On Wed, May 30, 2012 at 7:39 AM, Nikodemus Siivola
> On 29 May 2012 18:30, Paul Khuong <pvk@...> wrote:
>> Comments/review? functional-may-escape-p looks like a probable source
>> of bugs; it is very new, and much more aggressive than nikodemus's
>> initial code.
> Not a proper review, but a quick look so far:
> ;; We only descend on calls
> (unless (combination-p dest)
> (return t))
> What about SET nodes?
Returning T means that the function potentially escapes, so it's
always safe. Handling SET nodes would mean tracking whether values
bound to each lambda-var may escape as well, which needs another
(mutually recursive) function. It would be nice, but I don't know what
% of real code would benefit.