#260: More suggestions to bindctl
-------------------------+--------------------------------------------------
Reporter: shane | Owner:
Type: enhancement | Status: new
Priority: minor | Milestone: feature backlog item
Component: bind-ctl | Resolution:
Keywords: | Sensitive: 0
-------------------------+--------------------------------------------------
Comment(by jelte):
Let me add to that for a minute, having something that catches any
exception is not necessarily a bad thing (and sometimes it's a very good
thing), however, losing stacktraces certainly is. For programs like
bindctl, that are run from the command line, i personally don't care if it
aborts with a stack trace, and i certainly prefer that above printing a
simple error (that does not show what exactly went wrong there) and
running on.
So I think that a good rule would be that, in Python, if we put in a
catch-all (except Exception), we *always* print or log the full stack
trace. If we 'expect' a specific exception and know how to handle it, we
should check for that specific exception, and not a general one.
For c++ this might be a bit harder (since there's no portable way to get
the stacktrace), but the problem there might be less bad since our own
exceptions can already provide source file and line.
--
Ticket URL: <https://bind10.isc.org/ticket/260#comment:2>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development