nordic-dev.net/rust
nordic-dev.net/rust
2
0
Fork 0
mirror of https://github.com/rust-lang/rust.git synced 2024-11-25 16:24:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

Rollup merge of #118822 - Nadrieril:librarify, r=compiler-errors

Browse Source 
Extract exhaustiveness into its own crate

It now makes sense to extract exhaustiveness into its own crate! This was much-requested by rust-analyzer (they currently maintain by hand a copy of the algorithm), and I hope this can serve other projects e.g. clippy.

This is the churny PR: it exclusively moves code around. It's not yet useable outside of rustc but I wanted the churny parts to be out of the way.

r? `@compiler-errors`
This commit is contained in:
Matthias Krüger 2023-12-11 20:46:50 +01:00 committed by GitHub
commit dd0887c75c
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
19 changed files with 2614 additions and 2460 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
Cargo.lock
View File

@ -3756,6 +3756,7 @@ dependencies = [
"rustc_monomorphize",
"rustc_parse",
"rustc_passes",
"rustc_pattern_analysis",
"rustc_privacy",
"rustc_query_system",
"rustc_resolve",
@ -4229,6 +4230,7 @@ dependencies = [
"rustc_infer",
"rustc_macros",
"rustc_middle",
"rustc_pattern_analysis",
"rustc_session",
"rustc_span",
"rustc_target",
@ -4364,6 +4366,26 @@ dependencies = [
"tracing",
]
[[package]]
name = "rustc_pattern_analysis"
version = "0.0.0"
dependencies = [
"rustc_apfloat",
"rustc_arena",
"rustc_data_structures",
"rustc_errors",
"rustc_fluent_macro",
"rustc_hir",
"rustc_index",
"rustc_macros",
"rustc_middle",
"rustc_session",
"rustc_span",
"rustc_target",
"smallvec",
"tracing",
]
[[package]]
name = "rustc_privacy"
version = "0.0.0"

1
compiler/rustc_driver_impl/Cargo.toml
View File

@ -38,6 +38,7 @@ rustc_mir_transform = { path = "../rustc_mir_transform" }
rustc_monomorphize = { path = "../rustc_monomorphize" }
rustc_parse = { path = "../rustc_parse" }
rustc_passes = { path = "../rustc_passes" }
rustc_pattern_analysis = { path = "../rustc_pattern_analysis" }
rustc_privacy = { path = "../rustc_privacy" }
rustc_query_system = { path = "../rustc_query_system" }
rustc_resolve = { path = "../rustc_resolve" }

1
compiler/rustc_driver_impl/src/lib.rs
View File

@ -128,6 +128,7 @@ pub static DEFAULT_LOCALE_RESOURCES: &[&str] = &[
rustc_monomorphize::DEFAULT_LOCALE_RESOURCE,
rustc_parse::DEFAULT_LOCALE_RESOURCE,
rustc_passes::DEFAULT_LOCALE_RESOURCE,
rustc_pattern_analysis::DEFAULT_LOCALE_RESOURCE,
rustc_privacy::DEFAULT_LOCALE_RESOURCE,
rustc_query_system::DEFAULT_LOCALE_RESOURCE,
rustc_resolve::DEFAULT_LOCALE_RESOURCE,

1
compiler/rustc_mir_build/Cargo.toml
View File

@ -17,6 +17,7 @@ rustc_index = { path = "../rustc_index" }
rustc_infer = { path = "../rustc_infer" }
rustc_macros = { path = "../rustc_macros" }
rustc_middle = { path = "../rustc_middle" }
rustc_pattern_analysis = { path = "../rustc_pattern_analysis" }
rustc_session = { path = "../rustc_session" }
rustc_span = { path = "../rustc_span" }
rustc_target = { path = "../rustc_target" }

20
compiler/rustc_mir_build/messages.ftl
View File

@ -237,15 +237,6 @@ mir_build_non_const_path = runtime values cannot be referenced in patterns
mir_build_non_exhaustive_match_all_arms_guarded =
match arms with guards don't count towards exhaustivity
mir_build_non_exhaustive_omitted_pattern = some variants are not matched explicitly
.help = ensure that all variants are matched explicitly by adding the suggested match arms
.note = the matched value is of type `{$scrut_ty}` and the `non_exhaustive_omitted_patterns` attribute was found
mir_build_non_exhaustive_omitted_pattern_lint_on_arm = the lint level must be set on the whole match
.help = it no longer has any effect to set the lint level on an individual match arm
.label = remove this attribute
.suggestion = set the lint level on the whole match
mir_build_non_exhaustive_patterns_type_not_empty = non-exhaustive patterns: type `{$ty}` is non-empty
.def_note = `{$peeled_ty}` defined here
.type_note = the matched value is of type `{$ty}`
@ -260,10 +251,6 @@ mir_build_non_partial_eq_match =
mir_build_nontrivial_structural_match =
to use a constant of type `{$non_sm_ty}` in a pattern, the constant's initializer must be trivial or `{$non_sm_ty}` must be annotated with `#[derive(PartialEq, Eq)]`
mir_build_overlapping_range_endpoints = multiple patterns overlap on their endpoints
.range = ... with this range
.note = you likely meant to write mutually exclusive ranges
mir_build_pattern_not_covered = refutable pattern in {$origin}
.pattern_ty = the matched value is of type `{$pattern_ty}`
@ -317,13 +304,6 @@ mir_build_unconditional_recursion = function cannot return without recursing
mir_build_unconditional_recursion_call_site_label = recursive call site
mir_build_uncovered = {$count ->
[1] pattern `{$witness_1}`
[2] patterns `{$witness_1}` and `{$witness_2}`
[3] patterns `{$witness_1}`, `{$witness_2}` and `{$witness_3}`
*[other] patterns `{$witness_1}`, `{$witness_2}`, `{$witness_3}` and {$remainder} more
} not covered
mir_build_union_field_requires_unsafe =
access to union field is unsafe and requires unsafe block
.note = the field may not be properly initialized: using uninitialized data will cause undefined behavior

95
compiler/rustc_mir_build/src/errors.rs
View File

@ -1,15 +1,12 @@
use crate::{
fluent_generated as fluent,
thir::pattern::{deconstruct_pat::WitnessPat, MatchCheckCtxt},
};
use crate::fluent_generated as fluent;
use rustc_errors::DiagnosticArgValue;
use rustc_errors::{
error_code, AddToDiagnostic, Applicability, Diagnostic, DiagnosticBuilder, ErrorGuaranteed,
Handler, IntoDiagnostic, MultiSpan, SubdiagnosticMessage,
};
use rustc_macros::{Diagnostic, LintDiagnostic, Subdiagnostic};
use rustc_middle::thir::Pat;
use rustc_middle::ty::{self, Ty};
use rustc_pattern_analysis::{cx::MatchCheckCtxt, errors::Uncovered};
use rustc_span::symbol::Symbol;
use rustc_span::Span;
@ -812,94 +809,6 @@ pub struct NonPartialEqMatch<'tcx> {
pub non_peq_ty: Ty<'tcx>,
}
#[derive(LintDiagnostic)]
#[diag(mir_build_overlapping_range_endpoints)]
#[note]
pub struct OverlappingRangeEndpoints<'tcx> {
#[label(mir_build_range)]
pub range: Span,
#[subdiagnostic]
pub overlap: Vec<Overlap<'tcx>>,
}
pub struct Overlap<'tcx> {
pub span: Span,
pub range: Pat<'tcx>,
}
impl<'tcx> AddToDiagnostic for Overlap<'tcx> {
fn add_to_diagnostic_with<F>(self, diag: &mut Diagnostic, _: F)
where
F: Fn(&mut Diagnostic, SubdiagnosticMessage) -> SubdiagnosticMessage,
{
let Overlap { span, range } = self;
// FIXME(mejrs) unfortunately `#[derive(LintDiagnostic)]`
// does not support `#[subdiagnostic(eager)]`...
let message = format!("this range overlaps on `{range}`...");
diag.span_label(span, message);
}
}
#[derive(LintDiagnostic)]
#[diag(mir_build_non_exhaustive_omitted_pattern)]
#[help]
#[note]
pub(crate) struct NonExhaustiveOmittedPattern<'tcx> {
pub scrut_ty: Ty<'tcx>,
#[subdiagnostic]
pub uncovered: Uncovered<'tcx>,
}
#[derive(LintDiagnostic)]
#[diag(mir_build_non_exhaustive_omitted_pattern_lint_on_arm)]
#[help]
pub(crate) struct NonExhaustiveOmittedPatternLintOnArm {
#[label]
pub lint_span: Span,
#[suggestion(code = "#[{lint_level}({lint_name})]\n", applicability = "maybe-incorrect")]
pub suggest_lint_on_match: Option<Span>,
pub lint_level: &'static str,
pub lint_name: &'static str,
}
#[derive(Subdiagnostic)]
#[label(mir_build_uncovered)]
pub(crate) struct Uncovered<'tcx> {
#[primary_span]
span: Span,
count: usize,
witness_1: Pat<'tcx>,
witness_2: Pat<'tcx>,
witness_3: Pat<'tcx>,
remainder: usize,
}
impl<'tcx> Uncovered<'tcx> {
pub fn new<'p>(
span: Span,
cx: &MatchCheckCtxt<'p, 'tcx>,
witnesses: Vec<WitnessPat<'tcx>>,
) -> Self {
let witness_1 = witnesses.get(0).unwrap().to_diagnostic_pat(cx);
Self {
span,
count: witnesses.len(),
// Substitute dummy values if witnesses is smaller than 3. These will never be read.
witness_2: witnesses
.get(1)
.map(|w| w.to_diagnostic_pat(cx))
.unwrap_or_else(|| witness_1.clone()),
witness_3: witnesses
.get(2)
.map(|w| w.to_diagnostic_pat(cx))
.unwrap_or_else(|| witness_1.clone()),
witness_1,
remainder: witnesses.len().saturating_sub(3),
}
}
}
#[derive(Diagnostic)]
#[diag(mir_build_pattern_not_covered, code = "E0005")]
pub(crate) struct PatternNotCovered<'s, 'tcx> {

32
compiler/rustc_mir_build/src/thir/pattern/check_match.rs
View File

@ -1,7 +1,9 @@
use super::deconstruct_pat::{Constructor, DeconstructedPat, WitnessPat};
use super::usefulness::{
compute_match_usefulness, MatchArm, MatchCheckCtxt, Usefulness, UsefulnessReport,
};
use rustc_pattern_analysis::constructor::Constructor;
use rustc_pattern_analysis::cx::MatchCheckCtxt;
use rustc_pattern_analysis::errors::Uncovered;
use rustc_pattern_analysis::pat::{DeconstructedPat, WitnessPat};
use rustc_pattern_analysis::usefulness::{Usefulness, UsefulnessReport};
use rustc_pattern_analysis::{analyze_match, MatchArm};
use crate::errors::*;
@ -284,7 +286,7 @@ impl<'thir, 'p, 'tcx> MatchVisitor<'thir, 'p, 'tcx> {
check_borrow_conflicts_in_at_patterns(self, pat);
check_for_bindings_named_same_as_variants(self, pat, refutable);
});
Ok(cx.pattern_arena.alloc(DeconstructedPat::from_pat(cx, pat)))
Ok(cx.pattern_arena.alloc(cx.lower_pat(pat)))
}
}
@ -433,7 +435,7 @@ impl<'thir, 'p, 'tcx> MatchVisitor<'thir, 'p, 'tcx> {
}
let scrut_ty = scrut.ty;
let report = compute_match_usefulness(&cx, &tarms, scrut_ty);
let report = analyze_match(&cx, &tarms, scrut_ty);
match source {
// Don't report arm reachability of desugared `match $iter.into_iter() { iter => .. }`
@ -547,7 +549,7 @@ impl<'thir, 'p, 'tcx> MatchVisitor<'thir, 'p, 'tcx> {
let cx = self.new_cx(refutability, None, scrut, pat.span);
let pat = self.lower_pattern(&cx, pat)?;
let arms = [MatchArm { pat, hir_id: self.lint_level, has_guard: false }];
let report = compute_match_usefulness(&cx, &arms, pat.ty());
let report = analyze_match(&cx, &arms, pat.ty());
Ok((cx, report))
}
@ -924,7 +926,7 @@ fn report_non_exhaustive_match<'p, 'tcx>(
pattern = if witnesses.len() < 4 {
witnesses
.iter()
.map(|witness| witness.to_diagnostic_pat(cx).to_string())
.map(|witness| cx.hoist_witness_pat(witness).to_string())
.collect::<Vec<String>>()
.join(" | ")
} else {
@ -948,7 +950,7 @@ fn report_non_exhaustive_match<'p, 'tcx>(
if !is_empty_match {
let mut non_exhaustive_tys = FxHashSet::default();
// Look at the first witness.
collect_non_exhaustive_tys(cx.tcx, &witnesses[0], &mut non_exhaustive_tys);
collect_non_exhaustive_tys(cx, &witnesses[0], &mut non_exhaustive_tys);
for ty in non_exhaustive_tys {
if ty.is_ptr_sized_integral() {
@ -1083,13 +1085,13 @@ fn joined_uncovered_patterns<'p, 'tcx>(
witnesses: &[WitnessPat<'tcx>],
) -> String {
const LIMIT: usize = 3;
let pat_to_str = |pat: &WitnessPat<'tcx>| pat.to_diagnostic_pat(cx).to_string();
let pat_to_str = |pat: &WitnessPat<'tcx>| cx.hoist_witness_pat(pat).to_string();
match witnesses {
[] => bug!(),
[witness] => format!("`{}`", witness.to_diagnostic_pat(cx)),
[witness] => format!("`{}`", cx.hoist_witness_pat(witness)),
[head @ .., tail] if head.len() < LIMIT => {
let head: Vec<_> = head.iter().map(pat_to_str).collect();
format!("`{}` and `{}`", head.join("`, `"), tail.to_diagnostic_pat(cx))
format!("`{}` and `{}`", head.join("`, `"), cx.hoist_witness_pat(tail))
}
_ => {
let (head, tail) = witnesses.split_at(LIMIT);
@ -1100,7 +1102,7 @@ fn joined_uncovered_patterns<'p, 'tcx>(
}
fn collect_non_exhaustive_tys<'tcx>(
tcx: TyCtxt<'tcx>,
cx: &MatchCheckCtxt<'_, 'tcx>,
pat: &WitnessPat<'tcx>,
non_exhaustive_tys: &mut FxHashSet<Ty<'tcx>>,
) {
@ -1108,13 +1110,13 @@ fn collect_non_exhaustive_tys<'tcx>(
non_exhaustive_tys.insert(pat.ty());
}
if let Constructor::IntRange(range) = pat.ctor() {
if range.is_beyond_boundaries(pat.ty(), tcx) {
if cx.is_range_beyond_boundaries(range, pat.ty()) {
// The range denotes the values before `isize::MIN` or the values after `usize::MAX`/`isize::MAX`.
non_exhaustive_tys.insert(pat.ty());
}
}
pat.iter_fields()
.for_each(|field_pat| collect_non_exhaustive_tys(tcx, field_pat, non_exhaustive_tys))
.for_each(|field_pat| collect_non_exhaustive_tys(cx, field_pat, non_exhaustive_tys))
}
fn report_adt_defined_here<'tcx>(

1964
compiler/rustc_mir_build/src/thir/pattern/deconstruct_pat.rs
View File

File diff suppressed because it is too large Load Diff

3
compiler/rustc_mir_build/src/thir/pattern/mod.rs
View File

@ -2,11 +2,8 @@
mod check_match;
mod const_to_pat;
pub(crate) mod deconstruct_pat;
mod usefulness;
pub(crate) use self::check_match::check_match;
pub(crate) use self::usefulness::MatchCheckCtxt;
use crate::errors::*;
use crate::thir::util::UserAnnotatedTyHelpers;

22
compiler/rustc_pattern_analysis/Cargo.toml Normal file
View File

@ -0,0 +1,22 @@
[package]
name = "rustc_pattern_analysis"
version = "0.0.0"
edition = "2021"
[dependencies]
# tidy-alphabetical-start
rustc_apfloat = "0.2.0"
rustc_arena = { path = "../rustc_arena" }
rustc_data_structures = { path = "../rustc_data_structures" }
rustc_errors = { path = "../rustc_errors" }
rustc_fluent_macro = { path = "../rustc_fluent_macro" }
rustc_hir = { path = "../rustc_hir" }
rustc_index = { path = "../rustc_index" }
rustc_macros = { path = "../rustc_macros" }
rustc_middle = { path = "../rustc_middle" }
rustc_session = { path = "../rustc_session" }
rustc_span = { path = "../rustc_span" }
rustc_target = { path = "../rustc_target" }
smallvec = { version = "1.8.1", features = ["union", "may_dangle"] }
tracing = "0.1"
# tidy-alphabetical-end

