[ https://issues.apache.org/jira/browse/GEODE-2191?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15742799#comment-15742799 ]
ASF GitHub Bot commented on GEODE-2191: --------------------------------------- Github user galen-pivotal commented on the issue: https://github.com/apache/geode/pull/306 @kirklund I disagree. We expect string serialization to work reliably on *all* inputs, and since we can't test all of them, we may as well test a bunch picked at random. If we fail, it's because of a corner case we haven't thought of. I would rather catch it randomly in testing than when a customer wants to store strings we hadn't anticipated. The seed is captured and printed so that we can reproduce the failure. There are cases where random input could show deficiencies in our test or testing environment rather than in the code under test; I don't think this is one of those. > Add more randomized tests for PDX string serialization/deserialization. > ----------------------------------------------------------------------- > > Key: GEODE-2191 > URL: https://issues.apache.org/jira/browse/GEODE-2191 > Project: Geode > Issue Type: Test > Reporter: Galen O'Sullivan > Assignee: Mark Bretl > > There are currently a number of tests for `DataSerializer.writeString` and > `DataSerializer.readString`, which cover many of the corner cases we expect > for string serialization/deserialization, but it would be nice to have > randomized testing that generated Unicode strings and verified that they were > the same after being serialized and deserialized. -- This message was sent by Atlassian JIRA (v6.3.4#6332)