Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
// Copyright 2014 The Rust Project Developers. See the COPYRIGHT
|
|
|
|
// file at the top-level directory of this distribution and at
|
|
|
|
// http://rust-lang.org/COPYRIGHT.
|
|
|
|
//
|
|
|
|
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
|
|
|
|
// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
|
|
|
|
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
|
|
|
|
// option. This file may not be copied, modified, or distributed
|
|
|
|
// except according to those terms.
|
|
|
|
|
|
|
|
//! A pass that annotates every item and method with its stability level,
|
|
|
|
//! propagating default levels lexically from parent to children ast nodes.
|
|
|
|
|
2015-01-12 20:40:19 -06:00
|
|
|
use session::Session;
|
2015-01-01 20:41:44 -06:00
|
|
|
use middle::ty;
|
|
|
|
use metadata::csearch;
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
use syntax::codemap::Span;
|
|
|
|
use syntax::{attr, visit};
|
2014-07-10 13:17:40 -05:00
|
|
|
use syntax::ast;
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
use syntax::ast::{Attribute, Block, Crate, DefId, FnDecl, NodeId, Variant};
|
2014-08-05 21:44:21 -05:00
|
|
|
use syntax::ast::{Item, RequiredMethod, ProvidedMethod, TraitItem};
|
2014-11-11 14:46:47 -06:00
|
|
|
use syntax::ast::{TypeMethod, Method, Generics, StructField, TypeTraitItem};
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
use syntax::ast_util::is_local;
|
2015-01-12 20:40:19 -06:00
|
|
|
use syntax::attr::{Stability, AttrMetaMethods};
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
use syntax::visit::{FnKind, FkMethod, Visitor};
|
2015-01-01 20:41:44 -06:00
|
|
|
use util::nodemap::{NodeMap, DefIdMap};
|
|
|
|
use util::ppaux::Repr;
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
|
2014-09-12 05:10:30 -05:00
|
|
|
use std::mem::replace;
|
|
|
|
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
/// A stability index, giving the stability level for items and methods.
|
|
|
|
pub struct Index {
|
2015-01-12 20:40:19 -06:00
|
|
|
// Indicates whether this crate has #![staged_api]
|
|
|
|
staged_api: bool,
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
// stability for crate-local items; unmarked stability == no entry
|
|
|
|
local: NodeMap<Stability>,
|
|
|
|
// cache for extern-crate items; unmarked stability == entry with None
|
|
|
|
extern_cache: DefIdMap<Option<Stability>>
|
|
|
|
}
|
|
|
|
|
|
|
|
// A private tree-walker for producing an Index.
|
2015-01-12 20:40:19 -06:00
|
|
|
struct Annotator<'a> {
|
|
|
|
sess: &'a Session,
|
2014-09-12 05:10:30 -05:00
|
|
|
index: Index,
|
|
|
|
parent: Option<Stability>
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
}
|
|
|
|
|
2015-01-12 20:40:19 -06:00
|
|
|
impl<'a> Annotator<'a> {
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
// Determine the stability for a node based on its attributes and inherited
|
2014-09-12 05:10:30 -05:00
|
|
|
// stability. The stability is recorded in the index and used as the parent.
|
2014-12-17 22:12:41 -06:00
|
|
|
fn annotate<F>(&mut self, id: NodeId, use_parent: bool,
|
|
|
|
attrs: &Vec<Attribute>, f: F) where
|
2014-12-08 19:26:43 -06:00
|
|
|
F: FnOnce(&mut Annotator),
|
|
|
|
{
|
2015-01-12 20:40:19 -06:00
|
|
|
match attr::find_stability(self.sess.diagnostic(), attrs.as_slice()) {
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
Some(stab) => {
|
|
|
|
self.index.local.insert(id, stab.clone());
|
2014-11-11 14:46:47 -06:00
|
|
|
|
2015-01-12 20:40:19 -06:00
|
|
|
// Don't inherit #[stable(feature = "grandfathered", since = "1.0.0")]
|
2014-11-11 14:46:47 -06:00
|
|
|
if stab.level != attr::Stable {
|
|
|
|
let parent = replace(&mut self.parent, Some(stab));
|
|
|
|
f(self);
|
|
|
|
self.parent = parent;
|
|
|
|
} else {
|
|
|
|
f(self);
|
|
|
|
}
|
2014-09-12 05:10:30 -05:00
|
|
|
}
|
|
|
|
None => {
|
2014-12-17 22:12:41 -06:00
|
|
|
if use_parent {
|
|
|
|
self.parent.clone().map(|stab| self.index.local.insert(id, stab));
|
|
|
|
}
|
2014-09-12 05:10:30 -05:00
|
|
|
f(self);
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-01-12 20:40:19 -06:00
|
|
|
impl<'a, 'v> Visitor<'v> for Annotator<'a> {
|
2014-09-12 05:10:30 -05:00
|
|
|
fn visit_item(&mut self, i: &Item) {
|
2014-12-17 22:12:41 -06:00
|
|
|
// FIXME (#18969): the following is a hack around the fact
|
|
|
|
// that we cannot currently annotate the stability of
|
|
|
|
// `deriving`. Basically, we do *not* allow stability
|
|
|
|
// inheritance on trait implementations, so that derived
|
|
|
|
// implementations appear to be unannotated. This then allows
|
|
|
|
// derived implementations to be automatically tagged with the
|
|
|
|
// stability of the trait. This is WRONG, but expedient to get
|
|
|
|
// libstd stabilized for the 1.0 release.
