docs: Changed docs for size_of to describe size as a stride offset

[cramertj]

Current documentation for std::mem::size_of is ambiguous, and the documentation for std::intrinsics::size_of incorrectly defines size.

This fix re-defines size as the offset in bytes between successive instances of a type, as described in LLVM's getTypeAllocSize.

Fixes: #33266

[rust-highfive]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @alexcrichton (or someone else) soon.

If any changes to this PR are deemed necessary, please add them as extra commits. This ensures that the reviewer can see what has changed since they last reviewed the code. Due to the way GitHub handles out-of-date commits, this should also make it reasonably obvious what issues have or haven't been addressed. Large or tricky changes may require several passes of review and changes.

Please see the contribution instructions for more information.

[durka]

If I only have one T, what does "between successive values" mean? Does Box::new(T) allocate mem::size_of::<T>() bytes or fewer?

[cramertj]

@durka I can't give you a definitive answer on that-- perhaps @pnkfelix can shed some light.

I would speculate, however, that it would allocate exactly LLVM's getTypeAllocSize, seeing as this seems to be literally the allocation size of the type. This would guarantee alignment for future allocations of the type, although I'd imagine the actual heap-alignment is more heavily dependent upon the allocator in question. This would also fall in line with @pnkfelix's comment that the behavior of std::mem::size_of resembles that of sizeof(T) in C/C++, where it is common to see expressions such as malloc(sizeof(T)).

Note, too, that LLVM also provides a getTypeStoreSize specifying the number of bytes overwritten by the object in question, a more traditional/intuitive understanding of size_of.

[alexcrichton]

@bors: r+ 38c8836

Thanks!

Does Box::new(T) allocate mem::size_of::() bytes or fewer?

Box allocates size_of::<T>, so it'll allocate the padding bytes as well

[durka]

@alexcrichton I think the answer to my question should go into the size_of docs as well

[alexcrichton]

Sounds like something that'd be more appropriate for Box rather than the core mem::size_of as it's an implementation detail of that particular pointer type.