.help revisions Jun88 math .nf gsurfit/gs_deval.gx iminterp/mrider.x iminterp/mrieval.x Fixed some procedure calls being closed with a ']' insted of a ')' (2/17/08, MJF) gsurfit/gs_chomat.gx The test for singularity would fail with certain kinds of problems because the test used EPSILON (should have been EPSILOND for the double precision) but this is for distinguishing numbers small numbers from 1 and not from each other. The test is now done with a comparison against the smallest real or double difference. The place where this was found to be a problem was with CCSETWCS. (8/31/04, Valdes) ======== V2.12.2a ======== curfit/cvinit.gx If one of the error checks caused an error return the cv pointer would have been allocated (cv != NULL) but some of the pointer fields could have garbage values since a malloc was used instead of a calloc. A later call to cvfree could result in a segmentation error. This was changed so that 1) a null cv pointer is returned in the initial error checks cause an error return and 2) the cv pointer is initially allocated with calloc so that no pointer fields will be non-NULL until explicitly set. (11/18/02, Valdes) ======= V2.12.1 ======= gsurfit/gsder.gx gsurfit/gsderr.x gsurfit/gsderd.x gsurfit/gs_fder.gx gsurfit/gs_fderr.x gsurfit/gs_fderd.x gsurfit/gs_deval.gx gsurfit/gs_devalr.x gsurfit/gs_devald.x Rewrote the routines which compute the derivatives of the polynomial surfaces. There were some renormalization problems in this code for non-linear chebyshev and legendre polynomial surfaces. The new code uses recursion relations for computation the derivatives which could cause small errors if they are used to do flux conservation. (27/10/99 LED) curfit/mkpkg deboor/mkpkg iminterp/mkpkg Added some missing file dependencies to the above mkpkg files. (20/9/99 LED) slalib Upgraded slalib from Version 1.6.3 to Version 2.3.0. Seventeen new routines were added including 4 dealing with FK5 to ICRS coordinate conversions (17/6/99 LED). arbpix.x arider.x arieval.x asider.x asieval.x asifit.x asifree.x asigeti.x asigetr.x asigrl.x asiinit.x asirestore.x asisave.x asisinit.x asitype.x asivector.x ii_1dinteg.x ii_bieval.x ii_cubspl.f ii_eval.x ii_greval.x ii_pc1deval.x ii_pc2deval.x ii_polterp.x ii_sinctable.x ii_spline.x ii_spline2d.x im1interpdef.h im2interpdef.h mkpkg mrider.x mrieval.x msider.x msieval.x msifit.x msifree.x msigeti.x msigetr.x msigrid.x msigrl.x msiinit.x msirestore.x msisave.x msisinit.x msisqgrl.x msitype.x msivector.x The 1D interpolation routines were replaced with new versions which support look-up table since and drizzle interpolation. Minor modifications were also made to the existing 1D sinc routines. The 2D image interpolation routines were replaced with new versions modified which support sinc, look-up table sinc and drizzle interpolation. (12/28/98 Davis) iminterp.hd iminterp.men iminterp.spc arbpix.hlp arider.hlp arieval.hlp asider.hlp asieval.hlp asifit.hlp asifree.hlp asigeti.hlp asigetr.hlp asigrl.hlp asiinit.hlp asirestore.hlp asisave.hlp asisinit.hlp asitype.hlp asivector.hlp iminterp.hlp mrider.hlp mrieval.hlp msider.hlp msieval.hlp msifit.hlp msifree.hlp msigeti.hlp msigetr.hlp msigrid.hlp msigrl.hlp msiinit.hlp msirestore.hlp msisave.hlp msisinit.hlp msisqgrl.hlp msitype.hlp msivector.hlp The 1D and 2D image interpolation routines help pages were updated to reflect the changes in both packages. (12/28/98 Davis) math$nlfit/nlchomat.gx math$nlfit/nlchomatr.x math$nlfit/nlchomatd.x Modified the singular matrix test to make a comparison against EPSILON instead of 0.0 to avoid floating point problems. (8/1/98 Davis) math$gsurfit/ Installed a completely new version of gsurfit which support "half" or "diagonal" cross terms. All the .gx and .x files were replaced. (4/30/97 Davis) math$gsurfit/gsrestore.gx math$gsurfit/gsrestorer.x math$gsurfit/gsrestored.x Changed the type declaration of the xmin, xmax, ymin, ymax variables from real to PIXEL to avoid machine precision problems. (3/21/96 Davis) math$nlfit/nlfit.gx math$nlfit/nlfitr.x math$nlfit/nlfitd.x Fixed a divide by zero