Recent patch(1) update

Hello!

I've recently fixed a bug in patch(1) that caused patches with limited
context to be misapplied under some circumstances. Patches that were
candidates to be misapplied either had low context with respect to the
default max fuzzing (-U0 / -U1) or had lower context due to being near the
beginning/end of a file.

The fix was applied in base r326084; please test port updates with an
updated patch(1) to avoid build failures on -HEAD.

A sweep of the ports tree was done by portmgr_at_ to correct current cases of
these candidate patches.

Thanks,

Kyle Evans