bug#12215: CSET is unnecessarily confusing

[Tom Tromey <tromey@redhat.com>]

>>>>> "Paul" == Paul Eggert <eggert@cs.ucla.edu> writes:

Paul> How exactly does that work?  I just now looked at the concurrency branch
Paul> and its BVAR is the same as in the trunk.

Paul> Is BVAR a speculative change, that was put in because eventually we
Paul> thought we'd need it for concurrency?  If so, perhaps we should revert
Paul> it until the need is demonstrated.  After all, it's been many months
Paul> since BVAR went in, and the concurrency branch still isn't really
Paul> using it.

Yeah.

Originally I had planned a somewhat complicated scheme for handling
per-thread let bindings.  This scheme required some changes like BVAR.
I lobbied to get these changes into the trunk because carrying big
changes like this on a branch makes merging very hard -- this is partly
what sunk the first concurrency branch.

My planned bindings approach was tricky, though, and due to unrelated
changes in my life, I no longer have the time or energy to think through
all the details to make them work with other changes that have happened
in Emacs since then.  So on the new concurrency branch I opted for a
much simpler approach (which may still be wrong :-).

BVAR isn't needed right now.  It may be if someone else resurrects the
other approach.  It's unlikely I ever will.

Tom