Fix JDBC non-supported getUpdateCount

[eyalkoren]

Fixes #1006

• Implement
• Added tests
• Add to CHANGELOG

[codecov-io]

Codecov Report

❗ No coverage uploaded for pull request base (master@25b949a). Click here to learn what that means.
The diff coverage is 55.61%.

@@            Coverage Diff            @@
##             master    #1008   +/-   ##
=========================================
  Coverage          ?   62.62%           
  Complexity        ?       85           
=========================================
  Files             ?      271           
  Lines             ?    11010           
  Branches          ?     1463           
=========================================
  Hits              ?     6895           
  Misses            ?     3679           
  Partials          ?      436

Impacted Files	Coverage Δ	Complexity Δ
...va/co/elastic/apm/agent/impl/ElasticApmTracer.java	70.33% <ø> (ø)	0 <0> (?)
...m/agent/kafka/BaseKafkaHeadersInstrumentation.java	0% <ø> (ø)	0 <0> (?)
.../kafka/ConsumerRecordsIteratorInstrumentation.java	0% <0%> (ø)	0 <0> (?)
...afka/ConsumerRecordsRecordListInstrumentation.java	0% <0%> (ø)	0 <0> (?)
...gent/okhttp/OkHttp3ClientAsyncInstrumentation.java	21.73% <0%> (ø)	0 <0> (?)
.../apm/agent/okhttp/OkHttpClientInstrumentation.java	17.85% <0%> (ø)	0 <0> (?)
...esttemplate/SpringRestTemplateInstrumentation.java	36% <0%> (ø)	0 <0> (?)
.../ApacheHttpAsyncClientRedirectInstrumentation.java	50% <0%> (ø)	0 <0> (?)
...t/opentracing/impl/SpanContextInstrumentation.java	0% <0%> (ø)	0 <0> (?)
...acing/impl/ExternalSpanContextInstrumentation.java	0% <0%> (ø)	0 <0> (?)
... and 21 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 25b949a...21664eb. Read the comment docs.

[SylvainJuge]

I'm wondering if we could have something simpler in the advice code. Maybe moving the call to getUpdateCount alone in the try/catch would make it a bit more readable, and maybe moving this logic to a dedicated helper class.

Also, it's definitely missing one comment that explain why we do this in this case, maybe a unit test would also help to understand here.

[SylvainJuge]

this should never happen, we can remove the null check here.

[SylvainJuge]

Looks good to me, only the readme update is missing.