# version 2.2.2  - If expected variance is negative for some obscure reasons in likelihood.x, likelihood_2c.x or mlikelihood.x, it is automatically
#		   set to sigma_bkg**2. This eliminates annoying floating exception errors 05/06/01
#		   ICF enclosed flux scaling is applied only if encflux > 0.0. Modified t_gimfit2d.x, t_c2gimfit2d.x, t_mgimfit2d.x 05/05/01
#		   More stability modifications. Due to improperly checked boundary reflections in icf_param, parameter values could be outside
#		   min-max range (e.g. negative r_e!). Fixed in icf_param.x 05/07/01
#		   GIMFIT2D, C2GIMFIT2D and MGIMFIT2D now all check if the output residual image, the output model image and the logfile already exist before
#		   trying to start the fit for an object. More robust to the "racing problem" 05/07/01
#		   icx and icy in ginit_param.x cannot be less than im_dim/4 or greater than im_dim*(3/4). If they are, then dx and dy are both set to
#		   zero. Faint objects may have bad dx and dy determinations due to large errors in estimated total flux. Modified ginit_param.x 05/07/01
#		   Fixed background bookkeeping error in MGIMFIT2D when dobkg=no 06/20/01
#		   Modified symmetrization routine so that it produces correct symmetrized mask images! Modified GALPARAM so that when dosym=yes, an 
#		   output symmetrized mask image is also created. 07/01/01
#		   Modified GSCRIPTER and GGROUPER so that the output scripts can be written out in PYRAF form if dopyraf=yes or in the usual IRAF CL 
#		   form if dopyraf=no. 09/04/01
#		   Timestamp commands have been disabled in GIMFIT2D, C2GIMFIT2D, and MGIMFIT2D until PYRAF can deal with them. The timestamps are
#		   now in the scripts. 09/04/01
#		   Now possible to reject saturated pixels from fits by setting gim2d.rejsat to yes and gim2d.satlevel to the pixel saturation value 09/29/01
#		   PYRAF scripts can now also be produced by GSIMUL 10/09/01
#		   Changed names of all likelihood routines from *likelihood* to *lk* to avoid multiply defined symbols on some architectures. 05/21/02
#		   Improved "racing protection" with ifnoerr conditions. Modified GIMFIT2D, C2GIMFIT2D, and MGIMFIT2D. 07/10/02
#		   Modified GSIMUL so that it puts XGAL-like header keywords in simulated image and so that the segmentation image of the tmpl + simulated image
#		   is kept in xgalroot+i+_tmpl_m 07/29/02
#		   Modified XGAL so that it does not load all of the input science, mask and sigma images into memory to work. It now loads only the relevant image
#		   sections for a given object. Needed to do this for really large images like GOODS mosaics or CFHTLS 02/06/03
#
# version 3.0  -  Temperature is now frozen when convergence has been achieved - Mar02/2004
#		   When bkg_npix requirement is not met and original image must be used, bkg buffers are now (xmax-xmin)*(ymax-ymin) instead of full original image dimensions - Mar02/2004
#		   Fixed out of bound error in XGAL - Mar02/2004
#		   Fixed out of bound error in G2DMODEL - Apr14/2004
#		   Asymmetric indices are now given to three decimal places - May12/2004
#		   Eliminated absolute lower total flux limit of 1 DU - 26July2004
# version 3.1  -   Fixed two typos in C2GIMFIT2D in the calculations of the best-fit values for ftot,bt and sky - 26Oct2004
#		   Maximum number of Metropolis trials (NMETMAX=50000) now implemented in GIMFIT2D, C2GIMFIT2D, and MGIMFIT2D - 8Nov2004
#		   New bkgparam=yes behavior. If temperature for db parameter is positive, same as before. If T(db) < 0.0 and min(db)=max(db), then
#		   bkg level is as specified in mdpar file. It is not recalculated. bkgsig is recalculated as before. and T(db)=0. If T(db) < 0.0 and min(db)
#		   != max(db), then bkg level is as specified in mdpar file. It is not recalculated. bkgsig is recalculated as before. and T(db)=
#		   -T(db,mdparfile). Changes made in BKG_PARAM routine - 23Nov2004
#		   If disk input T is negative, then disk PA is forced to be equal to bulge PA. Implemented GIMFIT2D, C2GIMFIT2D, and MGIMFIT2D - 23Nov2004
#		   Initial seed value now written to output model and residual image headers. Implemented in GIMFIT2D, C2GIMFIT2D, and MGIMFIT2D - 23Oct2005
#		   Fixed some extern statements so that compiling on mac OSX would work - 8Feb2008
# version 3.2  -   If mdpar parameter values are supposed to remain fixed, they remain so through the pure bulge/pure disk tests in GIMFIT2D, C2GIMFIT2D, and MGIMFIT2D - 12Jun2008
#		   Increased bkg area on parent image to avoid problems for smaller galaxies in the wings of larger ones,i.e.,exarea = float(nx*ny)*(float(bkg_npix)/float(nbkg))*10.0 - 21Nov2008