-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Monday 02 June 2008 18:12 pm, Peter Dimov wrote:
The pseudocode isn't even O(N); it's O(NlnN) at best, maybe worse. :-) Do note that it doesn't necessarily dispatch a single request per wait_for_any call though.
Your pseudocode putting the futures in a set does bring up a missing feature in both the proposals. They don't provide an operator<() that is a strict weak ordering so they can be used as keys. A future_handle should have no problem providing an ordering where two future_handles are equivalent if and only if they both reference the same promise. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) iD8DBQFIRIYN5vihyNWuA4URAukHAJwNo5McItLrdn28lIq2cDEZR8UITQCfdqOt alZDXfzr+2bV77IniW4LdNU= =jzgI -----END PGP SIGNATURE-----