Re:The "C is Efficient" Language Fallacy (2006)
by AshleysBrain on Monday January 09, 2012 @10:32AM
(#3442854)
I'm not sure it's as rare as you make it out to be - AFAIK most compilers that support restrict declare malloc() and new to return restrict pointers, so
followed by some loop should be able to spot that x and y are not aliased. It should also be able to propagate the restrict as pointers are passed around, at least to some extent.
[ Parent
| on HN ]
Re:The "C is Efficient" Language Fallacy (2006)
by lukesandberg on Monday January 09, 2012 @11:40PM
(#3445712)
would the c compiler actually pull that restrict annotation forward without declaring the locals as restrict?