[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: More on float-consing

Thanks for the explanation.  If we have to get by with a heuristic for now
(until some dynamic analysis goes in), I would say that the heuristic
should be something that makes this case come out right.  Perhaps, "ignore
setq's to specials in deciding which representation to choose for a local"
or maybe just "when in doubt (evidence both ways), always use the
non-consing representation".

Any chance we can get some such fix installed soon?

-- Scott