When bugzilla is accessed via XML-RPC, it will strip fields from the response if they contain a NULL value .
One example of this is the 'deadline' field. If there is no data in the field, it will be absent from the response.
When python-bugzilla doesn't receive a field it expects, it seems to go and query the bug again. This can cause a problem when doing a bug search as python-bugzilla will spend a great deal of time querying every bug again for a field that it will never receive.
Thanks for the report, I didn't know that about NULL fields.
The pybz behavior has been around forever, and is convenient as a one off but is really prone to misuse as you point out. In pybz 1.1.0 I added a bug_autorefresh value to turn this off, but code needs to opt in:
I think for the next release I'll flip the default to 'off', but I'll need to make some noise about it first and add better docs. We can use this bug to track that change.
This bug appears to have been reported against 'rawhide' during the Fedora 24 development cycle.
Changing version to '24'.
More information and reason for this action is here:
This is changed in git now. And I sent an email announcing the semantic change, and some other API breaks:
python-bugzilla 2.0.0 is queued for f24+ now. Might let it bake in updates-testing for a while though