nssalian commented on code in PR #1182:
URL: https://github.com/apache/iceberg-go/pull/1182#discussion_r3405083327
##########
internal/recipe/local_spark.go:
##########
@@ -140,3 +142,33 @@ func ExecuteSpark(t *testing.T, scriptPath string, args
...string) (string, erro
return string(output), nil
}
+
+var (
+ sparkMajorOnce sync.Once
+ sparkMajor int
+ sparkMajorErr error
+)
+
+// SparkMajorVersion returns the major pyspark version (3 or 4) of the
Review Comment:
Was able to fix this.
switched to a mutex that only writes `sparkMajor` inside the success
branches, so transient exec failures leave it at 0 and the next caller
retries. On the race I couldn't reproduce it (`sync.Once.Do` runs the closure
once during the first caller's lifetime, `ExecuteSpark` has no goroutines, so
the captured `t` is held but never re-invoked) but the decoupling is still the
right cleanup.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
