Skip to content

stabilization template, docs #2219

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

Open
wants to merge 19 commits into
base: master
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from 2 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 19 additions & 0 deletions src/implementing_new_features.md
Original file line number Diff line number Diff line change
Expand Up @@ -204,3 +204,22 @@ tests/ui/feature-gates/ --bless`.
[here]: ./stabilization_guide.md
[tracking issue]: #tracking-issues
[add-feature-gate]: ./feature-gates.md#adding-a-feature-gate

## Call for testing

Once the implementation is complete, the feature will be available to nightly users, but not yet part of stable Rust. This is a good time to write a blog post on [one of the Rust blogs](https://github.com/rust-lang/blog.rust-lang.org/) and issue a call for testing (here are three [example](https://blog.rust-lang.org/2021/08/03/GATs-stabilization-push.html) [blog](https://blog.rust-lang.org/inside-rust/2024/08/09/async-closures-call-for-testing.html) [posts](https://blog.rust-lang.org/2024/09/05/impl-trait-capture-rules.html) to give you the idea). The post should highlight how the feature works, what areas you'd like people to play with, and how they can supply feedback.

## Affiliated work

Once the feature is supported by rustc, there is other associated work that needs to be done to give users a complete experience:

* Extending rustfmt to format any new syntax;
* Extending rust-analyzer;
* Documenting the feature in the Rust reference;
* ...

## Stabilization

The final step in the feature lifecycle is [stabilization][stab], which is when the feature becomes available to all Rust users. At this point, backwards incompatible changes are no longer permitted (modulo soundness bugs and inference changes; see the lang team's [defined semver policies](https://rust-lang.github.io/rfcs/1122-language-semver.html) for full details). To learn more about stabilization, see the [stabilization guide][stab].

[stab]: ./stabilization_guide.md
62 changes: 26 additions & 36 deletions src/stabilization_guide.md
Copy link
Member

Choose a reason for hiding this comment

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

This file contains the following:

Once we have decided to stabilize a feature, we need to have
a PR that actually makes that stabilization happen. These kinds
of PRs are a great way to get involved in Rust, as they take
you on a little tour through the source code.

And I'm not sure how accurate this is.

First of all the "Once we have decided to stabilize a feature" part is too vague. How would a new contributor know which features we decided to stabilize? Additionally the decision and stabilization report need to be done by someone with a good knowledge of the feature, a new contributor just isn't suitable for that.

I feel like this paragraph is misguiding, as stabilization is usually an intertwined process involving multiple teams, sometimes confusing changes (like changing lints that don't make sense anymore), checking many components, etc. If for library features it might be the case that a novice can craft a stabilization PR, for language features, more often than not that's not the case.

Copy link
Member

Choose a reason for hiding this comment

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

For this initial template, I will drop that paragraph entirely, and refrain from making substantial wording additions to make this PR less contentious to land. I have some Thoughts:tm: on this matter as well.

Original file line number Diff line number Diff line change
Expand Up @@ -43,44 +43,14 @@ has completed. Meanwhile, we can proceed to the next step.

## Write a stabilization report

Find the tracking issue of the feature, and create a short
stabilization report. Essentially this would be a brief summary
of the feature plus some links to test cases showing it works
as expected, along with a list of edge cases that came up
and were considered. This is a minimal "due diligence" that
we do before stabilizing.

The report should contain:

- A summary, showing examples (e.g. code snippets) what is
enabled by this feature.
- Links to test cases in our test suite regarding this feature
and describe the feature's behavior on encountering edge cases.
- Links to the documentations (the PRs we have made in the
previous steps).
- Any other relevant information.
- The resolutions of any unresolved questions if the stabilization
is for an RFC.

Examples of stabilization reports can be found in
[rust-lang/rust#44494][report1] and [rust-lang/rust#28237][report2] (these links
will bring you directly to the comment containing the stabilization report).

