On Sun, 30 Apr 2000, Dan Ingalls wrote:
[snip]
> It is unfortunate that this happened.  Of course there are many ways
> that it could have been prevented, but the absoluteness of the failure
> is out of the ordinary.  I do not think it would be all that hard to
> allow Squeak to fail from the save primitive, and keep running.

IIRC, the problem was one of running out of disk space. Presumably, it
wouldn't be difficult to check for free space *before* initiating the
save, and to provide a warning when that is less than, oh, the memory used
by the image plus some reasonable buffer.

Cheers,
Bijan Parsia.