On 08/14/2012 07:10 AM, Rob Stryker wrote:
I've tried to get an estimate of how long the full suite takes to run, 
but I'm having difficulty. I've added the following to my test's pom.xml:


             <plugin>
                 <groupId>org.eclipse.tycho</groupId>
<artifactId>tycho-surefire-plugin</artifactId>
                 <version>${tychoVersion}</version>
                 <configuration>
+ <surefire.timeout>10800</surefire.timeout>
<testSuite>org.jboss.tools.as.test.core</testSuite>
<testClass>org.jboss.tools.as.test.core.ASToolsTestSuite</testClass>
                 </configuration>
             </plugin>
With surefire, the right argument for timeout is forkedProcessTimeoutInSeconds ( http://www.eclipse.org/tycho/sitedocs/tycho-surefire/tycho-surefire-plugin/test-mojo.html#forkedProcessTimeoutInSeconds ). But with JBIDE-12422, I made the necessary changes so you don't have to modify pom.xml anymore to try new timeouts. Passing -Dsurefire.timeout=10800 to command-line will now work (property surefire.timeout is the default value for forkedProcessTimeoutInSeconds when not set in configuration). When you think you found a timeout that is fine for most cases, you can modify the value of the surefire.timeout property (not argument in tycho-surefire-plugin configure) to the value you think is the best.
Default value, set in parent pom, for surefire.timeout is 2400 seconds (40 minutes), but it is often overriden in project-specific pom.
Once again, let me insist that you should use the surefire.timeout property rather than the forkedProcessTimeoutInSeconds argument.

You should perform a local update to leverage these changes. Feedback is, as always, welcome.

If I set the timeout to something large, the suite still shuts off 
exactly at 42 minutes. If I set the timeout to something small, like 
500, it cuts off early as expected. So the problem here seems I can 
override the setting, but only to make it shorter (??), not longer.  If 
I'm wrong, then I have absolutely no way to explain the current behaviour.
That's weird, but I think this issue disappeared with changes for JBIDE-12422

--
Mickael Istria
Eclipse developer at JBoss, by Red Hat
My blog - My Tweets