Re: [Chicken-users] hash-table-walk + hash-table-delete!

[Kon Lovett]

On Mar 5, 2007, at 12:12 PM, Graham Fawcett wrote:

> On 3/5/07, Kon Lovett <address@hidden> wrote:
> > On Mar 5, 2007, at 11:23 AM, Graham Fawcett wrote:
> > >> However I think a '(hash-table-remove! hash-table (-> key value
> > >> boolean))' procedure should be added to "extras.scm" just to be
> > >> "safe".
> > >
> > > Could you please explain that signature? I don't know how to  
> > interpret
> > > your (-> ...) form.
> >
> > A strict prefix form:
> >
> > (-> key value boolean) is (key value -> boolean) but should be
> > (object object -> boolean) since there are no key & value types.
>
> Thanks. I thought it looked like Haskell, and I wasn't that far off.
> :-) Wouldn't it be '(hash-table-remove! (-> hash-table key value
> boolean)) ? Or do you write it as you did because the procedure
> mutates the hash-table?
>
> And lastly, are you passing both key and value in for correctness'
> sake (in lieu of passing in the actual bucket, to which we don't have
> a reference in a hash-table-walk), or would your implementation
> actually test for equality on both before removing?

The idea was hash-table-remove! will take 2 parameters, a hash-table  
& a predicate.
The predicate, (object object -> boolean), will return #t when the  
entry is to be deleted; kinda like SRFI-1 remove!. This allows  
conditioning the delete operation on all safely available data.

With hash-table-remove! weaker operations can be defined,
like (hash-table-remove!/keys hash-table KEY0 ...).

> Graham