rust/compiler/rustc_type_ir/src/lib.rs

611 lines
21 KiB
Rust

#![feature(never_type)]
#![feature(const_panic)]
#![feature(control_flow_enum)]
#[macro_use]
extern crate bitflags;
#[macro_use]
extern crate rustc_macros;
use rustc_data_structures::stable_hasher::{HashStable, StableHasher};
use rustc_data_structures::unify::{EqUnifyValue, UnifyKey};
use std::fmt;
use std::mem::discriminant;
bitflags! {
/// Flags that we track on types. These flags are propagated upwards
/// through the type during type construction, so that we can quickly check
/// whether the type has various kinds of types in it without recursing
/// over the type itself.
pub struct TypeFlags: u32 {
// Does this have parameters? Used to determine whether substitution is
// required.
/// Does this have `Param`?
const HAS_TY_PARAM = 1 << 0;
/// Does this have `ReEarlyBound`?
const HAS_RE_PARAM = 1 << 1;
/// Does this have `ConstKind::Param`?
const HAS_CT_PARAM = 1 << 2;
const NEEDS_SUBST = TypeFlags::HAS_TY_PARAM.bits
| TypeFlags::HAS_RE_PARAM.bits
| TypeFlags::HAS_CT_PARAM.bits;
/// Does this have `Infer`?
const HAS_TY_INFER = 1 << 3;
/// Does this have `ReVar`?
const HAS_RE_INFER = 1 << 4;
/// Does this have `ConstKind::Infer`?
const HAS_CT_INFER = 1 << 5;
/// Does this have inference variables? Used to determine whether
/// inference is required.
const NEEDS_INFER = TypeFlags::HAS_TY_INFER.bits
| TypeFlags::HAS_RE_INFER.bits
| TypeFlags::HAS_CT_INFER.bits;
/// Does this have `Placeholder`?
const HAS_TY_PLACEHOLDER = 1 << 6;
/// Does this have `RePlaceholder`?
const HAS_RE_PLACEHOLDER = 1 << 7;
/// Does this have `ConstKind::Placeholder`?
const HAS_CT_PLACEHOLDER = 1 << 8;
/// `true` if there are "names" of regions and so forth
/// that are local to a particular fn/inferctxt
const HAS_FREE_LOCAL_REGIONS = 1 << 9;
/// `true` if there are "names" of types and regions and so forth
/// that are local to a particular fn
const HAS_FREE_LOCAL_NAMES = TypeFlags::HAS_TY_PARAM.bits
| TypeFlags::HAS_CT_PARAM.bits
| TypeFlags::HAS_TY_INFER.bits
| TypeFlags::HAS_CT_INFER.bits
| TypeFlags::HAS_TY_PLACEHOLDER.bits
| TypeFlags::HAS_CT_PLACEHOLDER.bits
| TypeFlags::HAS_FREE_LOCAL_REGIONS.bits;
/// Does this have `Projection`?
const HAS_TY_PROJECTION = 1 << 10;
/// Does this have `Opaque`?
const HAS_TY_OPAQUE = 1 << 11;
/// Does this have `ConstKind::Unevaluated`?
const HAS_CT_PROJECTION = 1 << 12;
/// Could this type be normalized further?
const HAS_PROJECTION = TypeFlags::HAS_TY_PROJECTION.bits
| TypeFlags::HAS_TY_OPAQUE.bits
| TypeFlags::HAS_CT_PROJECTION.bits;
/// Is an error type/const reachable?
const HAS_ERROR = 1 << 13;
/// Does this have any region that "appears free" in the type?
