AVRO-2944: Fix read pointer in java reader to avoid hang on partial read input stream

[MickJermsurawong]

Make sure you have checked all steps below.

Jira

• My PR addresses the https://issues.apache.org/jira/browse/AVRO-2944

Tests

• My PR adds the following one unit test to first illustrate failure

Commits

• My commits all reference Jira issues in their subject lines. In addition, my commits follow the guidelines from "How to write a good git commit message":
  1. Subject is separated from body by a blank line
  2. Subject is limited to 50 characters (not including Jira issue reference)
  3. Subject does not end with a period
  4. Subject uses the imperative mood ("add", not "adding")
  5. Body wraps at 72 characters
  6. Body explains "what" and "why", not "how"

Documentation

• In case of new functionality, my PR adds documentation that describes how to use it.
  • All the public functions and the classes in the PR contain Javadoc that explain what it does

[MickJermsurawong]

The CI failure seems unrelated to the change in this PR.
For java 8:

Trevni Specification ............................... FAILURE [  3.264 s]

For java 11:

Updating dependencies
Your requirements could not be resolved to an installable set of packages.
  Problem 1
    - beberlei/composer-monorepo-plugin[v0.12, ..., v0.12.1] require composer-plugin-api ^1.0 -> found composer-plugin-api[2.0.0] but it does not match the constraint.
    - Root composer.json requires beberlei/composer-monorepo-plugin ^0.12 -> satisfiable by beberlei/composer-monorepo-plugin[v0.12, v0.12.1].

[RyanSkraba]

LGTM! Good catch, and nice unit test, thanks for the contribution!

[mickjermsurawong-stripe]

Hi @RyanSkraba just wondering if this patch could be back-ported to older versions too? We are using Iceberg that still uses Avro 1.9, and Hadoop with Avro 1.7.

[RyanSkraba]

Hello! I got caught up taking a look at the build failure -- it's obviously nothing to do with your PR, but I would have liked to see it green.

There isn't an official 1.7.x release or even 1.9.3 planned as far as I know. We'd have to bring it up on the mailing list!

[Fokko]

@MickJermsurawong can you rebase? The CI has been fixed on master

[MickJermsurawong]

Ah thank you Fokko for the note. Ryan, it's green now. Thank you for the review!
Will send out a note to dev list to see if folks have run into this problem.

[RyanSkraba]

Cherry-picked to branch-1.10! Thanks again for this fix!

[noslowerdna]

@MickJermsurawong / @RyanSkraba I asked a question on the JIRA, curious if this fix could be improved to cover the read returning -1 unexpectedly as the issue description indicates it ought to.