Constify `TypeId` ordering impls by raldone01 · Pull Request #101698 · rust-lang/rust

raldone01 · 2022-09-11T18:17:14Z

Tracking issue: #101871

Adding const ordering to TypeId allows rtti crates to optimize some casting scenarios (without transmuting to u64). This would also prevent these crates from breaking if the underlying type is changed from u64 to something different.

Feature gate: #![feature(const_cmp_type_id)]

rustbot · 2022-09-11T18:17:17Z

Hey! It looks like you've submitted a new PR for the library teams!

If this PR contains changes to any rust-lang/rust public library APIs then please comment with @rustbot label +T-libs-api -T-libs to tag it appropriately. If this PR contains changes to any unstable APIs please edit the PR description to add a link to the relevant API Change Proposal or create one if you haven't already. If you're unsure where your change falls no worries, just leave it as is and the reviewer will take a look and make a decision to forward on if necessary.

Examples of T-libs-api changes:

Stabilizing library features
Introducing insta-stable changes such as new implementations of existing stable traits on existing stable types
Introducing new or changing existing unstable library APIs (excluding permanently unstable features / features without a tracking issue)
Changing public documentation in ways that create new stability guarantees
Changing observable runtime behavior of library APIs

rust-highfive · 2022-09-11T18:17:17Z

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @scottmcm (or someone else) soon.

Please see the contribution instructions for more information.

fmease · 2022-09-11T22:22:00Z

Isn't this a bit weird? “impl defined here but it is not const” but it is const as shown?

Yes that is because this test doesn't use the const_trait_impl feature flag.
That always leads to this error if used this way.
The same error happens in src/test/ui/consts/const_cmp_type_id.rs if #![feature(const_trait_impl)] is removed, this is probably more of an issue in const_trait_impl than here.

onestacked · 2022-09-14T12:41:42Z

Makes public stable implementations of Ordering traits unstably const, so

@rustbot label +T-libs-api -T-libs

jhpratt

Diff LGTM.

scottmcm · 2022-09-24T04:09:36Z

Nominating for team discussions. The ordering of TypeIds becoming const-visible in particular feels potentially scary to me.

raldone01 · 2022-09-24T10:51:54Z

Some additional background that may help in the discussion:

I have created a crate called trait_cast_rs. It provides a custom Any type. That allows casts to dyn Traits.

Internally it has a const slice of function pointers with their associated TypeIds. Making TypeId const comparable would allow the crate to sort the slice at compile time. When downcasting at runtime a binary_search is performed to find the correct function pointer.

This has the advantage that the array must not be sorted on every program launch and the lookup slice can be a constant.

I currently use transmute<_, u64>(type_id) to sort the slice at compile time.

scottmcm · 2022-09-28T16:22:25Z

Per T-libs (see #102371 (comment)) we're going to wait on derive_const (or similar) before accepting more changes like this to library.
@rustbot label +S-Blocked

Please do not make TypeId "structural match" - instead, that needs to be removed - as per e.g. #77125 (comment) - sadly the one PR trying to do it was rejected and I'm not seeing a replacement just removing "structural match":
Pad size of TypeId and remove structural equality #99189

the impl StructualPartialEq was added because the test for issue-73976 failed as it used matches!.
It could be changed to a normal == though, but removing the structural match would be a breaking change (nightly only since const_type_id isn't stable yet).

I've added a commit that does those changes.

scottmcm · 2022-09-28T16:57:55Z

We discussed this in the lang meeting this week. It didn't result in anything that needs to block this for nightly, but I've added a note to the tracking issue that this precludes some possible implementations, and thus stabilizing it would need to be considered very carefully.

m-ou-se · 2022-10-04T15:17:18Z

Nominating this for the compiler team for input, since the compiler is the one generating the type ids. So question for the compiler team: will this get in the way of any future plans there might be with future typeid implementations? (And if so, is that a problem?)

apiraino · 2022-10-06T16:05:55Z

Issue was discussed in T-compiler meeting, I'm leaving here some notes; more will probably follow-up in this issue to reach a wider audience.

