MINOR: Avoid double null check in KStream::transform()

[cadonna]

A call to KStream::transform() checked twice the parameter
transformerSupplier for null. The double null check occurred
because KStream::transform() is expressed in terms of
KStream::flatTransform() and both methods performed the null check.

More detailed description of your change,
if necessary. The PR title and PR message become
the squashed commit message, so use a separate
comment to ping reviewers.

Summary of testing strategy (including rationale)
for the feature or bug fix. Unit and/or integration
tests are expected for any behaviour change and
system tests should be considered for larger changes.

Committer Checklist (excluded from commit message)

• Verify design and implementation
• Verify test coverage and CI build status
• Verify documentation (including upgrade notes)

[bbejeck]

\cc @vvcephei and @ableegoldman for reviews

[mjsax]

I see the intent, but the disadvantage is, that we get deeper stack traces, ie, the user will see doFlatTransform() throwing a NPE, however, user calls either transform() or flatTransform(). I personally prefer the current code, because it's give better error messages to users IMHO. It's subtle of course, especially for this case, as the stack trace will be only one level deeper. I still think, it's easier for users to see

NPE
  at ...transform()
...

instead of

NPE
  at ...doTransform()
  at ...transform()
...

Also, the PR only allows to share one line of code (so the benefits of the rewrite is small).

Thoughts?

[cadonna]

@mjsax thank you for your comment.

The NPEs are still thrown by transform and flatTransform. So the stack trace will not get deeper. Furthermore, the PR removes one line from the shared code rather than adding one, namely the null check in flatTransform (see line 452 in the original code).

[mjsax]

Ack.

Guess -- maybe it was too late to read a diff...