Rollup merge of rust-lang#108200 - jhpratt:restricted-damerau-levenshtein-distance, r=tmiasko

Use restricted Damerau-Levenshtein distance for diagnostics

This replaces the existing Levenshtein algorithm with the Damerau-Levenshtein algorithm. This means that "ab" to "ba" is one change (a transposition) instead of two (a deletion and insertion). More specifically, this is a _restricted_ implementation, in that "ca" to "abc" cannot be performed as "ca" → "ac" → "abc", as there is an insertion in the middle of a transposition. I believe that errors like that are sufficiently rare that it's not worth taking into account.

This was first brought up [on IRLO](https://internals.rust-lang.org/t/18227) when it was noticed that the diagnostic for `prinltn!` (transposed L and T) was `print!` and not `println!`. Only a single existing UI test was effected, with the result being an objective improvement.

~~I have left the method name and various other references to the Levenshtein algorithm untouched, as the exact manner in which the edit distance is calculated should not be relevant to the caller.~~

r? ``@estebank``

``@rustbot`` label +A-diagnostics +C-enhancement

Author: matthiaskrgr

compiler/rustc_ast_lowering/src/item.rs

@@ -13,7 +13,7 @@ use rustc_hir::def_id::{LocalDefId, CRATE_DEF_ID};
 use rustc_hir::PredicateOrigin;
 use rustc_index::vec::{Idx, IndexVec};
 use rustc_middle::ty::{DefIdTree, ResolverAstLowering, TyCtxt};
-use rustc_span::lev_distance::find_best_match_for_name;
+use rustc_span::edit_distance::find_best_match_for_name;
 use rustc_span::source_map::DesugaringKind;
 use rustc_span::symbol::{kw, sym, Ident};
 use rustc_span::{Span, Symbol};

compiler/rustc_hir_analysis/src/astconv/errors.rs

@@ -7,7 +7,7 @@ use rustc_hir::def_id::DefId;
 use rustc_infer::traits::FulfillmentError;
 use rustc_middle::ty::{self, Ty};
 use rustc_session::parse::feature_err;
-use rustc_span::lev_distance::find_best_match_for_name;
+use rustc_span::edit_distance::find_best_match_for_name;
 use rustc_span::symbol::{sym, Ident};
 use rustc_span::{Span, Symbol, DUMMY_SP};
 

compiler/rustc_hir_analysis/src/astconv/mod.rs

@@ -37,8 +37,8 @@ use rustc_middle::ty::DynKind;
 use rustc_middle::ty::GenericParamDefKind;
 use rustc_middle::ty::{self, Const, DefIdTree, IsSuggestable, Ty, TyCtxt, TypeVisitable};
 use rustc_session::lint::builtin::{AMBIGUOUS_ASSOCIATED_ITEMS, BARE_TRAIT_OBJECTS};
+use rustc_span::edit_distance::find_best_match_for_name;
 use rustc_span::edition::Edition;
-use rustc_span::lev_distance::find_best_match_for_name;
 use rustc_span::symbol::{kw, Ident, Symbol};
 use rustc_span::{sym, Span, DUMMY_SP};
 use rustc_target::spec::abi;

compiler/rustc_hir_typeck/src/expr.rs

@@ -45,8 +45,8 @@ use rustc_middle::ty::subst::SubstsRef;
 use rustc_middle::ty::{self, AdtKind, Ty, TypeVisitable};
 use rustc_session::errors::ExprParenthesesNeeded;
 use rustc_session::parse::feature_err;
+use rustc_span::edit_distance::find_best_match_for_name;
 use rustc_span::hygiene::DesugaringKind;
-use rustc_span::lev_distance::find_best_match_for_name;
 use rustc_span::source_map::{Span, Spanned};
 use rustc_span::symbol::{kw, sym, Ident, Symbol};
 use rustc_target::spec::abi::Abi::RustIntrinsic;

compiler/rustc_hir_typeck/src/method/probe.rs

@@ -24,8 +24,8 @@ use rustc_middle::ty::{InternalSubsts, SubstsRef};
 use rustc_session::lint;
 use rustc_span::def_id::DefId;
 use rustc_span::def_id::LocalDefId;
-use rustc_span::lev_distance::{
-    find_best_match_for_name_with_substrings, lev_distance_with_substrings,
+use rustc_span::edit_distance::{
+    edit_distance_with_substrings, find_best_match_for_name_with_substrings,
 };
 use rustc_span::symbol::sym;
 use rustc_span::{symbol::Ident, Span, Symbol, DUMMY_SP};
@@ -69,7 +69,7 @@ struct ProbeContext<'a, 'tcx> {
     impl_dups: FxHashSet<DefId>,
 
