Enhance IcebergProbeImpl with jar source logging and version fallback

[gerashegalov]

Fixes #14487

Requires #14494

Description

Fallback on the version in the jar file name

Checklists

• This PR has added documentation for new or modified features or behaviors.
• This PR has added new tests or modified existing tests to cover new code paths.
  (Please explain in the PR description how the new code paths are tested, such as names of the new/existing tests that cover them.)
• Performance testing has been performed and its results are added in the PR description. Or, an issue has been filed with a link in the PR description.

[greptile-apps]

Greptile Summary

This PR enhances IcebergProbeImpl with a fallback mechanism that extracts the Iceberg version from the jar file name when a git commit ID is not found in the hard-coded commitToVersion map, and adds Logging to emit a warning when the fallback is used.

Key changes:

• Adds with Logging to IcebergProbeImpl so Spark's standard log infrastructure is available.
• Adds jarVersionPattern (a lazy Regex with .unanchored) matching the canonical iceberg-spark-runtime jar naming scheme (e.g. iceberg-spark-runtime-3.5_2.12-1.10.0-...jar).
• Adds extractVersionFromJarPath(commitId), which defensively chains Option(getProtectionDomain).flatMap(…getCodeSource).flatMap(…getLocation) to avoid NPEs documented for all three calls, then uses collectFirst with the pattern to extract the version string.
• getDetectedVersion now calls extractVersionFromJarPath as the second-level fallback when the commit map lookup misses; the ultimate fallback is still the raw commit ID, preserving the previous UnsupportedOperationException behaviour in shimPackage.
• The previous review concerns about null safety (getProtectionDomain, getCodeSource, getLocation) and the missing isSupportedSparkVersion in the trait are fully addressed in this revision.

Confidence Score: 4/5

Safe to merge after considering a minor duplicate-warning issue; no correctness or data-loss risks.

All previously flagged P0/P1 concerns (null safety, trait mismatch) are resolved. The one remaining finding is a P2 style issue: the fallback warning can fire twice due to independent lazy val computations in IcebergProvider. This is cosmetic and does not affect correctness, but keeping the score at 4 to prompt the author to consider caching getDetectedVersion.

iceberg/common/src/main/scala/com/nvidia/spark/rapids/iceberg/IcebergProbeImpl.scala — duplicate warning log in fallback path.

Important Files Changed

Filename	Overview
iceberg/common/src/main/scala/com/nvidia/spark/rapids/iceberg/IcebergProbeImpl.scala	Adds jar-path-based version fallback with proper null-safety for getProtectionDomain/getCodeSource/getLocation, and mixes in Spark's Logging trait; the warning log can fire twice when both IcebergProvider.detectedVersion and IcebergProvider.shimPackage are accessed.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[getDetectedVersion] --> B{commitId in commitToVersion?}
    B -- yes --> C[return version string]
    B -- no --> D[extractVersionFromJarPath]
    D --> E{getProtectionDomain != null?}
    E -- no --> F[return commitId as-is]
    E -- yes --> G{getCodeSource != null?}
    G -- no --> F
    G -- yes --> H{getLocation != null?}
    H -- no --> F
    H -- yes --> I[url.getPath]
    I --> J{jarVersionPattern matches?}
    J -- no --> F
    J -- yes --> K[logWarning + return version]

Loading

_{Reviews (7): Last reviewed commit: "Refactor version extraction logic in Ice..." | Re-trigger Greptile}

[gerashegalov]

build