fix CopySource unquoting of special + char

[bentsku]

Motivation

As reported with #9990, we had an issue when copying a file with a space in the name with Go. This issue was not present with Python however. When looking at the sample provided by the user, it appeared that Go in URL encoding space character with + when Python is using %20.
When calling unquote with Python, it won't replace an encoded URL + character with a space. We will need to manually replace it before unquote the URL to not accidentally replaced an encoded +.

Changes

Manually replace an encoded URL + when parsing the CopySource header.
Added 2 different AWS tests to verify exactly the fix:

• one global with weird key encoding as we didn't have one
• one very specific test for this use case using the HTTP client because botocore will automatically encode space as %20 so we couldn't reproduce exactly the issue. This test fails without the fix, confirming it. This specific test is xfail for the v2 legacy provider as moto does not handle this edge case (so not a regression from the new provider)

fixes #9990

[github-actions]

S3 Image Test Results (AMD64 / ARM64)

  2 files    2 suites   3m 9s ⏱️
385 tests 335 ✅  50 💤 0 ❌
770 runs  670 ✅ 100 💤 0 ❌

Results for commit 155e7eb.

♻️ This comment has been updated with latest results.

[github-actions]

LocalStack Community integration with Pro

    2 files      2 suites   1h 15m 35s ⏱️
2 434 tests 2 208 ✅ 226 💤 0 ❌
2 435 runs  2 208 ✅ 227 💤 0 ❌

Results for commit 155e7eb.

♻️ This comment has been updated with latest results.

[steffyP]

Thanks for fixing, @bentsku!
LGTM! Nice to have additional tests 👍