|
|
|
|
let use_parent = match i.node {
|
2014-12-28 16:33:18 -06:00
|
|
|
ast::ItemImpl(_, _, _, Some(_), _, _) => false,
|
2014-12-17 22:12:41 -06:00
|
|
|
_ => true,
|
|
|
|
};
|
|
|
|
|
|
|
|
self.annotate(i.id, use_parent, &i.attrs, |v| visit::walk_item(v, i));
|
2014-11-11 14:46:47 -06:00
|
|
|
|
2014-11-29 15:41:21 -06:00
|
|
|
if let ast::ItemStruct(ref sd, _) = i.node {
|
|
|
|
sd.ctor_id.map(|id| {
|
2014-12-17 22:12:41 -06:00
|
|
|
self.annotate(id, true, &i.attrs, |_| {})
|
2014-11-29 15:41:21 -06:00
|
|
|
});
|
2014-11-11 14:46:47 -06:00
|
|
|
}
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
}
|
|
|
|
|
2014-11-17 17:24:27 -06:00
|
|
|
fn visit_fn(&mut self, fk: FnKind<'v>, _: &'v FnDecl,
|
|
|
|
_: &'v Block, _: Span, _: NodeId) {
|
2014-11-29 15:41:21 -06:00
|
|
|
if let FkMethod(_, _, meth) = fk {
|
|
|
|
// Methods are not already annotated, so we annotate it
|
2014-12-17 22:12:41 -06:00
|
|
|
self.annotate(meth.id, true, &meth.attrs, |_| {});
|
2014-09-12 05:10:30 -05:00
|
|
|
}
|
2014-11-17 04:37:07 -06:00
|
|
|
// Items defined in a function body have no reason to have
|
|
|
|
// a stability attribute, so we don't recurse.
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
}
|
|
|
|
|
2014-09-12 05:10:30 -05:00
|
|
|
fn visit_trait_item(&mut self, t: &TraitItem) {
|
|
|
|
let (id, attrs) = match *t {
|
|
|
|
RequiredMethod(TypeMethod {id, ref attrs, ..}) => (id, attrs),
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
|
|
|
|
// work around lack of pattern matching for @ types
|
2014-08-05 21:44:21 -05:00
|
|
|
ProvidedMethod(ref method) => {
|
|
|
|
match **method {
|
2014-10-05 19:36:53 -05:00
|
|
|
Method {ref attrs, id, ..} => (id, attrs),
|
2014-08-05 21:44:21 -05:00
|
|
|
}
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
}
|
2014-08-05 21:44:21 -05:00
|
|
|
|
2014-10-28 13:48:52 -05:00
|
|
|
TypeTraitItem(ref typedef) => (typedef.ty_param.id, &typedef.attrs),
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
};
|
2014-12-17 22:12:41 -06:00
|
|
|
self.annotate(id, true, attrs, |v| visit::walk_trait_item(v, t));
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
}
|
|
|
|
|
2014-09-09 17:54:36 -05:00
|
|
|
fn visit_variant(&mut self, var: &Variant, g: &'v Generics) {
|
2014-12-17 22:12:41 -06:00
|
|
|
self.annotate(var.node.id, true, &var.node.attrs,
|
|
|
|
|v| visit::walk_variant(v, var, g))
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
}
|
|
|
|
|
2014-09-12 05:10:30 -05:00
|
|
|
fn visit_struct_field(&mut self, s: &StructField) {
|
2014-12-17 22:12:41 -06:00
|
|
|
self.annotate(s.node.id, true, &s.node.attrs,
|
|
|
|
|v| visit::walk_struct_field(v, s));
|
2014-07-10 13:17:40 -05:00
|
|
|
}
|
2014-12-20 12:08:16 -06:00
|
|
|
|
|
|
|
fn visit_foreign_item(&mut self, i: &ast::ForeignItem) {
|
2014-12-21 02:12:56 -06:00
|
|
|
self.annotate(i.id, true, &i.attrs, |_| {});
|
2014-12-20 12:08:16 -06:00
|
|
|
}
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
impl Index {
|
|
|
|
/// Construct the stability index for a crate being compiled.
