Stabilization metabug: 1.0-alpha

[aturon]

This is a tracking metabug for stabilization of both APIs and gated features.

API stabilization for libstd

First pass stabilization/reform

For simpler/mostly stable APIs (core types, iterators), stabilization is mostly a matter of aligning with conventions and deprecating marginal methods and takes place directly under the supervision of the core team. For deeper redesign, RFCs are needed. Where applicable, the list below points to the relevant RFC.

Most APIs with a chechmark are still #unstable, either due to developing conventions, lack of language features (e.g. unboxed closures), or some other problem. We plan to do a rapid second pass closer to the 1.0.0 beta.

(You can track the following list quantitatively, though the fact that trait impls count exaggerates the amount of #[experimental] items.)

Items are marked complete when a PR/RFC has been posted -- this is just tracking the design work.

• any - std: Stabilize unit, bool, ty, tuple, arc, any #15936
• ascii (see RFC)
• bitflags (see RFC)
• bool - std: Stabilize unit, bool, ty, tuple, arc, any #15936
• borrow (see RFC)
• boxed - Stabilization for owned (now boxed) and cell #15591
• c_str (see RFC)
• c_vec (see RFC)
• cell - Stabilization for owned (now boxed) and cell #15591
• char - Some stabilization and conventions changes to std::char #18603
• clone - core: Add stability attributes to Clone #15135
• cmp - Stability for core::cmp, core::default, alloc::rc, std::task #15797 (see RFC)
• collections (see RFC)
• comm - Apply stability attributes to std::comm #15286
• default - Stability for core::cmp, core::default, alloc::rc, std::task #15797
• dynamic_lib (#[unstable])
• error (see RFC)
• f32, f64 (see RFC)
• finally (#[unstable])
• fmt (see RFC)
• hash - std: Stabilize the std::hash module #19673
• i8, i16, i32, i64, int (see RFC)
• intrinsics (#[unstable])
• io (see RFC)
• iter - libs: stabilize iter module #19176
• kinds (punted to second pass)
• thread_local (see RFC)
• macros (punted to second pass)
• mem - core: Finish stabilizing the mem module. #14392
• num (see RFC)
• ops (see RFC)
• option - stabilize core::{option, result} #16664
• os (see RFC)
• path (see RFC)
• prelude (see RFC)
• rand (#[unstable])
• raw (#[unstable])
• rc - Stability for core::cmp, core::default, alloc::rc, std::task #15797
• result - stabilize core::{option, result} #16664
• rt (#[unstable])
• simd (#[unstable])
• slice - Cleanup and stability for core::slice #16332
• str - std: Stabilize the std::str module #19741, (see RFC)
• string - collections: Stabilize String #17438
• sync - std: Rewrite the sync module #19274
• task (see RFC)
• time (punted to third pass)
• tuple - (likely to be deprecated before 1.0), std: Stabilize unit, bool, ty, tuple, arc, any #15936
• u8, u16, u32, u64, uint (see RFC)
• unit - std: Stabilize unit, bool, ty, tuple, arc, any #15936
• vec - collections: Stabilize Vec #17029

Second pass stabilization: for alpha

The second pass is about actually moving APIs to #[stable] status; it may involve a bit of minor shuffling, but any major design work should have already taken place. In many cases APIs were blocked on language features (like unboxed closures) before moving to this status.

• any (#[unstable for alpha)
• ascii (#[unstable] for alpha)
• bitflags (see RFC)
• bool - acrichto std: Remove public bool,tuple,unit modules #20006
• borrow (#[unstable] for alpha)
• boxed - acrichto std: Second pass stabilization for boxed #20420
• c_str - acrichto std: Redesign c_str and c_vec #20507
• c_vec - acrichto std: Redesign c_str and c_vec #20507
• cell - aturon Stabilize cell #20066 -- std module mistakenly left experimental
• char - huonw Second stabilisation pass of *::char #20395
• clone - aturon Stabilize clone #20070
• cmp - aturon Stabilize cmp #20065
• collections - aturon / @gankro stabilize more of collections #20356
• comm - acrichto std: Second pass stabilization for comm #20273
• default - acrichto std: Second-pass stabilization of mem/default #19902
• dynamic_lib (#[unstable])
• error - aturon Stabilize std::error #20647
• f32, f64 - aturon (note: already stable; the num::Float/FloatMath traits are the problem point)
• finally (#[unstable])
• fmt - acrichto
• hash - acrichto std: Stabilize the std::hash module #20654
• i8, i16, i32, i64, int - aturon Stabilize integer modules #20062
• intrinsics (#[unstable] for alpha)
• io (#[unstable] for alpha)
• iter - aturon Final pre-alpha stabilization of: iter, ops, slice, collections #20560
• kinds - aturon/nrc, flatten to markers Change std::kinds to std::markers; flatten std::kinds::marker #20607
• macros - acrichto std: Stablize the macros module #20657
• mem - acrichto std: Second-pass stabilization of mem/default #19902
• num - aturon/huonw Second stabilisation pass of *::num #20573
• ops - aturon Final pre-alpha stabilization of: iter, ops, slice, collections #20560
• option - acrichto std: Fully stabilize Option<T> #19849
• os (#[unstable] for alpha)
• path (#[unstable] for alpha)
• prelude - acrichto std: Stabilize the prelude module #20157
• ptr - acrichto std: Second pass stabilization for ptr #20042
• rand (#[unstable] for alpha)
• raw (#[unstable] for alpha)
• rc - aturon (complete from pass 1)
• result - acrichto std: Second pass stabilization of Result<T, E> #19961
• rt (#[unstable])
• simd (#[unstable])
• slice - aturon Second pass stabilization: slice and vec #20061, Final pre-alpha stabilization of: iter, ops, slice, collections #20560
• str - acrichto std: Stabilize the std::str module #19741 / kimundi WIP Partially stabilize StrExt for pattern-using methods #20058
• string - acrichto std: Second pass stabilization for string #20306
• sync - acrichto std: Return Result from RWLock/Mutex methods #19661, std: Second pass stabilization of sync #20315
• task (#[deprecated])
• thread - aturon Stabilize std::thread #20615
• thread_local - acrichto std: Second pass stabilization for thread_local #20354
• thunk (#[unstable])
• tuple - acrichto std: Remove public bool,tuple,unit modules #20006
• u8, u16, u32, u64, uint - aturon Stabilize integer modules #20062
• unit - acrichto std: Remove public bool,tuple,unit modules #20006
• vec - aturon Second pass stabilization: slice and vec #20061

