I don't know if this is a feature or not, but when I use KAPPA/SLIDE to
realign a spectrum, the result has an AXIS structure that has bad pixels
in it. When it comes time to display it with LINPLOT, I get the error:
!! One or more of the input AXIS arrays is non-monotonic.
! LINPLOT: Failed to display a plot of a 1-dimensional data set.
! Application exit status SAI__ERROR, Error
This only occurs if the slide amount is not a whole number of pixels.
I have a test file available at
http://www.jach.hawaii.edu/~bradc/starlink/test_preslide.sdf
Command that triggers it:
% slide in=test_preslide out=test_postslide abs=0.07 method=sincsinc \
? params="[2,1.5]"
Using SincSinc interpolation.
Using WCS information read from the FITS extension of
'/export/data/bradc/oracdr/reduced/uist/20050726/test_preslide'.
% linplot test_postslide
!! One or more of the input AXIS arrays is non-monotonic.
! LINPLOT: Failed to display a plot of a 1-dimensional data set.
! Application exit status SAI__ERROR, Error
! test_postslide
I don't seem to recall this happening in the past. I'm getting around it
by recalculating the AXIS structure using SETAXIS. This isn't really ideal
because I can't get the AXIS structure to be the same afterwards, I'm
always off by a few nanometers. I guess that doesn't really matter much as
this is an estimated wavelength axis, but still...
I have hazy memories of this in the past, but I can't find any mention of
them in the stardev mailing list archives.
Also, LINPLOT appears to suffer from the GRP identifier leak problem, as
when I display a spectrum successfully I get:
% linplot test_postslide
DEVICE - Name of graphics device /@xw/ >
WARNING: The number of active GRP identifiers increased from 0 to 2 during
execution of LINPLOT (KAPPA programming error).
Cheers,
Brad.
|