ecl patch
Roland Roberts wrote on Jan 07, 2009
I would have sworn I posted this a couple of years ago, but I can't find it now...
It is very convenient to be able to wrap ecl with a shell script that feeds in some command to be done. I have a custom ccxymatch task that I feed in like so
Of course, updating/upgrading IRAF breaks this every time because the distributed ecl ends with
Changing about three lines makes this work. The most obvious is
but to be completely correct, there are two places earlier where you have to shift argv. Here's the patch.
Can I "lobby" to have this change become permanent?
roland
It is very convenient to be able to wrap ecl with a shell script that feeds in some command to be done. I have a custom ccxymatch task that I feed in like so
ecl < /path/to/ccxymatch.cl FITS-FILEOf course, updating/upgrading IRAF breaks this every time because the distributed ecl ends with
exec $fileChanging about three lines makes this work. The most obvious is
exec $file $*but to be completely correct, there are two places earlier where you have to shift argv. Here's the patch.
595 roland> diff -uw cl.csh.~1~ cl.csh
--- cl.csh.~1~ 2007-10-11 00:22:57.000000000 -0400
+++ cl.csh 2009-01-06 22:41:18.000000000 -0500
@@ -19,8 +19,10 @@
else if ($#argv > 0) then
if ("$argv[1]" == "-old" || "$argv[1]" == "-o") then
set cl_binary = "cl.e"
+ shift
else if ("$argv[1]" == "-vo" || "$argv[1]" == "-o") then
set cl_binary = "vocl.e"
+ shift
else if ("$argv[1]:e" == "c") then
# Workaround for autoconf scripts attempting to use this command as
# a valid compiler option. On some systems (mostly Debian) a valid
@@ -180,4 +182,4 @@
set file = ${IRAFBIN}$cl_binary
# Run the desired CL.
-exec $file
+exec $file $*Can I "lobby" to have this change become permanent?
roland
Mike Fitzpatrick wrote on Jan 07, 2009
The preferred method for something like this is the so-called "#!cl" script, i.e. a way to make a .cl file executable using the CL as the interpreter. See for example
http://iraf.noao.edu/iraf/web/new_stuff/cl_host.html
for details. In that was your ccxymatch.cl file can become a standard host command.
-Mike
http://iraf.noao.edu/iraf/web/new_stuff/cl_host.html
for details. In that was your ccxymatch.cl file can become a standard host command.
-Mike
Roland Roberts wrote on Jan 07, 2009
I've gone through this with previous releases and I have never been able to get that to work. Consider the "simple" case of using images. One of the examples says
What other "magic" variables will I need to make various packages work? I don't know, I can't find documentation on "to use package X in a host script, include the following..." and the type of error messages I get back are cryptic.
Why is it trying to read my login.cl? Why can't it connect to subprocess imcoord$x_images.e which I didn't even know it was doing for the working version. Searching iraf.net shows no hits on the word "subprocess".
As you can tell, I find this process very frustratng and non-intuitive.
Below is my script, after I commented out the !rm which I also can't figure out how to get to work when running in this mode.
logver = "IRAF V2.11 May 1997" # needed for IMAGES packageWhat other "magic" variables will I need to make various packages work? I don't know, I can't find documentation on "to use package X in a host script, include the following..." and the type of error messages I get back are cryptic.
./foo.cl G5-rroberts.raso-b-139P-E-240-C-10-X1-Clear-001-UT040039
args = G5-rroberts.raso-b-139P-E-240-C-10-X1-Clear-001-UT040039
ERROR: Cannot open connected subprocess (imcoords$x_images.e)
cl ()
cl ()
cl ()
tv ()
cl ()
cl ()
cl ()
imutil ()
cl ()
immatch ()
cl ()
imgeom ()
cl ()
imfit ()
cl ()
imfilter ()
cl ()
imcoords ()
images ()
cl ()
cl ()
Error while reading login.cl file - may need to rebuild with mkiraf
Fatal startup error. CL dies.
Why is it trying to read my login.cl? Why can't it connect to subprocess imcoord$x_images.e which I didn't even know it was doing for the working version. Searching iraf.net shows no hits on the word "subprocess".
As you can tell, I find this process very frustratng and non-intuitive.
Below is my script, after I commented out the !rm which I also can't figure out how to get to work when running in this mode.
