Quoth Mark Walters on Jun 09 at 10:16 am:
> 
> Both of these patches look good to me +1. I was able to reproduce both
> bugs pretty reliably (the first one always unless masked by the second
> one which occurred about half the time). With these patches I cannot
> trigger either. Also all tests pass.
> 
> My only tiny concern is I couldn't find any documentation on whether the
> return value of the filter-function matters at all. Austin's original
> fix (via irc) returned t and this returns nil in the failing case (i.e.,
> when results-buf is dead).

Mm, interesting.  To be fair, my choice of "t" for the original fix
was completely arbitrary.  I think you're right that the Emacs
documentation doesn't have anything to say about the return values of
filter functions.  Furthermore, the example filter functions they give
don't have meaningful return values, so I'm pretty sure this is safe.
Also the code that calls the filter discards its result.