fix(python): Don't modify user-supplied storage_options dict (take a shallow-copy)

[alexander-beedie]

Closes #15849.
Closes #15850.

[stinodego]

Can you add a test for this?

Ideally, we don't always copy, but we create a copy when we add the encoding entry. But that might be micromanaging things 😬

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.37%. Comparing base (7eadccb) to head (46156a2).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #15859      +/-   ##
==========================================
+ Coverage   80.39%   81.37%   +0.98%     
==========================================
  Files        1263     1378     +115     
  Lines      165387   176827   +11440     
  Branches        0     3032    +3032     
==========================================
+ Hits       132957   143898   +10941     
- Misses      32430    32444      +14     
- Partials        0      485     +485     

Flag	Coverage Δ
python	74.76% <100.00%> (?)
rust	78.50% <100.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[alexander-beedie]

Can you add a test for this?

Done (I think; slightly hard to prove locally as I can't run the AWS tests from here, but the amended test should trigger the relevant codepath).

Ideally, we don't always copy, but we create a copy when we add the encoding entry. But that might be micromanaging things 😬

It's probably safer to do it for everything that comes into this function; otherwise we may end up having the same error again if we extend use of storage_options (and a shallow-copy is cheap for that peace of mind ;))

[stinodego]

Looks good, thanks!

I'll merge when I've fixed the CI.