/// Basically anything but `ReLateBound` and `ReErased`.
const HAS_FREE_REGIONS = 1 << 14;
/// Does this have any `ReLateBound` regions? Used to check
/// if a global bound is safe to evaluate.
const HAS_RE_LATE_BOUND = 1 << 15;
/// Does this have any `ReErased` regions?
const HAS_RE_ERASED = 1 << 16;
/// Does this value have parameters/placeholders/inference variables which could be
/// replaced later, in a way that would change the results of `impl` specialization?
const STILL_FURTHER_SPECIALIZABLE = 1 << 17;
}
}
rustc_index::newtype_index! {
/// A [De Bruijn index][dbi] is a standard means of representing
/// regions (and perhaps later types) in a higher-ranked setting. In
/// particular, imagine a type like this:
///
/// for<'a> fn(for<'b> fn(&'b isize, &'a isize), &'a char)
/// ^ ^ | | |
/// | | | | |
/// | +------------+ 0 | |
/// | | |
/// +----------------------------------+ 1 |
/// | |
/// +----------------------------------------------+ 0
///
/// In this type, there are two binders (the outer fn and the inner
/// fn). We need to be able to determine, for any given region, which
/// fn type it is bound by, the inner or the outer one. There are
/// various ways you can do this, but a De Bruijn index is one of the
/// more convenient and has some nice properties. The basic idea is to
/// count the number of binders, inside out. Some examples should help
/// clarify what I mean.
///
/// Let's start with the reference type `&'b isize` that is the first
/// argument to the inner function. This region `'b` is assigned a De
/// Bruijn index of 0, meaning "the innermost binder" (in this case, a
/// fn). The region `'a` that appears in the second argument type (`&'a
/// isize`) would then be assigned a De Bruijn index of 1, meaning "the
/// second-innermost binder". (These indices are written on the arrows
/// in the diagram).
///
/// What is interesting is that De Bruijn index attached to a particular
/// variable will vary depending on where it appears. For example,
/// the final type `&'a char` also refers to the region `'a` declared on
/// the outermost fn. But this time, this reference is not nested within
/// any other binders (i.e., it is not an argument to the inner fn, but
/// rather the outer one). Therefore, in this case, it is assigned a
/// De Bruijn index of 0, because the innermost binder in that location
/// is the outer fn.
///
/// [dbi]: https://en.wikipedia.org/wiki/De_Bruijn_index
pub struct DebruijnIndex {
DEBUG_FORMAT = "DebruijnIndex({})",
const INNERMOST = 0,
}
}
impl DebruijnIndex {
/// Returns the resulting index when this value is moved into
/// `amount` number of new binders. So, e.g., if you had
///
/// for<'a> fn(&'a x)
///
/// and you wanted to change it to
///
/// for<'a> fn(for<'b> fn(&'a x))
///
/// you would need to shift the index for `'a` into a new binder.
#[must_use]
pub fn shifted_in(self, amount: u32) -> DebruijnIndex {
DebruijnIndex::from_u32(self.as_u32() + amount)
}
/// Update this index in place by shifting it "in" through
/// `amount` number of binders.
pub fn shift_in(&mut self, amount: u32) {
*self = self.shifted_in(amount);
}
/// Returns the resulting index when this value is moved out from
/// `amount` number of new binders.
#[must_use]
pub fn shifted_out(self, amount: u32) -> DebruijnIndex {
DebruijnIndex::from_u32(self.as_u32() - amount)
}
/// Update in place by shifting out from `amount` binders.
pub fn shift_out(&mut self, amount: u32) {
*self = self.shifted_out(amount);
}
/// Adjusts any De Bruijn indices so as to make `to_binder` the
/// innermost binder. That is, if we have something bound at `to_binder`,
/// it will now be bound at INNERMOST. This is an appropriate thing to do
/// when moving a region out from inside binders:
///
/// ```
/// for<'a> fn(for<'b> for<'c> fn(&'a u32), _)
/// // Binder: D3 D2 D1 ^^
/// ```
///
/// Here, the region `'a` would have the De Bruijn index D3,
/// because it is the bound 3 binders out. However, if we wanted
/// to refer to that region `'a` in the second argument (the `_`),
/// those two binders would not be in scope. In that case, we
/// might invoke `shift_out_to_binder(D3)`. This would adjust the
/// De Bruijn index of `'a` to D1 (the innermost binder).
