dependencies: ginkgo v2.15.0, gomega v1.31.0

The main reason for updating is support for reporting the cause of context
cancellation: Ginkgo provides that information when canceling a context and
Gomega polling code includes that when generating a failure message.

vendor/github.com/onsi/ginkgo/v2/internal/suite.go

@@ -79,7 +79,7 @@ func NewSuite() *Suite {
 
 func (suite *Suite) Clone() (*Suite, error) {
 	if suite.phase != PhaseBuildTopLevel {
-		return nil, fmt.Errorf("cnanot clone suite after tree has been built")
+		return nil, fmt.Errorf("cannot clone suite after tree has been built")
 	}
 	return &Suite{
 		tree:                    &TreeNode{},
@@ -858,7 +858,7 @@ func (suite *Suite) runNode(node Node, specDeadline time.Time, text string) (typ
 	}
 
 	sc := NewSpecContext(suite)
-	defer sc.cancel()
+	defer sc.cancel(fmt.Errorf("spec has finished"))
 
 	suite.selectiveLock.Lock()
 	suite.currentSpecContext = sc
@@ -958,7 +958,7 @@ func (suite *Suite) runNode(node Node, specDeadline time.Time, text string) (typ
 
 			// tell the spec to stop.  it's important we generate the progress report first to make sure we capture where
 			// the spec is actually stuck
-			sc.cancel()
+			sc.cancel(fmt.Errorf("%s timeout occurred", timeoutInPlay))
 			//and now we wait for the grace period
 			gracePeriodChannel = time.After(gracePeriod)
 		case <-interruptStatus.Channel:
@@ -985,7 +985,7 @@ func (suite *Suite) runNode(node Node, specDeadline time.Time, text string) (typ
 			}
 
 			progressReport = progressReport.WithoutOtherGoroutines()
-			sc.cancel()
+			sc.cancel(fmt.Errorf(interruptStatus.Message()))
 
 			if interruptStatus.Level == interrupt_handler.InterruptLevelBailOut {
 				if interruptStatus.ShouldIncludeProgressReport() {