> > You are right, this would foul up the nextPut primitive for WriteStream, so
> you would 
> > have to add the new field separately after those defined in ReadStream and
> WriteStream
> >  rather than doing it in one place in PositionableStream.
> Duh, talk about missing the obvious one! Still, the prim recovery code
> handles it just fine....
Err, it turns out that the method for primitiveNextPut uses
StreamReadLimitIndex rather than StreamWriteLimitIndex.
So, in fact the changes I mis-made won't break anything; but that's only
because of another bug!  I wonder how this will manifest in normal use?
Doesn't look like it can actually write beyond the collection's bounds
assuming #commonVariable:at:put:cacheIndex: but it could  cause some fun if
the writeLimit is less than readLimit.

tim

-- 
Useful random insult:- "Body by Fisher -- brains by Mattel."
Tim Rowledge:  rowledge at interval.com (w)  +1 (650) 842-6110 (w)
 tim at sumeru.stanford.edu (h)  <http://sumeru.stanford.edu/tim>