On 01/30/2016 07:16 AM, clipka wrote:
> Am 29.01.2016 um 14:24 schrieb William F Pokorny:
>
>> If you are somewhat routinely getting a core dump:
>
> No, not really. "Operation timed out." is the usual deal.
>

Just a thought, but is the "Operation timed out." message localized 
enough that you could modify the code to core dump by adding abort()?

If the time out period is really long you might be able to send an 
abort via the kill command to get a core dump. Running with 1-2 threads 
would be essential for this attempt otherwise you'd not know what 
parallel process to kill.

(And sorry to hear about that desk spill... :-) )

Bill P.

Post a reply to this message