///
/// If we invoke `shift_out_to_binder` and the region is in fact
/// bound by one of the binders we are shifting out of, that is an
/// error (and should fail an assertion failure).
pub fn shifted_out_to_binder(self, to_binder: DebruijnIndex) -> Self {
self.shifted_out(to_binder.as_u32() - INNERMOST.as_u32())
}
}
#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash, Debug)]
#[derive(Encodable, Decodable)]
pub enum IntTy {
Isize,
I8,
I16,
I32,
I64,
I128,
}
impl IntTy {
pub fn name_str(&self) -> &'static str {
match *self {
IntTy::Isize => "isize",
IntTy::I8 => "i8",
IntTy::I16 => "i16",
IntTy::I32 => "i32",
IntTy::I64 => "i64",
IntTy::I128 => "i128",
}
}
pub fn bit_width(&self) -> Option<u64> {
Some(match *self {
IntTy::Isize => return None,
IntTy::I8 => 8,
IntTy::I16 => 16,
IntTy::I32 => 32,
IntTy::I64 => 64,
IntTy::I128 => 128,
})
}
pub fn normalize(&self, target_width: u32) -> Self {
match self {
IntTy::Isize => match target_width {
16 => IntTy::I16,
32 => IntTy::I32,
64 => IntTy::I64,
_ => unreachable!(),
},
_ => *self,
}
}
}
#[derive(Clone, PartialEq, Eq, PartialOrd, Ord, Hash, Copy, Debug)]
#[derive(Encodable, Decodable)]
pub enum UintTy {
Usize,
U8,
U16,
U32,
U64,
U128,
}
impl UintTy {
pub fn name_str(&self) -> &'static str {
match *self {
UintTy::Usize => "usize",
UintTy::U8 => "u8",
UintTy::U16 => "u16",
UintTy::U32 => "u32",
UintTy::U64 => "u64",
UintTy::U128 => "u128",
}
}
pub fn bit_width(&self) -> Option<u64> {
Some(match *self {
UintTy::Usize => return None,
UintTy::U8 => 8,
UintTy::U16 => 16,
UintTy::U32 => 32,
UintTy::U64 => 64,
UintTy::U128 => 128,
})
}
pub fn normalize(&self, target_width: u32) -> Self {
match self {
UintTy::Usize => match target_width {
16 => UintTy::U16,
32 => UintTy::U32,
64 => UintTy::U64,
_ => unreachable!(),
},
_ => *self,
}
}
}
#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash, Debug)]
#[derive(Encodable, Decodable)]
pub enum FloatTy {
F32,
F64,
}
impl FloatTy {
pub fn name_str(self) -> &'static str {
match self {
FloatTy::F32 => "f32",
FloatTy::F64 => "f64",
}
}
pub fn bit_width(self) -> u64 {
match self {
FloatTy::F32 => 32,
FloatTy::F64 => 64,
}
}
}
#[derive(Clone, Copy, PartialEq, Eq)]
pub enum IntVarValue {
IntType(IntTy),
UintType(UintTy),
}
#[derive(Clone, Copy, PartialEq, Eq)]
pub struct FloatVarValue(pub FloatTy);
/// A **ty**pe **v**ariable **ID**.
#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash, Encodable, Decodable)]
pub struct TyVid {
pub index: u32,
}
/// An **int**egral (`u32`, `i32`, `usize`, etc.) type **v**ariable **ID**.
#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash, Encodable, Decodable)]
pub struct IntVid {
pub index: u32,
}
/// An **float**ing-point (`f32` or `f64`) type **v**ariable **ID**.
#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash, Encodable, Decodable)]
pub struct FloatVid {
pub index: u32,
}
/// A placeholder for a type that hasn't been inferred yet.
///
/// E.g., if we have an empty array (`[]`), then we create a fresh
/// type variable for the element type since we won't know until it's
/// used what the element type is supposed to be.
#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash, Encodable, Decodable)]
pub enum InferTy {
/// A type variable.
TyVar(TyVid),
/// An integral type variable (`{integer}`).
///
/// These are created when the compiler sees an integer literal like
/// `1` that could be several different types (`u8`, `i32`, `u32`, etc.).
/// We don't know until it's used what type it's supposed to be, so
/// we create a fresh type variable.
IntVar(IntVid),
/// A floating-point type variable (`{float}`).