     /// When probing for names, include names that are close to the
-    /// requested name (by Levenshtein distance)
+    /// requested name (by edit distance)
     allow_similar_names: bool,
 
     /// Some(candidate) if there is a private candidate
@@ -1793,7 +1793,7 @@ impl<'a, 'tcx> ProbeContext<'a, 'tcx> {
 
     /// Similarly to `probe_for_return_type`, this method attempts to find the best matching
     /// candidate method where the method name may have been misspelled. Similarly to other
-    /// Levenshtein based suggestions, we provide at most one such suggestion.
+    /// edit distance based suggestions, we provide at most one such suggestion.
     fn probe_for_similar_candidate(&mut self) -> Result<Option<ty::AssocItem>, MethodError<'tcx>> {
         debug!("probing for method names similar to {:?}", self.method_name);
 
@@ -2024,8 +2024,11 @@ impl<'a, 'tcx> ProbeContext<'a, 'tcx> {
                         if self.matches_by_doc_alias(x.def_id) {
                             return true;
                         }
-                        match lev_distance_with_substrings(name.as_str(), x.name.as_str(), max_dist)
-                        {
+                        match edit_distance_with_substrings(
+                            name.as_str(),
+                            x.name.as_str(),
+                            max_dist,
+                        ) {
                             Some(d) => d > 0,
                             None => false,
                         }

compiler/rustc_hir_typeck/src/method/suggest.rs

@@ -31,7 +31,7 @@ use rustc_middle::ty::{self, DefIdTree, GenericArgKind, Ty, TyCtxt, TypeVisitabl
 use rustc_middle::ty::{IsSuggestable, ToPolyTraitRef};
 use rustc_span::symbol::{kw, sym, Ident};
 use rustc_span::Symbol;
-use rustc_span::{lev_distance, source_map, ExpnKind, FileName, MacroKind, Span};
+use rustc_span::{edit_distance, source_map, ExpnKind, FileName, MacroKind, Span};
 use rustc_trait_selection::traits::error_reporting::on_unimplemented::OnUnimplementedNote;
 use rustc_trait_selection::traits::error_reporting::on_unimplemented::TypeErrCtxtExt as _;
 use rustc_trait_selection::traits::query::evaluate_obligation::InferCtxtExt as _;
@@ -1014,7 +1014,7 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {
         // that had unsatisfied trait bounds
         if unsatisfied_predicates.is_empty() && rcvr_ty.is_enum() {
             let adt_def = rcvr_ty.ty_adt_def().expect("enum is not an ADT");
-            if let Some(suggestion) = lev_distance::find_best_match_for_name(
+            if let Some(suggestion) = edit_distance::find_best_match_for_name(
                 &adt_def.variants().iter().map(|s| s.name).collect::<Vec<_>>(),
                 item_name.name,
                 None,

compiler/rustc_hir_typeck/src/pat.rs

@@ -14,8 +14,8 @@ use rustc_infer::infer::type_variable::{TypeVariableOrigin, TypeVariableOriginKi
 use rustc_middle::middle::stability::EvalResult;
 use rustc_middle::ty::{self, Adt, BindingMode, Ty, TypeVisitable};
 use rustc_session::lint::builtin::NON_EXHAUSTIVE_OMITTED_PATTERNS;
+use rustc_span::edit_distance::find_best_match_for_name;
 use rustc_span::hygiene::DesugaringKind;
-use rustc_span::lev_distance::find_best_match_for_name;
 use rustc_span::source_map::{Span, Spanned};
 use rustc_span::symbol::{kw, sym, Ident};
 use rustc_span::{BytePos, DUMMY_SP};

compiler/rustc_interface/src/util.rs

@@ -14,8 +14,8 @@ use rustc_session::filesearch::sysroot_candidates;
 use rustc_session::lint::{self, BuiltinLintDiagnostics, LintBuffer};
 use rustc_session::parse::CrateConfig;
 use rustc_session::{early_error, filesearch, output, Session};
+use rustc_span::edit_distance::find_best_match_for_name;
 use rustc_span::edition::Edition;
-use rustc_span::lev_distance::find_best_match_for_name;
 use rustc_span::source_map::FileLoader;
 use rustc_span::symbol::{sym, Symbol};
 use session::CompilerIO;