[report1]: https://github.com/rust-lang/rust/issues/44494#issuecomment-360191474
[report2]: https://github.com/rust-lang/rust/issues/28237#issuecomment-363374130

## FCP

If any member of the team responsible for tracking this
feature agrees with stabilizing this feature, they will
start the FCP (final-comment-period) process by commenting
Author a stabilization report using the [template found in this repository][srt].
Stabilization reports summarize the work that has been done since the RFC.
The [template][srt] includes a series of questions that aim to surface interconnections between this feature and the various Rust teams (lang, types, etc) and also to identify items that are commonly overlooked.

```text
@rfcbot fcp merge
```
[srt]: ./stabilization_report_template.md

The rest of the team members will review the proposal. If the final
decision is to stabilize, we proceed to do the actual code modification.
The stabilization report is typically posted as the main comment on the stabilization PR (see the next section).
If you'd like to develop the stabilization report incrementally, we recommend adding it to
Copy link
Member

Choose a reason for hiding this comment

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

incomplete: "adding it to"...

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Oh yeah, I never finished. I was going to suggest the practice of adding the stabilization report to the unsafe book and updating it incrementally (e.g., with a list of PRs, if nothing else).

Copy link
Member

Choose a reason for hiding this comment

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

Dropped this partial sentence from the initial version


## Stabilization PR

Expand Down Expand Up @@ -194,3 +164,23 @@ if something { /* XXX */ }
[Rust by Example]: https://github.com/rust-lang/rust-by-example
[`Unstable Book`]: https://doc.rust-lang.org/unstable-book/index.html
[`src/doc/unstable-book`]: https://github.com/rust-lang/rust/tree/master/src/doc/unstable-book

## Lang team nomination
Copy link
Member

Choose a reason for hiding this comment

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

Maybe just change this section to team nominations, and put the lang team as the first item (and emphasize)?

Also, a discussion for another place, but this made me think: Is it all that great to nominate things immediately? There are usually plenty of comments on things and it might be worth delaying the teams discussing things until those immediate comments get addressed.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

+1 to the more general title. I wondered the same thing about when to nominate. I don't know the right answer.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

I guess saying "when comments die down, or if you don't get any comments, nominate"

Copy link
Member

Choose a reason for hiding this comment

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

Fixed title, and tried to add a sentence to that effect.


When you feel the PR is ready for consideration by the lang team, you can [nominate the PR](https://lang-team.rust-lang.org/how_to/nominate.html) to get it on the list for discussion in the next meeting. You should also cc the other interacting teams to review the report:

* `@rust-lang/types`, to look for type system interactions
* `@rust-lang/compiler`, to vouch for implementation quality
Copy link
Member

Choose a reason for hiding this comment

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

Maybe this should be an explicit step? For either the compiler team as a whole, or at least just the "owner" of the implementation to be on board with stabilization. Otherwise, "everything" should be compiler nominated?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Yes, this is a good question. My feeling is that both compiler and types ought to have a process of assigning one person to prepare a summary for others to read, but I felt that needed more discussion before I wrote it in here. I should probably open a tracking issue on this overall topic.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

To be honest, lang should have that process too :)

Copy link
Member

Choose a reason for hiding this comment

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

Left this unaddressed, not for an initial version

* `@rust-lang/opsem`, but only if this feature interacts with unsafe code and can create undefined behavior
* `@rust-lang/libs-api`, but only if there are additions to the standard library

## FCP proposed on the PR

Finally, some member of the team responsible for tracking this feature agrees with stabilizing this feature, will
start the FCP (final-comment-period) process by commenting

```text
@rfcbot fcp merge
```

The rest of the team members will review the proposal. If the final decision is to stabilize, the PR will be reviewed by the compiler team like any other PR.
54 changes: 54 additions & 0 deletions src/stabilization_report_template.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,54 @@
# Stabilization report template

> **What is this?** This is a template to use for [stabilization reports](./stabilization_guide.md). It consists of a series of questions that aim to provide the information most commonly needed and to help reviewers be more likely to identify potential problems up front. Not all parts of the template will apply to all stabilizations. Feel free to put N/A if a question doesn't seem to apply to your case.

## General design

### What is the RFC for this feature and what changes have occurred to the user-facing design since the RFC was finalized?

### What behavior are we committing to that has been controversial? Summarize the major arguments pro/con.

### Are there extensions to this feature that remain unstable? How do we know that we are not accidentally committing to those?

