Nonparametric dropck (tracking issue for RFC 1238)

d2a2588
Opened by Niko Matsakis at 2021-11-10 18:50:31

Tracking issue for rust-lang/rfcs#1238.

  • [x] subtask: #[may_dangle] attribute #34761
  • [ ] can't close this until support for we have stable support for 3rd party collection types

cc @pnkfelix

  1. Tracking issue for rust-lang/rfcs#1238.

    • [x] subtask: #[may_dangle] attribute #34761
    • [ ] can't close this until we have stable support for 3rd party collection types

    cc @pnkfelix

    Felix S Klock II at 2019-12-30 16:18:33

  2. triage: P-high

    Niko Matsakis at 2015-09-18 19:06:37

  3. triage: P-medium

    Niko Matsakis at 2015-10-29 11:53:21

  4. I think this was since implemented, so closing (but please reopen if there's something left!)

    Alex Crichton at 2016-02-18 18:56:14

  5. I think this should be reopened with the B-unstable label. The RFC is implemented, but the #[unsafe_destructor_blind_to_params] attribute is still behind the dropck_parametricity feature gate which points to this as its tracking issue.

    #[unsafe_destructor_blind_to_params] is used for Vec and a number of other types defined in the standard library. It isn’t any less useful for types outside the standard library. I’ve needed it for an arena allocator.

    Simon Sapin at 2016-05-04 14:13:23

  6. I agree.

    Niko Matsakis at 2016-05-04 14:21:54

  7. Triage: feature gate still active, not aware of any work in this area.

    Steve Klabnik at 2019-03-12 17:38:50

  8. (i doubt I'm going to be attacking this problem at any point in the foreseeable future. Unassigning self.)

    Felix S Klock II at 2019-05-17 10:22:46

  9. #62568 has removed the feature gate since, but this issue says:

    can't close this until we have stable support for 3rd party collection types

    est31 at 2020-09-18 09:29:51

  10. From today's @rust-lang/lang meeting: The feature seems to be working, but it may need further refinement in order to expose it as stable.

    Josh Triplett at 2021-11-10 18:50:31