There are some other spots in the compiler that still refer to OIBITs,
but I don't think changing those now is worth it since they are internal
and not particularly relevant to this PR.

How extensive is the usage of the name? I think if we're doing the name change, we should do it everywhere

There are some other spots in the compiler that still refer to OIBITs,
but I don't think changing those now is worth it since they are internal
and not particularly relevant to this PR.

How extensive is the usage of the name? I think if we're doing the name change, we should do it everywhere

Let me know what you want me to update!

$ rg -i '(oibit|opt-?in,?[ |_]?built-?in)' -g '!src/tools/rust-analyzer'
src/test/ui/phantom-oibit.stderr:  --> $DIR/phantom-oibit.rs:21:12
src/test/ui/phantom-oibit.stderr:  --> $DIR/phantom-oibit.rs:26:12
src/test/ui/phantom-oibit.rs:// Ensure that OIBIT checks `T` when it encounters a `PhantomData<T>` field, instead of checking
src/test/ui/traits/traits-inductive-overflow-supertrait-oibit.rs:// OIBIT-based version of #29859, supertrait version. Test that using
src/test/ui/traits/traits-inductive-overflow-supertrait-oibit.rs:// a simple OIBIT `..` impl alone still doesn't allow arbitrary bounds
src/test/ui/traits/traits-inductive-overflow-supertrait-oibit.stderr:  --> $DIR/traits-inductive-overflow-supertrait-oibit.rs:8:19
src/test/ui/traits/traits-inductive-overflow-supertrait-oibit.stderr:  --> $DIR/traits-inductive-overflow-supertrait-oibit.rs:16:23
src/test/rustdoc/auxiliary/rustdoc-impl-parts-crosscrate.rs:pub auto trait AnOibit {}
src/test/rustdoc/impl-parts.rs:pub auto trait AnOibit {}
src/test/rustdoc/impl-parts.rs://     "impl<T: Clone> !AnOibit for Foo<T> where T: Sync,"
src/test/rustdoc/impl-parts.rs:// @has impl_parts/trait.AnOibit.html '//*[@class="item-list"]//code' \
src/test/rustdoc/impl-parts.rs://     "impl<T: Clone> !AnOibit for Foo<T> where T: Sync,"
src/test/rustdoc/impl-parts.rs:impl<T: Clone> !AnOibit for Foo<T> where T: Sync {}
src/test/rustdoc/impl-parts-crosscrate.rs:// @has implementors/rustdoc_impl_parts_crosscrate/trait.AnOibit.js Bar
src/test/rustdoc/impl-parts-crosscrate.rs:// @has - !AnOibit
src/test/rustdoc/impl-parts-crosscrate.rs:impl<T: Send> !rustdoc_impl_parts_crosscrate::AnOibit for Bar<T>
compiler/rustc_middle/src/mir/mod.rs:    /// therefore don't affect the OIBIT or outlives properties of the
compiler/rustc_feature/src/removed.rs:    /// Allows features specific to OIBIT (now called auto traits).
compiler/rustc_feature/src/removed.rs:    (removed, optin_builtin_traits, "1.0.0", Some(13231), None,
compiler/rustc_feature/src/active.rs:    /// Renamed from `optin_builtin_traits`.
compiler/rustc_middle/src/traits/mod.rs:    /// There are some exceptions, e.g., around OIBITS and
compiler/rustc_mir_build/src/build/expr/as_rvalue.rs:                // and therefore is not considered during generator OIBIT
compiler/rustc_error_codes/src/error_codes/E0380.md:the [opt-in builtin traits RFC][RFC 19].
compiler/rustc_error_codes/src/error_codes/E0192.md:information see the [opt-in builtin traits RFC][RFC 19].
RELEASES.md:* The final bits of [OIBIT landed][oibit-final], meaning that traits
RELEASES.md:[oibit-final]: https://github.com/rust-lang/rust/pull/21689
RELEASES.md:      footguns and are collectively known as [opt-in built-in
RELEASES.md:      traits][oibit] (though `Sync` and `Send` will soon become pure
RELEASES.md:[oibit]: https://github.com/rust-lang/rfcs/blob/master/text/0019-opt-in-builtin-traits.md
library/rtstartup/rsbegin.rs:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]
library/rtstartup/rsend.rs:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]
library/proc_macro/src/lib.rs:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]
library/std/src/lib.rs:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]
library/core/src/lib.rs:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]
library/alloc/src/lib.rs:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]

Squashed.

