On 09/06/2009, Rainer Jung <rainer.j...@kippdata.de> wrote:
> I wanted to add another non-critical testing observation concering the
>  1.5 release.
>
>  I tested on Solaris Sparc on a somewhat slow/old system (V125, 1GHz
>  UltraSparc IIIi). The failed tests were testEviction() and
>  testEviction2() in TestGKOP. Both failures were due to timing issues.
>
>  The call to evict() took about 1 second to return, so the sleeps between
>  the tests were to short.
>
>  After changing the timing in testEviction() from 1000ms/600ms to
>  1500ms/1200ms and in testEviction2() from 1100ms/600ms to 1800ms/1400ms
>  the tests succeeded.
>
>  It will be hard to find "universal" constants working for all systems. :(

It might be possible to add a calibration test in order to find out
what timings are needed.

Identifying which tests are timing sensitive would be a start. Is it
possible to write additional information to the Surefire log files? If
so, this would be a good place to document timing dependencies. Or
maybe just add something like "(time-sensitive)" to all the fail
messages.

>  Regards,
>
>
>  Rainer
>
>
>  ---------------------------------------------------------------------
>  To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
>  For additional commands, e-mail: dev-h...@commons.apache.org
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org

Reply via email to