Expose malloc_usable_size in the high-level crate.
#18
Conversation
|
Or perhaps it is enough for users to depend on the |
src/lib.rs
Outdated
| /// Any discrepancy between the requested allocation size | ||
| /// and the size reported by this function should not be depended on, | ||
| /// since such behavior is entirely implementation-dependent. | ||
| pub unsafe fn usable_size<T: Sized>(ptr: *const T) -> usize { |
There was a problem hiding this comment.
Is : Sized needed here?
There was a problem hiding this comment.
I guess it’s not in the sense that it’s the default, but ?Sized would cause the cast to *const c_void to not work… won’t it? Hmm, I just tried and it looks like an casting unsized raw pointer to sized is allowed, though documenting it is "TODO" in https://doc.rust-lang.org/nightly/nomicon/casts.html. I don’t know if it’s desirable here, though.
There was a problem hiding this comment.
Oh sorry I just mean that : Sized is the default, so the choices here are basically : ?Sized or no bounds at all, but I like the idea of using a sized pointer as jemalloc isn't the one allocating vtables.
| /// rather it is provided solely as a tool for introspection purposes. | ||
| /// Any discrepancy between the requested allocation size | ||
| /// and the size reported by this function should not be depended on, | ||
| /// since such behavior is entirely implementation-dependent. |
There was a problem hiding this comment.
Mind adding an # Unsafety section here as well?
There was a problem hiding this comment.
Done, though I couldn’t find info on undefined behavior with this function in http://jemalloc.net/jemalloc.3.html so I based it on what I recall hearing about allocators in general.
SimonSapin
force-pushed
the
usable_size
branch
from
6de2ff5 to
d24b998
Compare
No description provided.