Leaving I-compiler-nominated for the moment.

apiraino · 2022-10-13T09:25:12Z

@scottmcm during T-compiler meeting (see previous comment), some questions arised (here, for example). Can you provide a bit of feedback to complement the discussion there? thanks

@rustbot label -i-compiler-nominated

eddyb · 2022-10-13T09:39:40Z

Adding const ordering to TypeId allows rtti crates to optimize some casting scenarios (without transmuting to u64).

If you want the hash, you should have never transmuted, but instead written a Hasher that is most efficient for write_u64 (ideally it should accept any data then truncate and/or mix but we found out the hard way that we can't really break people requiringwrite_u64 because it's kind of common in code that doesn't misuse transmute).

You can do this on stable Rust, and it should work all the way back to 1.0 AFAIK, so it's a good option to consider.

onestacked · 2023-01-16T19:23:21Z

Why did the CI not run?

scottmcm · 2023-01-16T19:24:11Z

Hmm, I'll try to re-kick it.

@bors r-

WaffleLapkin · 2023-01-16T19:34:09Z

@bors r=scottmcm
github is/was having some issues with github actions

bors · 2023-01-16T19:34:11Z

📌 Commit 73f8dc788b0e63d3ed4fd88fbbfd9ffa70c8acb9 has been approved by scottmcm

It is now in the queue for this repository.

onestacked · 2023-01-16T19:42:52Z

Should probably r- untill CI passes

scottmcm · 2023-01-16T19:44:48Z

@bors r-

onestacked · 2023-01-16T21:12:38Z

@scottmcm can you r+?

WaffleLapkin · 2023-01-17T07:34:47Z

@bors r=scottmcm

bors · 2023-01-17T07:34:48Z

📌 Commit 7355ab3 has been approved by scottmcm

It is now in the queue for this repository.

Rollup of 5 pull requests Successful merges: - rust-lang#101698 (Constify `TypeId` ordering impls) - rust-lang#106148 (Fix unused_parens issue for higher ranked function pointers) - rust-lang#106922 (Avoid unsafe code in `to_ascii_[lower/upper]case()`) - rust-lang#106951 (Remove ineffective run of SimplifyConstCondition) - rust-lang#106962 (Fix use suggestion span) Failed merges: r? `@ghost` `@rustbot` modify labels: rollup

`core::any::TypeId` implements `const PartialEq` as of [rust-lang/rust#101698][1]. [1]: rust-lang/rust#101698

…atch, r=dtolnay Remove structural match from `TypeId` As per rust-lang#99189 (comment). > Removing the structural equality might make sense, but is a breaking change that'd require a libs-api FCP. rust-lang#99189 (comment) > Landing this PR now (well, mainly the "remove structural equality" part) would unblock `const fn` `TypeId::of`, since we only postponed that because we were guaranteeing too much. See also rust-lang#99189, rust-lang#101698

…ch, r=dtolnay Remove structural match from `TypeId` As per rust-lang#99189 (comment). > Removing the structural equality might make sense, but is a breaking change that'd require a libs-api FCP. rust-lang#99189 (comment) > Landing this PR now (well, mainly the "remove structural equality" part) would unblock `const fn` `TypeId::of`, since we only postponed that because we were guaranteeing too much. See also rust-lang#99189, rust-lang#101698

rustbot added the T-libs Relevant to the library team, which will review and decide on the PR/issue. label Sep 11, 2022

rust-highfive assigned scottmcm Sep 11, 2022

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Sep 11, 2022

This comment has been minimized.
Sign in to view

fmease reviewed Sep 11, 2022
View reviewed changes

raldone01 changed the title ~~Consitfy TypeId ordering impls~~Constify TypeId ordering implsSep 11, 2022

rustbot added T-libs-api [DEPRECATED; DO NOT USE] and removed T-libs Relevant to the library team, which will review and decide on the PR/issue. labels Sep 14, 2022

