Reduce S3 Calls (GSI-2490)

[TheByronHimes]

Re-evaluated every S3 provider operation and removed defensive existence checks where the main S3 call already fails with a translatable error.

Removed prechecks

• delete_bucket and list_all_object_ids rely on NoSuchBucket
• list_parts, abort_multipart_upload, and complete_multipart_upload rely on NoSuchUpload
• get_object_etag/get_object_size/get_object_metadata rely on head_object failing

Optimized

• _assert_object_(not_)exists now checks the object first and only checks the bus one call instead of two

Kept (with comments explaining why)

• create_bucket - S3 silently succeeds re-creating an owned bucket, so we let the app handle that error
• Presigned URL ops - signing is local, nothing server-side would raise if the upload/bucket didn't exist
• init_multipart_upload - to prevent duplicate uploads
• delete_object - so the application decides how to interpret missing objects
• copy_object dest check - to prevent overwriting things by accident

Note for near future

We should consider removing MultipleActiveUploadsError. The built-in check for multiple active uploads doesn't give us much. The methods already target specific uploads via upload_id, and if an application needs enforce upload exclusivity then that should be done application-side. That would mandate a major version bump though.

Version bumped to 8.5.0

[coveralls]

Coverage Report for CI Build 29416503475

Warning

No base build found for commit c26d86f on main.
Coverage changes can't be calculated without a base build.
If a base build is processing, this comment will update automatically when it completes.

Coverage: 93.664%

Details

• Patch coverage: 1 uncovered change across 1 file (28 of 29 lines covered, 96.55%).

Uncovered Changes

File	Changed	Covered	%
src/hexkit/providers/s3/provider/objstorage.py	16	15	93.75%
Total (3 files)	29	28	96.55%

Coverage Regressions

Requires a base build to compare against. How to fix this →

---

Coverage Stats

Relevant Lines:	3977
Covered Lines:	3725
Line Coverage:	93.66%
Coverage Strength:	0.94 hits per line

---

💛 - Coveralls

[mephenor]

Looks good, I just some additional places that could use slight improvement.

Do the existing tests not cover all the failure modes for the provider?
If so, should we extend them?