[SPARK-47285][SQL] AdaptiveSparkPlanExec should always use the contex…

…t.session

### What changes were proposed in this pull request?

Use `context.session` instead of `session` to avoid potential issue. For example, a cached plan may re-instance `AdaptiveSparkPlanExec` with a different active session.

### Why are the changes needed?

avoid potential issue

### Does this PR introduce _any_ user-facing change?

avoid potential issue

### How was this patch tested?

avoid potential issue

### Was this patch authored or co-authored using generative AI tooling?

no

Closes apache#45388 from ulysses-you/aqe.

Authored-by: ulysses-you <[email protected]>
Signed-off-by: Kent Yao <[email protected]>

sql/core/src/main/scala/org/apache/spark/sql/execution/adaptive/AdaptiveSparkPlanExec.scala

@@ -87,7 +87,7 @@ case class AdaptiveSparkPlanExec(
 
   // The logical plan optimizer for re-optimizing the current logical plan.
   @transient private val optimizer = new AQEOptimizer(conf,
-    session.sessionState.adaptiveRulesHolder.runtimeOptimizerRules)
+    context.session.sessionState.adaptiveRulesHolder.runtimeOptimizerRules)
 
   // `EnsureRequirements` may remove user-specified repartition and assume the query plan won't
   // change its output partitioning. This assumption is not true in AQE. Here we check the
@@ -103,7 +103,8 @@ case class AdaptiveSparkPlanExec(
 
   @transient private val costEvaluator =
     conf.getConf(SQLConf.ADAPTIVE_CUSTOM_COST_EVALUATOR_CLASS) match {
-      case Some(className) => CostEvaluator.instantiate(className, session.sparkContext.getConf)
+      case Some(className) =>
+        CostEvaluator.instantiate(className, context.session.sparkContext.getConf)
       case _ => SimpleCostEvaluator(conf.getConf(SQLConf.ADAPTIVE_FORCE_OPTIMIZE_SKEWED_JOIN))
     }