Record segfault

[Joel Brobecker]

Michael,

> This seems fine to me.  Please check it into the main branch,
> but don't add it to the release branch unless Joel says it's OK.

It seems like a worthwhile patch to have, since it fixes a SEGV
(right?). Are you confident about the patch? It seems relatively
straightforward. My only comment is that it eliminates a query
when the user types a record stop - is that desirable? I guess
it is: What are the chances of someone typing "record stop" and
not wanting to stop the recording session?

In other words: I am ok with the patch being applied to the branch.
If you are, can you give Hui the go-ahead?

-- 
Joel