[ZEPPELIN-6259] Add null safety checks in SparkInterpreterLauncher.detectSparkScalaVersionByReplClass

[renechoi]

What is this PR for?

This PR adds null safety checks to the detectSparkScalaVersionByReplClass method in SparkInterpreterLauncher.java to prevent NullPointerException and provide clear error messages when the Spark jars directory is inaccessible.

Current Issues Fixed:

1. NullPointerException Risk: listFiles() returns null when directory doesn't exist or is inaccessible
2. Poor Error Messages: Users get cryptic NPE instead of meaningful error messages
3. Missing Validation: No checks for directory existence or type

What type of PR is it?

Bug Fix / Improvement

Todos

• - Code review
• - CI build verification

What is the Jira issue?

• https://issues.apache.org/jira/browse/ZEPPELIN-6259

How should this be tested?

• Unit tests added:

  • testDetectSparkScalaVersionByReplClassWithNonExistentDirectory - Verifies error when directory doesn't exist
  • testDetectSparkScalaVersionByReplClassWithFileInsteadOfDirectory - Verifies error when path is a file
  • testDetectSparkScalaVersionByReplClassWithValidDirectory - Verifies normal operation works
  • testDetectSparkScalaVersionByReplClassWithEmptyDirectory - Verifies error when no spark-repl jars found

• Manual testing:

  • Set invalid SPARK_HOME and verify clear error messages
  • Remove read permissions on SPARK_HOME/jars and verify permission error

• CI: All existing tests pass

Screenshots (if appropriate)

N/A

Questions:

• Does the license files need to update? No
• Is there breaking changes for older versions? No
• Does this needs documentation? No

Implementation Details

• Added directory existence check with clear error message
• Added directory type validation (ensures it's not a file)
• Added null check for listFiles() result with permission hint
• All error messages include the problematic path for easier debugging
• Used IOException for file system related errors (consistent with Java conventions)

Error Message Examples

Before (NPE):
java.lang.NullPointerException
at java.util.stream.Stream.of(Stream.java:1012)

After (Clear messages):
java.io.IOException: Spark jars directory does not exist: /invalid/path/jars. Please check your SPARK_HOME setting.
java.io.IOException: Spark jars path is not a directory: /some/file/jars
java.io.IOException: Cannot access Spark jars directory: /restricted/jars. Please check permissions.

Benefits

1. Better User Experience: Clear error messages help users quickly identify and fix configuration issues
2. Defensive Programming: Prevents crashes from null pointer exceptions
3. Easier Debugging: Specific error messages with paths make troubleshooting straightforward
4. Production Ready: Handles edge cases that can occur in various deployment environments

[Reamer]

Perhaps it is time to use the Files and Paths classes here?