WIP: Use final to make Error::type_id un-overridable

[mu001999]

Tracking: #60784

After #151783, we can declare final assoc functions. And Error::type_id is one motivation of RFC-3678. This PR uses final to make Error::type_id un-overridable and this should helps the stabilization of Error::type_id.

[rustbot]

r? @madsmtm

rustbot has assigned @madsmtm.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: compiler
• compiler expanded to 69 candidates
• Random selection from 16 candidates

[jieyouxu]

This PR uses final to make Error::type_id un-overridable and this should helps the stabilization of Error::type_id.

(Wouldn't that be publicly exposing unstable language surface via a stable API? Anyway, not a library reviewer)

[madsmtm]

I don't think so, Error::type_id is not a stable API?

[rust-log-analyzer]

The job pr-check-2 failed! Check out the build log: (web) (plain enhanced) (plain)

Click to see the possible cause of the failure (guessed by this bot)

/dev/sda15      105M  6.2M   99M   6% /boot/efi
tmpfs           1.6G   12K  1.6G   1% /run/user/1001
================================================================================

Sufficient disk space available (95787708KB >= 52428800KB). Skipping cleanup.
##[group]Run echo "[CI_PR_NUMBER=$num]"
echo "[CI_PR_NUMBER=$num]"
shell: /usr/bin/bash --noprofile --norc -e -o pipefail {0}
---
test num::bignum::test_mul_small_overflow ... ok
test num::bignum::test_sub ... ok
test num::bignum::test_sub_underflow_2 ... ok
test num::bignum::test_sub_underflow_1 ... ok
test num::carryless_mul::carrying_carryless_mul ... ok
test num::carryless_mul::carryless_mul_u16 ... ok
test num::carryless_mul::carryless_mul_u128 ... ok
test num::carryless_mul::carryless_mul_u32 ... ok
test num::carryless_mul::carryless_mul_u64 ... ok
test num::carryless_mul::carryless_mul_u8 ... ok
test num::carryless_mul::widening_carryless_mul ... ok
test num::dec2flt::decimal::check_fast_path_f16 ... ok
test num::const_from::from ... ok
test num::dec2flt::decimal::check_fast_path_f64 ... ok
test num::dec2flt::decimal::check_fast_path_f32 ... ok
test num::dec2flt::decimal_seq::test_parse ... ok
---
test time::div ... ok
test time::div_duration_f32 ... ok
test time::div_duration_f64 ... ok
test time::duration_const ... ok
test time::duration_fp_boundaries ... ok
test time::duration_fp_div_nan ... ok
test time::duration_fp_div_negative ... ok
test time::duration_fp_div_negzero ... ok
test time::duration_fp_div_overflow ... ok
test time::duration_fp_div_poszero ... ok
test time::duration_fp_mul_nan ... ok
test time::duration_fp_mul_negative ... ok
test time::duration_fp_mul_neginfinity ... ok
test time::duration_fp_mul_overflow ... ok
test time::duration_fp_mul_posinfinity ... ok
test time::from_days_overflow ... ok
test time::from_hours_overflow ... ok
test time::from_mins_overflow ... ok
test time::from_nanos_u128_overflow ... ok
test time::from_neg_zero ... ok
---
failures:

---- io::error::tests::test_custom_error_packing stdout ----

thread 'io::error::tests::test_custom_error_packing' (43508) panicked at library/std/src/io/error/tests.rs:133:5:
assertion `left matches right` failed
  left: Custom(Custom { kind: Uncategorized, error: Bojji(true) })
 right: ErrorData::Custom(Custom { kind: ErrorKind::Uncategorized, error, }) if
error.downcast_ref::<Bojji>().as_deref() == Some(&Bojji(true))
stack backtrace:
   0: __rustc::rust_begin_unwind
   1: core::panicking::panic_fmt
   2: core::panicking::assert_failed_inner
   3: core::panicking::assert_matches_failed::<std::io::error::ErrorData<&std::io::error::Custom>>
   4: <std::io::error::tests::test_custom_error_packing::{closure#0} as core::ops::function::FnOnce<()>>::call_once
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
---- io::error::tests::test_custom_error_packing stdout end ----
---- io::error::tests::test_downcasting stdout ----

thread 'io::error::tests::test_downcasting' (43510) panicked at library/std/src/io/error/tests.rs:51:5:
assertion failed: err.get_ref().unwrap().is::<TestError>()
stack backtrace:
   0: __rustc::rust_begin_unwind
   1: core::panicking::panic_fmt
   2: core::panicking::panic
   3: <std::io::error::tests::test_downcasting::{closure#0} as core::ops::function::FnOnce<()>>::call_once
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
---- io::error::tests::test_downcasting stdout end ----
---- io::error::tests::test_std_io_error_downcast stdout ----

thread 'io::error::tests::test_std_io_error_downcast' (43515) panicked at library/std/src/io/error/tests.rs:157:40:
called `Result::unwrap()` on an `Err` value: Custom { kind: Other, error: Bojji(true) }
stack backtrace:
   0: __rustc::rust_begin_unwind
   1: core::panicking::panic_fmt
   2: core::result::unwrap_failed
   3: <std::io::error::tests::test_std_io_error_downcast::{closure#0} as core::ops::function::FnOnce<()>>::call_once

[mu001999]

Well, current implementation of final will exclude final methods from the vtable. So TypeId::of::<Self>() will always return the type id of dyn Error.

cc @joshtriplett, what do you think about this?

---

Maybe we can make it in the vtable if it is dyn-compatible, but I think this may make the behavior of final unpredictable.

[madsmtm]

Not a library reviewer either:
r? libs

[mu001999]

This should work after #153696

@rustbot blocked