Re: [q-lang-users] Libqint and the stability of qexps

[John C. <co...@cc...>]

Albert Graef scripsit:

> Yes, they are. Expressions are never copied, always shared. And the 
> garbage collector will never move expressions around, it only adds them 
> to the freelist when they become inaccessible. 

Excellent.

> Note that if you keep pointers to temporary Q expressions (i.e., with a 
> reference count of 0) in your Scheme data structures then those 
> expressions may well be garbage-collected at any time during an 
> evaluation. 

Good point.  This by itself wouldn't be a problem, since the associations
between Scheme objects and Q objects are only used when recursively
decomposing the qexp and constructing the corresponding qsxp (Q-language
S-expression).  In this process, I never attempt to do anything with a
saved qexp except compare it with a returned qexp, so if the saved qexp
is garbage, no harm is done.

The real danger is that a qexp formerly referred to may be discarded
by Q and a newly created qexp may wind up at exactly the same C-level
address, causing false positive matches.  Because of this, I will be
sure to qnewref() my qexps.

> Therefore, to avoid nasty surprises, you should always count a reference
> (qnewref) to Q expressions which are stored somewhere.

Is it necessary to call qnewref() on each newly constructed qexp, or only
on the top-level qexp?

-- 
John Cowan    co...@cc...    http://ccil.org/~cowan
        Sound change operates regularly to produce irregularities;
        analogy operates irregularly to produce regularities.
                --E.H. Sturtevant, ca. 1945, probably at Yale