fix(storage): do not close object descriptor span on reader completion

[bajajneha27]

Problem
AsyncObjectDescriptorConnectionTracing::Read() was wrapping the returned reader in MakeTracingReaderConnection(span_, ...). AsyncReaderConnectionTracing ends the span as soon as the reader stream finishes. Because span_ represents the storage::AsyncConnection::Open span, it was ending prematurely on the first range completion instead of staying active for the lifetime of ObjectDescriptorConnection.

Fix
Returned impl_->Read(p) directly in AsyncObjectDescriptorConnectionTracing::Read(). Individual range reads are already traced by ObjectDescriptorImpl::Read(), which creates MakeTracingObjectDescriptorReader() when tracing is enabled.

[gemini-code-assist]

Code Review

This pull request simplifies the Read method in AsyncObjectDescriptorConnectionTracing by returning the result of impl_->Read(p) directly instead of wrapping it in a tracing reader connection. It also updates the corresponding unit tests to verify single and multiple read ranges, ensuring the open span is not prematurely ended. The review feedback suggests a safer and more idiomatic C++ approach in the tests by using std::make_unique and a mutable lambda capture instead of raw pointers to prevent potential memory leaks.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 92.24%. Comparing base (c49c86f) to head (09fd4ca).

Additional details and impacted files

@@           Coverage Diff           @@
##             main   #16321   +/-   ##
=======================================
  Coverage   92.23%   92.24%           
=======================================
  Files        2227     2227           
  Lines      209169   209203   +34     
=======================================
+ Hits       192929   192969   +40     
+ Misses      16240    16234    -6     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[v-pratap]

I was running a few experiments, and with the previous code I ran the Open of async and I got the trace, attached in the screenshot. I think it is correct and expected behaviour, Can you add a screenshot of trace before and after this PR change. And please let me know if my understanding is wrong. Also the error(red bar) is expected.
Trace: link

[bajajneha27]

Thanks for sharing the image.
Here's the image before the changes:

Notice that the Open span is ending before the last ReadRange span.

Image after the changes:

Let's discuss this in our sync today.