This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: Observation about hash-table.
- From: Dhruv Matani <dhruvbird at gmx dot net>
- To: libstdc++ <libstdc++ at gcc dot gnu dot org>
- Date: 07 May 2004 20:51:02 +0530
- Subject: Re: Observation about hash-table.
- Organization:
- References: <1083939365.1492.9.camel@localhost.localdomain>
On Fri, 2004-05-07 at 20:09, Dhruv Matani wrote:
> Hello everyone,
> I noticed that the hash-table uses chaining for its purposes. However,
> insertion is not thread-safe, but can be made so quite easily IMHO. Now,
> when the pointers for the vector and the link node are being set the
> following should ensure safety:
>
> struct node;
> vector<node*> vpn(bucket_size);
>
> void insert(const_ref value)
> {
> int bkt = //Bucket number.
> node* current = new node(value);
> node* tmp = current;
> atomic_exchange_after_deref(current, &vpn[bkt]);
> tmp->next = current;
> }
>
>
> Should this do the trick, or is something wrong?
Which I just realized is not as simple as I initially thought, because
the hash-table inserts values within a bucket in sorted order, as
opposed to just a stack like insertion as I initially thought :-(
>
--
-Dhruv Matani.
http://www.geocities.com/dhruvbird/
Proud to be a Vegetarian.
http://www.vegetarianstarterkit.com/
http://www.vegkids.com/vegkids/index.html