pipelinechanges

JMP 07/2004
-----------

Bug fix:

Old kbo_gen version was using minimum and maximum value of angle, but this
is not consistant with other programs, and with normal meaning of $ang and
$aw.
So changed kbo_gen to take mean angle and half width.

JMP 12/2003
-----------

Bug fixes:

bug in step2ajmp/plant combination. step2ajmp was writing the
transformation coefficient in free format, which resulted in output on 2
lines in some cases. match.csh (in plant.csh) assumed all coefficients were
on 1 line. Fixed by imposing output format in step2ajmp.


JMP 11/2003
-----------

Bug fixes:

bug in PSF making: bug actually in jmpmakepsf.cl: mkapfile was crashing
with floating point execption, no output, and was terminating jmpmakepsf.cl
and jmpprepimage.cl. Error handling code was by-passed. Added a crude
exception trapping mechanism. Should ask JJ how to do this better. Anyway,
solve the floating point exception problem by fitting more than only 3
paramters in the growth curve. So first try with the usual 3 parameters. If
this fails, then go for 4 parameters, and then 5 (the maximum in the
model). Should solve most of the problems, without impeding previous
success.

Unit fixes:

plant.csh was using pixel/hour while step3* was using ''/hour. Updated
plant.csh to use ''/hour. Had to modify both plant.csh and kbo_gen.c. Now
Object.planted generated with 1 extra column showing the rate as ''/hour.
Also added the possibility to give the rates and angles as argument to
plant.csh (and also work for control.pl).

Cosmetic changes:

Version control: I have added the version number into the *.mopheader
file. Also changed to version 1.2. Need to rerun all step*jmp programs to
reuse files. Can still skip rerunning makeallpsf.csh and plant.csh.

changed step2bjmp.f and step3jmp.f to make them faster by avoiding
unnecessary tests. Seems to work rather well.

added the possibility to force running jmpmakepsf.csh even when lock of
finished file exists. Do that by adding an optional fourth parameter:
"force".
Propogated that possibility into makeallpsf.sh that now accept the --force
parameter.


BG/LA 8/2002
------------

Changes to the pipeline:
(to build pipeline) 

1-change permissions on many files  : ) ("pre/postmarch" tars have wrong 
permissions)

2-make pipeline/lib, pipeline/lib/Linux, and pipeline/include
directories, since install-pipeline codes don't test for directory
presence before trying to write into directory.

3- Change PREFIX to ~/pipeline (or whatever) in install files and
change hardlinks from /h/f14a/hawaii to ~/pipeline and /usr/bin/perl

4- Move readq to hans directory from jmp directory) - maybe have to
make hans directory before jmp??

5- unoget.f-c and square.inc must be copied from hans to jmp directory
(to make pipeline work )

6- make ~/iraf directory (with login.cl) or just link ~/iraf to ~ as we did.

7- Change detrend.csh to use movesingle instead of cfh12k, because we
split apart the chips beforehand instead of using MEF files.  This is 
not the easiest way to do this (better to use mscred for preproc and 
then keep MEF files, use cfh12k), but this is the way it was used last time.
Watch parameter values into movesingle, b/c more than one version. Also change
extn# to extn=0.

Further hack (beyond basic workability) -
8- Added a change in step2jmp.f to modify adaptive evaluation of pix
saturation value (inten_max):
" c     BG/LA change to account for high pix values with CFH12k
         if (inten_max.gt.65536.0) then
            inten_max = 65536.0
         endif
         inten_max = 0.8*inten_max   "
Added intensity max cap b/c CFH12k has bad columns which can
reach over this value when processed; this then screws up
bright object selection, so columns are chosen instead of
stars. Then registration doesn't work.

9- Needed to make some changes to ccastrom - basically used Brett's version
instead of JJ's. Should work, as is, in current tar. 
I think these changes were all related to where to get the USNO data. 
Also had to modify the catalog.toc to use CD's instead of hard drive.