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

[project] Initial commit #1934

Open
wants to merge 33 commits into
base: main
Choose a base branch
from

Conversation

joshlf
Copy link
Member

@joshlf joshlf commented Oct 18, 2024

TODO: Could we support unsized types so long as we never call offset_of!
on the trailing unsized field?

TODO: Use Name type parameter trick to emit better error messages:
https://play.rust-lang.org/?version=stable&mode=debug&edition=2021&gist=97771667892860935bea60b218f8fdf4


This PR is on branch projection-sized.

joshlf and others added 26 commits October 12, 2024 14:06
We already have the merge queue; running on push is redundant.
Upgrade our MSRV to 1.65 and remove version detection logic prior to
that version.
* Release 0.9.0-alpha.0

Upgrade our MSRV to 1.65 and remove version detection logic prior to
that version.

* Enable clippy::missing_const_for_fn

While we're here, remove defensive programming against bug in
`Layout::from_size_align` which is no longer needed on our new MSRV.
Now that our MSRV is 1.65, we can clean up some code.

Makes progress on #67
Also clean up some code for 0.9.
* Upgrade some code for MSRV 1.65

Now that our MSRV is 1.65, we can clean up some code.

Makes progress on #67

* Upgrade versions of some dependencies

Now that our MSRV is 1.65, it unlocks upgrading some dependencies'
versions.
Bumps [github/codeql-action](https://github.com/github/codeql-action) from 3.26.12 to 3.26.13.
- [Release notes](https://github.com/github/codeql-action/releases)
- [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md)
- [Commits](github/codeql-action@c36620d...f779452)

---
updated-dependencies:
- dependency-name: github/codeql-action
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <[email protected]>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Bumps [Swatinem/rust-cache](https://github.com/swatinem/rust-cache) from 2.7.3 to 2.7.5.
- [Release notes](https://github.com/swatinem/rust-cache/releases)
- [Changelog](https://github.com/Swatinem/rust-cache/blob/master/CHANGELOG.md)
- [Commits](Swatinem/rust-cache@23bce25...82a92a6)

---
updated-dependencies:
- dependency-name: Swatinem/rust-cache
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <[email protected]>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Previously, we supported the `AtLeast` bound, which was used to describe
a subset relationship in which `I: AtLeast<J>` implied that `I` as at
least as restrictive as `J`. However, as described in #1866, this
incorrectly models invariants as monotonic. In reality, invariants both
provide guarantees but also *require* guarantees.

This commit takes a step in the direction of resolving #1866 by removing
`AtLeast`. Uses of `AtLeast<Shared>` are replaced by a new `Reference`
trait, which is implemented for `Shared` and `Exclusive`. This serves
two purposes: First, it makes it explicit what this bound means.
Previously, `AtLeast<Shared>` had an ambiguous meaning, while
`Reference` means precisely that an invariant is either `Shared` or
`Exclusive` and nothing else. Second, it paves the way for #1183, in
which we may add new aliasing invariants which convey ownership. In that
case, it will be important for existing methods to add `Reference`
bounds when those methods would not be sound in the face of ownership
semantics.

We also inline the items in the `invariant` module, which were
previously generated by macro. The addition of the `Reference` trait did
not play nicely with that macro, and we will likely need to go further
from the macro in order to fix #1839 – this fix will likely require
making aliasing invariants meaningfully different than other invariants,
for example by adding an associated type.

Makes progress on #1866
Explain why we only support concrete types so that future authors won't
spuriously add support for them.
`PtrInner` carries all invariants which are not controlled by type
parameters. Since `PtrInner` does not promise to uphold aliasing,
alignment, or validity, we can move some utility methods to `PtrInner`
which previously were responsible for maintaining invariants orthogonal
to their purpose.

Makes progress on #1892 (still needs to be fixed on v0.8.x)
Closes #1890
This prepares us for future changes which will significantly increase
the amount of code in the `invariant` module.