RFCs for reform

The following RFCs tackle deeper API redesigns:

Accepted

• Error interoperation
• Runtime reform
• Collections reform
• Numerics reform
• Formatting stabilization
• cmp/ops reform
• Path reform
• Thread-local storage reform
• Split Show
• Prelude stabilization
• Collections reform part 2
• Macro reform
• c_str, c_vec reform

Pending

• io and os reform

Non-std crates

See #18585 (comment)

Just need #[experimental]

We will ship these crates, but will they will participate in #![staged_api]
and will not be accessible in the stable channel.

• alloc
• arena
• collections
• core
• flate
• fmt_macros
• graphviz
• rand
• rbml
• rustc
• rustc_back
• rustc_llvm
• rustc_trans
• rustdoc
• syntax
• unicode

Need Cargo.toml and a publishing strategy

These crates will all be distributed, but participate in #![staged_api]. This
means they will not be accessible in the stable channel's standard
distribution, but they will all be available through crates.io.

• getopts - https://github.com/rust-lang/getopts
• libc - libc: Prepare for movement to crates.io #20606
• log - https://github.com/rust-lang/log
• regex - https://github.com/rust-lang/regex
• regex_macros - https://github.com/rust-lang/regex
• term - https://github.com/rust-lang/term

Need cooperation with the standard library due to compiler internals

• serialize - rustc: Start the deprecation of libserialize #19755, https://github.com/rust-lang/rustc-serialize
• test - Finalize story for libtest #20603

Removed or to remove

• green - Finish runtime removal #18967
• native - Finish runtime removal #18967
• rustrt - Replace std::task with std::thread, remove librustrt, make final "runtime" features lazily initialize #19654
• sync - Merge libsync into libstd #19255
• time - time: Deprecate the library in the distribution #18858

Feature stabilization

Ungating

• Globs
• Tuple indexing (RFC), Remove feature gates for if let, while let, and tuple indexing #19472
• Struct variants (RFC), Un-feature gate struct variants #18994
• if let/while let (RFC), Remove feature gates for if let, while let, and tuple indexing #19472
• Macros
• Associated types
• Default type parameters

Removal

• Slicing syntax: should be removed in favor of ops reform

CLI Tooling

• command line flags - Stability story of compiler flags #19051, rustc: Start "stabilizing" some flags #19900

[aturon]

cc @alexcrichton @brson @nikomatsakis @huonw @pcwalton

@alexcrichton, can you populate the non-std crates section with status/plans?

Everyone: please help me fill in the "Feature stabilization" section.

[aturon]

A note about the 1.0.0 beta period: we are planning to run at least two beta cycles.

During the first cycle, we plan to turn on the stability lints, but not to add the #[staged_api] attribute. That will mean that the beta channel will allow you to use #[experimental] and #[unstable] features, but will warn you when you do so. This cycle will serve to discover gaps in stabilization -- important APIs/features that are not yet #[stable] for the initial beta -- and also give us a little bit more time on the trickiest #[unstable] APIs.

The second cycle will then add #[staged_api], so that the 1.0.0-beta2 will serve as a true release candidate.

By the time we release the first beta, we expect the bulk of std to be #[stable], with the possible exception of std::io, as the tracking above indicates.

By the second cycle, there should be very little large-scale API revision. Thus, the two cycles should serve as a period of relative calm (compared to today's very rapid breakage) during which the crates.io ecosystem can stabilize and grow before the stable 1.0 release.

[aturon]

cc @reem @bstrie @steveklabnik @kmcallister

[Gankra]

cc me

[sanxiyn]

Default type parameter (RFC 213) should be in the feature stabilization list.