handle noop correctly

[mattspataro]

We recently migrated our codebase from OpenTracing to OpenTelemetry and were delighted to take advantage of the OpenTracingShim. Unfortunately we ran into issues with noop spans.

There are a few instances in which we register new spans as children of a io.opentracing.noop.NoopSpan. Unfortunately, when this happens, the OpenTelemetryShim attempts to turn the parent NoopSpan into a SpanShim, which throws an exception. This issue can be mitigated by performing a noop if we attempt to register a NoopSpan as a parent. It appears that this was meant to be handled anyways, judging from the following TODO comment that already existed in the codebase:
// TODO - Verify we handle a no-op Span

This PR fixes this issue and provides a unit test to verify that it works correctly.

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: mattspataro / name: Matt Spataro (9529a74)

[codecov]

Codecov Report

Base: 91.26% // Head: 91.26% // No change to project coverage 👍

Coverage data is based on head (ff4338c) compared to base (ac2ba4a).
Patch coverage: 100.00% of modified lines in pull request are covered.

Additional details and impacted files

@@            Coverage Diff            @@
##               main    #4968   +/-   ##
=========================================
  Coverage     91.26%   91.26%           
- Complexity     4885     4886    +1     
=========================================
  Files           552      552           
  Lines         14431    14431           
  Branches       1373     1373           
=========================================
  Hits          13170    13170           
  Misses          874      874           
  Partials        387      387           

Impacted Files	Coverage Δ
...opentelemetry/opentracingshim/SpanBuilderShim.java	93.79% <100.00%> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[jkwatson]

@carlosalberto can you take a look at this one?

[carlosalberto]

So this is an undocumented/corner-case of the OpenTracing API - for safety reasons, though, I'm happy to accept it - (I will add a note in the Spec's section for this, so other languages support this scenario as well).

(OpenTracing's No-op Span is not quite the same as our OpenTelemetry's Span.getInvalid())

[mattspataro]

@carlosalberto

So this is an undocumented/corner-case of the OpenTracing API - for safety reasons, though, I'm happy to accept it - (I will add a note in the Spec's section for this, so other languages support this scenario as well).

(OpenTracing's No-op Span is not quite the same as our OpenTelemetry's Span.getInvalid())

I see you what you mean. I realized later that the java doc for io.opentracing.Tracer.SpanBuilder.asChildOf says: "If parent==null, this is a noop." So our code-base should've been passing in null for noop spans anyways. But still, I think this could be helpful for those that make the same mistake as us and accidentally pass in a noop span.