error in the nlscatter routine which occurs if the fit has no degrees of freedom and the weighting type is WTS_SCATTER. (6/13/94 Davis) math/gsurfit/gssub.gx math/gsurfit/gssubd.x The gsurfit double precision routine gssubd was incorrectly calling the real precision routine gsgeti instead of the double precision routine dgsgeti due to an error in the gssub.gx file. This bug results in the wrong value of the parameter GSNSAVE being returned to the calling routine gssubd in the double precision case. (1/17/94 Davis) math/iminterp/arbpix.x math/interp/arbpix.x Changed all if (x == INDEF) constructs to IS_INDEF(x) constructs. (1/15/93 Davis) math/surfit/iseval.x Two indices were not being initialized before being used in a pointer offset computation in the spline1 and spline3 single input point code producing incorrect results. (12/17/92 Davis) math/nlfit/nlfit.gx math/nlfit/nlfitr.x math/nlfit/nlfitd.x math/nlfit/nliter.gx math/nlfit/nliterr.x math/nlfit/nliterd.x Added a test for lambda<=0.0 in the nlfit loop and in the iteration loop. Precision problems could cause the code to go into an infinite loop in some circumstances. The code returns a NOT_DONE status message if the convergence criteria have not been met in this circumstance. Made some other minor changes to the convergence criteria. (7/20/92 Davis) math/nlfit/nlacpts.gx math/nlfit/nlerrors.gx Fixed a bug in the routine nlacpts where the variable sum was not being initialized properly for each iteration. This could cause a slow drift in repeated fits. Also fixed a related problem in the error computation. Note this was not a problem with the version of nlfit in apphot in versions 2.9 or in testphot. (1/9/92 Davis) math/nlfit/nlfit.gx math/nlfit/nlfitr.x math/nlfit/nlfitd.x Added some code to force the nlfit package to do at least MINITER iterations whether or not the solution actually converged; MINITER + 1 if the weighting type is WTS_SCATTER. (1/3/91 Davis) math/nlfit/nliter.gx math/nlfit/nliterr.x math/nlfit/nliterd.x Simplied the test for determining whether the lambda factor needs to be increased or decreased. (1/3/91 Davis) math/nlfit/nlzero.gx math/nlfit/nlfitdefr.h math/nlfit/nlfitdefd.h Removed an extra argument from two aclr$t routines. Also these routines were being called with the pointer instead of the array as the argument. (10/3/91 Davis) lib/math/iminterp.h math/iminterp/arbpix.x math/iminterp/arider.x math/iminterp/arieval.x math/iminterp/asider.x math/iminterp/asieval.x math/iminterp/asifit.x math/iminterp/asiffree.x math/iminterp/asigrl.x math/iminterp/asiinit.x math/iminterp/asirestore.x math/iminterp/asisave.x math/iminterp/asivector.x math/iminterp/ii_1dinteg.x math/iminterp/ii_eval.x math/iminterp/ii_pc1deval.x math/iminterp/im1interpdef.h math/iminterp/doc/arbpix.hlp math/iminterp/doc/arider.hlp math/iminterp/doc/arieval.hlp math/iminterp/doc/asider.hlp math/iminterp/doc/asieval.hlp math/iminterp/doc/asifit.hlp math/iminterp/doc/asigrl.hlp math/iminterp/doc/asiinit.hlp math/iminterp/doc/asivector.hlp The 1D image interpolation routines were replaced with new versions modified by Frank Valdes to support sinc interpolation. (7/26/91 Davis) lib/math/curfit.h lib/math/gsurfit.h lib/math/iminterp.h lib/math/interp.h lib/math/surfit.h Added dictionary string definitions for the interpolator types and weights. Use of these strings insures that strdic/clgwrd will return the correct type code. Currently one has to assume the code definitions will not change or put a lookup table with a data statement containing the macro definitions. The code would also have to be modified to add new types. The dates were restored to their original values to avoid recompilation. (5/9, Valdes) math$curfit/cverrors.gx: Davis, May 6, 1991 math$curfit/cvpower.gx math$curfit/cvrefit.gx math$curfit/cvrestore.gx Finished cleanup of curfit package .gx files. math$gsurfit/gs_f1deval.gx: Davis, May 6, 1991 Changed the line "call amulk$t (,,,2.,,,)" to "call amulk$t (,,,2$f,,,) to remove a type dependence problem found by ftoc on the Mac. math$curfit/cvpower.gx: Davis, May 6, 1991 