",group="all",combine="median",reject="none",first=yes,weight="none")
----- task cl - line 0 ( ) -----
0 0+5: call scombine
0 5+11: pushconst `sg0001.ms.fits,sg0002.ms.fits,sg0003.ms.fits'
0 16+4: absargset input
0 20+7: pushconst `test.fits'
0 27+4: absargset output
0 31+6: pushconst `all'
0 37+4: absargset group
0 41+6: pushconst `median'
0 47+4: absargset combine
0 51+6: pushconst `none'
0 57+4: absargset reject
0 61+5: pushconst yes
0 66+4: absargset first
0 70+6: pushconst `none'
0 76+4: absargset weight
0 80+3: exec
----- exec external task scombine - line 0 -----
----- connect to noaobin$x_scombine.e -----
scombine\n
input="sg0001.ms.fits,sg0002.ms.fits,sg0003.ms.fits"\n
output="test.fits"\nnoutput=""\nlogfile="STDOUT"\n
apertures=""\ngroup="all"\ncombine="median"\nreject="none"\n
first=yes\nw1=INDEF\nw2=INDEF\ndw=INDEF\nnw=INDEF\nlog=no\n
scale="none"\nzero="none"\nweight="none"\nsample=""\n
lthreshold=INDEF\nhthreshold=INDEF\nnlow=1\nnhigh=1\nnkeep=1\n
mclip=yes\nlsigma=3.\nhsigma=3.\nrdnoise="0."\ngain="1."\n
snoise="0."\nsigscale=0.1\npclip=-0.5\ngrow=0\nblank=0.\n
mode="ql"\n$nargs=0\n
--------------------------------
----- task scombine - line 85 ( ) -----
85 86+4: inspect interp
85 90+3: end
----- task scombine - line 86 ( ) -----
86 86+4: call error
86 90+5: pushconst 501
86 95+4: posargset 0
86 99+8: pushconst `segmentation violation'
86 107+4: posargset 1
86 111+3: exec
----- exec builtin error : line 86 -----
----- disconnect noaobin$x_scombine.e -----
ERROR: segmentation violation
cl_error: abort=1 beep=1 trace=1 flpr=1
cl_error: code=501 do_err=1 errtype=1/1 task='cl'
******************************************************************
I can also supply the data files if needed, however this seems to occur every time I run scombine. I've tried with different datasets and get the same error.
Thanks for any insights anyone can provide.
Jon Saken
sakenjm@appstate.edu
I've tried to unlearn scombine and reran mkiraf. No luck.
On this particular machine I am running kernel version 2.6.26-2-amd64 and
glibc 2.11.
Here is the command and the results from d_trace:
************************************************
ecl> scombine(input="sg0001.ms.fits,sg0002.ms.fits,sg0003.ms.fits",output="test.fits",group="all",combine="median",reject="none",first=yes,weight="none")
----- task cl - line 0 ( ) -----
0 0+5: call scombine
0 5+11: pushconst `sg0001.ms.fits,sg0002.ms.fits,sg0003.ms.fits'
0 16+4: absargset input
0 20+7: pushconst `test.fits'
0 27+4: absargset output
0 31+6: pushconst `all'
0 37+4: absargset group
0 41+6: pushconst `median'
0 47+4: absargset combine
0 51+6: pushconst `none'
0 57+4: absargset reject
0 61+5: pushconst yes
0 66+4: absargset first
0 70+6: pushconst `none'
0 76+4: absargset weight
0 80+3: exec
----- exec external task scombine - line 0 -----
----- connect to noaobin$x_scombine.e -----
scombine\n
input="sg0001.ms.fits,sg0002.ms.fits,sg0003.ms.fits"\n
output="test.fits"\nnoutput=""\nlogfile="STDOUT"\n
apertures=""\ngroup="all"\ncombine="median"\nreject="none"\n
first=yes\nw1=INDEF\nw2=INDEF\ndw=INDEF\nnw=INDEF\nlog=no\n
scale="none"\nzero="none"\nweight="none"\nsample=""\n
lthreshold=INDEF\nhthreshold=INDEF\nnlow=1\nnhigh=1\nnkeep=1\n
mclip=yes\nlsigma=3.\nhsigma=3.\nrdnoise="0."\ngain="1."\n
snoise="0."\nsigscale=0.1\npclip=-0.5\ngrow=0\nblank=0.\n
mode="ql"\n$nargs=0\n
--------------------------------
----- task scombine - line 85 ( ) -----
85 86+4: inspect interp
85 90+3: end
----- task scombine - line 86 ( ) -----
86 86+4: call error
86 90+5: pushconst 501
86 95+4: posargset 0
86 99+8: pushconst `segmentation violation'
86 107+4: posargset 1
86 111+3: exec
----- exec builtin error : line 86 -----
----- disconnect noaobin$x_scombine.e -----
ERROR: segmentation violation
cl_error: abort=1 beep=1 trace=1 flpr=1
cl_error: code=501 do_err=1 errtype=1/1 task='cl'
******************************************************************
I can also supply the data files if needed, however this seems to occur every time I run scombine. I've tried with different datasets and get the same error.
Thanks for any insights anyone can provide.
Jon Saken
sakenjm@appstate.edu
Francisco Valdes wrote on May 04, 2011
This problem was diagnosed and fix was made. I don't know if the IRAF manager produced a patch or update. You can submit a request for the fix as a separate post.
