From 1e78db019f029707032821be283fd7f1372d6c4c Mon Sep 17 00:00:00 2001
From: Andy Grove <agrove@apache.org>
Date: Mon, 27 Jul 2026 08:15:30 -0600
Subject: [PATCH] feat: expose Comet version as spark.comet.version runtime
 config

Expose the loaded Comet build version through the Spark config
spark.comet.version (set by the Comet driver plugin) so it can be
queried at runtime via spark.conf.get or SET. Also document the
existing org.apache.comet.COMET_VERSION programmatic accessor.
---
 docs/source/user-guide/latest/installation.md | 27 +++++++++++++++++++
 .../main/scala/org/apache/spark/Plugins.scala | 12 ++++++++-
 .../org/apache/spark/CometPluginsSuite.scala  |  9 +++++++
 3 files changed, 47 insertions(+), 1 deletion(-)

diff --git a/docs/source/user-guide/latest/installation.md b/docs/source/user-guide/latest/installation.md
index ea3cc348bd..f762a2938c 100644
--- a/docs/source/user-guide/latest/installation.md
+++ b/docs/source/user-guide/latest/installation.md
@@ -166,6 +166,33 @@ CometNativeColumnarToRow
    +- CometNativeScan parquet [a#6] Batched: true, DataFilters: [isnotnull(a#6), (a#6 > 5)], Format: CometParquet, Location: InMemoryFileIndex(1 paths)[file:/tmp/test], PartitionFilters: [], PushedFilters: [IsNotNull(a), GreaterThan(a,5)], ReadSchema: struct<a:int>
 ```
 
+## Checking the Comet Version
+
+When the Comet plugin is loaded, it exposes its build version as the Spark config
+`spark.comet.version`. This can be queried at runtime from any supported language, for example:
+
+```scala
+scala> spark.conf.get("spark.comet.version")
+```
+
+```sql
+SET spark.comet.version;
+```
+
+The same value is available programmatically on the JVM classpath, along with additional build
+metadata that is useful when reporting issues:
+
+```scala
+scala> import org.apache.comet.{COMET_VERSION, COMET_BRANCH, COMET_REVISION}
+scala> println(COMET_VERSION)
+```
+
+Comet also logs its version when the native library is initialized:
+
+```shell
+INFO core/src/lib.rs: Comet native library version <version> initialized
+```
+
 ## Additional Configuration
 
 Depending on your deployment mode you may also need to set the driver & executor class path(s) to
diff --git a/spark/src/main/scala/org/apache/spark/Plugins.scala b/spark/src/main/scala/org/apache/spark/Plugins.scala
index 43945b97fd..e726cef378 100644
--- a/spark/src/main/scala/org/apache/spark/Plugins.scala
+++ b/spark/src/main/scala/org/apache/spark/Plugins.scala
@@ -28,7 +28,7 @@ import org.apache.spark.internal.Logging
 import org.apache.spark.internal.config.{EXECUTOR_MEMORY, EXECUTOR_MEMORY_OVERHEAD, EXECUTOR_MEMORY_OVERHEAD_FACTOR}
 import org.apache.spark.sql.internal.StaticSQLConf
 
-import org.apache.comet.{CometSparkSessionExtensions, NativeBase}
+import org.apache.comet.{COMET_VERSION, CometSparkSessionExtensions, NativeBase}
 import org.apache.comet.CometConf.{COMET_METRICS_ENABLED, COMET_ONHEAP_ENABLED}
 
 /**
@@ -48,6 +48,12 @@ class CometDriverPlugin extends DriverPlugin with Logging with ShimCometDriverPl
   override def init(sc: SparkContext, pluginContext: PluginContext): ju.Map[String, String] = {
     logInfo("CometDriverPlugin init")
 
+    // Expose the Comet build version as a Spark config so it can be queried at runtime, e.g.
+    // `spark.conf.get("spark.comet.version")` or `SET spark.comet.version` in SQL. This is set
+    // before the off-heap check below so the version is reported even when Comet is otherwise
+    // disabled.
+    sc.conf.set(CometDriverPlugin.COMET_VERSION_CONFIG, COMET_VERSION)
+
     if (!CometSparkSessionExtensions.isOffHeapEnabled(sc.getConf) &&
       !sc.getConf.getBoolean(COMET_ONHEAP_ENABLED.key, false)) {
       logWarning("Comet plugin is disabled because Spark is not running in off-heap mode.")
@@ -106,6 +112,10 @@ class CometDriverPlugin extends DriverPlugin with Logging with ShimCometDriverPl
 }
 
 object CometDriverPlugin extends Logging {
+
+  /** Spark config key under which the loaded Comet version is exposed at runtime. */
+  val COMET_VERSION_CONFIG = "spark.comet.version"
+
   def registerCometMetrics(sc: SparkContext): Unit = {
     if (sc.getConf.getBoolean(
         COMET_METRICS_ENABLED.key,
diff --git a/spark/src/test/scala/org/apache/spark/CometPluginsSuite.scala b/spark/src/test/scala/org/apache/spark/CometPluginsSuite.scala
index fa5f368e33..68ddb9b8e7 100644
--- a/spark/src/test/scala/org/apache/spark/CometPluginsSuite.scala
+++ b/spark/src/test/scala/org/apache/spark/CometPluginsSuite.scala
@@ -24,6 +24,8 @@ import java.io.File
 import org.apache.spark.sql.{CometTestBase, SaveMode}
 import org.apache.spark.sql.internal.StaticSQLConf
 
+import org.apache.comet.COMET_VERSION
+
 class CometPluginsSuite extends CometTestBase {
   override protected def sparkConf: SparkConf = {
     val conf = new SparkConf()
@@ -83,6 +85,13 @@ class CometPluginsSuite extends CometTestBase {
     }
   }
 
+  test("Comet version is exposed as a Spark config") {
+    // The driver plugin sets spark.comet.version, which is then visible both on the SparkContext
+    // conf and through the session runtime config (SET / spark.conf.get).
+    assert(spark.sparkContext.conf.get(CometDriverPlugin.COMET_VERSION_CONFIG) == COMET_VERSION)
+    assert(spark.conf.get(CometDriverPlugin.COMET_VERSION_CONFIG) == COMET_VERSION)
+  }
+
   test("CometSource metrics are recorded") {
     val nativeBefore = CometSource.NATIVE_OPERATORS.getCount
     val queriesBefore = CometSource.QUERIES_PLANNED.getCount