bq. You correctly describe the behavior; is there a concrete problem that it solves? In general you don't want to write code with fixed seeds. There are use cases for deterministic randomness. For example, parallel algorithms that depend on each actor having the same random sequence.
On 6/27/11, Hudson (JIRA) <[email protected]> wrote: > > [ > https://issues.apache.org/jira/browse/MAHOUT-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13055539#comment-13055539 > ] > > Hudson commented on MAHOUT-743: > ------------------------------- > > Integrated in Mahout-Quality #908 (See > [https://builds.apache.org/job/Mahout-Quality/908/]) > MAHOUT-743 use fixed seed in tests > > srowen : > http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1140097 > Files : > * > /mahout/trunk/math/src/test/java/org/apache/mahout/common/RandomUtilsTest.java > * > /mahout/trunk/math/src/main/java/org/apache/mahout/common/RandomWrapper.java > > >> Allow use of random seeds during unit tests >> ------------------------------------------- >> >> Key: MAHOUT-743 >> URL: https://issues.apache.org/jira/browse/MAHOUT-743 >> Project: Mahout >> Issue Type: Improvement >> Reporter: Lance Norskog >> Assignee: Sean Owen >> Fix For: 0.6 >> >> Attachments: MAHOUT-743.patch, RandomWrapper.patch >> >> >> The RandomUtils package has a trick for running unit tests repeatably. All >> random generators created with RandomUtils.getRandom() are started with >> seed 0. Unfortunately, all random generators started with >> RandomUtils.getRandom(seed) effectively run from a seed of 0. Thus, code >> which deliberately creates Random objects with seeds does not work >> correctly under unit tests. >> The problem is that the logic in RandomWrapper tests for unit test mode >> and settable seed mode in the wrong order. > > -- > This message is automatically generated by JIRA. > For more information on JIRA, see: http://www.atlassian.com/software/jira > > > -- Lance Norskog [email protected]