19
compiler/rustc_pattern_analysis/messages.ftl Normal file
View File

@ -0,0 +1,19 @@
pattern_analysis_non_exhaustive_omitted_pattern = some variants are not matched explicitly
.help = ensure that all variants are matched explicitly by adding the suggested match arms
.note = the matched value is of type `{$scrut_ty}` and the `non_exhaustive_omitted_patterns` attribute was found
pattern_analysis_non_exhaustive_omitted_pattern_lint_on_arm = the lint level must be set on the whole match
.help = it no longer has any effect to set the lint level on an individual match arm
.label = remove this attribute
.suggestion = set the lint level on the whole match
pattern_analysis_overlapping_range_endpoints = multiple patterns overlap on their endpoints
.label = ... with this range
.note = you likely meant to write mutually exclusive ranges
pattern_analysis_uncovered = {$count ->
[1] pattern `{$witness_1}`
[2] patterns `{$witness_1}` and `{$witness_2}`
[3] patterns `{$witness_1}`, `{$witness_2}` and `{$witness_3}`
*[other] patterns `{$witness_1}`, `{$witness_2}`, `{$witness_3}` and {$remainder} more
} not covered

987
compiler/rustc_pattern_analysis/src/constructor.rs Normal file
View File

@ -0,0 +1,987 @@
//! As explained in [`crate::usefulness`], values and patterns are made from constructors applied to
//! fields. This file defines a `Constructor` enum and various operations to manipulate them.
//!
//! There are two important bits of core logic in this file: constructor inclusion and constructor
//! splitting. Constructor inclusion, i.e. whether a constructor is included in/covered by another,
//! is straightforward and defined in [`Constructor::is_covered_by`].
//!
//! Constructor splitting is mentioned in [`crate::usefulness`] but not detailed. We describe it
//! precisely here.
//!
//!
//!
//! # Constructor grouping and splitting
//!
//! As explained in the corresponding section in [`crate::usefulness`], to make usefulness tractable
//! we need to group together constructors that have the same effect when they are used to
//! specialize the matrix.
//!
//! Example:
//! ```compile_fail,E0004
//! match (0, false) {
//! (0 ..=100, true) => {}
//! (50..=150, false) => {}
//! (0 ..=200, _) => {}
//! }
//! ```
//!
//! In this example we can restrict specialization to 5 cases: `0..50`, `50..=100`, `101..=150`,
//! `151..=200` and `200..`.
//!
//! In [`crate::usefulness`], we had said that `specialize` only takes value-only constructors. We
//! now relax this restriction: we allow `specialize` to take constructors like `0..50` as long as
//! we're careful to only do that with constructors that make sense. For example, `specialize(0..50,
//! (0..=100, true))` is sensible, but `specialize(50..=200, (0..=100, true))` is not.
//!
//! Constructor splitting looks at the constructors in the first column of the matrix and constructs
//! such a sensible set of constructors. Formally, we want to find a smallest disjoint set of
//! constructors:
//! - Whose union covers the whole type, and
//! - That have no non-trivial intersection with any of the constructors in the column (i.e. they're
//! each either disjoint with or covered by any given column constructor).
//!
//! We compute this in two steps: first [`crate::cx::MatchCheckCtxt::ctors_for_ty`] determines the
//! set of all possible constructors for the type. Then [`ConstructorSet::split`] looks at the
//! column of constructors and splits the set into groups accordingly. The precise invariants of
//! [`ConstructorSet::split`] is described in [`SplitConstructorSet`].
//!
//! Constructor splitting has two interesting special cases: integer range splitting (see
//! [`IntRange::split`]) and slice splitting (see [`Slice::split`]).
//!
//!
//!
//! # The `Missing` constructor
//!
//! We detail a special case of constructor splitting that is a bit subtle. Take the following:
//!
//! ```
//! enum Direction { North, South, East, West }
//! # let wind = (Direction::North, 0u8);
//! match wind {
//! (Direction::North, 50..) => {}
//! (_, _) => {}
//! }
//! ```
//!
//! Here we expect constructor splitting to output two cases: `North`, and "everything else". This
//! "everything else" is represented by [`Constructor::Missing`]. Unlike other constructors, it's a
//! bit contextual: to know the exact list of constructors it represents we have to look at the
//! column. In practice however we don't need to, because by construction it only matches rows that
//! have wildcards. This is how this constructor is special: the only constructor that covers it is
//! `Wildcard`.
//!
//! The only place where we care about which constructors `Missing` represents is in diagnostics
//! (see `crate::usefulness::WitnessMatrix::apply_constructor`).
//!
//! We choose whether to specialize with `Missing` in
//! `crate::usefulness::compute_exhaustiveness_and_usefulness`.
//!
//!
//!
//! ## Empty types, empty constructors, and the `exhaustive_patterns` feature
//!
//! An empty type is a type that has no valid value, like `!`, `enum Void {}`, or `Result<!, !>`.
//! They require careful handling.
//!
//! First, for soundness reasons related to the possible existence of invalid values, by default we
//! don't treat empty types as empty. We force them to be matched with wildcards. Except if the
//! `exhaustive_patterns` feature is turned on, in which case we do treat them as empty. And also
//! except if the type has no constructors (like `enum Void {}` but not like `Result<!, !>`), we
//! specifically allow `match void {}` to be exhaustive. There are additionally considerations of
//! place validity that are handled in `crate::usefulness`. Yes this is a bit tricky.
//!
//! The second thing is that regardless of the above, it is always allowed to use all the
//! constructors of a type. For example, all the following is ok:
//!
//! ```rust,ignore(example)
//! # #![feature(never_type)]
//! # #![feature(exhaustive_patterns)]
//! fn foo(x: Option<!>) {
//! match x {
//! None => {}
//! Some(_) => {}
//! }
//! }
//! fn bar(x: &[!]) -> u32 {
//! match x {
//! [] => 1,
//! [_] => 2,
//! [_, _] => 3,
//! }
//! }
//! ```
//!
//! Moreover, take the following:
//!
//! ```rust
//! # #![feature(never_type)]
//! # #![feature(exhaustive_patterns)]
//! # let x = None::<!>;
//! match x {
//! None => {}
//! }
//! ```
//!
//! On a normal type, we would identify `Some` as missing and tell the user. If `x: Option<!>`
//! however (and `exhaustive_patterns` is on), it's ok to omit `Some`. When listing the constructors
//! of a type, we must therefore track which can be omitted.
//!
//! Let's call "empty" a constructor that matches no valid value for the type, like `Some` for the
//! type `Option<!>`. What this all means is that `ConstructorSet` must know which constructors are
//! empty. The difference between empty and nonempty constructors is that empty constructors need
//! not be present for the match to be exhaustive.
//!
//! A final remark: empty constructors of arity 0 break specialization, we must avoid them. The
//! reason is that if we specialize by them, nothing remains to witness the emptiness; the rest of
//! the algorithm can't distinguish them from a nonempty constructor. The only known case where this
//! could happen is the `[..]` pattern on `[!; N]` with `N > 0` so we must take care to not emit it.
//!
//! This is all handled by [`crate::cx::MatchCheckCtxt::ctors_for_ty`] and
//! [`ConstructorSet::split`]. The invariants of [`SplitConstructorSet`] are also of interest.
//!
//!
//!
//! ## Opaque patterns
//!
//! Some patterns, such as constants that are not allowed to be matched structurally, cannot be
//! inspected, which we handle with `Constructor::Opaque`. Since we know nothing of these patterns,
//! we assume they never cover each other. In order to respect the invariants of
//! [`SplitConstructorSet`], we give each `Opaque` constructor a unique id so we can recognize it.
use std::cmp::{self, max, min, Ordering};
use std::fmt;
use std::iter::once;
use smallvec::SmallVec;
use rustc_apfloat::ieee::{DoubleS, IeeeFloat, SingleS};
use rustc_data_structures::fx::FxHashSet;
use rustc_hir::RangeEnd;
use rustc_index::IndexVec;
use rustc_middle::mir::Const;
use rustc_target::abi::VariantIdx;
use self::Constructor::*;
use self::MaybeInfiniteInt::*;
use self::SliceKind::*;
use crate::usefulness::PatCtxt;
/// Whether we have seen a constructor in the column or not.
#[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord)]
enum Presence {
Unseen,
Seen,
}
/// A possibly infinite integer. Values are encoded such that the ordering on `u128` matches the
/// natural order on the original type. For example, `-128i8` is encoded as `0` and `127i8` as
/// `255`. See `signed_bias` for details.
#[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord)]
pub enum MaybeInfiniteInt {
NegInfinity,
/// Encoded value. DO NOT CONSTRUCT BY HAND; use `new_finite`.
#[non_exhaustive]
Finite(u128),
/// The integer after `u128::MAX`. We need it to represent `x..=u128::MAX` as an exclusive range.
JustAfterMax,
PosInfinity,
}
impl MaybeInfiniteInt {
pub fn new_finite_uint(bits: u128) -> Self {
Finite(bits)
}
pub fn new_finite_int(bits: u128, size: u64) -> Self {
// Perform a shift if the underlying types are signed, which makes the interval arithmetic
// type-independent.
let bias = 1u128 << (size - 1);
Finite(bits ^ bias)
}
pub fn as_finite_uint(self) -> Option<u128> {
match self {
Finite(bits) => Some(bits),
_ => None,
}
}
pub fn as_finite_int(self, size: u64) -> Option<u128> {
// We decode the shift.
match self {
Finite(bits) => {
let bias = 1u128 << (size - 1);
Some(bits ^ bias)
}
_ => None,
}
}
/// Note: this will not turn a finite value into an infinite one or vice-versa.
pub fn minus_one(self) -> Self {
match self {
Finite(n) => match n.checked_sub(1) {
Some(m) => Finite(m),
None => bug!(),
},
JustAfterMax => Finite(u128::MAX),
x => x,
}
}
/// Note: this will not turn a finite value into an infinite one or vice-versa.
pub fn plus_one(self) -> Self {
match self {
Finite(n) => match n.checked_add(1) {
Some(m) => Finite(m),
None => JustAfterMax,
},
JustAfterMax => bug!(),
x => x,
}
}
}
/// An exclusive interval, used for precise integer exhaustiveness checking. `IntRange`s always
/// store a contiguous range.
///
/// `IntRange` is never used to encode an empty range or a "range" that wraps around the (offset)
/// space: i.e., `range.lo < range.hi`.
#[derive(Clone, Copy, PartialEq, Eq)]
pub struct IntRange {
pub lo: MaybeInfiniteInt, // Must not be `PosInfinity`.
pub hi: MaybeInfiniteInt, // Must not be `NegInfinity`.
}
impl IntRange {
/// Best effort; will not know that e.g. `255u8..` is a singleton.
pub(crate) fn is_singleton(&self) -> bool {
// Since `lo` and `hi` can't be the same `Infinity` and `plus_one` never changes from finite
// to infinite, this correctly only detects ranges that contain exacly one `Finite(x)`.
self.lo.plus_one() == self.hi
}
#[inline]
pub fn from_singleton(x: MaybeInfiniteInt) -> IntRange {
IntRange { lo: x, hi: x.plus_one() }
}
#[inline]
pub fn from_range(lo: MaybeInfiniteInt, mut hi: MaybeInfiniteInt, end: RangeEnd) -> IntRange {
if end == RangeEnd::Included {
hi = hi.plus_one();
}
if lo >= hi {
// This should have been caught earlier by E0030.
bug!("malformed range pattern: {lo:?}..{hi:?}");
}
IntRange { lo, hi }
}
fn is_subrange(&self, other: &Self) -> bool {
other.lo <= self.lo && self.hi <= other.hi
}
fn intersection(&self, other: &Self) -> Option<Self> {
if self.lo < other.hi && other.lo < self.hi {
Some(IntRange { lo: max(self.lo, other.lo), hi: min(self.hi, other.hi) })
} else {
None
}
}
/// Partition a range of integers into disjoint subranges. This does constructor splitting for
/// integer ranges as explained at the top of the file.
///
/// This returns an output that covers `self`. The output is split so that the only
/// intersections between an output range and a column range are inclusions. No output range
/// straddles the boundary of one of the inputs.
///
/// Additionally, we track for each output range whether it is covered by one of the column ranges or not.
///
/// The following input:
/// ```text
/// (--------------------------) // `self`
/// (------) (----------) (-)
/// (------) (--------)
/// ```
/// is first intersected with `self`:
/// ```text
/// (--------------------------) // `self`
/// (----) (----------) (-)
/// (------) (--------)
/// ```
/// and then iterated over as follows:
/// ```text
/// (-(--)-(-)-(------)-)--(-)-
/// ```
/// where each sequence of dashes is an output range, and dashes outside parentheses are marked
/// as `Presence::Missing`.
///
/// ## `isize`/`usize`
///
/// Whereas a wildcard of type `i32` stands for the range `i32::MIN..=i32::MAX`, a `usize`
/// wildcard stands for `0..PosInfinity` and a `isize` wildcard stands for
/// `NegInfinity..PosInfinity`. In other words, as far as `IntRange` is concerned, there are
/// values before `isize::MIN` and after `usize::MAX`/`isize::MAX`.
/// This is to avoid e.g. `0..(u32::MAX as usize)` from being exhaustive on one architecture and
/// not others. This was decided in <https://github.com/rust-lang/rfcs/pull/2591>.
///
/// These infinities affect splitting subtly: it is possible to get `NegInfinity..0` and
/// `usize::MAX+1..PosInfinity` in the output. Diagnostics must be careful to handle these
/// fictitious ranges sensibly.
fn split(
&self,
column_ranges: impl Iterator<Item = IntRange>,
) -> impl Iterator<Item = (Presence, IntRange)> {
// The boundaries of ranges in `column_ranges` intersected with `self`.
// We do parenthesis matching for input ranges. A boundary counts as +1 if it starts
// a range and -1 if it ends it. When the count is > 0 between two boundaries, we
// are within an input range.
let mut boundaries: Vec<(MaybeInfiniteInt, isize)> = column_ranges
.filter_map(|r| self.intersection(&r))
.flat_map(|r| [(r.lo, 1), (r.hi, -1)])
.collect();
// We sort by boundary, and for each boundary we sort the "closing parentheses" first. The
// order of +1/-1 for a same boundary value is actually irrelevant, because we only look at
// the accumulated count between distinct boundary values.
boundaries.sort_unstable();
// Accumulate parenthesis counts.
let mut paren_counter = 0isize;
// Gather pairs of adjacent boundaries.
let mut prev_bdy = self.lo;
boundaries
.into_iter()
// End with the end of the range. The count is ignored.
.chain(once((self.hi, 0)))
// List pairs of adjacent boundaries and the count between them.
.map(move |(bdy, delta)| {
// `delta` affects the count as we cross `bdy`, so the relevant count between
// `prev_bdy` and `bdy` is untouched by `delta`.
let ret = (prev_bdy, paren_counter, bdy);
prev_bdy = bdy;
paren_counter += delta;
ret
})
// Skip empty ranges.
.filter(|&(prev_bdy, _, bdy)| prev_bdy != bdy)
// Convert back to ranges.
.map(move |(prev_bdy, paren_count, bdy)| {
use Presence::*;
let presence = if paren_count > 0 { Seen } else { Unseen };
let range = IntRange { lo: prev_bdy, hi: bdy };
(presence, range)
})
}
}
/// Note: this will render signed ranges incorrectly. To render properly, convert to a pattern
/// first.
impl fmt::Debug for IntRange {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
if let Finite(lo) = self.lo {
write!(f, "{lo}")?;
}
write!(f, "{}", RangeEnd::Excluded)?;
if let Finite(hi) = self.hi {
write!(f, "{hi}")?;
}
Ok(())
}
}
#[derive(Copy, Clone, Debug, PartialEq, Eq)]
pub enum SliceKind {
/// Patterns of length `n` (`[x, y]`).
FixedLen(usize),
/// Patterns using the `..` notation (`[x, .., y]`).
/// Captures any array constructor of `length >= i + j`.
/// In the case where `array_len` is `Some(_)`,
/// this indicates that we only care about the first `i` and the last `j` values of the array,
/// and everything in between is a wildcard `_`.
VarLen(usize, usize),
}
impl SliceKind {
fn arity(self) -> usize {
match self {
FixedLen(length) => length,
VarLen(prefix, suffix) => prefix + suffix,
}
}
/// Whether this pattern includes patterns of length `other_len`.
fn covers_length(self, other_len: usize) -> bool {
match self {
FixedLen(len) => len == other_len,
VarLen(prefix, suffix) => prefix + suffix <= other_len,
}
}
}
/// A constructor for array and slice patterns.
#[derive(Copy, Clone, Debug, PartialEq, Eq)]
pub struct Slice {
/// `None` if the matched value is a slice, `Some(n)` if it is an array of size `n`.
pub(crate) array_len: Option<usize>,
/// The kind of pattern it is: fixed-length `[x, y]` or variable length `[x, .., y]`.
pub(crate) kind: SliceKind,
}
impl Slice {
pub fn new(array_len: Option<usize>, kind: SliceKind) -> Self {
let kind = match (array_len, kind) {
// If the middle `..` has length 0, we effectively have a fixed-length pattern.
(Some(len), VarLen(prefix, suffix)) if prefix + suffix == len => FixedLen(len),
(Some(len), VarLen(prefix, suffix)) if prefix + suffix > len => bug!(
"Slice pattern of length {} longer than its array length {len}",
prefix + suffix
),
_ => kind,
};
Slice { array_len, kind }
}
pub(crate) fn arity(self) -> usize {
self.kind.arity()
}
/// See `Constructor::is_covered_by`
fn is_covered_by(self, other: Self) -> bool {
other.kind.covers_length(self.arity())
}
/// This computes constructor splitting for variable-length slices, as explained at the top of
/// the file.
///
/// A slice pattern `[x, .., y]` behaves like the infinite or-pattern `[x, y] | [x, _, y] | [x,
/// _, _, y] | etc`. The corresponding value constructors are fixed-length array constructors of
/// corresponding lengths. We obviously can't list this infinitude of constructors.
/// Thankfully, it turns out that for each finite set of slice patterns, all sufficiently large
/// array lengths are equivalent.
///
/// Let's look at an example, where we are trying to split the last pattern:
/// ```
/// # fn foo(x: &[bool]) {
/// match x {
/// [true, true, ..] => {}
/// [.., false, false] => {}
/// [..] => {}
/// }
/// # }
/// ```
/// Here are the results of specialization for the first few lengths:
/// ```
/// # fn foo(x: &[bool]) { match x {
