No subject

that no "audit" or formal proof of any sort is needed.
Just that the author believes that any use of a
potentially unsafe feature is OK.

So by my reading, it is enough just to look over the
API to make sure nothing apparently unsafe is
exported, plus include in our email to the package
maintainer a request to contact the list before adding
the pragma if there is a suspicion that something
unsafe might be going on.

Anyway, in real life, most of us are pretty familiar
with most of the packages in the platform.
We've identified a few that need some thought.
If there are any other problems, I hope someone
here will let us know.

Thanks,
Yitz