On Sep 16, 2011, at 5:14 AM, Andreas Rossberg wrote:
> ...
> Consider writing a server. If I cannot rely on tail call optimization
> then writing its message loop as a recursive function (e.g. actors
> style) would be incorrect. If I cannot rely on GC, then allocating an
> object for each received message would be incorrect.
But at the extreme you can't rely on GC. Any allocation may be the one that
exhausts actual available storage or triggers a interminable GC thrashing of
virtual memory. In the real world these are quality of implementation issues
that developers must deal with via load testing on real implementations.
> If I cannot rely
> on weak maps, then, say, mapping every message object to a return IP
> would be incorrect.
A WeakMap reliably maintains a mapping from an object to a value. If you have
access to an object key then you can rely upon a WeakMap preserving any
mappings it has for that key. What you can not rely upon is the timely
recycling of any resources that are used to represent inaccessible WeakMap
mappings. This is the exact situation that exists for any ECMAScript dynamic
allocation. There are no guarantees in the ES spec. that resources allocated
with an inaccessible object will be freed in a timely manner or ever. Given
the state of the art of memory management isn't clear how you could make such
guarantees without unreasonably limiting implementors.
Allen
_______________________________________________
es-discuss mailing list
[email protected]
https://mail.mozilla.org/listinfo/es-discuss