///
/// These are created when the compiler sees an float literal like
/// `1.0` that could be either an `f32` or an `f64`.
/// We don't know until it's used what type it's supposed to be, so
/// we create a fresh type variable.
FloatVar(FloatVid),
/// A [`FreshTy`][Self::FreshTy] is one that is generated as a replacement
/// for an unbound type variable. This is convenient for caching etc. See
/// `rustc_infer::infer::freshen` for more details.
///
/// Compare with [`TyVar`][Self::TyVar].
FreshTy(u32),
/// Like [`FreshTy`][Self::FreshTy], but as a replacement for [`IntVar`][Self::IntVar].
FreshIntTy(u32),
/// Like [`FreshTy`][Self::FreshTy], but as a replacement for [`FloatVar`][Self::FloatVar].
FreshFloatTy(u32),
}
/// Raw `TyVid` are used as the unification key for `sub_relations`;
/// they carry no values.
impl UnifyKey for TyVid {
type Value = ();
fn index(&self) -> u32 {
self.index
}
fn from_index(i: u32) -> TyVid {
TyVid { index: i }
}
fn tag() -> &'static str {
"TyVid"
}
}
impl EqUnifyValue for IntVarValue {}
impl UnifyKey for IntVid {
type Value = Option<IntVarValue>;
fn index(&self) -> u32 {
self.index
}
fn from_index(i: u32) -> IntVid {
IntVid { index: i }
}
fn tag() -> &'static str {
"IntVid"
}
}
impl EqUnifyValue for FloatVarValue {}
impl UnifyKey for FloatVid {
type Value = Option<FloatVarValue>;
fn index(&self) -> u32 {
self.index
}
fn from_index(i: u32) -> FloatVid {
FloatVid { index: i }
}
fn tag() -> &'static str {
"FloatVid"
}
}
#[derive(Copy, Clone, PartialEq, Decodable, Encodable)]
pub enum Variance {
Covariant, // T<A> <: T<B> iff A <: B -- e.g., function return type
Invariant, // T<A> <: T<B> iff B == A -- e.g., type of mutable cell
Contravariant, // T<A> <: T<B> iff B <: A -- e.g., function param type
Bivariant, // T<A> <: T<B> -- e.g., unused type parameter
}
impl Variance {
/// `a.xform(b)` combines the variance of a context with the
/// variance of a type with the following meaning. If we are in a
/// context with variance `a`, and we encounter a type argument in
/// a position with variance `b`, then `a.xform(b)` is the new
/// variance with which the argument appears.
///
/// Example 1:
///
/// *mut Vec<i32>
///
/// Here, the "ambient" variance starts as covariant. `*mut T` is
/// invariant with respect to `T`, so the variance in which the
/// `Vec<i32>` appears is `Covariant.xform(Invariant)`, which
/// yields `Invariant`. Now, the type `Vec<T>` is covariant with
/// respect to its type argument `T`, and hence the variance of
/// the `i32` here is `Invariant.xform(Covariant)`, which results
/// (again) in `Invariant`.
///
/// Example 2:
///
/// fn(*const Vec<i32>, *mut Vec<i32)
///
/// The ambient variance is covariant. A `fn` type is
/// contravariant with respect to its parameters, so the variance
/// within which both pointer types appear is
/// `Covariant.xform(Contravariant)`, or `Contravariant`. `*const
/// T` is covariant with respect to `T`, so the variance within
/// which the first `Vec<i32>` appears is
/// `Contravariant.xform(Covariant)` or `Contravariant`. The same
/// is true for its `i32` argument. In the `*mut T` case, the
/// variance of `Vec<i32>` is `Contravariant.xform(Invariant)`,
/// and hence the outermost type is `Invariant` with respect to
/// `Vec<i32>` (and its `i32` argument).
///
/// Source: Figure 1 of "Taming the Wildcards:
/// Combining Definition- and Use-Site Variance" published in PLDI'11.
pub fn xform(self, v: Variance) -> Variance {
match (self, v) {
// Figure 1, column 1.
