Re: Resizing hash tables in Guile

[Mikael Djurfeldt]

Marius Vollmer <address@hidden> writes:

> Mikael Djurfeldt <address@hidden> writes:
>
>> Is this a good thing?  Should we keep it?
>
> I think resizing hash tables are important.

I've been thinking that maybe we should continue the move and let the
resizing tables entirely replace the fixed size ones.  It seems a
little silly to have to explain to Guile users that there are two
(sorry, eight) different kinds of hash tables...  Also, I think the
opacity of the resizing table objects is an advantage rather than a
disadvantage.  If they are opaque, we can any time modify the
underlying implementation (the well-known data abstraction argument).

What do you say?

>> The implementation isn't quite finished.  Removal should be rewritten
>> for the new tables.  Also, weak hash tables need to be handled in
>> hashtab.c in order to maintain a correct item count.
>
> If you are not going to finish this soonish, please make a note in
> workbook/tasks/TODO about what needs to still be done.

Right.  Removal is fixed now.  Weak tables could be fixed on a train
ride during the weekend. :)

(BTW, at that time it would be good if we have come to a decision with
regard to dropping the fixed-size vector of alist tables.)

M