use passed timeout argument instead of global timeouts for host/check

[inqrphl]

handle_host_check and handle_svc_check get a data pointer of type nebstruct_host/service_check_data.

this struct has another field called 'object_ptr' , that points to the raw host/service check struct

in a planned naemon feature, these host/service check structs will get their own 'check_timeout' field. this field is initialized to the default global value for the service/host check timeouts, purely found within the naemon source code.users can override this default global value by specifying another check_timeout in the config. when a neb request is fired, it now uses this check_timeout set in the host/service timeout

additionally the check_result struct now has a 'timeout' field, this is due to timeouts being customizable. it makes sense to add them to a check_result

first change is for mod_gearman to stop writing timeout in the gearman data wrongly. instead of looking into nebstruct_host/service_data 'timeout' field it was using global host/service timeouts from the naemon source code, which it is including. after this data is written and sent out to a gearman queue, a worker parses the lines for available fields, including 'timeout'

second change is for setting the check_result->timeout to 0 on orphaned checks, just making sure that its zeroed since it might have a random value.

Dont merge it before changes in naemon.