/// // length 0
/// [] => {}
/// // length 1
/// [_] => {}
/// // length 2
/// [true, true] => {}
/// [false, false] => {}
/// [_, _] => {}
/// // length 3
/// [true, true, _ ] => {}
/// [_, false, false] => {}
/// [_, _, _ ] => {}
/// // length 4
/// [true, true, _, _ ] => {}
/// [_, _, false, false] => {}
/// [_, _, _, _ ] => {}
/// // length 5
/// [true, true, _, _, _ ] => {}
/// [_, _, _, false, false] => {}
/// [_, _, _, _, _ ] => {}
/// # _ => {}
/// # }}
/// ```
///
/// We see that above length 4, we are simply inserting columns full of wildcards in the middle.
/// This means that specialization and witness computation with slices of length `l >= 4` will
/// give equivalent results regardless of `l`. This applies to any set of slice patterns: there
/// will be a length `L` above which all lengths behave the same. This is exactly what we need
/// for constructor splitting.
///
/// A variable-length slice pattern covers all lengths from its arity up to infinity. As we just
/// saw, we can split this in two: lengths below `L` are treated individually with a
/// fixed-length slice each; lengths above `L` are grouped into a single variable-length slice
/// constructor.
///
/// For each variable-length slice pattern `p` with a prefix of length `plₚ` and suffix of
/// length `slₚ`, only the first `plₚ` and the last `slₚ` elements are examined. Therefore, as
/// long as `L` is positive (to avoid concerns about empty types), all elements after the
/// maximum prefix length and before the maximum suffix length are not examined by any
/// variable-length pattern, and therefore can be ignored. This gives us a way to compute `L`.
///
/// Additionally, if fixed-length patterns exist, we must pick an `L` large enough to miss them,
/// so we can pick `L = max(max(FIXED_LEN)+1, max(PREFIX_LEN) + max(SUFFIX_LEN))`.
/// `max_slice` below will be made to have this arity `L`.
///
/// If `self` is fixed-length, it is returned as-is.
///
/// Additionally, we track for each output slice whether it is covered by one of the column slices or not.
fn split(
self,
column_slices: impl Iterator<Item = Slice>,
) -> impl Iterator<Item = (Presence, Slice)> {
// Range of lengths below `L`.
let smaller_lengths;
let arity = self.arity();
let mut max_slice = self.kind;
// Tracks the smallest variable-length slice we've seen. Any slice arity above it is
// therefore `Presence::Seen` in the column.
let mut min_var_len = usize::MAX;
// Tracks the fixed-length slices we've seen, to mark them as `Presence::Seen`.
let mut seen_fixed_lens = FxHashSet::default();
match &mut max_slice {
VarLen(max_prefix_len, max_suffix_len) => {
// A length larger than any fixed-length slice encountered.
// We start at 1 in case the subtype is empty because in that case the zero-length
// slice must be treated separately from the rest.
let mut fixed_len_upper_bound = 1;
// We grow `max_slice` to be larger than all slices encountered, as described above.
// `L` is `max_slice.arity()`. For diagnostics, we keep the prefix and suffix
// lengths separate.
for slice in column_slices {
match slice.kind {
FixedLen(len) => {
fixed_len_upper_bound = cmp::max(fixed_len_upper_bound, len + 1);
seen_fixed_lens.insert(len);
}
VarLen(prefix, suffix) => {
*max_prefix_len = cmp::max(*max_prefix_len, prefix);
*max_suffix_len = cmp::max(*max_suffix_len, suffix);
min_var_len = cmp::min(min_var_len, prefix + suffix);
}
}
}
// If `fixed_len_upper_bound >= L`, we set `L` to `fixed_len_upper_bound`.
if let Some(delta) =
fixed_len_upper_bound.checked_sub(*max_prefix_len + *max_suffix_len)
{
*max_prefix_len += delta
}
// We cap the arity of `max_slice` at the array size.
match self.array_len {
Some(len) if max_slice.arity() >= len => max_slice = FixedLen(len),
_ => {}
}
smaller_lengths = match self.array_len {
// The only admissible fixed-length slice is one of the array size. Whether `max_slice`
// is fixed-length or variable-length, it will be the only relevant slice to output
// here.
Some(_) => 0..0, // empty range
// We need to cover all arities in the range `(arity..infinity)`. We split that
// range into two: lengths smaller than `max_slice.arity()` are treated
// independently as fixed-lengths slices, and lengths above are captured by
// `max_slice`.
None => self.arity()..max_slice.arity(),
};
}
FixedLen(_) => {
// No need to split here. We only track presence.
for slice in column_slices {
match slice.kind {
FixedLen(len) => {
if len == arity {
seen_fixed_lens.insert(len);
}
}
VarLen(prefix, suffix) => {
min_var_len = cmp::min(min_var_len, prefix + suffix);
}
}
}
smaller_lengths = 0..0;
}
};
smaller_lengths.map(FixedLen).chain(once(max_slice)).map(move |kind| {
let arity = kind.arity();
let seen = if min_var_len <= arity || seen_fixed_lens.contains(&arity) {
Presence::Seen
} else {
Presence::Unseen
};
(seen, Slice::new(self.array_len, kind))
})
}
}
/// A globally unique id to distinguish `Opaque` patterns.
#[derive(Clone, Debug, PartialEq, Eq)]
pub struct OpaqueId(u32);
impl OpaqueId {
pub fn new() -> Self {
use std::sync::atomic::{AtomicU32, Ordering};
static OPAQUE_ID: AtomicU32 = AtomicU32::new(0);
OpaqueId(OPAQUE_ID.fetch_add(1, Ordering::SeqCst))
}
}
/// A value can be decomposed into a constructor applied to some fields. This struct represents
/// the constructor. See also `Fields`.
///
/// `pat_constructor` retrieves the constructor corresponding to a pattern.
/// `specialize_constructor` returns the list of fields corresponding to a pattern, given a
/// constructor. `Constructor::apply` reconstructs the pattern from a pair of `Constructor` and
/// `Fields`.
#[derive(Clone, Debug, PartialEq)]
pub enum Constructor<'tcx> {
/// The constructor for patterns that have a single constructor, like tuples, struct patterns,
/// and references. Fixed-length arrays are treated separately with `Slice`.
Single,
/// Enum variants.
Variant(VariantIdx),
/// Booleans
Bool(bool),
/// Ranges of integer literal values (`2`, `2..=5` or `2..5`).
IntRange(IntRange),
/// Ranges of floating-point literal values (`2.0..=5.2`).
F32Range(IeeeFloat<SingleS>, IeeeFloat<SingleS>, RangeEnd),
F64Range(IeeeFloat<DoubleS>, IeeeFloat<DoubleS>, RangeEnd),
/// String literals. Strings are not quite the same as `&[u8]` so we treat them separately.
Str(Const<'tcx>),
/// Array and slice patterns.
Slice(Slice),
/// Constants that must not be matched structurally. They are treated as black boxes for the
/// purposes of exhaustiveness: we must not inspect them, and they don't count towards making a
/// match exhaustive.
/// Carries an id that must be unique within a match. We need this to ensure the invariants of
/// [`SplitConstructorSet`].
Opaque(OpaqueId),
/// Or-pattern.
Or,
/// Wildcard pattern.
Wildcard,
/// Fake extra constructor for enums that aren't allowed to be matched exhaustively. Also used
/// for those types for which we cannot list constructors explicitly, like `f64` and `str`.
NonExhaustive,
/// Fake extra constructor for variants that should not be mentioned in diagnostics.
/// We use this for variants behind an unstable gate as well as
/// `#[doc(hidden)]` ones.
Hidden,
/// Fake extra constructor for constructors that are not seen in the matrix, as explained at the
/// top of the file.
Missing,
}
impl<'tcx> Constructor<'tcx> {
pub(crate) fn is_non_exhaustive(&self) -> bool {
matches!(self, NonExhaustive)
}
pub(crate) fn as_variant(&self) -> Option<VariantIdx> {
match self {
Variant(i) => Some(*i),
_ => None,
}
}
fn as_bool(&self) -> Option<bool> {
match self {
Bool(b) => Some(*b),
_ => None,
}
}
pub(crate) fn as_int_range(&self) -> Option<&IntRange> {
match self {
IntRange(range) => Some(range),
_ => None,
}
}
fn as_slice(&self) -> Option<Slice> {
match self {
Slice(slice) => Some(*slice),
_ => None,
}
}
/// The number of fields for this constructor. This must be kept in sync with
/// `Fields::wildcards`.
pub(crate) fn arity(&self, pcx: &PatCtxt<'_, '_, 'tcx>) -> usize {
pcx.cx.ctor_arity(self, pcx.ty)
}
/// Returns whether `self` is covered by `other`, i.e. whether `self` is a subset of `other`.
/// For the simple cases, this is simply checking for equality. For the "grouped" constructors,
/// this checks for inclusion.
// We inline because this has a single call site in `Matrix::specialize_constructor`.
#[inline]
pub(crate) fn is_covered_by<'p>(&self, pcx: &PatCtxt<'_, 'p, 'tcx>, other: &Self) -> bool {
match (self, other) {
(Wildcard, _) => {
span_bug!(
pcx.cx.scrut_span,
"Constructor splitting should not have returned `Wildcard`"
)
}
// Wildcards cover anything
(_, Wildcard) => true,
// Only a wildcard pattern can match these special constructors.
(Missing { .. } | NonExhaustive | Hidden, _) => false,
(Single, Single) => true,
(Variant(self_id), Variant(other_id)) => self_id == other_id,
(Bool(self_b), Bool(other_b)) => self_b == other_b,
(IntRange(self_range), IntRange(other_range)) => self_range.is_subrange(other_range),
(F32Range(self_from, self_to, self_end), F32Range(other_from, other_to, other_end)) => {
self_from.ge(other_from)
&& match self_to.partial_cmp(other_to) {
Some(Ordering::Less) => true,
Some(Ordering::Equal) => other_end == self_end,
_ => false,
}
}
(F64Range(self_from, self_to, self_end), F64Range(other_from, other_to, other_end)) => {
self_from.ge(other_from)
&& match self_to.partial_cmp(other_to) {
Some(Ordering::Less) => true,
Some(Ordering::Equal) => other_end == self_end,
_ => false,
}
}
(Str(self_val), Str(other_val)) => {
// FIXME Once valtrees are available we can directly use the bytes
// in the `Str` variant of the valtree for the comparison here.
self_val == other_val
}
(Slice(self_slice), Slice(other_slice)) => self_slice.is_covered_by(*other_slice),
// Opaque constructors don't interact with anything unless they come from the
// syntactically identical pattern.
(Opaque(self_id), Opaque(other_id)) => self_id == other_id,
(Opaque(..), _) | (_, Opaque(..)) => false,
_ => span_bug!(
pcx.cx.scrut_span,
"trying to compare incompatible constructors {:?} and {:?}",
self,
other
),
}
}
}
#[derive(Debug, Clone, Copy)]
pub enum VariantVisibility {
/// Variant that doesn't fit the other cases, i.e. most variants.
Visible,
/// Variant behind an unstable gate or with the `#[doc(hidden)]` attribute. It will not be
/// mentioned in diagnostics unless the user mentioned it first.
Hidden,
/// Variant that matches no value. E.g. `Some::<Option<!>>` if the `exhaustive_patterns` feature
/// is enabled. Like `Hidden`, it will not be mentioned in diagnostics unless the user mentioned
/// it first.
Empty,
}
/// Describes the set of all constructors for a type. For details, in particular about the emptiness
/// of constructors, see the top of the file.
///
/// In terms of division of responsibility, [`ConstructorSet::split`] handles all of the
/// `exhaustive_patterns` feature.
#[derive(Debug)]
pub enum ConstructorSet {
/// The type has a single constructor, e.g. `&T` or a struct. `empty` tracks whether the
/// constructor is empty.
Single { empty: bool },
/// This type has the following list of constructors. If `variants` is empty and
/// `non_exhaustive` is false, don't use this; use `NoConstructors` instead.
Variants { variants: IndexVec<VariantIdx, VariantVisibility>, non_exhaustive: bool },
/// Booleans.
Bool,
/// The type is spanned by integer values. The range or ranges give the set of allowed values.
/// The second range is only useful for `char`.
Integers { range_1: IntRange, range_2: Option<IntRange> },
/// The type is matched by slices. `array_len` is the compile-time length of the array, if
/// known. If `subtype_is_empty`, all constructors are empty except possibly the zero-length
/// slice `[]`.
Slice { array_len: Option<usize>, subtype_is_empty: bool },
/// The constructors cannot be listed, and the type cannot be matched exhaustively. E.g. `str`,
/// floats.
Unlistable,
/// The type has no constructors (not even empty ones). This is `!` and empty enums.
NoConstructors,
}
/// Describes the result of analyzing the constructors in a column of a match.
///
/// `present` is morally the set of constructors present in the column, and `missing` is the set of
/// constructors that exist in the type but are not present in the column.
///
/// More formally, if we discard wildcards from the column, this respects the following constraints:
/// 1. the union of `present`, `missing` and `missing_empty` covers all the constructors of the type
/// 2. each constructor in `present` is covered by something in the column
/// 3. no constructor in `missing` or `missing_empty` is covered by anything in the column
/// 4. each constructor in the column is equal to the union of one or more constructors in `present`
/// 5. `missing` does not contain empty constructors (see discussion about emptiness at the top of
/// the file);
/// 6. `missing_empty` contains only empty constructors
/// 7. constructors in `present`, `missing` and `missing_empty` are split for the column; in other
/// words, they are either fully included in or fully disjoint from each constructor in the
/// column. In yet other words, there are no non-trivial intersections like between `0..10` and
/// `5..15`.
///
/// We must be particularly careful with weird constructors like `Opaque`: they're not formally part
/// of the `ConstructorSet` for the type, yet if we forgot to include them in `present` we would be
/// ignoring any row with `Opaque`s in the algorithm. Hence the importance of point 4.
#[derive(Debug)]
pub(crate) struct SplitConstructorSet<'tcx> {
pub(crate) present: SmallVec<[Constructor<'tcx>; 1]>,
pub(crate) missing: Vec<Constructor<'tcx>>,
pub(crate) missing_empty: Vec<Constructor<'tcx>>,
}
impl ConstructorSet {
/// This analyzes a column of constructors to 1/ determine which constructors of the type (if
/// any) are missing; 2/ split constructors to handle non-trivial intersections e.g. on ranges
/// or slices. This can get subtle; see [`SplitConstructorSet`] for details of this operation
/// and its invariants.
#[instrument(level = "debug", skip(self, pcx, ctors), ret)]
pub(crate) fn split<'a, 'tcx>(
&self,
pcx: &PatCtxt<'_, '_, 'tcx>,
ctors: impl Iterator<Item = &'a Constructor<'tcx>> + Clone,
) -> SplitConstructorSet<'tcx>
where
'tcx: 'a,
{
let mut present: SmallVec<[_; 1]> = SmallVec::new();
// Empty constructors found missing.
let mut missing_empty = Vec::new();
// Nonempty constructors found missing.
let mut missing = Vec::new();
// Constructors in `ctors`, except wildcards and opaques.
let mut seen = Vec::new();
for ctor in ctors.cloned() {
match ctor {
Opaque(..) => present.push(ctor),
Wildcard => {} // discard wildcards
_ => seen.push(ctor),
}
}
match self {
ConstructorSet::Single { empty } => {
if !seen.is_empty() {
present.push(Single);
} else if *empty {
missing_empty.push(Single);
} else {
missing.push(Single);
}
}
ConstructorSet::Variants { variants, non_exhaustive } => {
let seen_set: FxHashSet<_> = seen.iter().map(|c| c.as_variant().unwrap()).collect();
let mut skipped_a_hidden_variant = false;
for (idx, visibility) in variants.iter_enumerated() {
let ctor = Variant(idx);
if seen_set.contains(&idx) {
present.push(ctor);
} else {
// We only put visible variants directly into `missing`.
match visibility {
VariantVisibility::Visible => missing.push(ctor),
VariantVisibility::Hidden => skipped_a_hidden_variant = true,
VariantVisibility::Empty => missing_empty.push(ctor),
}
}
}
if skipped_a_hidden_variant {
missing.push(Hidden);
}
if *non_exhaustive {
missing.push(NonExhaustive);
}
}
ConstructorSet::Bool => {
let mut seen_false = false;
let mut seen_true = false;
for b in seen.iter().map(|ctor| ctor.as_bool().unwrap()) {
if b {
seen_true = true;
} else {
seen_false = true;
}
}
if seen_false {
present.push(Bool(false));
} else {
missing.push(Bool(false));
}
if seen_true {
present.push(Bool(true));
} else {
missing.push(Bool(true));
}
}
ConstructorSet::Integers { range_1, range_2 } => {
let seen_ranges: Vec<_> =
seen.iter().map(|ctor| ctor.as_int_range().unwrap().clone()).collect();
for (seen, splitted_range) in range_1.split(seen_ranges.iter().cloned()) {
match seen {
Presence::Unseen => missing.push(IntRange(splitted_range)),
Presence::Seen => present.push(IntRange(splitted_range)),
}
}
if let Some(range_2) = range_2 {
for (seen, splitted_range) in range_2.split(seen_ranges.into_iter()) {
match seen {
Presence::Unseen => missing.push(IntRange(splitted_range)),
Presence::Seen => present.push(IntRange(splitted_range)),
}
}
}
}
ConstructorSet::Slice { array_len, subtype_is_empty } => {
let seen_slices = seen.iter().map(|c| c.as_slice().unwrap());
let base_slice = Slice::new(*array_len, VarLen(0, 0));
for (seen, splitted_slice) in base_slice.split(seen_slices) {
let ctor = Slice(splitted_slice);
match seen {
Presence::Seen => present.push(ctor),
Presence::Unseen => {
if *subtype_is_empty && splitted_slice.arity() != 0 {
// We have subpatterns of an empty type, so the constructor is
// empty.
missing_empty.push(ctor);
} else {
missing.push(ctor);
}
}
}
}
}
ConstructorSet::Unlistable => {
// Since we can't list constructors, we take the ones in the column. This might list
// some constructors several times but there's not much we can do.
present.extend(seen);
missing.push(NonExhaustive);
}
ConstructorSet::NoConstructors => {
// In a `MaybeInvalid` place even an empty pattern may be reachable. We therefore
// add a dummy empty constructor here, which will be ignored if the place is
// `ValidOnly`.
missing_empty.push(NonExhaustive);
}
}
// We have now grouped all the constructors into 3 buckets: present, missing, missing_empty.
// In the absence of the `exhaustive_patterns` feature however, we don't count nested empty
// types as empty. Only non-nested `!` or `enum Foo {}` are considered empty.
if !pcx.cx.tcx.features().exhaustive_patterns
&& !(pcx.is_top_level && matches!(self, Self::NoConstructors))
{
// Treat all missing constructors as nonempty.
missing.extend(missing_empty.drain(..));
}
SplitConstructorSet { present, missing, missing_empty }
}
}