compiler/rustc_lint/src/context.rs

@@ -39,7 +39,7 @@ use rustc_middle::ty::{self, print::Printer, subst::GenericArg, RegisteredTools,
 use rustc_session::lint::{BuiltinLintDiagnostics, LintExpectationId};
 use rustc_session::lint::{FutureIncompatibleInfo, Level, Lint, LintBuffer, LintId};
 use rustc_session::Session;
-use rustc_span::lev_distance::find_best_match_for_name;
+use rustc_span::edit_distance::find_best_match_for_name;
 use rustc_span::symbol::{sym, Ident, Symbol};
 use rustc_span::{BytePos, Span};
 use rustc_target::abi;

compiler/rustc_parse/src/parser/item.rs

@@ -19,8 +19,8 @@ use rustc_errors::{
     struct_span_err, Applicability, DiagnosticBuilder, ErrorGuaranteed, IntoDiagnostic, PResult,
     StashKey,
 };
+use rustc_span::edit_distance::edit_distance;
 use rustc_span::edition::Edition;
-use rustc_span::lev_distance::lev_distance;
 use rustc_span::source_map::{self, Span};
 use rustc_span::symbol::{kw, sym, Ident, Symbol};
 use rustc_span::DUMMY_SP;
@@ -459,7 +459,8 @@ impl<'a> Parser<'a> {
                 // Maybe the user misspelled `macro_rules` (issue #91227)
                 if self.token.is_ident()
                     && path.segments.len() == 1
-                    && lev_distance("macro_rules", &path.segments[0].ident.to_string(), 3).is_some()
+                    && edit_distance("macro_rules", &path.segments[0].ident.to_string(), 2)
+                        .is_some()
                 {
                     err.span_suggestion(
                         path.span,

compiler/rustc_resolve/src/diagnostics.rs

@@ -21,9 +21,9 @@ use rustc_session::lint::builtin::ABSOLUTE_PATHS_NOT_STARTING_WITH_CRATE;
 use rustc_session::lint::builtin::MACRO_EXPANDED_MACRO_EXPORTS_ACCESSED_BY_ABSOLUTE_PATHS;
 use rustc_session::lint::BuiltinLintDiagnostics;
 use rustc_session::Session;
+use rustc_span::edit_distance::find_best_match_for_name;
 use rustc_span::edition::Edition;
 use rustc_span::hygiene::MacroKind;
-use rustc_span::lev_distance::find_best_match_for_name;
 use rustc_span::source_map::SourceMap;
 use rustc_span::symbol::{kw, sym, Ident, Symbol};
 use rustc_span::{BytePos, Span, SyntaxContext};

compiler/rustc_resolve/src/imports.rs

@@ -21,8 +21,8 @@ use rustc_middle::span_bug;
 use rustc_middle::ty;
 use rustc_session::lint::builtin::{PUB_USE_OF_PRIVATE_EXTERN_CRATE, UNUSED_IMPORTS};
 use rustc_session::lint::BuiltinLintDiagnostics;
+use rustc_span::edit_distance::find_best_match_for_name;
 use rustc_span::hygiene::LocalExpnId;
-use rustc_span::lev_distance::find_best_match_for_name;
 use rustc_span::symbol::{kw, Ident, Symbol};
 use rustc_span::Span;
 

compiler/rustc_resolve/src/late/diagnostics.rs

@@ -25,9 +25,9 @@ use rustc_middle::ty::DefIdTree;
 use rustc_session::lint;
 use rustc_session::parse::feature_err;
 use rustc_session::Session;
+use rustc_span::edit_distance::find_best_match_for_name;
 use rustc_span::edition::Edition;
 use rustc_span::hygiene::MacroKind;
-use rustc_span::lev_distance::find_best_match_for_name;
 use rustc_span::symbol::{kw, sym, Ident, Symbol};
 use rustc_span::{BytePos, Span};
 
@@ -542,7 +542,7 @@ impl<'a: 'ast, 'ast, 'tcx> LateResolutionVisitor<'a, '_, 'ast, 'tcx> {
             }
         }
 
-        // Try Levenshtein algorithm.
+        // Try finding a suitable replacement.
         let typo_sugg =
             self.lookup_typo_candidate(path, source.namespace(), is_expected).to_opt_suggestion();
         if path.len() == 1 && self.self_type_is_available() {
@@ -770,7 +770,7 @@ impl<'a: 'ast, 'ast, 'tcx> LateResolutionVisitor<'a, '_, 'ast, 'tcx> {
                 _ => {}
             }
 
-            // If the trait has a single item (which wasn't matched by Levenshtein), suggest it
+            // If the trait has a single item (which wasn't matched by the algorithm), suggest it
             let suggestion = self.get_single_associated_item(&path, &source, is_expected);
             if !self.r.add_typo_suggestion(err, suggestion, ident_span) {
                 fallback = !self.let_binding_suggestion(err, ident_span);