From: Hugo Duncan (hugoduncan_at_[hidden])
Date: 2004-10-28 06:37:51
Many thanks for dealing with this!
On Mon, 25 Oct 2004 Doug Gregor wrote:
> Fixed when EdgeListS has stable iterators (e.g., when it uses the
> default listS). Hugo, I would appreciate if you could verify that the
> horrible performance you were getting is improved by this patch.
It is somewhat improved. The speed isn't completely what it was (to
be expected), but I have not yet convinced myself that there is no
I have been testing with a pre-mpl update cvs version with your latest
Sorry to be imprecise, but I'd like to test with a self consistent
boost version which will probably take until next week. Having said
that, if it passes the regression tests it is definitely a speed
> Unfortunately, adjacency_list is broken in this area and we aren't
> going to fix it for release.
It would also be nice to have a specialisation for the no edge property
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk