WG14: C pointer values are explicitly *not* simple bit patterns; they can well be path-dependent http://www.open-std.org/JTC1/SC22/WG14/www/docs/n2012.htm#clarifying-the-c-memory-object-model-pointer-provenance …
-
-
Replying to @whitequark
@whitequark ... then it must treat all pointers as if aliased, and not try to do any fancy/stupid/annoying optimizations.1 reply 0 retweets 0 likes -
Replying to @whitequark
@whitequark at least, I don't think? But still, even if -fno-strict-aliasing were mandated in the standard, would that be bad?1 reply 0 retweets 0 likes -
Replying to @DrPizza
@DrPizza@whitequark Yes. GCC should rename -fno-strict-aliasing to -fno-vectorization so people stop repeating this mistake.3 replies 0 retweets 0 likes -
Replying to @RichFelker
@RichFelker@whitequark if I want vectorization I have SIMD intrinsics.1 reply 0 retweets 0 likes -
Replying to @DrPizza
@DrPizza@whitequark Uhg. The whole purpose of optimizing compilers is to keep people from writing asm or equivalent monstrosities.2 replies 0 retweets 0 likes -
Replying to @RichFelker
@RichFelker@whitequark it's (demonstrably) hard enough to write bug-free code without the compiler subverting our intentions.1 reply 0 retweets 0 likes
@DrPizza @whitequark It's pretty easy to write code that doesn't access memory using the wrong types. Just don't write bogus casts.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.