You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Jeff Jirsa (JIRA)" <ji...@apache.org> on 2017/09/14 00:01:37 UTC
[jira] [Commented] (CASSANDRA-13866) Clock-dependent integer
overflow in tests CellTest and RowsTest
[ https://issues.apache.org/jira/browse/CASSANDRA-13866?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16165494#comment-16165494 ]
Jeff Jirsa commented on CASSANDRA-13866:
----------------------------------------
lgtm.
> Clock-dependent integer overflow in tests CellTest and RowsTest
> ---------------------------------------------------------------
>
> Key: CASSANDRA-13866
> URL: https://issues.apache.org/jira/browse/CASSANDRA-13866
> Project: Cassandra
> Issue Type: Bug
> Components: Testing
> Reporter: Joel Knighton
> Assignee: Joel Knighton
> Priority: Trivial
> Fix For: 3.0.x, 3.11.x, 4.x
>
>
> These tests create timestamps from Unix time, but this is done as int math with the result stored in a long. This means that if the test is run at certain times, like 1505177731, corresponding to Tuesday, September 12, 2017, 12:55:31, the test can have two timestamps separated by a single second that reverse their ordering when multiplied by 1000000, such as 1505177731 -> 2147149504 and 1505177732 -> -2146817792. This causes a variety of test failures, since it changes the reconciliation order of these cells.
> Note that I've tagged this as trivial because the problem is in the manual construction of timestamps in the test; I know of nowhere that we make this mistake with real data.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org