[m-dev.] map.merge argument order

[Julien Fischer]

On Mon, Dec 9, 2013 at 4:48 PM, Paul Bone <paul at bone.id.au> wrote:

>
>
> Since in Mercury we don't have destructive update and usually have to
> reconstruct the parts of a tree that refer to the parts that actually
> changed.  Is there any significant extra cost to storing the size of the
> subtrees within each node?  (I don't expect to find current
> Mercury-specific
> benchmarks, but I wonder if some general benchmarks exist.)
>

For sets we have set_tree234 and set_ctree234, the latter of which stores
the size of the tree at the root node only rather than in every node.

Cheers,
Julien.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mercurylang.org/archives/developers/attachments/20131209/9c692fde/attachment.html>