856
compiler/rustc_pattern_analysis/src/cx.rs Normal file
View File

@ -0,0 +1,856 @@
use std::fmt;
use std::iter::once;
use rustc_arena::TypedArena;
use rustc_data_structures::captures::Captures;
use rustc_hir::def_id::DefId;
use rustc_hir::{HirId, RangeEnd};
use rustc_index::Idx;
use rustc_index::IndexVec;
use rustc_middle::middle::stability::EvalResult;
use rustc_middle::mir;
use rustc_middle::mir::interpret::Scalar;
use rustc_middle::thir::{FieldPat, Pat, PatKind, PatRange, PatRangeBoundary};
use rustc_middle::ty::layout::IntegerExt;
use rustc_middle::ty::{self, Ty, TyCtxt, VariantDef};
use rustc_span::{Span, DUMMY_SP};
use rustc_target::abi::{FieldIdx, Integer, VariantIdx, FIRST_VARIANT};
use smallvec::SmallVec;
use crate::constructor::{
Constructor, ConstructorSet, IntRange, MaybeInfiniteInt, OpaqueId, Slice, SliceKind,
VariantVisibility,
};
use crate::pat::{DeconstructedPat, WitnessPat};
use Constructor::*;
pub struct MatchCheckCtxt<'p, 'tcx> {
pub tcx: TyCtxt<'tcx>,
/// The module in which the match occurs. This is necessary for
/// checking inhabited-ness of types because whether a type is (visibly)
/// inhabited can depend on whether it was defined in the current module or
/// not. E.g., `struct Foo { _private: ! }` cannot be seen to be empty
/// outside its module and should not be matchable with an empty match statement.
pub module: DefId,
pub param_env: ty::ParamEnv<'tcx>,
pub pattern_arena: &'p TypedArena<DeconstructedPat<'p, 'tcx>>,
/// Lint level at the match.
pub match_lint_level: HirId,
/// The span of the whole match, if applicable.
pub whole_match_span: Option<Span>,
/// Span of the scrutinee.
pub scrut_span: Span,
/// Only produce `NON_EXHAUSTIVE_OMITTED_PATTERNS` lint on refutable patterns.
pub refutable: bool,
/// Whether the data at the scrutinee is known to be valid. This is false if the scrutinee comes
/// from a union field, a pointer deref, or a reference deref (pending opsem decisions).
pub known_valid_scrutinee: bool,
}
impl<'p, 'tcx> MatchCheckCtxt<'p, 'tcx> {
pub(super) fn is_uninhabited(&self, ty: Ty<'tcx>) -> bool {
!ty.is_inhabited_from(self.tcx, self.module, self.param_env)
}
/// Returns whether the given type is an enum from another crate declared `#[non_exhaustive]`.
pub fn is_foreign_non_exhaustive_enum(&self, ty: Ty<'tcx>) -> bool {
match ty.kind() {
ty::Adt(def, ..) => {
def.is_enum() && def.is_variant_list_non_exhaustive() && !def.did().is_local()
}
_ => false,
}
}
pub(crate) fn alloc_wildcard_slice(
&self,
tys: impl IntoIterator<Item = Ty<'tcx>>,
) -> &'p [DeconstructedPat<'p, 'tcx>] {
self.pattern_arena
.alloc_from_iter(tys.into_iter().map(|ty| DeconstructedPat::wildcard(ty, DUMMY_SP)))
}
// In the cases of either a `#[non_exhaustive]` field list or a non-public field, we hide
// uninhabited fields in order not to reveal the uninhabitedness of the whole variant.
// This lists the fields we keep along with their types.
pub(crate) fn list_variant_nonhidden_fields<'a>(
&'a self,
ty: Ty<'tcx>,
variant: &'a VariantDef,
) -> impl Iterator<Item = (FieldIdx, Ty<'tcx>)> + Captures<'p> + Captures<'a> {
let cx = self;
let ty::Adt(adt, args) = ty.kind() else { bug!() };
// Whether we must not match the fields of this variant exhaustively.
let is_non_exhaustive = variant.is_field_list_non_exhaustive() && !adt.did().is_local();
variant.fields.iter().enumerate().filter_map(move |(i, field)| {
let ty = field.ty(cx.tcx, args);
// `field.ty()` doesn't normalize after substituting.
let ty = cx.tcx.normalize_erasing_regions(cx.param_env, ty);
let is_visible = adt.is_enum() || field.vis.is_accessible_from(cx.module, cx.tcx);
let is_uninhabited = cx.tcx.features().exhaustive_patterns && cx.is_uninhabited(ty);
if is_uninhabited && (!is_visible || is_non_exhaustive) {
None
} else {
Some((FieldIdx::new(i), ty))
}
})
}
pub(crate) fn variant_index_for_adt(
ctor: &Constructor<'tcx>,
adt: ty::AdtDef<'tcx>,
) -> VariantIdx {
match *ctor {
Variant(idx) => idx,
Single => {
assert!(!adt.is_enum());
FIRST_VARIANT
}
_ => bug!("bad constructor {:?} for adt {:?}", ctor, adt),
}
}
/// Creates a new list of wildcard fields for a given constructor. The result must have a length
/// of `ctor.arity()`.
#[instrument(level = "trace", skip(self))]
pub(crate) fn ctor_wildcard_fields(
&self,
ctor: &Constructor<'tcx>,
ty: Ty<'tcx>,
) -> &'p [DeconstructedPat<'p, 'tcx>] {
let cx = self;
match ctor {
Single | Variant(_) => match ty.kind() {
ty::Tuple(fs) => cx.alloc_wildcard_slice(fs.iter()),
ty::Ref(_, rty, _) => cx.alloc_wildcard_slice(once(*rty)),
ty::Adt(adt, args) => {
if adt.is_box() {
// The only legal patterns of type `Box` (outside `std`) are `_` and box
// patterns. If we're here we can assume this is a box pattern.
cx.alloc_wildcard_slice(once(args.type_at(0)))
} else {
let variant =
&adt.variant(MatchCheckCtxt::variant_index_for_adt(&ctor, *adt));
let tys = cx.list_variant_nonhidden_fields(ty, variant).map(|(_, ty)| ty);
cx.alloc_wildcard_slice(tys)
}
}
_ => bug!("Unexpected type for `Single` constructor: {:?}", ty),
},
Slice(slice) => match *ty.kind() {
ty::Slice(ty) | ty::Array(ty, _) => {
let arity = slice.arity();
cx.alloc_wildcard_slice((0..arity).map(|_| ty))
}
_ => bug!("bad slice pattern {:?} {:?}", ctor, ty),
},
Bool(..)
| IntRange(..)
| F32Range(..)
| F64Range(..)
| Str(..)
| Opaque(..)
| NonExhaustive
| Hidden
| Missing { .. }
| Wildcard => &[],
Or => {
bug!("called `Fields::wildcards` on an `Or` ctor")
}
}
}
/// The number of fields for this constructor. This must be kept in sync with
/// `Fields::wildcards`.
pub(crate) fn ctor_arity(&self, ctor: &Constructor<'tcx>, ty: Ty<'tcx>) -> usize {
match ctor {
Single | Variant(_) => match ty.kind() {
ty::Tuple(fs) => fs.len(),
ty::Ref(..) => 1,
ty::Adt(adt, ..) => {
if adt.is_box() {
// The only legal patterns of type `Box` (outside `std`) are `_` and box
// patterns. If we're here we can assume this is a box pattern.
1
} else {
let variant =
&adt.variant(MatchCheckCtxt::variant_index_for_adt(&ctor, *adt));
self.list_variant_nonhidden_fields(ty, variant).count()
}
}
_ => bug!("Unexpected type for `Single` constructor: {:?}", ty),
},
Slice(slice) => slice.arity(),
Bool(..)
| IntRange(..)
| F32Range(..)
| F64Range(..)
| Str(..)
| Opaque(..)
| NonExhaustive
| Hidden
| Missing { .. }
| Wildcard => 0,
Or => bug!("The `Or` constructor doesn't have a fixed arity"),
}
}
/// Creates a set that represents all the constructors of `ty`.
///
/// See [`crate::constructor`] for considerations of emptiness.
#[instrument(level = "debug", skip(self), ret)]
pub fn ctors_for_ty(&self, ty: Ty<'tcx>) -> ConstructorSet {
let cx = self;
let make_uint_range = |start, end| {
IntRange::from_range(
MaybeInfiniteInt::new_finite_uint(start),
MaybeInfiniteInt::new_finite_uint(end),
RangeEnd::Included,
)
};
// This determines the set of all possible constructors for the type `ty`. For numbers,
// arrays and slices we use ranges and variable-length slices when appropriate.
match ty.kind() {
ty::Bool => ConstructorSet::Bool,
ty::Char => {
// The valid Unicode Scalar Value ranges.
ConstructorSet::Integers {
range_1: make_uint_range('\u{0000}' as u128, '\u{D7FF}' as u128),
range_2: Some(make_uint_range('\u{E000}' as u128, '\u{10FFFF}' as u128)),
}
}
&ty::Int(ity) => {
let range = if ty.is_ptr_sized_integral() {
// The min/max values of `isize` are not allowed to be observed.
IntRange {
lo: MaybeInfiniteInt::NegInfinity,
hi: MaybeInfiniteInt::PosInfinity,
}
} else {
let size = Integer::from_int_ty(&cx.tcx, ity).size().bits();
let min = 1u128 << (size - 1);
let max = min - 1;
let min = MaybeInfiniteInt::new_finite_int(min, size);
let max = MaybeInfiniteInt::new_finite_int(max, size);
IntRange::from_range(min, max, RangeEnd::Included)
};
ConstructorSet::Integers { range_1: range, range_2: None }
}
&ty::Uint(uty) => {
let range = if ty.is_ptr_sized_integral() {
// The max value of `usize` is not allowed to be observed.
let lo = MaybeInfiniteInt::new_finite_uint(0);
IntRange { lo, hi: MaybeInfiniteInt::PosInfinity }
} else {
let size = Integer::from_uint_ty(&cx.tcx, uty).size();
let max = size.truncate(u128::MAX);
make_uint_range(0, max)
};
ConstructorSet::Integers { range_1: range, range_2: None }
}
ty::Slice(sub_ty) => ConstructorSet::Slice {
array_len: None,
subtype_is_empty: cx.is_uninhabited(*sub_ty),
},
ty::Array(sub_ty, len) => {
// We treat arrays of a constant but unknown length like slices.
ConstructorSet::Slice {
array_len: len.try_eval_target_usize(cx.tcx, cx.param_env).map(|l| l as usize),
subtype_is_empty: cx.is_uninhabited(*sub_ty),
}
}
ty::Adt(def, args) if def.is_enum() => {
let is_declared_nonexhaustive = cx.is_foreign_non_exhaustive_enum(ty);
if def.variants().is_empty() && !is_declared_nonexhaustive {
ConstructorSet::NoConstructors
} else {
let mut variants =
IndexVec::from_elem(VariantVisibility::Visible, def.variants());
for (idx, v) in def.variants().iter_enumerated() {
let variant_def_id = def.variant(idx).def_id;
// Visibly uninhabited variants.
let is_inhabited = v
.inhabited_predicate(cx.tcx, *def)
.instantiate(cx.tcx, args)
.apply(cx.tcx, cx.param_env, cx.module);
// Variants that depend on a disabled unstable feature.
let is_unstable = matches!(
cx.tcx.eval_stability(variant_def_id, None, DUMMY_SP, None),
EvalResult::Deny { .. }
);
// Foreign `#[doc(hidden)]` variants.
let is_doc_hidden =
cx.tcx.is_doc_hidden(variant_def_id) && !variant_def_id.is_local();
let visibility = if !is_inhabited {
// FIXME: handle empty+hidden
VariantVisibility::Empty
} else if is_unstable || is_doc_hidden {
VariantVisibility::Hidden
} else {
VariantVisibility::Visible
};
variants[idx] = visibility;
}
ConstructorSet::Variants { variants, non_exhaustive: is_declared_nonexhaustive }
}
}
ty::Adt(..) | ty::Tuple(..) | ty::Ref(..) => {
ConstructorSet::Single { empty: cx.is_uninhabited(ty) }
}
ty::Never => ConstructorSet::NoConstructors,
// This type is one for which we cannot list constructors, like `str` or `f64`.
// FIXME(Nadrieril): which of these are actually allowed?
ty::Float(_)
| ty::Str
| ty::Foreign(_)
| ty::RawPtr(_)
| ty::FnDef(_, _)
| ty::FnPtr(_)
| ty::Dynamic(_, _, _)
| ty::Closure(_, _)
| ty::Coroutine(_, _, _)
| ty::Alias(_, _)
| ty::Param(_)
| ty::Error(_) => ConstructorSet::Unlistable,
ty::CoroutineWitness(_, _) | ty::Bound(_, _) | ty::Placeholder(_) | ty::Infer(_) => {
bug!("Encountered unexpected type in `ConstructorSet::for_ty`: {ty:?}")
}
}
}
pub(crate) fn lower_pat_range_bdy(
&self,
bdy: PatRangeBoundary<'tcx>,
ty: Ty<'tcx>,
) -> MaybeInfiniteInt {
match bdy {
PatRangeBoundary::NegInfinity => MaybeInfiniteInt::NegInfinity,
PatRangeBoundary::Finite(value) => {
let bits = value.eval_bits(self.tcx, self.param_env);
match *ty.kind() {
ty::Int(ity) => {
let size = Integer::from_int_ty(&self.tcx, ity).size().bits();
MaybeInfiniteInt::new_finite_int(bits, size)
}
_ => MaybeInfiniteInt::new_finite_uint(bits),
}
}
PatRangeBoundary::PosInfinity => MaybeInfiniteInt::PosInfinity,
}
}
/// Note: the input patterns must have been lowered through
/// `rustc_mir_build::thir::pattern::check_match::MatchVisitor::lower_pattern`.
pub fn lower_pat(&self, pat: &Pat<'tcx>) -> DeconstructedPat<'p, 'tcx> {
let singleton = |pat| std::slice::from_ref(self.pattern_arena.alloc(pat));
let cx = self;
let ctor;
let fields: &[_];
match &pat.kind {
PatKind::AscribeUserType { subpattern, .. }
| PatKind::InlineConstant { subpattern, .. } => return self.lower_pat(subpattern),
PatKind::Binding { subpattern: Some(subpat), .. } => return self.lower_pat(subpat),
PatKind::Binding { subpattern: None, .. } | PatKind::Wild => {
ctor = Wildcard;
fields = &[];
}
PatKind::Deref { subpattern } => {
ctor = Single;
fields = singleton(self.lower_pat(subpattern));
}
PatKind::Leaf { subpatterns } | PatKind::Variant { subpatterns, .. } => {
match pat.ty.kind() {
ty::Tuple(fs) => {
ctor = Single;
let mut wilds: SmallVec<[_; 2]> =
fs.iter().map(|ty| DeconstructedPat::wildcard(ty, pat.span)).collect();
for pat in subpatterns {
wilds[pat.field.index()] = self.lower_pat(&pat.pattern);
}
fields = cx.pattern_arena.alloc_from_iter(wilds);
}
ty::Adt(adt, args) if adt.is_box() => {
// The only legal patterns of type `Box` (outside `std`) are `_` and box
// patterns. If we're here we can assume this is a box pattern.
// FIXME(Nadrieril): A `Box` can in theory be matched either with `Box(_,
// _)` or a box pattern. As a hack to avoid an ICE with the former, we
// ignore other fields than the first one. This will trigger an error later
// anyway.
// See https://github.com/rust-lang/rust/issues/82772 ,
// explanation: https://github.com/rust-lang/rust/pull/82789#issuecomment-796921977
// The problem is that we can't know from the type whether we'll match
// normally or through box-patterns. We'll have to figure out a proper
// solution when we introduce generalized deref patterns. Also need to
// prevent mixing of those two options.
let pattern = subpatterns.into_iter().find(|pat| pat.field.index() == 0);
let pat = if let Some(pat) = pattern {
self.lower_pat(&pat.pattern)
} else {
DeconstructedPat::wildcard(args.type_at(0), pat.span)
};
ctor = Single;
fields = singleton(pat);
}
ty::Adt(adt, _) => {
ctor = match pat.kind {
PatKind::Leaf { .. } => Single,
PatKind::Variant { variant_index, .. } => Variant(variant_index),
_ => bug!(),
};
let variant =
&adt.variant(MatchCheckCtxt::variant_index_for_adt(&ctor, *adt));
// For each field in the variant, we store the relevant index into `self.fields` if any.
let mut field_id_to_id: Vec<Option<usize>> =
(0..variant.fields.len()).map(|_| None).collect();
let tys = cx
.list_variant_nonhidden_fields(pat.ty, variant)
.enumerate()
.map(|(i, (field, ty))| {
field_id_to_id[field.index()] = Some(i);
ty
});
let mut wilds: SmallVec<[_; 2]> =
tys.map(|ty| DeconstructedPat::wildcard(ty, pat.span)).collect();
for pat in subpatterns {
