> +def get_inst(ip, virt, cn, key):
> + inst = None +
> try:
> - ret_inst = assoc.AssociatorNames(host,esd_cn, class_name, virt,
> - InstanceID = id)
> + key_list = {"InstanceID" : key }
> +
> + inst = getInstance(ip, cn, key_list, virt)
> +
> + except Exception, details:
> + logger.error("Exception %s" % details)
> + return None +
> + if inst is None:
> + logger.error("Expected at least one %s instance" % cn)
> + return None +
> + return inst +
>
The above function can be moved to a library . We can probably pass the
key_list to get_inst() fn.
I think I'll leave this as is. Having a getInstance() function, and
then another, similar, get_inst() function is confusing.
The only thing this function adds is trapping of the exception and some
additional error checking. It's not much extra work to have the test
case trap the exception and test for error conditions.
It's possible that the calling function will want to trap the exception
differently or return a different error message.
--
Kaitlin Rupert
IBM Linux Technology Center
kaitlin(a)linux.vnet.ibm.com