Frank Valdes
Frank Valdes
Mike Fitzpatrick wrote on May 04, 2011
I had a feeling there was a report I forgot to follow up with.... Anyway, the fixed version of SCOMBINE can be found in
ftp://iraf.noao.edu/iraf/v215/support/linux64/x_scombine.e
Just download in binary mode, make sure its executable and place in the $iraf/noao/bin.linux64 directory. If you still have problems let me know (note: macintel binary is also available).
ftp://iraf.noao.edu/iraf/v215/support/linux64/x_scombine.e
Just download in binary mode, make sure its executable and place in the $iraf/noao/bin.linux64 directory. If you still have problems let me know (note: macintel binary is also available).
tsergori wrote on May 04, 2011
Hi
I'm also having problems with scombine, consistently getting "segmentation violation" errors when combining three or more spectra.
I'm running the following version of IRAF
NOAO/IRAF PC-IRAF Revision 2.15.1a EXPORT Mon Feb 21 18:54:16 MST 2011
on an Intel Mac 10.6.7
Following the post above, here is the command the results from d_trace
*********************************
echelle> scomb 8066* test
----- task cl - line 33 ( ) -----
33 30+4: call scomb
33 34+6: pushconst `8066*'
33 40+4: posargset 0
33 44+6: pushconst `test'
33 50+4: indirposset 1
33 54+3: exec
----- exec external task scombine - line 33 -----
----- connect to noaobin$x_scombine.e -----
scombine\ninput="8066*"\noutput="test"\nnoutput=""\n
logfile="STDOUT"\napertures=""\ngroup="apertures"\n
combine="sum"\nreject="none"\nfirst=no\nw1=INDEF\nw2=INDEF\n
dw=INDEF\nnw=INDEF\nlog=no\nscale="none"\nzero="none"\n
weight="none"\nsample=""\nlthreshold=INDEF\nhthreshold=INDEF\n
nlow=1\nnhigh=1\nnkeep=1\nmclip=yes\nlsigma=3.\nhsigma=3.\n
rdnoise="0."\ngain="1."\nsnoise="0."\nsigscale=0.1\n
pclip=-0.5\ngrow=0\nblank=0.\nmode="ql"\n$nargs=2\n
--------------------------------
----- task scombine - line 4 ( ) -----
4 60+4: call error
1: CALL error
4 64+5: pushconst 501
4 69+4: posargset 0
4 73+8: pushconst `segmentation violation'
4 81+4: posargset 1
4 85+3: exec
----- exec builtin error : line 4 -----
----- disconnect noaobin$x_scombine.e -----
ERROR: segmentation violation
cl_error: abort=1 beep=1 trace=1 flpr=1
cl_error: code=501 do_err=1 errtype=1/1 task='cl'
************************
I tried clicking on the ftp link above, but I got an Alert message "550 Failed to change directory". I couldn't work out what the link should be.
If this is also a scombine problem, could you please point me to a fixed version of SCOMBINE?
Thanks
Cheers
Dave
I'm also having problems with scombine, consistently getting "segmentation violation" errors when combining three or more spectra.
I'm running the following version of IRAF
NOAO/IRAF PC-IRAF Revision 2.15.1a EXPORT Mon Feb 21 18:54:16 MST 2011
on an Intel Mac 10.6.7
Following the post above, here is the command the results from d_trace
*********************************
echelle> scomb 8066* test
----- task cl - line 33 ( ) -----
33 30+4: call scomb
33 34+6: pushconst `8066*'
33 40+4: posargset 0
33 44+6: pushconst `test'
33 50+4: indirposset 1
33 54+3: exec
----- exec external task scombine - line 33 -----
----- connect to noaobin$x_scombine.e -----
scombine\ninput="8066*"\noutput="test"\nnoutput=""\n
logfile="STDOUT"\napertures=""\ngroup="apertures"\n
combine="sum"\nreject="none"\nfirst=no\nw1=INDEF\nw2=INDEF\n
dw=INDEF\nnw=INDEF\nlog=no\nscale="none"\nzero="none"\n
weight="none"\nsample=""\nlthreshold=INDEF\nhthreshold=INDEF\n
nlow=1\nnhigh=1\nnkeep=1\nmclip=yes\nlsigma=3.\nhsigma=3.\n
rdnoise="0."\ngain="1."\nsnoise="0."\nsigscale=0.1\n
pclip=-0.5\ngrow=0\nblank=0.\nmode="ql"\n$nargs=2\n
--------------------------------
----- task scombine - line 4 ( ) -----
4 60+4: call error
1: CALL error
4 64+5: pushconst 501
4 69+4: posargset 0
4 73+8: pushconst `segmentation violation'
4 81+4: posargset 1
4 85+3: exec
----- exec builtin error : line 4 -----
----- disconnect noaobin$x_scombine.e -----
ERROR: segmentation violation
cl_error: abort=1 beep=1 trace=1 flpr=1
cl_error: code=501 do_err=1 errtype=1/1 task='cl'
************************
I tried clicking on the ftp link above, but I got an Alert message "550 Failed to change directory". I couldn't work out what the link should be.
If this is also a scombine problem, could you please point me to a fixed version of SCOMBINE?
Thanks
Cheers
Dave
Mike Fitzpatrick wrote on May 04, 2011
Sorry for the typo, I've corrected the above link but for your system you would use 'macintel' instead of 'linux64', e.g.
ftp://iraf.noao.edu/iraf/v215/support/macintel/x_scombine.e
ftp://iraf.noao.edu/iraf/v215/support/macintel/x_scombine.e
tsergori wrote on May 04, 2011
Thanks for your help!
sakenjm wrote on May 04, 2011
Thank you for providing the fixed version of scombine. It seems to have completely solved the problem.
Jon Saken
Jon Saken
Last post on May 04, 2011