Let me know what you want me to update!

Ideally all of it, even the filenames. Though it's also fine to just paste your list in an E-easy issue if you prefer.

Should I change the name that the RFC is referenced by in the error codes? I feel like it might be confusing to see them called "auto traits" everywhere else, even though the RFC is called OIBIT. Maybe call it OIBIT, but mention that it's now called auto traits? Or something else?

These are the remaining references:

❯ rg -i '(oibit|opt-?in,?[ |_]?built-?in)' -g '!src/tools/rust-analyzer'
compiler/rustc_feature/src/removed.rs
74:    /// Allows features specific to OIBIT (now called auto traits).
76:    (removed, optin_builtin_traits, "1.0.0", Some(13231), None,

compiler/rustc_feature/src/active.rs
216:    /// Renamed from `optin_builtin_traits`.

compiler/rustc_span/src/symbol.rs
773:        optin_builtin_traits,

compiler/rustc_error_codes/src/error_codes/E0380.md
12:the [opt-in builtin traits RFC][RFC 19].

compiler/rustc_error_codes/src/error_codes/E0192.md
18:information see the [opt-in builtin traits RFC][RFC 19].

compiler/rustc_data_structures/src/lib.rs
18:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]

src/tools/cargo/tests/testsuite/custom_target.rs
65:        // Requires features no_core, lang_items, optin_builtin_traits
87:                #![feature(optin_builtin_traits)]

library/std/src/lib.rs
289:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]

library/core/src/lib.rs
122:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]

library/rtstartup/rsbegin.rs
17:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]

library/rtstartup/rsend.rs
5:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]

library/proc_macro/src/lib.rs
31:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]

library/alloc/src/lib.rs
115:#![cfg_attr(bootstrap, feature(optin_builtin_traits))]

RELEASES.md
7837:* The final bits of [OIBIT landed][oibit-final], meaning that traits
7941:[oibit-final]: https://github.com/rust-lang/rust/pull/21689
8117:      footguns and are collectively known as [opt-in built-in
8118:      traits][oibit] (though `Sync` and `Send` will soon become pure
8225:[oibit]: https://github.com/rust-lang/rfcs/blob/master/text/0019-opt-in-builtin-traits.md

By the way, should we mention that the name was changed in the relnotes? Probably not, mainly because it's a nightly feature.

Ah yes, that makes sense! Thanks

@bors r+

📌 Commit 82dc99b has been approved by oli-obk

⌛ Testing commit 82dc99b with merge ec039bd...

Ah yes, that makes sense!

Hmm, what do you mean? Were you referring to my suggestion that we mention this in the release notes?

Sorry, I responded to what I thought was the general statement you made

Should I change the name that the RFC is referenced by in the error codes?

No, let's not change history 😆

I feel like it might be confusing to see them called "auto traits" everywhere else, even though the RFC is called OIBIT. Maybe call it OIBIT, but mention that it's now called auto traits?

You do mention that on the feature gate docs, so I think that's fine. We have lots of RFCs that don't make sense by their name anymore and that is fine.

☀️ Test successful - checks-actions
Approved by: oli-obk
Pushing ec039bd to master...

📣 Toolstate changed by #79336!

Tested on commit ec039bd.
Direct link to PR: #79336

💔 rls on windows: test-pass → build-fail (cc @Xanewok).
💔 rls on linux: test-pass → build-fail (cc @Xanewok).
💔 rustfmt on windows: test-pass → build-fail (cc @topecongiro @calebcartwright).
💔 rustfmt on linux: test-pass → build-fail (cc @topecongiro @calebcartwright).

Okay, thanks for the explanation!