(Variance::Covariant, Variance::Covariant) => Variance::Covariant,
(Variance::Covariant, Variance::Contravariant) => Variance::Contravariant,
(Variance::Covariant, Variance::Invariant) => Variance::Invariant,
(Variance::Covariant, Variance::Bivariant) => Variance::Bivariant,
// Figure 1, column 2.
(Variance::Contravariant, Variance::Covariant) => Variance::Contravariant,
(Variance::Contravariant, Variance::Contravariant) => Variance::Covariant,
(Variance::Contravariant, Variance::Invariant) => Variance::Invariant,
(Variance::Contravariant, Variance::Bivariant) => Variance::Bivariant,
// Figure 1, column 3.
(Variance::Invariant, _) => Variance::Invariant,
// Figure 1, column 4.
(Variance::Bivariant, _) => Variance::Bivariant,
}
}
}
impl<CTX> HashStable<CTX> for DebruijnIndex {
fn hash_stable(&self, ctx: &mut CTX, hasher: &mut StableHasher) {
self.as_u32().hash_stable(ctx, hasher);
}
}
impl<CTX> HashStable<CTX> for IntTy {
fn hash_stable(&self, ctx: &mut CTX, hasher: &mut StableHasher) {
discriminant(self).hash_stable(ctx, hasher);
}
}
impl<CTX> HashStable<CTX> for UintTy {
fn hash_stable(&self, ctx: &mut CTX, hasher: &mut StableHasher) {
discriminant(self).hash_stable(ctx, hasher);
}
}
impl<CTX> HashStable<CTX> for FloatTy {
fn hash_stable(&self, ctx: &mut CTX, hasher: &mut StableHasher) {
discriminant(self).hash_stable(ctx, hasher);
}
}
impl<CTX> HashStable<CTX> for InferTy {
fn hash_stable(&self, ctx: &mut CTX, hasher: &mut StableHasher) {
use InferTy::*;
match self {
TyVar(v) => v.index.hash_stable(ctx, hasher),
IntVar(v) => v.index.hash_stable(ctx, hasher),
FloatVar(v) => v.index.hash_stable(ctx, hasher),
FreshTy(v) | FreshIntTy(v) | FreshFloatTy(v) => v.hash_stable(ctx, hasher),
}
}
}
impl<CTX> HashStable<CTX> for Variance {
fn hash_stable(&self, ctx: &mut CTX, hasher: &mut StableHasher) {
discriminant(self).hash_stable(ctx, hasher);
}
}
impl fmt::Debug for IntVarValue {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
match *self {
IntVarValue::IntType(ref v) => v.fmt(f),
IntVarValue::UintType(ref v) => v.fmt(f),
}
}
}
impl fmt::Debug for FloatVarValue {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
self.0.fmt(f)
}
}
impl fmt::Debug for TyVid {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
write!(f, "_#{}t", self.index)
}
}
impl fmt::Debug for IntVid {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
write!(f, "_#{}i", self.index)
}
}
impl fmt::Debug for FloatVid {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
write!(f, "_#{}f", self.index)
}
}
impl fmt::Debug for InferTy {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
use InferTy::*;
match *self {
TyVar(ref v) => v.fmt(f),
IntVar(ref v) => v.fmt(f),
FloatVar(ref v) => v.fmt(f),
FreshTy(v) => write!(f, "FreshTy({:?})", v),
FreshIntTy(v) => write!(f, "FreshIntTy({:?})", v),
FreshFloatTy(v) => write!(f, "FreshFloatTy({:?})", v),
}
}
}
impl fmt::Debug for Variance {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
f.write_str(match *self {
Variance::Covariant => "+",
Variance::Contravariant => "-",
Variance::Invariant => "o",
Variance::Bivariant => "*",
})
}
}
impl fmt::Display for InferTy {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
use InferTy::*;
match *self {
TyVar(_) => write!(f, "_"),
IntVar(_) => write!(f, "{}", "{integer}"),
FloatVar(_) => write!(f, "{}", "{float}"),
FreshTy(v) => write!(f, "FreshTy({})", v),
FreshIntTy(v) => write!(f, "FreshIntTy({})", v),
FreshFloatTy(v) => write!(f, "FreshFloatTy({})", v),
}
}
}