
2017.01.09 : EAM

  * ppMerge : comment out a free (this introduces a leak and should be conditional)
  * stacktool : sort the skycells used for a summary skycell by skycell_id to ensure lists match.
  * psLib : 
    psMetadata.c : F64 data needs to be printed with %g not %f
    psFitsScale.c : allow quartile stats to return a sigma based on +/- 1-sigma range
  * psastro : 
    fit a mosaic-wide distortion model, if it exists (e.g., RA---ZPN), otherwise maintain the mosaic fit model
    allow missing model for a given chip (e.g., model not defined for bad chips)
    save the calibrated magnitudes in MATCHED_REFS (using measured zero point)
    add ROT_PARITY concept to the astrometry model (defined sign of ROTANGLE)
    tpa->fpa transformation needs to have more polynomial terms than fpa->tpa (where the fit is performed)
  * ippScripts : add HSC to cameras not corrected for background (in camera_exp.pl, should be a recipe-thing) 
    
  * Ohana:
    various changes to relastro, relphot, fakeastro, addstar, delstar, uniphot, photdbc, dvomerge which do not affect nightly processing
    libdvo/coorops.c : upgrade ZPN to full terms (remove UKIRT exemption)
    libdvo : add test galaxy models; add object / measure bits for GAIA, 2MASS, TYCHO, SDSS, HSC, CFH, DEC
    libfits : do not allow gfits_modify, etc to match FOO to FOOBAR, etc
    libkapa & kapa2 : change some hardwired numbers to enum values, allow names for e.g., -pt and -lt
    opihi : 
      unify random number seeding; 
      allow user-defined names for header NAXISi keywords for imbox; 
      fix csystem conversion of motions
      allow sphere-based 2d matches (matches2d)
      define some physical constants as variables at startup (e.g., M_PI, M_c, M_c_cgs)
      
  * psModules:
    allow photometry-based match in astrometry : downweight photometrically discrepant values
    add ZPN model
    do not attempt to force CRPIX1,2 to be 0.0
    support for ROT_PARITY
    add some tracing for detrending
    
  * Nebulous & Nebulous-Server : changes to support proximity 
  
  * ippconfig : 
    GAIA and HSC photcodes
    support for ROT_PARITY and psastro photometry window
    accurate HSC zero points

2010.10.25 : EAM

  Merging development work from the past 2 months from my branch
  (branches/eam_branches/ipp-20100823) into the trunk.  This is work
  that covers a range of issues:

  Ohana:
    * remove some deprecated code (gfits_downsize_and_convert_$STRUCT)
    * add support for CMF_PS1_V3 files from psphot to addstar
    * clean up BYTE_SWAP existence tests
    * auto-init for RotFonts
    * kapa improvements: hex/float button, better tick marks & box
      labels; better positioning of graphs and images, clip contours
    * add %e,%g,%o,%i,%u,%x to opihi fprintf
    * add 'extra host' concept to pcontrol (additional processing
      restrictions)
    * plug some memory leaks
    * add opihi 'reindex' function for working with sorted indexes
    * re-include bond version of svd
    
  psLib:
    * add options for colorbar and scale label to psImageJpeg
    * plug memory leak in psMinimizeLMM	
    * add libkapa support for psImageJpeg (bDraw functions)

  psModules/imcombine:
    * added psf-match stats to output headers; clarified their meaning
    * replace pmSubtractionMaskInvalid with pmReadoutMaskInvalid (masks
      only unmasked NANs)
    * cleanup pmSubtractionMatch to remove various deprecated
      (if-def-ed out) blocks (especially the SEPARATE sections)
    * save fwhm independently from pmSubtractionParamsScale
    * repace central delta function re-zero with 0th gaussian re-zero
    * re-work matrix to only include the fitted terms (not the
      normalization or bg, which are done elsewhere)
    * replace diagonal penalties for DUAL with combined 2nd moments
      minimization
    * add function to calculate the moments (needed by DUAL
      minimization)
    * modify stack mask to generate OR of inputs for BAD outputs
      and fuzzy AND over 65% of inputs for not BAD outputs
    * clarify visualization functions
    * weighting window uses sampleMedian (but is not actually used)
    * determine flux normalization from a consistent pair of
      apertures, selected on the basis of the 1st radial moment
    * PSF envelope calculation does not apply a S/N limit for pixels

  psModules/objects:
    * add aperture fluxes (for insignificant and diff objects)
    * re-work pmSourceMoments to measure centroid with a tophat first
      then a gaussian window centered on the initial centroid
    * option to skip aperture magnitudes in pmSourceMag
    * finish CMF_PS1_V3 format
    * replace peak position with centroid position were available
    * replace peak position error with mean fwhm, scaled by S/N
    * add some fields to PS1_DV2
    * add S/N limit to type DEFECT assignment
    
  psModules/other:
    * update jpeg creation to select FLIPS, COLORBAR & SCALE options
      (ready to take object overlays in bDraw format)
    * add function pmReadoutMaskInvalid to replace
      pmReadoutMaskNonfinite in most cases
    * fix possible segfault in pmVisual.c
     
  psphot:
    * allow negative flux in src/psphotFitSourcesLinearStack.c
    * add header outputs for source size analysis metadata
    * simplify some visualization routines
    * use M_minor and Kron magnitude for CR classification
    * dynamically set CR classif. limits
    * use psf-kron mag for EXT classification
    * mask detected CRs
    * use full image for source classification (do not use sub-regions
      since we do not use the moments)
    * remember to subtract sources when doing extended source fitting
     (even failed sources)
    * plug some memory leaks
    * add flag to show that a source was found on another image (ie,
      this version of the source was supplied not detected) 

  ppSub
    * failure to re-scale is now an error, not a warning
    * recipe option to specify convolution target	
    * use pmReadoutMaskInvalid instead of pmReadoutMaskNonfinite
    * cleanup visualization and plug some leaks

  ppStack
    * allow recipe option to specify target PSF size
    * masking improvements (see psModules above)
    * avoid possible segfault in ppStackCombineFinal.c
    * set FWHMs separately from pmSubtractionParamsScale

  ippTools
    * difftool : handle case of multiple existing runs (rerun mode for
      definestackstack)
    * camtool : add -data_group selection
    * difftool : allow -bothways options for warpstack and stackstack; 
      	       -not-bothways for warpwarp

  ippconfig
    * add FPA.ZP to simtest camera
    * default output format: PS1_V3 (chip) & PS1_DV2 (diff)
    * add CR detection params
    * stack output options for RAW or COMP_IMG (filerules-simple.mdc only)
    * JPEG options
    * reduce ppSub kernel size, autoscale, use photpipe kernel sizes,
    * options for PSF manual scaling
    * adjust slope of bright-star mask to better match fainter
      saturated stars