Also merge `aliasing_safety` into this new file.
When the aliasing mode is `Any`, `Ptr<'a, T>` is invariant in `'a` and
`T`. When the aliasing mode is `Shared` or `Exclusive`, `Ptr` has the
same variance as `&'a T` and `&'a mut T` respectively.

Makes progress on #1839
`AliasingSafe` is really about whether a pointer permits unsynchronized
reads - either because the referent contains no `UnsafeCell`s or because
the aliasing mode is `Exclusive`. Previously, `AliasingSafe` was not
named consistent with this meaning, and was a function of a *pair* of
types rather than of a single type. This commit fixes both oversights.

While we're here, we also add `Read` bounds in some places, allowing us
to simplify many safety comments.
For aliasing, use `Inaccessible`. For alignment and validity, use
`Unknown`.
This commit adds a framework which supports encoding in the type system
any `I -> I` mapping where `I` is any `Invariant` type. This permits us
to make `cast_unsized`'s return value smarter, and as a result, allows
us to remove a lot of `unsafe` code.

Makes progress on #1122
gherrit-pr-id: I003d5360d1b7f7882a71490813eca50b39025f14
These tests depend on `src` being aligned to multiples of 2. With
this commit, that dependency is explicitly enforced.
We eventually hope to make use of `#[marker]` traits once they're
stable. This permits us to test to make sure the feature is as we expect
and that our intended usage works.

gherrit-pr-id: I3a111bf5647fdcc9805cbadf36f729ac69b28509
And allow `non_local_definitions`.
We eventually hope to make use of `#[marker]` traits once they're
stable. This permits us to test to make sure the feature is as we expect
and that our intended usage works.

gherrit-pr-id: I3a111bf5647fdcc9805cbadf36f729ac69b28509
@joshlf
Copy link
Member Author

joshlf commented Oct 18, 2024

This supersedes #1933, which has a soundness hole.

This only supports sized types for now, but maybe we'll figure out a way to lift that limitation eventually.

@joshlf joshlf force-pushed the I54eb863c96f056738f58727fc0fec4e1630bf97a branch from 874f368 to 79841ca Compare October 18, 2024 05:55
@joshlf joshlf force-pushed the I54eb863c96f056738f58727fc0fec4e1630bf97a branch from 79841ca to b39b4b1 Compare October 18, 2024 13:57
@joshlf joshlf force-pushed the I54eb863c96f056738f58727fc0fec4e1630bf97a branch 2 times, most recently from 451d820 to 21338e0 Compare October 19, 2024 01:55
Makes progress on #1931

gherrit-pr-id: If0e198c377137dd941ebd5dc68787766a593e1eb
TODO: Could we support unsized types so long as we never call offset_of!
on the trailing unsized field?

TODO: Use `Name` type parameter trick to emit better error messages:
https://play.rust-lang.org/?version=stable&mode=debug&edition=2021&gist=97771667892860935bea60b218f8fdf4

gherrit-pr-id: I54eb863c96f056738f58727fc0fec4e1630bf97a
@joshlf joshlf force-pushed the I54eb863c96f056738f58727fc0fec4e1630bf97a branch from 21338e0 to 5ef235b Compare October 21, 2024 20:03
impl<'a, T, I> Ptr<'a, T, I>
where
T: 'a + ?Sized,
I: Invariants,
Copy link
Contributor

Choose a reason for hiding this comment

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

Note that this can only ever be sound for invariants that structurally propagate to their fields. This doesn't apply for the Pin invariant.

// NOTE: Doing the actual cast inside this function makes the `project!`
// macro simpler, but it also means that we can't handle both const code and
// unsized types at the same time. In order to be const, we can't call trait
// methods, but in order to support unsized types, we'd need to provide a
Copy link
Contributor

Choose a reason for hiding this comment

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

For slice DSTs, all you need is the offset of the last field, which is the same for all sizes. This could be stored as a const.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants