Skip to content
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

RFC: Minimum Supported Rust Version #2495

Merged
merged 19 commits into from
Oct 10, 2019
Merged
Changes from 15 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
179 changes: 179 additions & 0 deletions text/0000-min-rust-version.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,179 @@
- Feature Name: min_rust_version
- Start Date: 2018-06-28
- RFC PR: (leave this empty)
- Rust Issue: (leave this empty)

# Summary
[summary]: #summary

Add `rust` field to the package section of `Cargo.toml` which will be used to
specify crate's Minimum Supported Rust Version (MSRV):
```toml
[package]
name = "foo"
version = "0.1.0"
rust = "1.30"
```

# Motivation
[motivation]: #motivation

Currently crates have no way to formally specify MSRV. As a result users can't
check if crate can be built on their toolchain without building it. It also
leads to the debate on how to handle crate version change on bumping MSRV,
conservative approach is to consider such changes as breaking ones, which can
hinder adoption of new features across ecosystem or result in version number
inflation, which makes it harder to keep downstream crates up-to-date. More
relaxed approach on another hand can result in broken crates for user of older
compiler versions.

# Guide-level explanation
[guide-level-explanation]: #guide-level-explanation

If you target a specific MSRV add a `rust` field to the `[package]` section of
your `Cargo.toml` with a value equal to the targeted Rust version. If you build
a crate with a dependency which has MSRV higher than the current version of your
toolchain, `cargo` will return a compilation error stating the dependency and
its MSRV. This behavior can be disabled by using `--no-msrv-check` flag.

# Reference-level explanation
[reference-level-explanation]: #reference-level-explanation

During build process (including `run`, `test`, `benchmark`, `verify` and `publish`
sub-commands) `cargo` will check MSRV requirements of all crates in a dependency
tree scheduled to be built or checked. Crates which are part of the dependency
tree, but will not be built are excluded from this check (e.g. target-dependent
or optional crates).

`rust` field should respect the following minimal requirements:
- Value should be a version in semver format **without** range operators. Note
that "1.50" is a valid value and implies "1.50.0".
- Version can not be bigger than a current stable toolchain (it will be checked
by crates.io during crate upload).
- Version can not be smaller than 1.27 (version in which `package.rust` field
became a warning instead of an error).
- Version can not be smaller than release version of a used edition, i.e.
combination of `rust = "1.27"` and `edition = "2018"` is an invalid one.

# Future work and extensions
[future-work]: #future-work
Copy link
Member

Choose a reason for hiding this comment

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

Some things in this section are written in the wrong tense, given that they are matters that this RFC does not decide upon. Because they’re hypothetical or tentative plans, “would” should be used instead of “will”, or else it’s easy to get confused about what’s in scope for implementation, when reading the RFC. For example, part way through reading it I was thinking that influencing version resolution was in scope due to the use of the word “will”, but it is in fact designated as possible future work.

(If you’re not sure what I’m talking about, https://english.stackexchange.com/questions/3657/when-should-i-use-would-would-have-will-and-will-have explains it decently. Also let it be known that I love the term subjunctive mood.)

Copy link
Contributor Author

@newpavlov newpavlov Oct 1, 2019

Choose a reason for hiding this comment

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

Ah, I forgot to change it after I moved this text to the "future work" section. I will try to fix it.


## Influencing version resolution

The value of `rust` field (explicit or automatically selected by `cargo`) will
be used to select appropriate dependency versions.

For example, let's imagine that your crate depends on crate `foo` with 10 published
versions from `0.1.0` to `0.1.9`, in versions from `0.1.0` to `0.1.5` `rust`
field in the `Cargo.toml` sent to crates.io equals to "1.30" and for others to
"1.40". Now if you'll build your project with e.g. Rust 1.33, `cargo` will select
`foo v0.1.5`. `foo v0.1.9` will be selected only if you'll build your project with
Rust 1.40 or higher. But if you'll try to build your project with Rust 1.29 cargo
will issue an error.

`rust` field value will be checked as well. During crate build `cargo` will check
if all upstream dependencies can be built with the specified MSRV. (i.e. it will
check if there is exists solution for given crates and Rust versions constraints)
Yanked crates will be ignored in this process.

Implementing this functionality hopefully will allow us to close the long-standing
debate regarding whether MSRV bump is a breaking change or not and will allow
crate authors to feel less restrictive about bumping their crate's MSRV. (though
it may be a useful convention for post-1.0 crates to bump minor version on MSRV
bump to allow publishing backports which fix serious issues using patch version)

Note that described MSRV constraints and checks for dependency versions resolution
can be disabled with the `--no-msrv-check` option.

## Checking MSRV during publishing

`cargo publish` will check that upload is done with a toolchain version specified
in the `rust` field. If toolchain version is different, `cargo` will refuse to
upload the crate. It will be a failsafe to prevent uses of incorrect `rust` values
due to unintended MSRV bumps. This check can be disabled by using the existing
`--no-verify` option.

## Making `rust` field mandatory

In future (probably in a next edition) we could make `rust` field mandatory for
a newly uploaded crates. MSRV for older crates will be determined by the `edition`
field. In other words `edition = "2018"` will imply `rust = "1.31"` and
`edition = "2015"` will imply `rust = "1.0"`.
Copy link
Member

Choose a reason for hiding this comment

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

Implying Rust 1.0 seems mostly incompatible with this statement from earlier in the document:

Version can not be smaller than 1.27 (version in which package.rust field became a warning instead of an error).

Copy link
Contributor Author

@newpavlov newpavlov Oct 1, 2019

Choose a reason for hiding this comment

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

The cited text is one of the requirements for the rust field value if it's provided explicitly, while the edition text is about how MSRV should be inferred if explicit value is not provided. I guess a better wording could be something like this:

In other words edition = "2018" will imply that MSRV for the crate is equal to 1.31 and edition = "2015" will imply that MSRV is equal to 1.0.


`cargo init` will use version of a currently used toolchain.
newpavlov marked this conversation as resolved.
Show resolved Hide resolved

## `cfg`-based MSRVs

Some crates can have different MSRVs depending on target architecture or enabled
features. In such cases it can be useful to describe how MSRV depends on them,
e.g. in the following way:
```toml
[package]
rust = "1.30"

[target.x86_64-pc-windows-gnu]
newpavlov marked this conversation as resolved.
Show resolved Hide resolved
rust = "1.35"

[target.'cfg(feature = "foo")']
newpavlov marked this conversation as resolved.
Show resolved Hide resolved
rust = "1.33"
```

All `rust` values in the `target` sections should be equal or bigger to a `rust` value
specified in the `package` section.

If target condition is true, then `cargo ` will use `rust` value from this section.
If several target section conditions are true, then maximum value will be used.

## Nightly and stable versions

Some crates may prefer to target only the most recent stable or nightly toolchain.
In addition to the versions we could allow `stable` and `nightly` values to declare
that maintainers do not track MSRV for the crate.

For some bleeding-edge crates which experience frequent breaks on Nightly updates
(e.g. `rocket`) it can be useful to specify exact Nightly version(s) on which
crate can be built. One way to achieve this is by using the following syntax:
- auto-select: "nightly" This variant will behave in the same way as "stable", i.e.
Copy link
Member

Choose a reason for hiding this comment

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

This entire list is poorly formatted for reading, with this line being particularly problematic.

it will take a current nightly version and will use it in a "more or equal" constraint.
- single version: "nightly: 2018-01-01" (the main variant)
- enumeration: "nightly: 2018-01-01, 2018-01-15"
- semver-like conditions: "nightly: >=2018-01-01", "nightly: >=2018-01-01, <=2018-01-15",
"nightly: >=2018-01-01, <=2018-01-15, 2018-01-20". (the latter is interpreted as
"(version >= 2018-01-01 && version <= 2018-01-20) || version == 2018-01-20")

Such restrictions can be quite severe, but hopefully this functionality will be
used only by handful of crates.

# Drawbacks
[drawbacks]: #drawbacks

- Declaration of MSRV, even with the checks, does not guarantee that crate
will work correctly on the specified MSRV, only appropriate CI testing can do that.
- More complex dependency versions resolution algorithm.
- MSRV selected by `cargo publish` with `rust = "stable"` can be too
conservative.

# Alternatives
[alternatives]: #alternatives

- Automatically calculate MSRV.
Copy link
Member

Choose a reason for hiding this comment

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

This seems too vague to have much meaning; there are wildly different things it could mean, with completely different trade-offs. What is automatically calculating the MSRV? rustc, cargo, crates.io? How is it conveyed? Would it play well with the future work of MSRV influencing version resolution? I can think of two radically different approaches to this off the top of my head, one of which is completely different from this proposal and one of which boils down to just augmenting the package.rust field of this specification with some tooling to automate discovery of the MSRV.

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 didn't want to cover all potential approaches, as you do say yourself there are a lot of options to choose from. The main idea here is that instead of asking people to manually select MSRV via the rust field, we could rely on some automatic system (be it on rustc, cargo or crates.io side) to partially solve issues which have motivated this RFC.

- Do nothing and rely on [LTS releases](https://github.com/rust-lang/rfcs/pull/2483)
for bumping crate MSRVs.
newpavlov marked this conversation as resolved.
Show resolved Hide resolved
- Allow version and path based `cfg` attributes as proposed in [RFC 2523](https://github.com/rust-lang/rfcs/pull/2523).
Copy link
Member

Choose a reason for hiding this comment

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

RFC 2523 isn’t an alternative to this functionality. It’s just related functionality, and the two go together quite nicely. (RFC 2523 has been accepted now, BTW, so if mention of it is kept, the link should probably change.)

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Well, I think it's relatively common to list relevant proposals and approaches in the alternatives section, even though they are not alternatives per se. @Centril has asked me to do it here.


# Prior art
[prior-art]: #prior-art

Previous proposals:
- [RFC 1707](https://github.com/rust-lang/rfcs/pull/1707)
- [RFC 1709](https://github.com/rust-lang/rfcs/pull/1709)
- [RFC 1953](https://github.com/rust-lang/rfcs/pull/1953)
- [RFC 2182](https://github.com/rust-lang/rfcs/pull/2182) (arguably this one got off-track)

# Unresolved questions
[unresolved]: #unresolved-questions

- Name bike-shedding: `rust` vs `rustc` vs `min-rust-version`
- Additional checks?
- Better description of versions resolution algorithm.
- How nightly versions will work with "cfg based MSRV"?