[Development] CI broken again?

Nurmi J-P jpnurmi at digia.com
Wed Mar 5 10:00:22 CET 2014


On 05 Mar 2014, at 09:32, Rutledge Shawn <Shawn.Rutledge at digia.com> wrote:

> Sometimes though, we try to fix the autotests that fail the most frequently.  If you can't reproduce the failure on your own machine, with the same OS, often the cause seems to be heavy multi-tasking on the CI machines, which will slow down timing-sensitive tests to the point of failure.  But even that is hard to prove, since CI is a black box to most of us; and even if you get access to a CI machine to run tests on, they won't fail because it's not multi-tasking so heavily at that time.

A virtual machine with very limited resources can often reproduce such test failures. At least VBox lets you change CPU execution cap while the virtual machine is running.

--
J-P Nurmi




More information about the Development mailing list