Is this backportable?

Most likely yes, it is just one line, and #5053 came in v1.2 so it should be fine to backport to 2.0.

Codecov Report

Merging #8563 into master will increase coverage by <.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #8563      +/-   ##
==========================================
+ Coverage   86.81%   86.81%   +<.01%     
==========================================
  Files         386      386              
  Lines       58162    58161       -1     
  Branches     1060     1060              
==========================================
+ Hits        50493    50494       +1     
+ Misses       7054     7052       -2     
  Partials      615      615

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 8824789...cb08b9a. Read the comment docs.

Tests passed but 😬

@saimn , you marked this as refactoring, not a bug. Generally speaking, we don't backport "refactoring". And if this is a bug, please update the labels and add a change log entry. Thanks!

@pllim - This is not a bug for me, there is no issue caused by this, so I'm fine with not backporting.

I agree with the not-backporting. Just in case (this is io.fits after all) this has some weird edge cases where it breaks some code.

I think this is fine.

I think this is fine.

@MSeifert04 , do you want to formalize it and hit "approve"? 😄

Maybe we could leave a bread-crump comment there that the data setter actually updates the header? This is completely optional, just to avoid potential confusion?!

I updated the comment just above.

just to avoid potential confusion?!

Maybe, but I love how the comment says that bitpix/bscale/bzero will be updated when setting the data, and just after it sets again those values in case they changed 😬

🤞 🤞 🤞