Re: Illegal C++

[Erik Sandberg]

On Tuesday 11 October 2005 18.48, Nicolas Sceaux wrote:
> Erik Sandberg <address@hidden> writes:
> > To me, those property lists look like major bottlenecks (though I
> > haven't done any real profiling). Especially the grob property alists:
> > While I was debugging some time ago, I saw that
> > Grob::internal_set_property was called over 1000 times in a trivial
> > score.

BTW, I did a typo. The often-called function was Grob::internal_get_property, 
not .._set_property. 

> But what time a call to Grob::internal_set_property takes on an alist
> compared to eg a guile hash table?  I don't know of guile's hashtables
> and alists, but in Common Lisp, alists are said to be more efficient
> that hashtables for small tables[1].  So there are cases where alist is
> really a sound choice of table structure. (I don't know how many
> properties grobs typically have).

Then it could be sensible to use a 'resizable associative array' 
datastructure, which implements itself as an alist when it's small, and as a 
resizable hash table for larger sizes.

-- 
Erik