[
https://issues.apache.org/jira/browse/HBASE-8584?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13663407#comment-13663407
]
Aleksandr Shulman commented on HBASE-8584:
------------------------------------------
Although code coverage is a useful metric to tell us if we've skipped testing
areas of the code, I think adding methods that merely touch the code and
nothing else might actually do more harm that good.
In situations like these, a good question to ask might be what kind of bugs
would this test actually catch (that other tests would not)?
In my mind, the pros of adding these tests must outweigh the cons. The cons of
adding more tests just to improve coverage is:
1. More tests means longer running time and more log/report noise
2. If we don't have testing this in this area, let's add useful tests that will
catch regressions that will affect user workflows. Adding tests that merely
touch the code might mask a real need for testing and give us a false sense of
security.
That said, I'm open to hearing the upside of these tests and what benefits they
may offer us and I think it's great that we're paying attention to these
metrics.
> increase unit-test coverage of class
> org.apache.hadoop.hbase.coprocessor.BaseMasterObserver
> -------------------------------------------------------------------------------------------
>
> Key: HBASE-8584
> URL: https://issues.apache.org/jira/browse/HBASE-8584
> Project: HBase
> Issue Type: Test
> Affects Versions: 0.98.0, 0.95.1, 0.94.9
> Reporter: Ivan A. Veselovsky
> Attachments: HBASE-8584-0.94--N1.patch, HBASE-8584-trunk--N1.patch
>
>
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
