Noam Postavsky <npostavs@gmail.com> writes:

> By the way, while adding the test case I found an additional regression
> involving loop termination by a 'var = ...' clause.  I'll open another
> bug about it soon.

Actually, it's very closely connected (only observable after fixing the
original bug, since otherwise the necessary conditions trigger the
--cl-vec-- is void error), so I'll keep it here.  When there is a
(cl-loop for VAR across ARRAY and VAR2 = ...) the array index would get
incremented twice per loop.  Fixed by the patch below (applies on top of
the patch in my previous message).