Changed the constant from INDEFR to INDEF in the amov$t call in cvpower.gx. This was causing a problem for the Mac compiler. math$curfit/cv_b1eval.gx: Davis, April 23, 1991 math$curfit/cv_beval.gx math$curfit/cv_feval.gx math$curfit/cvaccum.gx math$curfit/cvacpts.gx math$curfit/cvchomat.gx math$curfit/cvfree.gx math$curfit/cvinit.gx Did some cleaning up in the following .gx files to make the code easier to read. math$nlfit/: Davis, Jan 31, 1991 Added the nlfit package to the math package. math$surfit/isreplace.x: Davis, September 18, 1990 1. Changed the int calls in isreplace.x to nint calls. This is a totally safe way to do the conversion from floating point to integer quantities in the surfit package and removes any potential precision problems for task which must read the surfit structure back from a text database file. math$gsurfit/gsrestore.gx,gsrestorer.x,gsrestored.x: Davis, September 18, 1990 1. Changed the int calls in gsrestore.gx to nint calls. This is a totally safe way to do the conversion from floating point to integer quantities in the gsurfit package and removes any potential precision problems for task which must read the gsurfit structure back from a text database file. math$curfit/cvrestore.gx,cvrestorer.x,cvrestored.x: Davis, September 18, 1990 1. Changed the int calls in cvrestore.gx to nint calls. This is a totally safe way to do the conversion from floating point to integer quantities in the curfit package and removes any potential precision problems for task which must read the curfit structure back from a text database file. math$gsurfit/gsder.gx,gsderd.x: Davis, August 1, 1990 1. Fixed a typo in the routine which computes the double precision derivatives of a surface. Due to this type the salloc routine was being passed a pointer instead of size. When the value of this pointer got very large this could cause an out of memory allocation error. math$iminterp/msigrl.x: Davis, Feb 8, 1989 1. There was a bug in the shift and wrap routines in the 2D routine which integrates in polygonal apertures. If the vertices of the polygon were listed in certain orders the routine would produce wrong results. The routine has been totally rewritten to use the polygon clipping code in polyphot. math$interp/asigrl.x: Davis, Jan 27, 1988 1. The routine ii_getpcoeff was duplicated in the files asigrl.x and ii_1dinteg.x. I removed the copy in asigrl.x and rebuilt the library. math$interp/arider.x: Davis, February 10, 1988 1. The routines iidr_poly3, iidr_poly5 and iidr_spine3 were declared as functions but called everywhere as subroutines. I removed the function declarations. math$gsurfit/gsder.gx: Davis, October 16, 1987 1. changed a real type conversion call to double in the gsder routine. This error was causing the derivative evaluations to fail on the sun. math$gsurfit/gsder.gx: Davis, August 11, 1987 1. A call in gsder to gscoeff was not being set to double precision. math$gsurfit: Davis, January 22, 1986 1. A new version of the gsurfit package was installed. A double precision gsurfit has been added. The double precision calls are the same as the single precision calls but prefixed by a d. Some of the evaluation routines were vectorized and run significantly faster than the old version. A power series polynomial function was added to the package for convenience in evaluating the simple rotation transformation functions. math$iminterp/arider.x: Davis, September 7, 1986 1. The routines ii_pcpoly3, ii_pcpoly5 and ii_pcspline3 were declared as functions but called everywhere as subroutines. Removed the function declaration. I also moved these routines into a file of their own ii_pc1deval.x and changed the prefix to ia_ from ii_ to avoid conflict with the 2D equivalents. math$interp/arider.x: Davis, September 7, 1986 1. The routines ii_pcpoly3, ii_pcpoly5 and ii_pcspline3 were declared as functions but called everywhere as subroutines. Removed the function declaration. From Davis Jan 8, 1986 1. Added documentation for surfit package. ---------------------------------------------------------------------------- .endhelp