fix: support ArrayBufferView and ArrayBuffer as body

[LiviaMedeiros]

Implementation-wise, do we really need to make a copy?

[codecov-commenter]

Codecov Report

Merging #1584 (f0da2be) into main (dd613ef) will decrease coverage by 0.02%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main    #1584      +/-   ##
==========================================
- Coverage   94.91%   94.89%   -0.03%     
==========================================
  Files          50       50              
  Lines        4724     4724              
==========================================
- Hits         4484     4483       -1     
- Misses        240      241       +1     

Impacted Files	Coverage Δ
lib/fetch/body.js	97.10% <100.00%> (ø)
lib/fetch/file.js	91.66% <0.00%> (-1.20%)	⬇️

---

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 dd613ef...f0da2be. Read the comment docs.

[ronag]

Needs a test

[KhafraDev]

Implementation-wise, do we really need to make a copy?

Yes, see extract:

BufferSource

    Set source to a copy of the bytes held by object.

[LiviaMedeiros]

Set source to a copy of the bytes held by object.

This is true spec-wise, but seems costly in implementation. I wonder if there is a workaround with better performance, considering that most other implementations do not expose Request.body directly.

Anyway, even if it's possible to optimize later, working code and spec compliance goes first.

[mcollina]

lgtm

[mcollina]

ping @ronag