BUG: Handle invalid masked_invalid mask

[pllim]

I came across this while investigating spacetelescope/jdaviz#1877 . np.ma.masked_invalid(img).mask could return False. This is not being recognized by specutils, which will then emit a confusing message about the mask having mismatch shape ().

Possible suspects:

• BUG: Make mask_invalid consistent with mask_where if copy is set to False numpy/numpy#22046
• DOC: Add changelog for masked_invalid change. numpy/numpy#22406

Also see:

• TST: Add dev numpy to devdeps, fix compat with numpy 1.24 #153

[codecov]

Codecov Report

Merging #151 (4d97ebd) into main (05a174d) will decrease coverage by 0.07%.
The diff coverage is 50.00%.

@@            Coverage Diff             @@
##             main     #151      +/-   ##
==========================================
- Coverage   75.46%   75.39%   -0.08%     
==========================================
  Files           9        9              
  Lines         693      695       +2     
==========================================
+ Hits          523      524       +1     
- Misses        170      171       +1     

Impacted Files	Coverage Δ
specreduce/core.py	73.17% <50.00%> (-1.19%)	⬇️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[tepickering]

looks like this change needs some test coverage to made codecov happy...

[pllim]

I don't know how to write that test. Suggestion welcome!

[pllim]

I folded this PR to #153