#!/iraf/iraf/bin.redhat/cl.e -f
# Automatically find the plate solution.
lists
images
imcoords
tv
printf("args = %s\n", args)
{
string sources
string field
string match
string platesol
string residuals
string image
sources = args+".sources_bright"
field = args+".usno_bright"
match = args+".ccxymatch"
platesol = args+".platesol"
residuals = args+".residuals"
image = args+".fit"
# Remove the old files for idempotency
# !rm -f match residuals
ccxymatch sources field match 6 12 matching=triangles nmatch=50 xmag=3 ymag=3 verbose=yes
ccmap input=match database=platesol results=residuals images=image lngcolumn=1 latcolumn=2 xcolumn=3 ycolumn=4 update=yes interactive=no verbose=no
}
logoutMike Fitzpatrick wrote on Jan 07, 2009
I think you may have simply mistyped the value of the 'arch' environment variable, which should be the same as the IRAFARCH except for a leading period. In your case IRAFARCH is 'redhat', and so the arch should be ".redhat".
This variable is used internally to find the bin directories and is normally set by the 'cl' script. Also, be sure to do a "limit stacksize unlimited" in your .cshrc file or on the command line for recent linux systems (or use "ulimit -s unlimited" for Bash shells, again something done for you by the 'cl' startup script). Otherwise your script looks fine and nominally works for me as expected.
-Mike
This variable is used internally to find the bin directories and is normally set by the 'cl' script. Also, be sure to do a "limit stacksize unlimited" in your .cshrc file or on the command line for recent linux systems (or use "ulimit -s unlimited" for Bash shells, again something done for you by the 'cl' startup script). Otherwise your script looks fine and nominally works for me as expected.
-Mike
Roland Roberts wrote on Jan 07, 2009
That does seem to be part of the problem. I'm running from bash, I never use csh....
I've set both IRAFARCH and arch as you indicated. I've now tried running using csh as well. I do get closer once the environment variable is set, or by setting it in the script. But which shell makes no difference:
What's wrong with the error is that the sources file is present. I can't figure out where it is looking. The files are all right in the current directory. I added some diagnostics but they aren't telling me anything. I've tried adding "cl < login.cl" followed by "keep" out of desperation 'cause I have no clue what is missing. I also tried copying login.cl into my script. Those both fail the same way. ulimit -a shows unlimited stack.
I have no idea what I'm missing.
I've set both IRAFARCH and arch as you indicated. I've now tried running using csh as well. I do get closer once the environment variable is set, or by setting it in the script. But which shell makes no difference:
% ./foo.cl G4-rroberts.raso-b-139P-E-240-C-10-X1-Clear-001-UT053827
args = G4-rroberts.raso-b-139P-E-240-C-10-X1-Clear-001-UT053827
sources = G4-rroberts.raso-b-139P-E-240-C-10-X1-Clear-001-UT053827.sources_bright
field = G4-rroberts.raso-b-139P-E-240-C-10-X1-Clear-001-UT053827.usno_bright
match = G4-rroberts.raso-b-139P-E-240-C-10-X1-Clear-001-UT053827.ccxymatch
platesol = G4-rroberts.raso-b-139P-E-240-C-10-X1-Clear-001-UT053827.platesol
residual = G4-rroberts.raso-b-139P-E-240-C-10-X1-Clear-001-UT053827.residuals
image = G4-rroberts.raso-b-139P-E-240-C-10-X1-Clear-001-UT053827.fit
ERROR: Cannot open file (sources)
cl ()
tv ()
cl ()
cl ()
cl ()
imutil ()
cl ()
immatch ()
cl ()
imgeom ()
cl ()
imfit ()
cl ()
imfilter ()
cl ()
imcoords ()
images ()
cl ()
cl ()
Error while reading login.cl file - may need to rebuild with mkiraf
Fatal startup error. CL dies.What's wrong with the error is that the sources file is present. I can't figure out where it is looking. The files are all right in the current directory. I added some diagnostics but they aren't telling me anything. I've tried adding "cl < login.cl" followed by "keep" out of desperation 'cause I have no clue what is missing. I also tried copying login.cl into my script. Those both fail the same way. ulimit -a shows unlimited stack.
