[icweaver]

Hi Chris, here are two changes to header parsing/construction that came up while I was porting Reproject.jl from WCS.jl to FITSWCS.jl.

1. Standard-compliant RADESYS defaulting

Previously, an absent RADESYS always defaulted to "ICRS". The FITS WCS standard (Paper II, Section 3.1, and wcslib's wcsset()) instead prescribes:

Header	wcs.radesys before	wcs.radesys now
Equatorial/ecliptic, no RADESYS, no EQUINOX	"ICRS"	"ICRS"
Equatorial/ecliptic, no RADESYS, EQUINOX < 1984	"ICRS"	"FK4"
Equatorial/ecliptic, no RADESYS, EQUINOX >= 1984	"ICRS"	"FK5"
Non-equatorial (galactic, linear, spectral, ...), no RADESYS	"ICRS"	""
Explicit RADESYS	kept	kept (unchanged)

2. pv keyword in WCS(naxis; ...)

The keyword constructor now accepts pv = [(axis, m, value), ...], mapped to the corresponding PVi_m header keywords. The same convention as WCS.jl's WCSTransform(naxis; pv = ...), closing a small migration gap:

WCS(2; ctype = ["RA---AIR", "DEC--AIR"], crval = [0.0, -90.0], pv = [(2, 1, 45.0)])

Housekeeping

• [compat] FITSFiles = "0.2, 0.3" widened to "0.2, 0.3, 0.4" for the upcoming FITSFiles 0.4.0 also needed for the Reproject.jl port
• Docstrings updated to match the new defaulting behavior.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 86.55%. Comparing base (6d33b92) to head (59adeaa).

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #19      +/-   ##
==========================================
+ Coverage   86.46%   86.55%   +0.09%     
==========================================
  Files          18       18              
  Lines        3376     3385       +9     
==========================================
+ Hits         2919     2930      +11     
+ Misses        457      455       -2     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[cgarling]

Thanks, I'm currently in the process of moving and starting a new job so I may not be able to review and merge this for a week or so. Let me know if that's a problem for you

[icweaver]

Not at all, congrats again on your new position! Hope the travels go smoothly

[cgarling]

Thanks, this looks good

One minor thing I noticed is that wcslib also clears equinox to UNDEFINED when RADESYS is "ICRS" or "GAPPT", and sets default equinox values for FK4/FK5 when equinox is undefined. This PR handles the main case (RADESYS defaulting from EQUINOX) correctly; the reverse-direction defaulting could be a follow-up if needed.