Persistence and Newton OS...

[Norman Culver]

Fare Rideau wrote:
> 
> >
> Actually, we *are* seriously considering persistence,
> but in a software, not hardware, implementation.
> 

Several years ago I wrote a persistent multi-heap malloc which worked
very well as long as the underlying OS would allocate virtual addresses
starting at the same place each time it ran. This unfortunately is
not the property of Msoft Win95 or WinNT (there is no guarantee) so
I don't use the implementation much these days. If/when tunes gets
to the stage of needing such a thing I would be happy to write it
for you. A copy of a non-persistent multi-heap malloc implemented
with skip-lists 'cmalloc.zip' is present on several archives including
Simtel. It works very well in a demand paged environment since it
does not keep bookkeeping info in the allocated space.

--------------------------------------------------------
Norman David Culver   ndc@icanect.net  (954)463-4754
Oxbow Software 1323 SE 17th St. Ft. Lauderdale, FL 33316
--------------------------------------------------------