Spark: fix flaky testReadingStreamFromFutureTimestamp by switching to Awaitility.await

[dramaticlly]

Fix the flaky test in
https://github.com/apache/iceberg/actions/runs/30405288650/job/90429251920#step:6:11506

TestStructuredStreamingRead3 > testReadingStreamFromFutureTimetsamp() > catalogName = testhive, implementation = org.apache.iceberg.spark.SparkCatalog, config = {type=hive, default-namespace=default}, async = true FAILED
    java.lang.AssertionError: 
    Expecting actual:
      []
    to contain exactly in any order:
      [{"id"=-2,"data"="minustwo"},
        {"id"=-1,"data"="minusone"},
        {"id"=0,"data"="zero"}]
    but could not find the following elements:
      [{"id"=-2,"data"="minustwo"},
        {"id"=-1,"data"="minusone"},
        {"id"=0,"data"="zero"}]
        at org.apache.iceberg.spark.source.TestStructuredStreamingRead3.testReadingStreamFromFutureTimetsamp(TestStructuredStreamingRead3.java:635)

switch to Awaitility.await for better polling. Verified 5 times locally on all spark 3.5/4.0/4.1

[dramaticlly]

FYI @amogh-jahagirdar this CI failure come from merge of 23b5ce8, in case you want to take a look