Comment 3 for bug 309448
Revision history for this message
| Nikodemus Siivola (nikodemus) wrote Re: compiler performance fiasco involving type inference and UNION-TYPE | #3 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
0e1f616
(Get the code!)
It is worth noting that we cannot simplify
(or (integer 0 0) (integer 2 2) (integer 4 4))
into eg.
(integer 0 4)
in general, because
(not (or (integer 0 0) (integer 2 2) (integer 4 4)))
=> (or (intger * -1) (integer 1 1) (integer 3 3) (integer 5 *))
whereas
(not (integer 0 4))
=> (or (integer * -1) (integer 5 *))
...so the place to make these simplification decisions is in the compiler, which hopefully has enough contextual knowledge to figure out when it is safe -- and not the type system.