Taylor R Campbell wrote:
that they don't behave nicely when the procedures passed to them modify the table.
Is there an efficient algorithm around that problem (i.e. short of copying the whole table or their keys at once first)?
Anyway, the same problem holds true for example for Perl's hash table iterators. It's just one of the caveats one has to know about.
Btw, there is at least a |table-for-each| procedure already, and documented in the Gambit docs. Actually it is even more restrictive than just explicit modification. (Wondering why, no time to check now.)
Christian.