|
2015-01-12 20:40:19 -06:00
|
|
|
pub fn build(sess: &Session, krate: &Crate) -> Index {
|
|
|
|
let mut staged_api = false;
|
|
|
|
for attr in krate.attrs.iter() {
|
|
|
|
if attr.name().get() == "staged_api" {
|
|
|
|
match attr.node.value.node {
|
|
|
|
ast::MetaWord(_) => {
|
|
|
|
attr::mark_used(attr);
|
|
|
|
staged_api = true;
|
|
|
|
}
|
|
|
|
_ => (/*pass*/)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
let index = Index {
|
|
|
|
staged_api: staged_api,
|
|
|
|
local: NodeMap(),
|
|
|
|
extern_cache: DefIdMap()
|
|
|
|
};
|
|
|
|
if !staged_api {
|
|
|
|
return index;
|
|
|
|
}
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
let mut annotator = Annotator {
|
2015-01-12 20:40:19 -06:00
|
|
|
sess: sess,
|
|
|
|
index: index,
|
2014-09-12 05:10:30 -05:00
|
|
|
parent: None
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
};
|
2014-12-17 22:12:41 -06:00
|
|
|
annotator.annotate(ast::CRATE_NODE_ID, true, &krate.attrs,
|
|
|
|
|v| visit::walk_crate(v, krate));
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
annotator.index
|
|
|
|
}
|
2014-06-26 13:37:39 -05:00
|
|
|
}
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
|
2014-06-26 13:37:39 -05:00
|
|
|
/// Lookup the stability for a node, loading external crate
|
|
|
|
/// metadata as necessary.
|
2014-08-04 15:56:56 -05:00
|
|
|
pub fn lookup(tcx: &ty::ctxt, id: DefId) -> Option<Stability> {
|
2015-01-01 20:41:44 -06:00
|
|
|
debug!("lookup(id={})",
|
|
|
|
id.repr(tcx));
|
|
|
|
|
2014-06-26 13:37:39 -05:00
|
|
|
// is this definition the implementation of a trait method?
|
2014-08-04 15:56:56 -05:00
|
|
|
match ty::trait_item_of_item(tcx, id) {
|
2015-01-01 20:41:44 -06:00
|
|
|
Some(ty::MethodTraitItemId(trait_method_id)) if trait_method_id != id => {
|
2014-12-20 02:09:35 -06:00
|
|
|
debug!("lookup: trait_method_id={:?}", trait_method_id);
|
2014-12-17 22:12:41 -06:00
|
|
|
return lookup(tcx, trait_method_id)
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
}
|
2014-12-17 22:12:41 -06:00
|
|
|
_ => {}
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
}
|
2014-12-17 22:12:41 -06:00
|
|
|
|
|
|
|
let item_stab = if is_local(id) {
|
|
|
|
tcx.stability.borrow().local.get(&id.node).cloned()
|
|
|
|
} else {
|
|
|
|
let stab = csearch::get_stability(&tcx.sess.cstore, id);
|
|
|
|
let mut index = tcx.stability.borrow_mut();
|
|
|
|
(*index).extern_cache.insert(id, stab.clone());
|
|
|
|
stab
|
|
|
|
};
|
|
|
|
|
|
|
|
item_stab.or_else(|| {
|
|
|
|
if let Some(trait_id) = ty::trait_id_of_impl(tcx, id) {
|
|
|
|
// FIXME (#18969): for the time being, simply use the
|
|
|
|
// stability of the trait to determine the stability of any
|
|
|
|
// unmarked impls for it. See FIXME above for more details.
|
|
|
|
|
2014-12-20 02:09:35 -06:00
|
|
|
debug!("lookup: trait_id={:?}", trait_id);
|
2014-12-17 22:12:41 -06:00
|
|
|
lookup(tcx, trait_id)
|
|
|
|
} else {
|
|
|
|
None
|
|
|
|
}
|
|
|
|
})
|
Add stability inheritance
This commit makes several changes to the stability index infrastructure:
* Stability levels are now inherited lexically, i.e., each item's
stability level becomes the default for any nested items.
* The computed stability level for an item is stored as part of the
metadata. When using an item from an external crate, this data is
looked up and cached.
* The stability lint works from the computed stability level, rather
than manual stability attribute annotations. However, the lint still
checks only a limited set of item uses (e.g., it does not check every
component of a path on import). This will be addressed in a later PR,
as part of issue #8962.
* The stability lint only applies to items originating from external
crates, since the stability index is intended as a promise to
downstream crates.
* The "experimental" lint is now _allow_ by default. This is because
almost all existing crates have been marked "experimental", pending
library stabilization. With inheritance in place, this would generate
a massive explosion of warnings for every Rust program.