## Has a call-for-testing period been conducted? If so, what feedback was received?

## Implementation quality

### Summarize the major parts of the implementation and provide links into the code (or to PRs)

An example for async closures: https://rustc-dev-guide.rust-lang.org/coroutine-closures.html

### Summarize existing test coverage of this feature

- What does the test coverage landscape for this feature look like?
- (Positive/negative) Behavioral tests?
- (Positive/negative) Interface tests? (e.g. compiler cli interface)
- Maybe link to test folders or individual tests (ui/codegen/assembly/run-make tests, etc.)
- Are there any (intentional/unintentional) gaps in test coverage?

Copy link
Contributor

Choose a reason for hiding this comment

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

Suggested change
Consider what the "edges" of this feature are. We're particularly interested in seeing tests that assure us about exactly what nearby things we're not stabilizing.

Copy link
Contributor

@traviscross traviscross Jan 25, 2025

Choose a reason for hiding this comment

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

Going even further:

Suggested change
Consider what the "edges" of this feature are. We're particularly interested in seeing tests that assure us about exactly what nearby things we're not stabilizing.
Within each test, include a comment at the top describing the purpose of the test and what set of invariants it intends to demonstrate. This is a great help to those reviewing the tests at stabilization time.
Similarly, please consider including, when appropriate, `//@ reference:` annotations to connect each test with the corresponding item in the Reference.

In reviewing the tests for the arbitrary self types stabilization, I'm reminded how helpful it is for each test to describe at the top what it is intending to demonstrate, so it's worth mentioning that.

It's also probably worth mentioning here the utility of the Reference annotations, but that raises an interesting ordering question. We merge tests ahead of the stabilization, generally, but then we don't merge the Reference PR until after the stabilization. So we'd either need to merge the tests with dangling references (to identifiers in unmerged Reference PRs) or perhaps these references could be added to the tests in the stabilization PR itself. Or they could be added later, but then these helpful things aren't there when reviewing the stabilization.

(Another wilder option is that we merge the Reference into rust-lang/rust itself, as was recently done with the dev guide, and then the Reference PR becomes a part of the stabilization PR, though we're probably not yet ready to do that in general.)

@ehuss, @nikomatsakis, what do you think?

Copy link
Member

Choose a reason for hiding this comment

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

it's often useful to us to directly push our own commits to a Reference PR branch rather than going back and forth with the author, but permissions on rust-lang/rust aren't currently set in a way that would enable that.)

AFAIK team members can push to PRs in rust-lang/rust as well, but individual PR authors can opt-out of that.

Copy link
Contributor

Choose a reason for hiding this comment

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

Thanks. I must have hit the odd case before then. Edited to correct.

Copy link
Member

Choose a reason for hiding this comment

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

Note that for this initial version, I intentionally left off the

Similarly, please consider including, when appropriate, //@ reference: annotations to connect each test with the corresponding item in the Reference.

I don't even know how to do that logistically myself.

### What outstanding bugs in the issue tracker involve this feature? Are they stabilization-blocking?

### Summarize contributors to the feature by name for recognition and assuredness that people involved in the feature agree with stabilization

### What FIXMEs are still in the code for that feature and why is it ok to leave them there?

### Which tools need to be adjusted to support this feature. Has this work been done?

*Consider rustdoc, clippy, rust-analyzer, rustfmt, rustup, docs.rs.*

## Type system and execution rules

### What compilation-time checks are done that are needed to prevent undefined behavior?
Copy link
Contributor

Choose a reason for hiding this comment

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

Similar question: Does the feature's implementation need checks to prevent UB or is it sound by default and needs opt in in places to perform the dangerous/unsafe operations? If it is not sound by default, what is the rationale?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

I've been thinking about generalizing this question -- like, "what type system rules are enforced and what is their purpose"

Copy link
Member

Choose a reason for hiding this comment

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

I added oli's question for the initial version


(Be sure to link to tests demonstrating that these tests are being done.)

### Can users use this feature to introduce undefined behavior? (Describe.)

### What updates are needed to the reference/specification? (link to PRs when they exist)

## Common interactions

### Does this feature introduce new expressions and can they produce temporaries? What are the lifetimes of those temporaries?

### What other unstable features may be exposed by this feature?

Loading