if let Some(i) = field_id_to_id[pat.field.index()] {
wilds[i] = self.lower_pat(&pat.pattern);
}
}
fields = cx.pattern_arena.alloc_from_iter(wilds);
}
_ => bug!("pattern has unexpected type: pat: {:?}, ty: {:?}", pat, pat.ty),
}
}
PatKind::Constant { value } => {
match pat.ty.kind() {
ty::Bool => {
ctor = match value.try_eval_bool(cx.tcx, cx.param_env) {
Some(b) => Bool(b),
None => Opaque(OpaqueId::new()),
};
fields = &[];
}
ty::Char | ty::Int(_) | ty::Uint(_) => {
ctor = match value.try_eval_bits(cx.tcx, cx.param_env) {
Some(bits) => {
let x = match *pat.ty.kind() {
ty::Int(ity) => {
let size = Integer::from_int_ty(&cx.tcx, ity).size().bits();
MaybeInfiniteInt::new_finite_int(bits, size)
}
_ => MaybeInfiniteInt::new_finite_uint(bits),
};
IntRange(IntRange::from_singleton(x))
}
None => Opaque(OpaqueId::new()),
};
fields = &[];
}
ty::Float(ty::FloatTy::F32) => {
ctor = match value.try_eval_bits(cx.tcx, cx.param_env) {
Some(bits) => {
use rustc_apfloat::Float;
let value = rustc_apfloat::ieee::Single::from_bits(bits);
F32Range(value, value, RangeEnd::Included)
}
None => Opaque(OpaqueId::new()),
};
fields = &[];
}
ty::Float(ty::FloatTy::F64) => {
ctor = match value.try_eval_bits(cx.tcx, cx.param_env) {
Some(bits) => {
use rustc_apfloat::Float;
let value = rustc_apfloat::ieee::Double::from_bits(bits);
F64Range(value, value, RangeEnd::Included)
}
None => Opaque(OpaqueId::new()),
};
fields = &[];
}
ty::Ref(_, t, _) if t.is_str() => {
// We want a `&str` constant to behave like a `Deref` pattern, to be compatible
// with other `Deref` patterns. This could have been done in `const_to_pat`,
// but that causes issues with the rest of the matching code.
// So here, the constructor for a `"foo"` pattern is `&` (represented by
// `Single`), and has one field. That field has constructor `Str(value)` and no
// fields.
// Note: `t` is `str`, not `&str`.
let subpattern = DeconstructedPat::new(Str(*value), &[], *t, pat.span);
ctor = Single;
fields = singleton(subpattern)
}
// All constants that can be structurally matched have already been expanded
// into the corresponding `Pat`s by `const_to_pat`. Constants that remain are
// opaque.
_ => {
ctor = Opaque(OpaqueId::new());
fields = &[];
}
}
}
PatKind::Range(patrange) => {
let PatRange { lo, hi, end, .. } = patrange.as_ref();
let ty = pat.ty;
ctor = match ty.kind() {
ty::Char | ty::Int(_) | ty::Uint(_) => {
let lo = cx.lower_pat_range_bdy(*lo, ty);
let hi = cx.lower_pat_range_bdy(*hi, ty);
IntRange(IntRange::from_range(lo, hi, *end))
}
ty::Float(fty) => {
use rustc_apfloat::Float;
let lo = lo.as_finite().map(|c| c.eval_bits(cx.tcx, cx.param_env));
let hi = hi.as_finite().map(|c| c.eval_bits(cx.tcx, cx.param_env));
match fty {
ty::FloatTy::F32 => {
use rustc_apfloat::ieee::Single;
let lo = lo.map(Single::from_bits).unwrap_or(-Single::INFINITY);
let hi = hi.map(Single::from_bits).unwrap_or(Single::INFINITY);
F32Range(lo, hi, *end)
}
ty::FloatTy::F64 => {
use rustc_apfloat::ieee::Double;
let lo = lo.map(Double::from_bits).unwrap_or(-Double::INFINITY);
let hi = hi.map(Double::from_bits).unwrap_or(Double::INFINITY);
F64Range(lo, hi, *end)
}
}
}
_ => bug!("invalid type for range pattern: {}", ty),
};
fields = &[];
}
PatKind::Array { prefix, slice, suffix } | PatKind::Slice { prefix, slice, suffix } => {
let array_len = match pat.ty.kind() {
ty::Array(_, length) => {
Some(length.eval_target_usize(cx.tcx, cx.param_env) as usize)
}
ty::Slice(_) => None,
_ => span_bug!(pat.span, "bad ty {:?} for slice pattern", pat.ty),
};
let kind = if slice.is_some() {
SliceKind::VarLen(prefix.len(), suffix.len())
} else {
SliceKind::FixedLen(prefix.len() + suffix.len())
};
ctor = Slice(Slice::new(array_len, kind));
fields = cx.pattern_arena.alloc_from_iter(
prefix.iter().chain(suffix.iter()).map(|p| self.lower_pat(&*p)),
)
}
PatKind::Or { .. } => {
ctor = Or;
let pats = expand_or_pat(pat);
fields =
cx.pattern_arena.alloc_from_iter(pats.into_iter().map(|p| self.lower_pat(p)))
}
PatKind::Never => {
// FIXME(never_patterns): handle `!` in exhaustiveness. This is a sane default
// in the meantime.
ctor = Wildcard;
fields = &[];
}
PatKind::Error(_) => {
ctor = Opaque(OpaqueId::new());
fields = &[];
}
}
DeconstructedPat::new(ctor, fields, pat.ty, pat.span)
}
/// Convert back to a `thir::PatRangeBoundary` for diagnostic purposes.
/// Note: it is possible to get `isize/usize::MAX+1` here, as explained in the doc for
/// [`IntRange::split`]. This cannot be represented as a `Const`, so we represent it with
/// `PosInfinity`.
pub(crate) fn hoist_pat_range_bdy(
&self,
miint: MaybeInfiniteInt,
ty: Ty<'tcx>,
) -> PatRangeBoundary<'tcx> {
use MaybeInfiniteInt::*;
let tcx = self.tcx;
match miint {
NegInfinity => PatRangeBoundary::NegInfinity,
Finite(_) => {
let size = ty.primitive_size(tcx);
let bits = match *ty.kind() {
ty::Int(_) => miint.as_finite_int(size.bits()).unwrap(),
_ => miint.as_finite_uint().unwrap(),
};
match Scalar::try_from_uint(bits, size) {
Some(scalar) => {
let value = mir::Const::from_scalar(tcx, scalar, ty);
PatRangeBoundary::Finite(value)
}
// The value doesn't fit. Since `x >= 0` and 0 always encodes the minimum value
// for a type, the problem isn't that the value is too small. So it must be too
// large.
None => PatRangeBoundary::PosInfinity,
}
}
JustAfterMax | PosInfinity => PatRangeBoundary::PosInfinity,
}
}
/// Whether the range denotes the fictitious values before `isize::MIN` or after
/// `usize::MAX`/`isize::MAX` (see doc of [`IntRange::split`] for why these exist).
pub fn is_range_beyond_boundaries(&self, range: &IntRange, ty: Ty<'tcx>) -> bool {
ty.is_ptr_sized_integral() && {
// The two invalid ranges are `NegInfinity..isize::MIN` (represented as
// `NegInfinity..0`), and `{u,i}size::MAX+1..PosInfinity`. `hoist_pat_range_bdy`
// converts `MAX+1` to `PosInfinity`, and we couldn't have `PosInfinity` in `range.lo`
// otherwise.
let lo = self.hoist_pat_range_bdy(range.lo, ty);
matches!(lo, PatRangeBoundary::PosInfinity)
|| matches!(range.hi, MaybeInfiniteInt::Finite(0))
}
}
/// Convert back to a `thir::Pat` for diagnostic purposes.
pub(crate) fn hoist_pat_range(&self, range: &IntRange, ty: Ty<'tcx>) -> Pat<'tcx> {
use MaybeInfiniteInt::*;
let cx = self;
let kind = if matches!((range.lo, range.hi), (NegInfinity, PosInfinity)) {
PatKind::Wild
} else if range.is_singleton() {
let lo = cx.hoist_pat_range_bdy(range.lo, ty);
let value = lo.as_finite().unwrap();
PatKind::Constant { value }
} else {
// We convert to an inclusive range for diagnostics.
let mut end = RangeEnd::Included;
let mut lo = cx.hoist_pat_range_bdy(range.lo, ty);
if matches!(lo, PatRangeBoundary::PosInfinity) {
// The only reason to get `PosInfinity` here is the special case where
// `hoist_pat_range_bdy` found `{u,i}size::MAX+1`. So the range denotes the
// fictitious values after `{u,i}size::MAX` (see [`IntRange::split`] for why we do
// this). We show this to the user as `usize::MAX..` which is slightly incorrect but
// probably clear enough.
let c = ty.numeric_max_val(cx.tcx).unwrap();
let value = mir::Const::from_ty_const(c, cx.tcx);
lo = PatRangeBoundary::Finite(value);
}
let hi = if matches!(range.hi, Finite(0)) {
// The range encodes `..ty::MIN`, so we can't convert it to an inclusive range.
end = RangeEnd::Excluded;
range.hi
} else {
range.hi.minus_one()
};
let hi = cx.hoist_pat_range_bdy(hi, ty);
PatKind::Range(Box::new(PatRange { lo, hi, end, ty }))
};
Pat { ty, span: DUMMY_SP, kind }
}
/// Convert back to a `thir::Pat` for diagnostic purposes. This panics for patterns that don't
/// appear in diagnostics, like float ranges.
pub fn hoist_witness_pat(&self, pat: &WitnessPat<'tcx>) -> Pat<'tcx> {
let cx = self;
let is_wildcard = |pat: &Pat<'_>| matches!(pat.kind, PatKind::Wild);
let mut subpatterns = pat.iter_fields().map(|p| Box::new(cx.hoist_witness_pat(p)));
let kind = match pat.ctor() {
Bool(b) => PatKind::Constant { value: mir::Const::from_bool(cx.tcx, *b) },
IntRange(range) => return self.hoist_pat_range(range, pat.ty()),
Single | Variant(_) => match pat.ty().kind() {
ty::Tuple(..) => PatKind::Leaf {
subpatterns: subpatterns
.enumerate()
.map(|(i, pattern)| FieldPat { field: FieldIdx::new(i), pattern })
.collect(),
},
ty::Adt(adt_def, _) if adt_def.is_box() => {
// Without `box_patterns`, the only legal pattern of type `Box` is `_` (outside
// of `std`). So this branch is only reachable when the feature is enabled and
// the pattern is a box pattern.
PatKind::Deref { subpattern: subpatterns.next().unwrap() }
}
ty::Adt(adt_def, args) => {
let variant_index =
MatchCheckCtxt::variant_index_for_adt(&pat.ctor(), *adt_def);
let variant = &adt_def.variant(variant_index);
let subpatterns = cx
.list_variant_nonhidden_fields(pat.ty(), variant)
.zip(subpatterns)
.map(|((field, _ty), pattern)| FieldPat { field, pattern })
.collect();
if adt_def.is_enum() {
PatKind::Variant { adt_def: *adt_def, args, variant_index, subpatterns }
} else {
PatKind::Leaf { subpatterns }
}
}
// Note: given the expansion of `&str` patterns done in `expand_pattern`, we should
// be careful to reconstruct the correct constant pattern here. However a string
// literal pattern will never be reported as a non-exhaustiveness witness, so we
// ignore this issue.
ty::Ref(..) => PatKind::Deref { subpattern: subpatterns.next().unwrap() },
_ => bug!("unexpected ctor for type {:?} {:?}", pat.ctor(), pat.ty()),
},
Slice(slice) => {
match slice.kind {
SliceKind::FixedLen(_) => PatKind::Slice {
prefix: subpatterns.collect(),
slice: None,
suffix: Box::new([]),
},
SliceKind::VarLen(prefix, _) => {
let mut subpatterns = subpatterns.peekable();
let mut prefix: Vec<_> = subpatterns.by_ref().take(prefix).collect();
if slice.array_len.is_some() {
// Improves diagnostics a bit: if the type is a known-size array, instead
// of reporting `[x, _, .., _, y]`, we prefer to report `[x, .., y]`.
// This is incorrect if the size is not known, since `[_, ..]` captures
// arrays of lengths `>= 1` whereas `[..]` captures any length.
while !prefix.is_empty() && is_wildcard(prefix.last().unwrap()) {
prefix.pop();
}
while subpatterns.peek().is_some()
&& is_wildcard(subpatterns.peek().unwrap())
{
subpatterns.next();
}
}
let suffix: Box<[_]> = subpatterns.collect();
let wild = Pat::wildcard_from_ty(pat.ty());
PatKind::Slice {
prefix: prefix.into_boxed_slice(),
slice: Some(Box::new(wild)),
suffix,
}
}
}
}
&Str(value) => PatKind::Constant { value },
Wildcard | NonExhaustive | Hidden => PatKind::Wild,
Missing { .. } => bug!(
"trying to convert a `Missing` constructor into a `Pat`; this is probably a bug,
`Missing` should have been processed in `apply_constructors`"
),
F32Range(..) | F64Range(..) | Opaque(..) | Or => {
bug!("can't convert to pattern: {:?}", pat)
}
};
Pat { ty: pat.ty(), span: DUMMY_SP, kind }
}
/// Best-effort `Debug` implementation.
pub(crate) fn debug_pat(
f: &mut fmt::Formatter<'_>,
pat: &DeconstructedPat<'p, 'tcx>,
) -> fmt::Result {
let mut first = true;
let mut start_or_continue = |s| {
if first {
first = false;
""
} else {
s
}
};
let mut start_or_comma = || start_or_continue(", ");
match pat.ctor() {
Single | Variant(_) => match pat.ty().kind() {
ty::Adt(def, _) if def.is_box() => {
// Without `box_patterns`, the only legal pattern of type `Box` is `_` (outside
// of `std`). So this branch is only reachable when the feature is enabled and
// the pattern is a box pattern.
let subpattern = pat.iter_fields().next().unwrap();
write!(f, "box {subpattern:?}")
}
ty::Adt(..) | ty::Tuple(..) => {
let variant = match pat.ty().kind() {
ty::Adt(adt, _) => Some(
adt.variant(MatchCheckCtxt::variant_index_for_adt(pat.ctor(), *adt)),
),
ty::Tuple(_) => None,
_ => unreachable!(),
};
if let Some(variant) = variant {
write!(f, "{}", variant.name)?;
}
// Without `cx`, we can't know which field corresponds to which, so we can't
// get the names of the fields. Instead we just display everything as a tuple
// struct, which should be good enough.
write!(f, "(")?;
for p in pat.iter_fields() {
write!(f, "{}", start_or_comma())?;
write!(f, "{p:?}")?;
}
write!(f, ")")
}
// Note: given the expansion of `&str` patterns done in `expand_pattern`, we should
// be careful to detect strings here. However a string literal pattern will never
// be reported as a non-exhaustiveness witness, so we can ignore this issue.
ty::Ref(_, _, mutbl) => {
let subpattern = pat.iter_fields().next().unwrap();
write!(f, "&{}{:?}", mutbl.prefix_str(), subpattern)
}
_ => write!(f, "_"),
},
Slice(slice) => {
let mut subpatterns = pat.iter_fields();
write!(f, "[")?;
match slice.kind {
SliceKind::FixedLen(_) => {
for p in subpatterns {
write!(f, "{}{:?}", start_or_comma(), p)?;
}
}
SliceKind::VarLen(prefix_len, _) => {
for p in subpatterns.by_ref().take(prefix_len) {
write!(f, "{}{:?}", start_or_comma(), p)?;
}
write!(f, "{}", start_or_comma())?;
write!(f, "..")?;
for p in subpatterns {
write!(f, "{}{:?}", start_or_comma(), p)?;
}
}
}
write!(f, "]")
}
Bool(b) => write!(f, "{b}"),
// Best-effort, will render signed ranges incorrectly
IntRange(range) => write!(f, "{range:?}"),
F32Range(lo, hi, end) => write!(f, "{lo}{end}{hi}"),
F64Range(lo, hi, end) => write!(f, "{lo}{end}{hi}"),
Str(value) => write!(f, "{value}"),
Opaque(..) => write!(f, "<constant pattern>"),
Or => {
for pat in pat.iter_fields() {
write!(f, "{}{:?}", start_or_continue(" | "), pat)?;
}
Ok(())
}
Wildcard | Missing { .. } | NonExhaustive | Hidden => write!(f, "_ : {:?}", pat.ty()),
}
}
}
/// Recursively expand this pattern into its subpatterns. Only useful for or-patterns.
fn expand_or_pat<'p, 'tcx>(pat: &'p Pat<'tcx>) -> Vec<&'p Pat<'tcx>> {
fn expand<'p, 'tcx>(pat: &'p Pat<'tcx>, vec: &mut Vec<&'p Pat<'tcx>>) {
if let PatKind::Or { pats } = &pat.kind {
for pat in pats.iter() {
expand(pat, vec);
}
} else {
vec.push(pat)
}
}
let mut pats = Vec::new();
expand(pat, &mut pats);
pats
}

