cc: squeak-dev

On Sun, Mar 15, 2015 at 3:21 PM, Chris Muller <asqueaker at gmail.com> wrote:
> On Sun, Mar 15, 2015 at 2:08 PM, Chris Muller <asqueaker at gmail.com> wrote:
>> Do test timeouts add any value, or do they only subtract value by
>> forcing humans to research non-problems only to discover the timeout
>> value needs to be tweaked?  Is there a payoff to continue doing that?
>
> I know these timeouts are supposed to solve the case that if some code
> is changed to cause one a Test to get stuck in a loop, then the test
> will fail after the timeout rather than sitting and spinning in the
> Jenkins server forever.
>
> But whether Jenkins reports a timeout, or reports a crash due to a
> stack overflow, or stops producing reports at all because its
> spinning; each of these cases requires some kind of human
> intervention.  Its just that with the individual timeouts solution,
> not only are the humans required to intervene when there isn't really
> a problem, but such intervention even involves making the code itself
> bear a cost too.
>
> It seems like one 5-minute timeout handler hard-coded at the top would
> handle the case equally sufficiently..?  5 minutes is ample long
> enough for most any regular test, while short enough to "wait" for
> output from Jenkins..