Unfortunately failures in the multi run can't be diagnosed at all from
the diff that they produce. Can you post a zip of the results of that
test to a new JIRA entry? Often the actual problem can only be
diagnosed either from the
derby.log or the client log of the test.
I believe we have seen a number of regressions in the multi tests, maybe
before we make a release we should analyze these. There have been a
number of changes in 10.3 to change latches and remove use of
synchronization on a data structures where it looked like it was not
needed. If there were any bugs in these changes they are likely only
to show up in multi-user testing and we don't have much multi-user
testing in the nightly regression suite. This test happens to be
one of them.
If anyone has a free machine, or free cycles overnight it might be
interesting to run this test over and over again and let the list know
how many times it fails.
[EMAIL PROTECTED] wrote:
[Auto-generated mail]
*tinderbox_trunk16* 540926/2007-05-23 13:22:47 CEST
Failed Tests OK Skip Duration Suite
-------------------------------------------------------
*Jvm: 1.6*
SunOS-5.10_i86pc-i386
1 310 309 0 257.95% derbyall
0 8054 8054 0 1532.26%
org.apache.derbyTesting.functionTests.suites.All
Details in http://dbtg.thresher.com/derby/test/tinderbox_trunk16/jvm1.6/testing/Limited/testSummary-540926.html
Attempted failure analysis in
http://dbtg.thresher.com/derby/test/tinderbox_trunk16/jvm1.6/FailReports/540926_bySig.html
-------------------------------------------------------
Changes in http://dbtg.thresher.com/derby/test/tinderbox_trunk16/UpdateInfo/540926.txt
( All results in http://dbtg.thresher.com/derby/test/ )