95
compiler/rustc_pattern_analysis/src/errors.rs Normal file
View File

@ -0,0 +1,95 @@
use crate::{cx::MatchCheckCtxt, pat::WitnessPat};
use rustc_errors::{AddToDiagnostic, Diagnostic, SubdiagnosticMessage};
use rustc_macros::{LintDiagnostic, Subdiagnostic};
use rustc_middle::thir::Pat;
use rustc_middle::ty::Ty;
use rustc_span::Span;
#[derive(Subdiagnostic)]
#[label(pattern_analysis_uncovered)]
pub struct Uncovered<'tcx> {
#[primary_span]
span: Span,
count: usize,
witness_1: Pat<'tcx>,
witness_2: Pat<'tcx>,
witness_3: Pat<'tcx>,
remainder: usize,
}
impl<'tcx> Uncovered<'tcx> {
pub fn new<'p>(
span: Span,
cx: &MatchCheckCtxt<'p, 'tcx>,
witnesses: Vec<WitnessPat<'tcx>>,
) -> Self {
let witness_1 = cx.hoist_witness_pat(witnesses.get(0).unwrap());
Self {
span,
count: witnesses.len(),
// Substitute dummy values if witnesses is smaller than 3. These will never be read.
witness_2: witnesses
.get(1)
.map(|w| cx.hoist_witness_pat(w))
.unwrap_or_else(|| witness_1.clone()),
witness_3: witnesses
.get(2)
.map(|w| cx.hoist_witness_pat(w))
.unwrap_or_else(|| witness_1.clone()),
witness_1,
remainder: witnesses.len().saturating_sub(3),
}
}
}
#[derive(LintDiagnostic)]
#[diag(pattern_analysis_overlapping_range_endpoints)]
#[note]
pub struct OverlappingRangeEndpoints<'tcx> {
#[label]
pub range: Span,
#[subdiagnostic]
pub overlap: Vec<Overlap<'tcx>>,
}
pub struct Overlap<'tcx> {
pub span: Span,
pub range: Pat<'tcx>,
}
impl<'tcx> AddToDiagnostic for Overlap<'tcx> {
fn add_to_diagnostic_with<F>(self, diag: &mut Diagnostic, _: F)
where
F: Fn(&mut Diagnostic, SubdiagnosticMessage) -> SubdiagnosticMessage,
{
let Overlap { span, range } = self;
// FIXME(mejrs) unfortunately `#[derive(LintDiagnostic)]`
// does not support `#[subdiagnostic(eager)]`...
let message = format!("this range overlaps on `{range}`...");
diag.span_label(span, message);
}
}
#[derive(LintDiagnostic)]
#[diag(pattern_analysis_non_exhaustive_omitted_pattern)]
#[help]
#[note]
pub(crate) struct NonExhaustiveOmittedPattern<'tcx> {
pub scrut_ty: Ty<'tcx>,
#[subdiagnostic]
pub uncovered: Uncovered<'tcx>,
}
#[derive(LintDiagnostic)]
#[diag(pattern_analysis_non_exhaustive_omitted_pattern_lint_on_arm)]
#[help]
pub(crate) struct NonExhaustiveOmittedPatternLintOnArm {
#[label]
pub lint_span: Span,
#[suggestion(code = "#[{lint_level}({lint_name})]\n", applicability = "maybe-incorrect")]
pub suggest_lint_on_match: Option<Span>,
pub lint_level: &'static str,
pub lint_name: &'static str,
}

56
compiler/rustc_pattern_analysis/src/lib.rs Normal file
View File

@ -0,0 +1,56 @@
//! Analysis of patterns, notably match exhaustiveness checking.
pub mod constructor;
pub mod cx;
pub mod errors;
pub(crate) mod lints;
pub mod pat;
pub mod usefulness;
#[macro_use]
extern crate tracing;
#[macro_use]
extern crate rustc_middle;
rustc_fluent_macro::fluent_messages! { "../messages.ftl" }
use lints::PatternColumn;
use rustc_hir::HirId;
use rustc_middle::ty::Ty;
use usefulness::{compute_match_usefulness, UsefulnessReport};
use crate::cx::MatchCheckCtxt;
use crate::lints::{lint_nonexhaustive_missing_variants, lint_overlapping_range_endpoints};
use crate::pat::DeconstructedPat;
/// The arm of a match expression.
#[derive(Clone, Copy, Debug)]
pub struct MatchArm<'p, 'tcx> {
/// The pattern must have been lowered through `check_match::MatchVisitor::lower_pattern`.
pub pat: &'p DeconstructedPat<'p, 'tcx>,
pub hir_id: HirId,
pub has_guard: bool,
}
/// The entrypoint for this crate. Computes whether a match is exhaustive and which of its arms are
/// useful, and runs some lints.
pub fn analyze_match<'p, 'tcx>(
cx: &MatchCheckCtxt<'p, 'tcx>,
arms: &[MatchArm<'p, 'tcx>],
scrut_ty: Ty<'tcx>,
) -> UsefulnessReport<'p, 'tcx> {
let pat_column = PatternColumn::new(arms);
let report = compute_match_usefulness(cx, arms, scrut_ty);
// Lint on ranges that overlap on their endpoints, which is likely a mistake.
lint_overlapping_range_endpoints(cx, &pat_column);
// Run the non_exhaustive_omitted_patterns lint. Only run on refutable patterns to avoid hitting
// `if let`s. Only run if the match is exhaustive otherwise the error is redundant.
if cx.refutable && report.non_exhaustiveness_witnesses.is_empty() {
lint_nonexhaustive_missing_variants(cx, arms, &pat_column, scrut_ty)
}
report
}

290
compiler/rustc_pattern_analysis/src/lints.rs Normal file
View File

@ -0,0 +1,290 @@
use smallvec::SmallVec;
use rustc_data_structures::captures::Captures;
use rustc_middle::ty::{self, Ty};
use rustc_session::lint;
use rustc_session::lint::builtin::NON_EXHAUSTIVE_OMITTED_PATTERNS;
use rustc_span::Span;
use crate::constructor::{Constructor, IntRange, MaybeInfiniteInt, SplitConstructorSet};
use crate::cx::MatchCheckCtxt;
use crate::errors::{
NonExhaustiveOmittedPattern, NonExhaustiveOmittedPatternLintOnArm, Overlap,
OverlappingRangeEndpoints, Uncovered,
};
use crate::pat::{DeconstructedPat, WitnessPat};
use crate::usefulness::PatCtxt;
use crate::MatchArm;
/// A column of patterns in the matrix, where a column is the intuitive notion of "subpatterns that
/// inspect the same subvalue/place".
/// This is used to traverse patterns column-by-column for lints. Despite similarities with the
/// algorithm in [`crate::usefulness`], this does a different traversal. Notably this is linear in
/// the depth of patterns, whereas `compute_exhaustiveness_and_usefulness` is worst-case exponential
/// (exhaustiveness is NP-complete). The core difference is that we treat sub-columns separately.
///
/// This must not contain an or-pattern. `specialize` takes care to expand them.
///
/// This is not used in the main algorithm; only in lints.
#[derive(Debug)]
pub(crate) struct PatternColumn<'p, 'tcx> {
patterns: Vec<&'p DeconstructedPat<'p, 'tcx>>,
}
impl<'p, 'tcx> PatternColumn<'p, 'tcx> {
pub(crate) fn new(arms: &[MatchArm<'p, 'tcx>]) -> Self {
let mut patterns = Vec::with_capacity(arms.len());
for arm in arms {
if arm.pat.is_or_pat() {
patterns.extend(arm.pat.flatten_or_pat())
} else {
patterns.push(arm.pat)
}
}
Self { patterns }
}
fn is_empty(&self) -> bool {
self.patterns.is_empty()
}
fn head_ty(&self) -> Option<Ty<'tcx>> {
if self.patterns.len() == 0 {
return None;
}
// If the type is opaque and it is revealed anywhere in the column, we take the revealed
// version. Otherwise we could encounter constructors for the revealed type and crash.
let is_opaque = |ty: Ty<'tcx>| matches!(ty.kind(), ty::Alias(ty::Opaque, ..));
let first_ty = self.patterns[0].ty();
if is_opaque(first_ty) {
for pat in &self.patterns {
let ty = pat.ty();
if !is_opaque(ty) {
return Some(ty);
}
}
}
Some(first_ty)
}
/// Do constructor splitting on the constructors of the column.
fn analyze_ctors(&self, pcx: &PatCtxt<'_, 'p, 'tcx>) -> SplitConstructorSet<'tcx> {
let column_ctors = self.patterns.iter().map(|p| p.ctor());
pcx.cx.ctors_for_ty(pcx.ty).split(pcx, column_ctors)
}
fn iter<'a>(&'a self) -> impl Iterator<Item = &'p DeconstructedPat<'p, 'tcx>> + Captures<'a> {
self.patterns.iter().copied()
}
/// Does specialization: given a constructor, this takes the patterns from the column that match
/// the constructor, and outputs their fields.
/// This returns one column per field of the constructor. They usually all have the same length
/// (the number of patterns in `self` that matched `ctor`), except that we expand or-patterns
/// which may change the lengths.
fn specialize(&self, pcx: &PatCtxt<'_, 'p, 'tcx>, ctor: &Constructor<'tcx>) -> Vec<Self> {
let arity = ctor.arity(pcx);
if arity == 0 {
return Vec::new();
}
// We specialize the column by `ctor`. This gives us `arity`-many columns of patterns. These
// columns may have different lengths in the presence of or-patterns (this is why we can't
// reuse `Matrix`).
let mut specialized_columns: Vec<_> =
(0..arity).map(|_| Self { patterns: Vec::new() }).collect();
let relevant_patterns =
self.patterns.iter().filter(|pat| ctor.is_covered_by(pcx, pat.ctor()));
for pat in relevant_patterns {
let specialized = pat.specialize(pcx, ctor);
for (subpat, column) in specialized.iter().zip(&mut specialized_columns) {
if subpat.is_or_pat() {
column.patterns.extend(subpat.flatten_or_pat())
} else {
column.patterns.push(subpat)
}
}
}
assert!(
!specialized_columns[0].is_empty(),
"ctor {ctor:?} was listed as present but isn't;
there is an inconsistency between `Constructor::is_covered_by` and `ConstructorSet::split`"
);
specialized_columns
}
}
/// Traverse the patterns to collect any variants of a non_exhaustive enum that fail to be mentioned
/// in a given column.
#[instrument(level = "debug", skip(cx), ret)]
fn collect_nonexhaustive_missing_variants<'p, 'tcx>(
cx: &MatchCheckCtxt<'p, 'tcx>,
column: &PatternColumn<'p, 'tcx>,
) -> Vec<WitnessPat<'tcx>> {
let Some(ty) = column.head_ty() else {
return Vec::new();
};
let pcx = &PatCtxt::new_dummy(cx, ty);
let set = column.analyze_ctors(pcx);
if set.present.is_empty() {
// We can't consistently handle the case where no constructors are present (since this would
// require digging deep through any type in case there's a non_exhaustive enum somewhere),
// so for consistency we refuse to handle the top-level case, where we could handle it.
return vec![];
}
let mut witnesses = Vec::new();
if cx.is_foreign_non_exhaustive_enum(ty) {
witnesses.extend(
set.missing
.into_iter()
// This will list missing visible variants.
.filter(|c| !matches!(c, Constructor::Hidden | Constructor::NonExhaustive))
.map(|missing_ctor| WitnessPat::wild_from_ctor(pcx, missing_ctor)),
)
}
// Recurse into the fields.
for ctor in set.present {
let specialized_columns = column.specialize(pcx, &ctor);
let wild_pat = WitnessPat::wild_from_ctor(pcx, ctor);
for (i, col_i) in specialized_columns.iter().enumerate() {
// Compute witnesses for each column.
let wits_for_col_i = collect_nonexhaustive_missing_variants(cx, col_i);
// For each witness, we build a new pattern in the shape of `ctor(_, _, wit, _, _)`,
// adding enough wildcards to match `arity`.
for wit in wits_for_col_i {
let mut pat = wild_pat.clone();
pat.fields[i] = wit;
witnesses.push(pat);
}
}
}
witnesses
}
pub(crate) fn lint_nonexhaustive_missing_variants<'p, 'tcx>(
cx: &MatchCheckCtxt<'p, 'tcx>,
arms: &[MatchArm<'p, 'tcx>],
pat_column: &PatternColumn<'p, 'tcx>,
scrut_ty: Ty<'tcx>,
) {
if !matches!(
cx.tcx.lint_level_at_node(NON_EXHAUSTIVE_OMITTED_PATTERNS, cx.match_lint_level).0,
rustc_session::lint::Level::Allow
) {
let witnesses = collect_nonexhaustive_missing_variants(cx, pat_column);
if !witnesses.is_empty() {
// Report that a match of a `non_exhaustive` enum marked with `non_exhaustive_omitted_patterns`
// is not exhaustive enough.
//
// NB: The partner lint for structs lives in `compiler/rustc_hir_analysis/src/check/pat.rs`.
cx.tcx.emit_spanned_lint(
NON_EXHAUSTIVE_OMITTED_PATTERNS,
cx.match_lint_level,
cx.scrut_span,
NonExhaustiveOmittedPattern {
scrut_ty,
uncovered: Uncovered::new(cx.scrut_span, cx, witnesses),
},
);
}
} else {
// We used to allow putting the `#[allow(non_exhaustive_omitted_patterns)]` on a match
// arm. This no longer makes sense so we warn users, to avoid silently breaking their
// usage of the lint.
for arm in arms {
let (lint_level, lint_level_source) =
cx.tcx.lint_level_at_node(NON_EXHAUSTIVE_OMITTED_PATTERNS, arm.hir_id);
if !matches!(lint_level, rustc_session::lint::Level::Allow) {
let decorator = NonExhaustiveOmittedPatternLintOnArm {
lint_span: lint_level_source.span(),
suggest_lint_on_match: cx.whole_match_span.map(|span| span.shrink_to_lo()),
lint_level: lint_level.as_str(),
lint_name: "non_exhaustive_omitted_patterns",
};
use rustc_errors::DecorateLint;
let mut err = cx.tcx.sess.struct_span_warn(arm.pat.span(), "");
err.set_primary_message(decorator.msg());
decorator.decorate_lint(&mut err);
err.emit();
}
}
}
}
/// Traverse the patterns to warn the user about ranges that overlap on their endpoints.
#[instrument(level = "debug", skip(cx))]
pub(crate) fn lint_overlapping_range_endpoints<'p, 'tcx>(
cx: &MatchCheckCtxt<'p, 'tcx>,
column: &PatternColumn<'p, 'tcx>,
) {
let Some(ty) = column.head_ty() else {
return;
};
let pcx = &PatCtxt::new_dummy(cx, ty);
let set = column.analyze_ctors(pcx);
if matches!(ty.kind(), ty::Char | ty::Int(_) | ty::Uint(_)) {
let emit_lint = |overlap: &IntRange, this_span: Span, overlapped_spans: &[Span]| {
let overlap_as_pat = cx.hoist_pat_range(overlap, ty);
let overlaps: Vec<_> = overlapped_spans
.iter()
.copied()
.map(|span| Overlap { range: overlap_as_pat.clone(), span })
.collect();
cx.tcx.emit_spanned_lint(
lint::builtin::OVERLAPPING_RANGE_ENDPOINTS,
cx.match_lint_level,
this_span,
OverlappingRangeEndpoints { overlap: overlaps, range: this_span },
);
};
// If two ranges overlapped, the split set will contain their intersection as a singleton.
let split_int_ranges = set.present.iter().filter_map(|c| c.as_int_range());
for overlap_range in split_int_ranges.clone() {
if overlap_range.is_singleton() {
let overlap: MaybeInfiniteInt = overlap_range.lo;
// Ranges that look like `lo..=overlap`.
let mut prefixes: SmallVec<[_; 1]> = Default::default();
// Ranges that look like `overlap..=hi`.
let mut suffixes: SmallVec<[_; 1]> = Default::default();
// Iterate on patterns that contained `overlap`.
for pat in column.iter() {
let this_span = pat.span();
let Constructor::IntRange(this_range) = pat.ctor() else { continue };
if this_range.is_singleton() {
// Don't lint when one of the ranges is a singleton.
continue;
}
if this_range.lo == overlap {
// `this_range` looks like `overlap..=this_range.hi`; it overlaps with any
// ranges that look like `lo..=overlap`.
if !prefixes.is_empty() {
emit_lint(overlap_range, this_span, &prefixes);
}
suffixes.push(this_span)
} else if this_range.hi == overlap.plus_one() {
// `this_range` looks like `this_range.lo..=overlap`; it overlaps with any
// ranges that look like `overlap..=hi`.
if !suffixes.is_empty() {
emit_lint(overlap_range, this_span, &suffixes);
}
prefixes.push(this_span)
}
}
}
}
} else {
// Recurse into the fields.
for ctor in set.present {
for col in column.specialize(pcx, &ctor) {
lint_overlapping_range_endpoints(cx, &col);
}
}
}
}