raldone01 mentioned this pull request Sep 15, 2022
Tracking Issue for comparing TypeId in const items #101871
Open
4 tasks

jhpratt approved these changes Sep 21, 2022
View reviewed changes

scottmcm added I-libs-api-nominated [DEPRECATED; DO NOT USE] I-lang-nominated Nominated for discussion during a lang team meeting. labels Sep 24, 2022

scottmcm reviewed Sep 27, 2022
View reviewed changes

Comment threadlibrary/core/src/any.rs Outdated

scottmcm reviewed Sep 28, 2022
View reviewed changes

rustbot added the S-blocked Status: Blocked on something else such as an RFC or other implementation work. label Sep 28, 2022

scottmcm removed the I-lang-nominated Nominated for discussion during a lang team meeting. label Sep 28, 2022

m-ou-se added the I-compiler-nominated Nominated for discussion during a compiler team meeting. label Oct 4, 2022

apiraino removed the I-compiler-nominated Nominated for discussion during a compiler team meeting. label Oct 13, 2022

bors added the S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. label Jan 16, 2023

scottmcm closed this Jan 16, 2023

bors added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Jan 16, 2023

scottmcm reopened this Jan 16, 2023

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jan 16, 2023

bors added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Jan 16, 2023

Constify TypeId ordering impls
7355ab3

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jan 17, 2023

Dylan-DPC mentioned this pull request Jan 17, 2023
Rollup of 5 pull requests #106984
Merged

bors merged commit 09faa26 into rust-lang:masterJan 17, 2023

rustbot added this to the 1.68.0 milestone Jan 17, 2023

onestacked deleted the feat/const_cmp_typeid branch January 18, 2023 09:06

yvt added a commit to r3-os/r3 that referenced this pull request Mar 18, 2023
refactor(core): remove the TypeId: const PartialEq wrapper
8b64feb
`core::any::TypeId` implements `const PartialEq` as of [rust-lang/rust#101698][1]. [1]: rust-lang/rust#101698

Uh oh!

Uh oh!

Conversation

raldone01 commented Sep 11, 2022• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Sep 11, 2022

Uh oh!

rust-highfive commented Sep 11, 2022

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

fmeaseSep 11, 2022

Choose a reason for hiding this comment

Uh oh!

onestackedSep 11, 2022

Choose a reason for hiding this comment

Uh oh!

onestacked commented Sep 14, 2022

Uh oh!

jhpratt left a comment

Choose a reason for hiding this comment

Uh oh!

scottmcm commented Sep 24, 2022

Uh oh!

raldone01 commented Sep 24, 2022• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

scottmcmSep 28, 2022

Choose a reason for hiding this comment

Uh oh!

eddybOct 13, 2022

Choose a reason for hiding this comment

Uh oh!

onestackedOct 13, 2022• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

onestackedOct 13, 2022

Choose a reason for hiding this comment

Uh oh!

scottmcm commented Sep 28, 2022

Uh oh!

m-ou-se commented Oct 4, 2022

Uh oh!

apiraino commented Oct 6, 2022

Uh oh!

apiraino commented Oct 13, 2022

Uh oh!

eddyb commented Oct 13, 2022

Uh oh!

onestacked commented Jan 16, 2023

Uh oh!

scottmcm commented Jan 16, 2023

Uh oh!

WaffleLapkin commented Jan 16, 2023

Uh oh!

bors commented Jan 16, 2023

Uh oh!

onestacked commented Jan 16, 2023

Uh oh!

scottmcm commented Jan 16, 2023

Uh oh!

onestacked commented Jan 16, 2023

Uh oh!

WaffleLapkin commented Jan 17, 2023

Uh oh!

bors commented Jan 17, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

14 participants

raldone01 commented Sep 11, 2022•
edited
Loading

raldone01 commented Sep 24, 2022•
edited
Loading

onestackedOct 13, 2022•
edited
Loading