Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Expand and organize offset_of! documentation. #138323

Merged
merged 2 commits into from
Mar 16, 2025
Merged
Changes from 1 commit
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
62 changes: 41 additions & 21 deletions library/core/src/mem/mod.rs
Original file line number Diff line number Diff line change
@@ -1249,42 +1249,60 @@ impl<T> SizedTypeProperties for T {}

/// Expands to the offset in bytes of a field from the beginning of the given type.
///
/// Structs, enums, unions and tuples are supported.
/// The type may be a `struct`, `enum`, `union`, or tuple.
///
/// Nested field accesses may be used, but not array indexes.
/// The field may be a nested field (`field1.field2`), but not an array index.
/// The field must be visible to the call site.
///
/// The offset is returned as a [`usize`].
///
/// If the nightly-only feature `offset_of_enum` is enabled,
/// variants may be traversed as if they were fields.
/// `enum` variants may be traversed as if they were fields.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggestion: maybe link to the tracking issue?

[`offset_of_enum`]: https://github.com/rust-lang/rust/issues/120141

But also, this is describing the interaction of this stable macro w/ an unstable feature. Maybe move this to a section like

# Interactions with unstable features

or maybe move this bit to docs for offset_of_enum? 🤔

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it's OK in this particular case to reference unstable features, even if we normally don't, since this seems like a natural place for people to look for that functionality. I don't think we have offset_of_enum! (yet?) - not sure what you mean by that.

Copy link
Member

@jieyouxu jieyouxu Mar 11, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Right.

I don't think we have offset_of_enum! (yet?) - not sure what you mean by that.

Not sure why I wrote this lmao.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I assumed you meant to put the documentation in the unstable book, and I think that makes sense. Should I not?

(Also while writing that documentation, I found a bug-or-design-question: #138327.)

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Right, I think that's what I wanted to say lol

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we should at minimum link to those docs, even if we don't have a stable thing to link to - otherwise people will handroll stuff I suspect and we've seen that cause UB, etc. in the past.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

IOW, probably do nothing here :D

Copy link
Contributor Author

@kpreid kpreid Mar 11, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I've now moved all the unstable feature documentation to the unstable book, except for links to it from offset_of, and summaries.

/// Variants themselves do not have an offset.
///
/// Visibility is respected - all types and fields must be visible to the call site:
///
/// ```
/// mod nested {
/// #[repr(C)]
/// pub struct Struct {
/// private: u8,
/// }
/// }
/// # Offsets of, and in, dynamically sized types
///
/// // assert_eq!(mem::offset_of!(nested::Struct, private), 0);
/// // ^^^ error[E0616]: field `private` of struct `Struct` is private
/// ```
/// The field’s type must be [`Sized`], but it may be located in a [dynamically sized] container.
/// If the field type is dynamically sized, then you cannot use `offset_of!` (since the field's
/// alignment, and therefore its offset, may also be dynamic) and must take the offset from an
/// actual pointer to the container instead.
///
/// Only [`Sized`] fields are supported, but the container may be unsized:
/// ```
/// # use core::mem;
/// # use core::fmt::Debug;
/// #[repr(C)]
/// pub struct Struct {
/// pub struct Struct<T: ?Sized> {
/// a: u8,
/// b: [u8],
/// b: T,
/// }
///
/// assert_eq!(mem::offset_of!(Struct, a), 0); // OK
/// // assert_eq!(mem::offset_of!(Struct, b), 1);
/// // ^^^ error[E0277]: doesn't have a size known at compile-time
/// #[derive(Debug)]
/// #[repr(C, align(4))]
/// struct Align4(u32);
///
/// assert_eq!(mem::offset_of!(Struct<dyn Debug>, a), 0); // OK — Sized field
/// assert_eq!(mem::offset_of!(Struct<Align4>, b), 4); // OK — not DST
///
/// // assert_eq!(mem::offset_of!(Struct<dyn Debug>, b), 1);
/// // ^^^ error[E0277]: ... cannot be known at compilation time
///
/// // To obtain the offset of a !Sized field, examine a concrete value
/// // instead of using offset_of!.
/// let value: Struct<Align4> = Struct { a: 1, b: Align4(2) };
/// let ref_unsized: &Struct<dyn Debug> = &value;
/// let offset_of_b = unsafe {
/// (&raw const ref_unsized.b).byte_offset_from_unsigned(ref_unsized)
/// };
/// assert_eq!(offset_of_b, 4);
/// ```
///
/// If you need to obtain the offset of a field of a `!Sized` type, then, since the offset may
/// depend on the particular value being stored (in particular, `dyn Trait` values have a
/// dynamically-determined alignment), you must retrieve the offset from a specific reference
/// or pointer, and so you cannot use `offset_of!` to work without one.
///
/// # Layout is subject to change
///
/// Note that type layout is, in general, [subject to change and
/// platform-specific](https://doc.rust-lang.org/reference/type-layout.html). If
/// layout stability is required, consider using an [explicit `repr` attribute].
@@ -1358,6 +1376,8 @@ impl<T> SizedTypeProperties for T {}
///
/// assert_eq!(mem::offset_of!(Option<&u8>, Some.0), 0);
/// ```
///
/// [dynamically sized]: https://doc.rust-lang.org/reference/dynamically-sized-types.html
#[stable(feature = "offset_of", since = "1.77.0")]
#[allow_internal_unstable(builtin_syntax)]
pub macro offset_of($Container:ty, $($fields:expr)+ $(,)?) {
Loading