205
compiler/rustc_pattern_analysis/src/pat.rs Normal file
View File

@ -0,0 +1,205 @@
//! As explained in [`crate::usefulness`], values and patterns are made from constructors applied to
//! fields. This file defines types that represent patterns in this way.
use std::cell::Cell;
use std::fmt;
use smallvec::{smallvec, SmallVec};
use rustc_data_structures::captures::Captures;
use rustc_middle::ty::{self, Ty};
use rustc_span::{Span, DUMMY_SP};
use self::Constructor::*;
use self::SliceKind::*;
use crate::constructor::{Constructor, SliceKind};
use crate::cx::MatchCheckCtxt;
use crate::usefulness::PatCtxt;
/// Values and patterns can be represented as a constructor applied to some fields. This represents
/// a pattern in this form.
/// This also uses interior mutability to keep track of whether the pattern has been found reachable
/// during analysis. For this reason they cannot be cloned.
/// A `DeconstructedPat` will almost always come from user input; the only exception are some
/// `Wildcard`s introduced during specialization.
///
/// Note that the number of fields may not match the fields declared in the original struct/variant.
/// This happens if a private or `non_exhaustive` field is uninhabited, because the code mustn't
/// observe that it is uninhabited. In that case that field is not included in `fields`. Care must
/// be taken when converting to/from `thir::Pat`.
pub struct DeconstructedPat<'p, 'tcx> {
ctor: Constructor<'tcx>,
fields: &'p [DeconstructedPat<'p, 'tcx>],
ty: Ty<'tcx>,
span: Span,
/// Whether removing this arm would change the behavior of the match expression.
useful: Cell<bool>,
}
impl<'p, 'tcx> DeconstructedPat<'p, 'tcx> {
pub fn wildcard(ty: Ty<'tcx>, span: Span) -> Self {
Self::new(Wildcard, &[], ty, span)
}
pub fn new(
ctor: Constructor<'tcx>,
fields: &'p [DeconstructedPat<'p, 'tcx>],
ty: Ty<'tcx>,
span: Span,
) -> Self {
DeconstructedPat { ctor, fields, ty, span, useful: Cell::new(false) }
}
pub(crate) fn is_or_pat(&self) -> bool {
matches!(self.ctor, Or)
}
/// Expand this (possibly-nested) or-pattern into its alternatives.
pub(crate) fn flatten_or_pat(&'p self) -> SmallVec<[&'p Self; 1]> {
if self.is_or_pat() {
self.iter_fields().flat_map(|p| p.flatten_or_pat()).collect()
} else {
smallvec![self]
}
}
pub fn ctor(&self) -> &Constructor<'tcx> {
&self.ctor
}
pub fn ty(&self) -> Ty<'tcx> {
self.ty
}
pub fn span(&self) -> Span {
self.span
}
pub fn iter_fields<'a>(
&'a self,
) -> impl Iterator<Item = &'p DeconstructedPat<'p, 'tcx>> + Captures<'a> {
self.fields.iter()
}
/// Specialize this pattern with a constructor.
/// `other_ctor` can be different from `self.ctor`, but must be covered by it.
pub(crate) fn specialize<'a>(
&'a self,
pcx: &PatCtxt<'_, 'p, 'tcx>,
other_ctor: &Constructor<'tcx>,
) -> SmallVec<[&'p DeconstructedPat<'p, 'tcx>; 2]> {
match (&self.ctor, other_ctor) {
(Wildcard, _) => {
// We return a wildcard for each field of `other_ctor`.
pcx.cx.ctor_wildcard_fields(other_ctor, pcx.ty).iter().collect()
}
(Slice(self_slice), Slice(other_slice))
if self_slice.arity() != other_slice.arity() =>
{
// The only tricky case: two slices of different arity. Since `self_slice` covers
// `other_slice`, `self_slice` must be `VarLen`, i.e. of the form
// `[prefix, .., suffix]`. Moreover `other_slice` is guaranteed to have a larger
// arity. So we fill the middle part with enough wildcards to reach the length of
// the new, larger slice.
match self_slice.kind {
FixedLen(_) => bug!("{:?} doesn't cover {:?}", self_slice, other_slice),
VarLen(prefix, suffix) => {
let (ty::Slice(inner_ty) | ty::Array(inner_ty, _)) = *self.ty.kind() else {
bug!("bad slice pattern {:?} {:?}", self.ctor, self.ty);
};
let prefix = &self.fields[..prefix];
let suffix = &self.fields[self_slice.arity() - suffix..];
let wildcard: &_ = pcx
.cx
.pattern_arena
.alloc(DeconstructedPat::wildcard(inner_ty, DUMMY_SP));
let extra_wildcards = other_slice.arity() - self_slice.arity();
let extra_wildcards = (0..extra_wildcards).map(|_| wildcard);
prefix.iter().chain(extra_wildcards).chain(suffix).collect()
}
}
}
_ => self.fields.iter().collect(),
}
}
/// We keep track for each pattern if it was ever useful during the analysis. This is used
/// with `redundant_spans` to report redundant subpatterns arising from or patterns.
pub(crate) fn set_useful(&self) {
self.useful.set(true)
}
pub(crate) fn is_useful(&self) -> bool {
if self.useful.get() {
true
} else if self.is_or_pat() && self.iter_fields().any(|f| f.is_useful()) {
// We always expand or patterns in the matrix, so we will never see the actual
// or-pattern (the one with constructor `Or`) in the column. As such, it will not be
// marked as useful itself, only its children will. We recover this information here.
self.set_useful();
true
} else {
false
}
}
/// Report the spans of subpatterns that were not useful, if any.
pub(crate) fn redundant_spans(&self) -> Vec<Span> {
let mut spans = Vec::new();
self.collect_redundant_spans(&mut spans);
spans
}
fn collect_redundant_spans(&self, spans: &mut Vec<Span>) {
// We don't look at subpatterns if we already reported the whole pattern as redundant.
if !self.is_useful() {
spans.push(self.span);
} else {
for p in self.iter_fields() {
p.collect_redundant_spans(spans);
}
}
}
}
/// This is mostly copied from the `Pat` impl. This is best effort and not good enough for a
/// `Display` impl.
impl<'p, 'tcx> fmt::Debug for DeconstructedPat<'p, 'tcx> {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
MatchCheckCtxt::debug_pat(f, self)
}
}
/// Same idea as `DeconstructedPat`, except this is a fictitious pattern built up for diagnostics
/// purposes. As such they don't use interning and can be cloned.
#[derive(Debug, Clone)]
pub struct WitnessPat<'tcx> {
ctor: Constructor<'tcx>,
pub(crate) fields: Vec<WitnessPat<'tcx>>,
ty: Ty<'tcx>,
}
impl<'tcx> WitnessPat<'tcx> {
pub(crate) fn new(ctor: Constructor<'tcx>, fields: Vec<Self>, ty: Ty<'tcx>) -> Self {
Self { ctor, fields, ty }
}
pub(crate) fn wildcard(ty: Ty<'tcx>) -> Self {
Self::new(Wildcard, Vec::new(), ty)
}
/// Construct a pattern that matches everything that starts with this constructor.
/// For example, if `ctor` is a `Constructor::Variant` for `Option::Some`, we get the pattern
/// `Some(_)`.
pub(crate) fn wild_from_ctor(pcx: &PatCtxt<'_, '_, 'tcx>, ctor: Constructor<'tcx>) -> Self {
let field_tys =
pcx.cx.ctor_wildcard_fields(&ctor, pcx.ty).iter().map(|deco_pat| deco_pat.ty());
let fields = field_tys.map(|ty| Self::wildcard(ty)).collect();
Self::new(ctor, fields, pcx.ty)
}
pub fn ctor(&self) -> &Constructor<'tcx> {
&self.ctor
}
pub fn ty(&self) -> Ty<'tcx> {
self.ty
}
pub fn iter_fields<'a>(&'a self) -> impl Iterator<Item = &'a WitnessPat<'tcx>> {
self.fields.iter()
}
}

401
compiler/rustc_mir_build/src/thir/pattern/usefulness.rs → compiler/rustc_pattern_analysis/src/usefulness.rs
View File

