Changelog

0.85:
	One tiny bit of GPL v2-only code was replaced with a BSD equivalent.

	The primary branch on Github has been changed to 'main'.

	Minor: an-pnmtofits adds a CTYPE3='RGB' card for the Aladin image
	viewer.

0.84:
	Minor: change the python 'multiproc' module to handle >2GB pickles.

0.83:
	Minor: re-add -K flag to build-astrometry-index to keep RA,Dec columns.

0.82:
	Simplify os-features tests by always using our vendored qsort_r.

0.81:
	All 'sextractor' command-line args renamed to 'source-extractor'.

	The word "blind" (in the ableist sense in which we were using it)
	has almost entirely been removed.

0.69:
	Tiny bugfixes over 0.68

0.68:
	Add extra driver programs for python executables (image2pnm, etc), named X rather than X.py.  Some python3 compatibility fixes.

0.67:
	Minor; python calls for WCS parsing from fitsio headers.  Tag for legacysurvey.org Data Release 3.

0.66:
	Minor; bugfix to enable building on Fedora Rawhide

0.65:
	Minor; add wcs_pv2sip; fixes for Solaris.

0.64:
	Minor; add sip_scale.  Tag for legacysurvey.org Data Release 2.

0.63:
	Additional minor changes for Debian packaging.

0.62:
	More relicensing with permission from contributors; remove radix.c

0.61:
	Minor changes for Debian packaging.

0.60:
	Relicense most of the Astrometry.net code proper under a BSD-3
	license.  Whole package still under GPL-3+ thanks to GNU gsl.

0.57:
	Makefile fixes: create dirs before install; propagate build flags
	Add man pages from Ole Streicher
	Some py3 friendliness
	Add timingpool.py

0.56:
	implement anwcs_write_to
	Makefile files for install-indexes
	add -std=gnu89 to default CFLAGS
	add clip_wcs() method

0.55:
	Fix bug in WCS numpy array iterators

0.54:
	Try importing pyfits and then astropy.io.fits

0.53:
	Update wcs-pv2sip and include in default build
	Fixes for git version strings
	Read and write linear SIP distortion terms
	multiindex: don't assert when reloading already-loaded file

0.52:
	Fix release version strings in released versions.

0.51:
	Fix half-pixel bug in "solve-field --crpix-center": the center was
	set to 1 + {W,H}/2, not 0.5 + {W,H}/2.  (The WCS was still
	correct, the CRPIX was set to a half pixel in each dimension away
	from the center pixel.