#!/iraf/iraf/bin.redhat/cl.e -f
# Automatically find the plate solution.
set arch = ".redhat"
lists
images
imcoords
tv
printf("args = %s\n", args)
string sources
string field
string match
string platesol
string residuals
string image
sources = args+".sources_bright"
field = args+".usno_bright"
match = args+".ccxymatch"
platesol = args+".platesol"
residuals = args+".residuals"
image = args+".fit"
printf("sources = %s\n", sources)
printf("field = %s\n", field)
printf("match = %s\n", match)
printf("platesol = %s\n", platesol)
printf("residual = %s\n", residual)
printf("image = %s\n", image)
# Remove the old files for idempotency
!rm -f match residuals
ccxymatch sources field match 6 12 matching=triangles nmatch=50 xmag=3 ymag=3 verbose=yes
ccmap input=match database=platesol results=residuals images=image lngcolumn=1 latcolumn=2 xcolumn=3 ycolumn=4 update=yes interactive=no verbose=no
logoutI have no idea what I'm missing.
Mike Fitzpatrick wrote on Jan 07, 2009
Try using the "program mode" syntax, i.e.
instead of
As written, the script is looking for a file names 'sources', not the filename stored in the 'sources' variable. It's a CL quirk, but you'd see the same thing at the interactive prompt:
-Mike
ccxymatch(sources, field, .....)instead of
ccxymatch sources field ....As written, the script is looking for a file names 'sources', not the filename stored in the 'sources' variable. It's a CL quirk, but you'd see the same thing at the interactive prompt:
ecl> s1 = "dev$pix"
ecl> imstat s1
# IMAGE NPIX MEAN STDDEV MIN MAX
Error reading image s1 ...
ecl> imstat (s1)
# IMAGE NPIX MEAN STDDEV MIN MAX
dev$pix 262144 108.3 131.3 -1. 19936.
-Mike
Roland Roberts wrote on Jan 07, 2009
Thank you! It's mostly working now. I don't understand why the line
is doing nothing (neither fails nor deletes the files). I can move the delete to the wrapper, but how do I get these type of shell tasks to run from inside cl?
!rm -f match residualsis doing nothing (neither fails nor deletes the files). I can move the delete to the wrapper, but how do I get these type of shell tasks to run from inside cl?
#!/iraf/iraf/bin.redhat/cl.e -f
# Automatically find the plate solution.
set arch = ".redhat"
lists
images
imcoords
tv
printf("args = %s\n", args)
string sources
string field
string match
string platesol
string residuals
string image
sources = args+".sources_bright"
field = args+".usno_bright"
match = args+".ccxymatch"
platesol = args+".platesol"
residuals = args+".residuals"
image = args+".fit"
printf("sources = %s\n", sources)
printf("field = %s\n", field)
printf("match = %s\n", match)
printf("platesol = %s\n", platesol)
printf("residual = %s\n", residual)
printf("image = %s\n", image)
# Remove the old files for idempotency
!rm -f match residuals
ccxymatch.matching = "triangles"
ccxymatch.nmatch = 50
ccxymatch.xmag = 3
ccxymatch.ymag = 3
ccxymatch.verbose = yes
ccxymatch(sources, field, match, 6, 12)
ccmap.results = residuals
ccmap.images = image
ccmap.lngcolumn = 1
ccmap.latcolumn = 2
ccmap.xcolumn = 3
ccmap.ycolumn = 4
ccmap.update = yes
ccmap.interactive = no
ccmap.verbose = no
ccmap(match, platesol)
logoutMike Fitzpatrick wrote on Jan 07, 2009
Same problem, 'match' and 'residuals' are local script variables, not actual filenames.
You can either use the DELETE command in IRAF to do the same thing, or if you need to construct a host command build the string and pipe it to the CL for execution, e.g.
-Mike
You can either use the DELETE command in IRAF to do the same thing, or if you need to construct a host command build the string and pipe it to the CL for execution, e.g.
printf ("!rm -f %s %s\n", match, residuals) | cl()-Mike
Roland Roberts wrote on Jan 07, 2009
Thanks Mike. My frustration level has receded into the noise :-)
I've been reading the "An Introductory Guide to IRAF Scripts" and it's helping me put some finishing touches on the script where I need less work done in the shell wrapper.
I've been reading the "An Introductory Guide to IRAF Scripts" and it's helping me put some finishing touches on the script where I need less work done in the shell wrapper.
Last post on Jan 07, 2009