Actually on second hand I think there's a separate issue here : whether we should remove the supportsQueryTimeout field.
Selectively disabling some methods like we are with supportsQueryTimeout is problematic and it's easy to get out of sync with JDBC vendors (like we did in this case). We could let the call go through and let the upstream callers handle the exception - rather than checking for supportsQueryTimeout before calling the method.
That's not a change I'm comfortable making in 1.2.x or 1.0.x though. Negating the call and logging could be done though.
There's a fair amount of precedent for checking before calling though (avoiding the trip to the database). Unless there are objections I'll open a sub-task / related issue to discuss whether we need the supportsQueryTimeout field.