[TRACE SDK] Batch span processor options now using env variables

[nikhilbhatia08]

Fixes #2085

Using this change we can get max_queue_size schedule_delay export_timeout max_export_batch_size using environment variables. BatchSpanProcessor still does not use export_timeout and tests have been added.

Changes

Please provide a brief description of the changes here.

For significant contributions please make sure you have completed the following items:

• CHANGELOG.md updated for non-trivial changes
• Unit tests have been added
• Changes in public API reviewed

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.18%. Comparing base (ecb155b) to head (f42ac79).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3661      +/-   ##
==========================================
+ Coverage   90.11%   90.18%   +0.07%     
==========================================
  Files         221      222       +1     
  Lines        7135     7154      +19     
==========================================
+ Hits         6429     6451      +22     
+ Misses        706      703       -3     

Files with missing lines	Coverage Δ
sdk/src/trace/batch_span_processor_options.cc	100.00% <100.00%> (ø)

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

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

[marcalff]

Some changes needed:

The header file exposes implementation details, which are compiled into the application binary.

Move every implementation details, including default values, helpers, names of environment variables, to batch_span_processor_options.cc

See existing options structures for the pattern.

[nikhilbhatia08]

@marcalff Applied the above changes, Thanks! But the BatchSpanProcessor does not have the field of export_timeout so it would be another issue for implementing it in BatchSpanProcessor ?

[marcalff]

Looking much better.

See minor cleanup,
I will approve once I can verify the changes.

[marcalff]

LGTM, thanks for the fix.

[georgthegreat]

@marcalff @nikhilbhatia08, this PR breaks a bunch of existing code without providing a sensible alternative.

Prior to that we were using the following syntax using C++20 aggregate initialization:

 auto processor = otel::BatchSpanProcessorFactory::Create(
            std::move(exporter),
            otel::BatchSpanProcessorOptions{
                .max_queue_size = config.batch().max_queue_size(),
                .schedule_delay_millis = std::chrono::milliseconds(config.batch().schedule_delay().MilliSeconds()),
                .max_export_batch_size = config.batch().max_export_batch_size(),
            }
        );

It does not seem that there is a sensible rewrite of this code.
Maybe this PR can be reverted?

[georgthegreat]

NB: #3687 has the same problem

[lalitb]

I believe the issue comes from adding a user-defined constructor, which means BatchSpanProcessorOptions is no longer an aggregate type. That disables C++20 designated initializers, so patterns like .max_queue_size = ... stop working even though they worked earlier.
We can fix this without reverting anything by removing the constructor and using default member initializers along with a small set of static helper functions for the environment variable logic. This restores aggregate initialization while keeping the original behavior intact.

[georgthegreat]

Seems like a solution to me!