The lint should be changed back to deny-by-default after library
stabilization is complete.
* The "deprecated" lint still warns by default.
The net result: we can begin tracking stability index for the standard
libraries as we stabilize, without impacting most clients.
Closes #13540.
2014-06-11 19:23:11 -05:00
|
|
|
}
|
Preliminary feature staging
This partially implements the feature staging described in the
[release channel RFC][rc]. It does not yet fully conform to the RFC as
written, but does accomplish its goals sufficiently for the 1.0 alpha
release.
It has three primary user-visible effects:
* On the nightly channel, use of unstable APIs generates a warning.
* On the beta channel, use of unstable APIs generates a warning.
* On the beta channel, use of feature gates generates a warning.
Code that does not trigger these warnings is considered 'stable',
modulo pre-1.0 bugs.
Disabling the warnings for unstable APIs continues to be done in the
existing (i.e. old) style, via `#[allow(...)]`, not that specified in
the RFC. I deem this marginally acceptable since any code that must do
this is not using the stable dialect of Rust.
Use of feature gates is itself gated with the new 'unstable_features'
lint, on nightly set to 'allow', and on beta 'warn'.
The attribute scheme used here corresponds to an older version of the
RFC, with the `#[staged_api]` crate attribute toggling the staging
behavior of the stability attributes, but the user impact is only
in-tree so I'm not concerned about having to make design changes later
(and I may ultimately prefer the scheme here after all, with the
`#[staged_api]` crate attribute).
Since the Rust codebase itself makes use of unstable features the
compiler and build system to a midly elaborate dance to allow it to
bootstrap while disobeying these lints (which would otherwise be
errors because Rust builds with `-D warnings`).
This patch includes one significant hack that causes a
regression. Because the `format_args!` macro emits calls to unstable
APIs it would trigger the lint. I added a hack to the lint to make it
not trigger, but this in turn causes arguments to `println!` not to be
checked for feature gates. I don't presently understand macro
expansion well enough to fix. This is bug #20661.
Closes #16678
[rc]: https://github.com/rust-lang/rfcs/blob/master/text/0507-release-channels.md
2015-01-06 08:26:08 -06:00
|
|
|
|
|
|
|
pub fn is_staged_api(tcx: &ty::ctxt, id: DefId) -> bool {
|
|
|
|
match ty::trait_item_of_item(tcx, id) {
|
|
|
|
Some(ty::MethodTraitItemId(trait_method_id))
|
|
|
|
if trait_method_id != id => {
|
|
|
|
is_staged_api(tcx, trait_method_id)
|
|
|
|
}
|
|
|
|
_ if is_local(id) => {
|
2015-01-12 20:40:19 -06:00
|
|
|
tcx.stability.borrow().staged_api
|
Preliminary feature staging
This partially implements the feature staging described in the
[release channel RFC][rc]. It does not yet fully conform to the RFC as
written, but does accomplish its goals sufficiently for the 1.0 alpha
release.
It has three primary user-visible effects:
* On the nightly channel, use of unstable APIs generates a warning.
* On the beta channel, use of unstable APIs generates a warning.
* On the beta channel, use of feature gates generates a warning.
Code that does not trigger these warnings is considered 'stable',
modulo pre-1.0 bugs.
Disabling the warnings for unstable APIs continues to be done in the
existing (i.e. old) style, via `#[allow(...)]`, not that specified in
the RFC. I deem this marginally acceptable since any code that must do
this is not using the stable dialect of Rust.
Use of feature gates is itself gated with the new 'unstable_features'
lint, on nightly set to 'allow', and on beta 'warn'.
The attribute scheme used here corresponds to an older version of the
RFC, with the `#[staged_api]` crate attribute toggling the staging
behavior of the stability attributes, but the user impact is only
in-tree so I'm not concerned about having to make design changes later
(and I may ultimately prefer the scheme here after all, with the
`#[staged_api]` crate attribute).
Since the Rust codebase itself makes use of unstable features the
compiler and build system to a midly elaborate dance to allow it to
bootstrap while disobeying these lints (which would otherwise be
errors because Rust builds with `-D warnings`).
This patch includes one significant hack that causes a
regression. Because the `format_args!` macro emits calls to unstable
APIs it would trigger the lint. I added a hack to the lint to make it
not trigger, but this in turn causes arguments to `println!` not to be
checked for feature gates. I don't presently understand macro
expansion well enough to fix. This is bug #20661.
Closes #16678
[rc]: https://github.com/rust-lang/rfcs/blob/master/text/0507-release-channels.md
2015-01-06 08:26:08 -06:00
|
|
|
}
|
|
|
|
_ => {
|
|
|
|
csearch::is_staged_api(&tcx.sess.cstore, id)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|