@ -97,8 +97,9 @@
//! - `matches!([v0], [p0, .., p1]) := false` (incompatible lengths)
//! - `matches!([v0, v1, v2], [p0, .., p1]) := matches!(v0, p0) && matches!(v2, p1)`
//!
//! Constructors, fields and relevant operations are defined in the [`super::deconstruct_pat`]
//! module. The question of whether a constructor is matched by another one is answered by
//! Constructors and relevant operations are defined in the [`crate::constructor`] module. A
//! representation of patterns that uses constructors is available in [`crate::pat`]. The question
//! of whether a constructor is matched by another one is answered by
//! [`Constructor::is_covered_by`].
//!
//! Note 1: variable bindings (like the `x` in `Some(x)`) match anything, so we treat them as wildcards.
@ -241,8 +242,8 @@
//! Therefore `usefulness(tp_1, tp_2, tq)` returns the single witness-tuple `[Variant2(Some(true), 0)]`.
//!
//!
//! Computing the set of constructors for a type is done in [`ConstructorSet::for_ty`]. See the
//! following sections for more accurate versions of the algorithm and corresponding links.
//! Computing the set of constructors for a type is done in [`MatchCheckCtxt::ctors_for_ty`]. See
//! the following sections for more accurate versions of the algorithm and corresponding links.
//!
//!
//!
@ -295,7 +296,7 @@
//! the same reasoning, we only need to try two cases: `North`, and "everything else".
//!
//! We call _constructor splitting_ the operation that computes such a minimal set of cases to try.
//! This is done in [`ConstructorSet::split`] and explained in [`super::deconstruct_pat`].
//! This is done in [`ConstructorSet::split`] and explained in [`crate::constructor`].
//!
//!
//!
@ -551,82 +552,33 @@
//! I (Nadrieril) prefer to put new tests in `ui/pattern/usefulness` unless there's a specific
//! reason not to, for example if they crucially depend on a particular feature like `or_patterns`.
use self::ValidityConstraint::*;
use super::deconstruct_pat::{
Constructor, ConstructorSet, DeconstructedPat, IntRange, MaybeInfiniteInt, SplitConstructorSet,
WitnessPat,
};
use crate::errors::{
NonExhaustiveOmittedPattern, NonExhaustiveOmittedPatternLintOnArm, Overlap,
OverlappingRangeEndpoints, Uncovered,
};
use rustc_data_structures::captures::Captures;
use rustc_arena::TypedArena;
use rustc_data_structures::stack::ensure_sufficient_stack;
use rustc_hir::def_id::DefId;
use rustc_hir::HirId;
use rustc_middle::ty::{self, Ty, TyCtxt};
use rustc_session::lint;
use rustc_session::lint::builtin::NON_EXHAUSTIVE_OMITTED_PATTERNS;
use rustc_span::{Span, DUMMY_SP};
use smallvec::{smallvec, SmallVec};
use std::fmt;
pub(crate) struct MatchCheckCtxt<'p, 'tcx> {
pub(crate) tcx: TyCtxt<'tcx>,
/// The module in which the match occurs. This is necessary for
/// checking inhabited-ness of types because whether a type is (visibly)
/// inhabited can depend on whether it was defined in the current module or
/// not. E.g., `struct Foo { _private: ! }` cannot be seen to be empty
/// outside its module and should not be matchable with an empty match statement.
pub(crate) module: DefId,
pub(crate) param_env: ty::ParamEnv<'tcx>,
pub(crate) pattern_arena: &'p TypedArena<DeconstructedPat<'p, 'tcx>>,
/// Lint level at the match.
pub(crate) match_lint_level: HirId,
/// The span of the whole match, if applicable.
pub(crate) whole_match_span: Option<Span>,
/// Span of the scrutinee.
pub(crate) scrut_span: Span,
/// Only produce `NON_EXHAUSTIVE_OMITTED_PATTERNS` lint on refutable patterns.
pub(crate) refutable: bool,
/// Whether the data at the scrutinee is known to be valid. This is false if the scrutinee comes
/// from a union field, a pointer deref, or a reference deref (pending opsem decisions).
pub(crate) known_valid_scrutinee: bool,
}
use rustc_data_structures::{captures::Captures, stack::ensure_sufficient_stack};
use rustc_middle::ty::{self, Ty};
use rustc_span::{Span, DUMMY_SP};
impl<'a, 'tcx> MatchCheckCtxt<'a, 'tcx> {
pub(super) fn is_uninhabited(&self, ty: Ty<'tcx>) -> bool {
!ty.is_inhabited_from(self.tcx, self.module, self.param_env)
}
use crate::constructor::{Constructor, ConstructorSet};
use crate::cx::MatchCheckCtxt;
use crate::pat::{DeconstructedPat, WitnessPat};
use crate::MatchArm;
/// Returns whether the given type is an enum from another crate declared `#[non_exhaustive]`.
pub(super) fn is_foreign_non_exhaustive_enum(&self, ty: Ty<'tcx>) -> bool {
match ty.kind() {
ty::Adt(def, ..) => {
def.is_enum() && def.is_variant_list_non_exhaustive() && !def.did().is_local()
}
_ => false,
}
}
}
use self::ValidityConstraint::*;
#[derive(Copy, Clone)]
pub(super) struct PatCtxt<'a, 'p, 'tcx> {
pub(super) cx: &'a MatchCheckCtxt<'p, 'tcx>,
pub(crate) struct PatCtxt<'a, 'p, 'tcx> {
pub(crate) cx: &'a MatchCheckCtxt<'p, 'tcx>,
/// Type of the current column under investigation.
pub(super) ty: Ty<'tcx>,
pub(crate) ty: Ty<'tcx>,
/// Whether the current pattern is the whole pattern as found in a match arm, or if it's a
/// subpattern.
pub(super) is_top_level: bool,
pub(crate) is_top_level: bool,
}
impl<'a, 'p, 'tcx> PatCtxt<'a, 'p, 'tcx> {
/// A `PatCtxt` when code other than `is_useful` needs one.
fn new_dummy(cx: &'a MatchCheckCtxt<'p, 'tcx>, ty: Ty<'tcx>) -> Self {
pub(crate) fn new_dummy(cx: &'a MatchCheckCtxt<'p, 'tcx>, ty: Ty<'tcx>) -> Self {
PatCtxt { cx, ty, is_top_level: false }
}
}
@ -643,7 +595,7 @@ impl<'a, 'p, 'tcx> fmt::Debug for PatCtxt<'a, 'p, 'tcx> {
/// - in the matrix, track whether a given place (aka column) is known to contain a valid value or
/// not.
#[derive(Debug, Copy, Clone, PartialEq, Eq)]
pub(super) enum ValidityConstraint {
enum ValidityConstraint {
ValidOnly,
MaybeInvalid,
/// Option for backwards compatibility: the place is not known to be valid but we allow omitting
@ -652,7 +604,7 @@ pub(super) enum ValidityConstraint {
}
impl ValidityConstraint {
pub(super) fn from_bool(is_valid_only: bool) -> Self {
fn from_bool(is_valid_only: bool) -> Self {
if is_valid_only { ValidOnly } else { MaybeInvalid }
}
@ -664,10 +616,10 @@ impl ValidityConstraint {
}
}
pub(super) fn is_known_valid(self) -> bool {
fn is_known_valid(self) -> bool {
matches!(self, ValidOnly)
}
pub(super) fn allows_omitting_empty_arms(self) -> bool {
fn allows_omitting_empty_arms(self) -> bool {
matches!(self, ValidOnly | MaybeInvalidButAllowOmittingArms)
}
@ -677,11 +629,7 @@ impl ValidityConstraint {
///
/// Pending further opsem decisions, the current behavior is: validity is preserved, except
/// inside `&` and union fields where validity is reset to `MaybeInvalid`.
pub(super) fn specialize<'tcx>(
self,
pcx: &PatCtxt<'_, '_, 'tcx>,
ctor: &Constructor<'tcx>,
) -> Self {
fn specialize<'tcx>(self, pcx: &PatCtxt<'_, '_, 'tcx>, ctor: &Constructor<'tcx>) -> Self {
// We preserve validity except when we go inside a reference or a union field.
if matches!(ctor, Constructor::Single)
&& (matches!(pcx.ty.kind(), ty::Ref(..))
@ -1072,7 +1020,7 @@ impl<'p, 'tcx> fmt::Debug for Matrix<'p, 'tcx> {
///
/// See the top of the file for more detailed explanations and examples.
#[derive(Debug, Clone)]
pub(crate) struct WitnessStack<'tcx>(Vec<WitnessPat<'tcx>>);
struct WitnessStack<'tcx>(Vec<WitnessPat<'tcx>>);
impl<'tcx> WitnessStack<'tcx> {
/// Asserts that the witness contains a single pattern, and returns it.
@ -1119,7 +1067,7 @@ impl<'tcx> WitnessStack<'tcx> {
/// Just as the `Matrix` starts with a single column, by the end of the algorithm, this has a single
/// column, which contains the patterns that are missing for the match to be exhaustive.
#[derive(Debug, Clone)]
pub struct WitnessMatrix<'tcx>(Vec<WitnessStack<'tcx>>);
struct WitnessMatrix<'tcx>(Vec<WitnessStack<'tcx>>);
impl<'tcx> WitnessMatrix<'tcx> {
/// New matrix with no witnesses.
@ -1246,7 +1194,9 @@ fn compute_exhaustiveness_and_usefulness<'p, 'tcx>(
// Analyze the constructors present in this column.
let ctors = matrix.heads().map(|p| p.ctor());
let split_set = ConstructorSet::for_ty(cx, ty).split(pcx, ctors);
let ctors_for_ty = &cx.ctors_for_ty(ty);
let is_integers = matches!(ctors_for_ty, ConstructorSet::Integers { .. }); // For diagnostics.
let split_set = ctors_for_ty.split(pcx, ctors);
let all_missing = split_set.present.is_empty();
// Build the set of constructors we will specialize with. It must cover the whole type.
@ -1261,7 +1211,7 @@ fn compute_exhaustiveness_and_usefulness<'p, 'tcx>(
}
// Decide what constructors to report.
let always_report_all = is_top_level && !IntRange::is_integral(pcx.ty);
let always_report_all = is_top_level && !is_integers;
// Whether we should report "Enum::A and Enum::C are missing" or "_ is missing".
let report_individual_missing_ctors = always_report_all || !all_missing;
// Which constructors are considered missing. We ensure that `!missing_ctors.is_empty() =>
@ -1318,233 +1268,9 @@ fn compute_exhaustiveness_and_usefulness<'p, 'tcx>(
ret
}
/// A column of patterns in the matrix, where a column is the intuitive notion of "subpatterns that
/// inspect the same subvalue/place".
/// This is used to traverse patterns column-by-column for lints. Despite similarities with
/// [`compute_exhaustiveness_and_usefulness`], this does a different traversal. Notably this is
/// linear in the depth of patterns, whereas `compute_exhaustiveness_and_usefulness` is worst-case
/// exponential (exhaustiveness is NP-complete). The core difference is that we treat sub-columns
/// separately.
///
/// This must not contain an or-pattern. `specialize` takes care to expand them.
///
/// This is not used in the main algorithm; only in lints.
#[derive(Debug)]
struct PatternColumn<'p, 'tcx> {
patterns: Vec<&'p DeconstructedPat<'p, 'tcx>>,
}
impl<'p, 'tcx> PatternColumn<'p, 'tcx> {
fn new(patterns: Vec<&'p DeconstructedPat<'p, 'tcx>>) -> Self {
Self { patterns }
}
fn is_empty(&self) -> bool {
self.patterns.is_empty()
}
fn head_ty(&self) -> Option<Ty<'tcx>> {
if self.patterns.len() == 0 {
return None;
}
// If the type is opaque and it is revealed anywhere in the column, we take the revealed
// version. Otherwise we could encounter constructors for the revealed type and crash.
let is_opaque = |ty: Ty<'tcx>| matches!(ty.kind(), ty::Alias(ty::Opaque, ..));
let first_ty = self.patterns[0].ty();
if is_opaque(first_ty) {
for pat in &self.patterns {
let ty = pat.ty();
if !is_opaque(ty) {
return Some(ty);
}
}
}
Some(first_ty)
}
/// Do constructor splitting on the constructors of the column.
fn analyze_ctors(&self, pcx: &PatCtxt<'_, 'p, 'tcx>) -> SplitConstructorSet<'tcx> {
let column_ctors = self.patterns.iter().map(|p| p.ctor());
ConstructorSet::for_ty(pcx.cx, pcx.ty).split(pcx, column_ctors)
}
fn iter<'a>(&'a self) -> impl Iterator<Item = &'p DeconstructedPat<'p, 'tcx>> + Captures<'a> {
self.patterns.iter().copied()
}
/// Does specialization: given a constructor, this takes the patterns from the column that match
/// the constructor, and outputs their fields.
/// This returns one column per field of the constructor. They usually all have the same length
/// (the number of patterns in `self` that matched `ctor`), except that we expand or-patterns
/// which may change the lengths.
fn specialize(&self, pcx: &PatCtxt<'_, 'p, 'tcx>, ctor: &Constructor<'tcx>) -> Vec<Self> {
let arity = ctor.arity(pcx);
if arity == 0 {
return Vec::new();
}
// We specialize the column by `ctor`. This gives us `arity`-many columns of patterns. These
// columns may have different lengths in the presence of or-patterns (this is why we can't
// reuse `Matrix`).
let mut specialized_columns: Vec<_> =
(0..arity).map(|_| Self { patterns: Vec::new() }).collect();
let relevant_patterns =
self.patterns.iter().filter(|pat| ctor.is_covered_by(pcx, pat.ctor()));
for pat in relevant_patterns {
let specialized = pat.specialize(pcx, ctor);
for (subpat, column) in specialized.iter().zip(&mut specialized_columns) {
if subpat.is_or_pat() {
column.patterns.extend(subpat.flatten_or_pat())
} else {
column.patterns.push(subpat)
}
}
}
assert!(
!specialized_columns[0].is_empty(),
"ctor {ctor:?} was listed as present but isn't;
there is an inconsistency between `Constructor::is_covered_by` and `ConstructorSet::split`"
);
specialized_columns
}
}
/// Traverse the patterns to collect any variants of a non_exhaustive enum that fail to be mentioned
/// in a given column.
#[instrument(level = "debug", skip(cx), ret)]
fn collect_nonexhaustive_missing_variants<'p, 'tcx>(
cx: &MatchCheckCtxt<'p, 'tcx>,
column: &PatternColumn<'p, 'tcx>,
) -> Vec<WitnessPat<'tcx>> {
let Some(ty) = column.head_ty() else {
return Vec::new();
};
let pcx = &PatCtxt::new_dummy(cx, ty);
let set = column.analyze_ctors(pcx);
if set.present.is_empty() {
// We can't consistently handle the case where no constructors are present (since this would
// require digging deep through any type in case there's a non_exhaustive enum somewhere),
// so for consistency we refuse to handle the top-level case, where we could handle it.
return vec![];
}
let mut witnesses = Vec::new();
if cx.is_foreign_non_exhaustive_enum(ty) {
witnesses.extend(
set.missing
.into_iter()
// This will list missing visible variants.
.filter(|c| !matches!(c, Constructor::Hidden | Constructor::NonExhaustive))
.map(|missing_ctor| WitnessPat::wild_from_ctor(pcx, missing_ctor)),
)
}
// Recurse into the fields.
for ctor in set.present {
let specialized_columns = column.specialize(pcx, &ctor);
let wild_pat = WitnessPat::wild_from_ctor(pcx, ctor);
for (i, col_i) in specialized_columns.iter().enumerate() {
// Compute witnesses for each column.
let wits_for_col_i = collect_nonexhaustive_missing_variants(cx, col_i);
// For each witness, we build a new pattern in the shape of `ctor(_, _, wit, _, _)`,
// adding enough wildcards to match `arity`.
for wit in wits_for_col_i {
let mut pat = wild_pat.clone();
pat.fields[i] = wit;
witnesses.push(pat);
}
}
}
witnesses
}
/// Traverse the patterns to warn the user about ranges that overlap on their endpoints.
#[instrument(level = "debug", skip(cx))]
fn lint_overlapping_range_endpoints<'p, 'tcx>(
cx: &MatchCheckCtxt<'p, 'tcx>,
column: &PatternColumn<'p, 'tcx>,
) {
let Some(ty) = column.head_ty() else {
return;
};
let pcx = &PatCtxt::new_dummy(cx, ty);
let set = column.analyze_ctors(pcx);
if IntRange::is_integral(ty) {
let emit_lint = |overlap: &IntRange, this_span: Span, overlapped_spans: &[Span]| {
let overlap_as_pat = overlap.to_diagnostic_pat(ty, cx.tcx);
let overlaps: Vec<_> = overlapped_spans
.iter()
.copied()
.map(|span| Overlap { range: overlap_as_pat.clone(), span })
.collect();
cx.tcx.emit_spanned_lint(
lint::builtin::OVERLAPPING_RANGE_ENDPOINTS,
cx.match_lint_level,
this_span,
OverlappingRangeEndpoints { overlap: overlaps, range: this_span },
);
};
// If two ranges overlapped, the split set will contain their intersection as a singleton.
let split_int_ranges = set.present.iter().filter_map(|c| c.as_int_range());
for overlap_range in split_int_ranges.clone() {
if overlap_range.is_singleton() {
let overlap: MaybeInfiniteInt = overlap_range.lo;
// Ranges that look like `lo..=overlap`.
let mut prefixes: SmallVec<[_; 1]> = Default::default();
// Ranges that look like `overlap..=hi`.
let mut suffixes: SmallVec<[_; 1]> = Default::default();
// Iterate on patterns that contained `overlap`.
for pat in column.iter() {
let this_span = pat.span();
let Constructor::IntRange(this_range) = pat.ctor() else { continue };
if this_range.is_singleton() {
// Don't lint when one of the ranges is a singleton.
continue;
}
if this_range.lo == overlap {
// `this_range` looks like `overlap..=this_range.hi`; it overlaps with any
// ranges that look like `lo..=overlap`.
if !prefixes.is_empty() {
emit_lint(overlap_range, this_span, &prefixes);
}
suffixes.push(this_span)
} else if this_range.hi == overlap.plus_one() {
// `this_range` looks like `this_range.lo..=overlap`; it overlaps with any
// ranges that look like `overlap..=hi`.
if !suffixes.is_empty() {
emit_lint(overlap_range, this_span, &suffixes);
}
prefixes.push(this_span)
}
}
}
}
} else {
// Recurse into the fields.
for ctor in set.present {
for col in column.specialize(pcx, &ctor) {
lint_overlapping_range_endpoints(cx, &col);
}
}
}
}
/// The arm of a match expression.
#[derive(Clone, Copy, Debug)]
pub(crate) struct MatchArm<'p, 'tcx> {
/// The pattern must have been lowered through `check_match::MatchVisitor::lower_pattern`.
pub(crate) pat: &'p DeconstructedPat<'p, 'tcx>,
pub(crate) hir_id: HirId,
pub(crate) has_guard: bool,
}
/// Indicates whether or not a given arm is useful.
#[derive(Clone, Debug)]
pub(crate) enum Usefulness {
pub enum Usefulness {
/// The arm is useful. This additionally carries a set of or-pattern branches that have been
/// found to be redundant despite the overall arm being useful. Used only in the presence of
/// or-patterns, otherwise it stays empty.
@ -1555,16 +1281,15 @@ pub(crate) enum Usefulness {
}
/// The output of checking a match for exhaustiveness and arm usefulness.
pub(crate) struct UsefulnessReport<'p, 'tcx> {
pub struct UsefulnessReport<'p, 'tcx> {
/// For each arm of the input, whether that arm is useful after the arms above it.
pub(crate) arm_usefulness: Vec<(MatchArm<'p, 'tcx>, Usefulness)>,
pub arm_usefulness: Vec<(MatchArm<'p, 'tcx>, Usefulness)>,
/// If the match is exhaustive, this is empty. If not, this contains witnesses for the lack of
/// exhaustiveness.
pub(crate) non_exhaustiveness_witnesses: Vec<WitnessPat<'tcx>>,
pub non_exhaustiveness_witnesses: Vec<WitnessPat<'tcx>>,
}
/// The entrypoint for this file. Computes whether a match is exhaustive and which of its arms are
/// useful.
/// Computes whether a match is exhaustive and which of its arms are useful.
#[instrument(skip(cx, arms), level = "debug")]
pub(crate) fn compute_match_usefulness<'p, 'tcx>(
cx: &MatchCheckCtxt<'p, 'tcx>,
@ -1590,59 +1315,5 @@ pub(crate) fn compute_match_usefulness<'p, 'tcx>(
(arm, usefulness)
})
.collect();
let report = UsefulnessReport { arm_usefulness, non_exhaustiveness_witnesses };
let pat_column = PatternColumn::new(matrix.heads().collect());
// Lint on ranges that overlap on their endpoints, which is likely a mistake.
lint_overlapping_range_endpoints(cx, &pat_column);
// Run the non_exhaustive_omitted_patterns lint. Only run on refutable patterns to avoid hitting
// `if let`s. Only run if the match is exhaustive otherwise the error is redundant.
if cx.refutable && report.non_exhaustiveness_witnesses.is_empty() {
if !matches!(
cx.tcx.lint_level_at_node(NON_EXHAUSTIVE_OMITTED_PATTERNS, cx.match_lint_level).0,
rustc_session::lint::Level::Allow
) {
let witnesses = collect_nonexhaustive_missing_variants(cx, &pat_column);
if !witnesses.is_empty() {
// Report that a match of a `non_exhaustive` enum marked with `non_exhaustive_omitted_patterns`
// is not exhaustive enough.
//
// NB: The partner lint for structs lives in `compiler/rustc_hir_analysis/src/check/pat.rs`.
cx.tcx.emit_spanned_lint(
NON_EXHAUSTIVE_OMITTED_PATTERNS,
cx.match_lint_level,
cx.scrut_span,
NonExhaustiveOmittedPattern {
scrut_ty,
uncovered: Uncovered::new(cx.scrut_span, cx, witnesses),
},
);
}
} else {
// We used to allow putting the `#[allow(non_exhaustive_omitted_patterns)]` on a match
// arm. This no longer makes sense so we warn users, to avoid silently breaking their
// usage of the lint.
for arm in arms {
let (lint_level, lint_level_source) =
cx.tcx.lint_level_at_node(NON_EXHAUSTIVE_OMITTED_PATTERNS, arm.hir_id);
if !matches!(lint_level, rustc_session::lint::Level::Allow) {
let decorator = NonExhaustiveOmittedPatternLintOnArm {
lint_span: lint_level_source.span(),
suggest_lint_on_match: cx.whole_match_span.map(|span| span.shrink_to_lo()),
lint_level: lint_level.as_str(),
lint_name: "non_exhaustive_omitted_patterns",
};
use rustc_errors::DecorateLint;
let mut err = cx.tcx.sess.struct_span_warn(arm.pat.span(), "");
err.set_primary_message(decorator.msg());
decorator.decorate_lint(&mut err);
err.emit();
}
}
}
}
report
UsefulnessReport { arm_usefulness, non_exhaustiveness_witnesses }
}

4
triagebot.toml
View File

@ -480,6 +480,10 @@ cc = ["@GuillaumeGomez"]
message = "Some changes might have occurred in exhaustiveness checking"
cc = ["@Nadrieril"]
[mentions."compiler/rustc_pattern_analysis"]
message = "Some changes might have occurred in exhaustiveness checking"
cc = ["@Nadrieril"]
[mentions."library/portable-simd"]
message = """
Portable SIMD is developed in its own repository. If possible, consider \