Add Spark 4.0.3 shim support

[firestarman]

Fixes #15065.

Description

• Add Spark 4.0.3 as a supported Scala 2.13 shim profile so the plugin can build and load against Spark 4.0.3.
• Add the Spark 4.0.3 shim service provider and generated support metadata so runtime shim discovery and qualification documentation include Spark 4.0.3.
• Split Spark 4.0.3 from the 4.0.1/4.0.2 SparkShims path so 4.0.3 can carry its Spark-specific behavior without affecting earlier 4.0.x shims.
• Share ACOSH/ASINH compatibility overrides and boundary tests with Spark 4.1.2 because Spark 4.0.3 has the same CPU behavior for large hyperbolic inputs.
• Update AST fallback handling for ACOSH/ASINH so Spark 4.0.3 uses the non-AST path where GPU AST semantics would not match Spark CPU results.
• Update docs/download.md so the published support matrix lists Apache Spark 4.0.3 and the Scala 2.13 Spark 4.0.3 support line.
• Validated with SPARK_HOME=/bigdata/work/tools/spark-4.0.3-bin-hadoop3 mvn -B -s /home/liangcail/.m2/settings_art.xml -f scala2.13/pom.xml -Dbuildver=403 -Dcuda.version=cuda13 verify: BUILD SUCCESS; integration tests reported 35271 passed, 2058 skipped, 513 xfailed, 895 xpassed; Scala tests reported 1759 succeeded, 0 failed.
• Validated review updates with SPARK_HOME=/bigdata/work/tools/spark-4.0.3-bin-hadoop3 mvn -B -s /home/liangcail/.m2/settings_art.xml -f scala2.13/pom.xml -Dbuildver=403 -Dcuda.version=cuda13 -Dmaven.scalastyle.skip=true -Drat.skip=true -DskipTests -Dmaven.scaladoc.skip=true -pl tests -am package: BUILD SUCCESS.
• Local NDS performance results show no overall performance regression observed for Spark 4.0.3.

Item	Value
Dataset	/bigdata/tpcds_data/parquet_100f
Format	Parquet
RAPIDS jar	rapids-4-spark_2.13-26.08.0-SNAPSHOT-cuda13-403-perf.jar
Spark 4.0.2	/bigdata/work/tools/spark-4.0.2-bin-hadoop3
Spark 4.0.3	/bigdata/work/tools/spark-4.0.3-bin-hadoop3
Runs	3 per shim
Compared runs	Warm runs only: run 2 and run 3
Query status	618 query JSON records checked, 0 non-Completed
Result	No overall performance regression observed for Spark 4.0.3

Metric	Spark 4.0.2	Spark 4.0.3	4.0.3 vs 4.0.2
Power run 2	450.000s	454.000s	+0.89%
Power run 3	464.000s	455.000s	-1.94%
Power run avg	457.000s	454.500s	-0.55%
Sum of per-query avg times	455.934s	453.577s	-0.52%

Checklists

Documentation

• Updated for new or modified user-facing features or behaviors
• No user-facing change

Testing

• Added or modified tests to cover new code paths
• Covered by existing tests
  (Please provide the names of the existing tests in the PR description.)
• Not required

Performance

• Tests ran and results are added in the PR description
• Issue filed with a link in the PR description
• Not required

[greptile-apps]

Too many files changed for review. (207 files found, 100 file limit)

[firestarman]

build

[firestarman]

build

[firestarman]

build

[res-life]

The Spark 4.0.3 shim metadata is missing from seven shared Scala test suites: ConcurrentWriterMetricsSuite, GpuIntervalUtilsTest, IntervalCastSuite, IntervalDivisionSuite, IntervalMultiplySuite, OrcEncryptionSuite, and RapidsShuffleThreadedWriterSuite. Please add {"spark": "403"} to each suite. Otherwise, -Dbuildver=403 silently excludes these tests from compilation and execution.

[res-life]

docs/download.md still lists Apache Spark support only through 4.0.2, including the Scala 2.13 support line. Since this PR adds Spark 4.0.3 support, please add 4.0.3 to both lists and update the Documentation checklist accordingly.

[firestarman]

Addressed the latest review feedback in 8bff69c8b:

• Added Spark 4.0.3 shim metadata to the seven shared Scala test suites so they compile for -Dbuildver=403.
• Updated docs/download.md to list Spark 4.0.3 in the supported Spark versions and Scala 2.13 support line.
• Updated the PR Documentation checklist and added the focused validation command to the description.

Validation: SPARK_HOME=/bigdata/work/tools/spark-4.0.3-bin-hadoop3 mvn -B -s /home/liangcail/.m2/settings_art.xml -f scala2.13/pom.xml -Dbuildver=403 -Dcuda.version=cuda13 -Dmaven.scalastyle.skip=true -Drat.skip=true -DskipTests -Dmaven.scaladoc.skip=true -pl tests -am package passed.

[firestarman]

build

[res-life]

LGTM

[NvTimLiu]

LGTM, +1