1 use crate::{EarlyContext, EarlyLintPass, LintContext};
3 use rustc_data_structures::fx::FxHashMap;
4 use rustc_span::symbol::Symbol;
7 /// The `non_ascii_idents` lint detects non-ASCII identifiers.
11 /// ```rust,compile_fail
12 /// # #![allow(unused)]
13 /// #![deny(non_ascii_idents)]
23 /// This lint allows projects that wish to retain the limit of only using
24 /// ASCII characters to switch this lint to "forbid" (for example to ease
25 /// collaboration or for security reasons).
26 /// See [RFC 2457] for more details.
28 /// [RFC 2457]: https://github.com/rust-lang/rfcs/blob/master/text/2457-non-ascii-idents.md
31 "detects non-ASCII identifiers",
36 /// The `uncommon_codepoints` lint detects uncommon Unicode codepoints in
42 /// # #![allow(unused)]
43 /// const µ: f64 = 0.000001;
50 /// This lint warns about using characters which are not commonly used, and may
51 /// cause visual confusion.
53 /// This lint is triggered by identifiers that contain a codepoint that is
54 /// not part of the set of "Allowed" codepoints as described by [Unicode®
55 /// Technical Standard #39 Unicode Security Mechanisms Section 3.1 General
56 /// Security Profile for Identifiers][TR39Allowed].
58 /// Note that the set of uncommon codepoints may change over time. Beware
59 /// that if you "forbid" this lint that existing code may fail in the
62 /// [TR39Allowed]: https://www.unicode.org/reports/tr39/#General_Security_Profile
63 pub UNCOMMON_CODEPOINTS,
65 "detects uncommon Unicode codepoints in identifiers",
70 /// The `confusable_idents` lint detects visually confusable pairs between
76 /// // Latin Capital Letter E With Caron
77 /// pub const Ě: i32 = 1;
78 /// // Latin Capital Letter E With Breve
79 /// pub const Ĕ: i32 = 2;
86 /// This lint warns when different identifiers may appear visually similar,
87 /// which can cause confusion.
89 /// The confusable detection algorithm is based on [Unicode® Technical
90 /// Standard #39 Unicode Security Mechanisms Section 4 Confusable
91 /// Detection][TR39Confusable]. For every distinct identifier X execute
92 /// the function `skeleton(X)`. If there exist two distinct identifiers X
93 /// and Y in the same crate where `skeleton(X) = skeleton(Y)` report it.
94 /// The compiler uses the same mechanism to check if an identifier is too
95 /// similar to a keyword.
97 /// Note that the set of confusable characters may change over time.
98 /// Beware that if you "forbid" this lint that existing code may fail in
101 /// [TR39Confusable]: https://www.unicode.org/reports/tr39/#Confusable_Detection
102 pub CONFUSABLE_IDENTS,
104 "detects visually confusable pairs between identifiers",
109 /// The `mixed_script_confusables` lint detects visually confusable
110 /// characters in identifiers between different [scripts].
112 /// [scripts]: https://en.wikipedia.org/wiki/Script_(Unicode)
117 /// // The Japanese katakana character エ can be confused with the Han character 工.
118 /// const エ: &'static str = "アイウ";
125 /// This lint warns when characters between different scripts may appear
126 /// visually similar, which can cause confusion.
128 /// If the crate contains other identifiers in the same script that have
129 /// non-confusable characters, then this lint will *not* be issued. For
130 /// example, if the example given above has another identifier with
131 /// katakana characters (such as `let カタカナ = 123;`), then this indicates
132 /// that you are intentionally using katakana, and it will not warn about
135 /// Note that the set of confusable characters may change over time.
136 /// Beware that if you "forbid" this lint that existing code may fail in
138 pub MIXED_SCRIPT_CONFUSABLES,
140 "detects Unicode scripts whose mixed script confusables codepoints are solely used",
144 declare_lint_pass!(NonAsciiIdents => [NON_ASCII_IDENTS, UNCOMMON_CODEPOINTS, CONFUSABLE_IDENTS, MIXED_SCRIPT_CONFUSABLES]);
146 impl EarlyLintPass for NonAsciiIdents {
147 fn check_crate(&mut self, cx: &EarlyContext<'_>, _: &ast::Crate) {
148 use rustc_session::lint::Level;
149 use rustc_span::Span;
150 use std::collections::BTreeMap;
151 use unicode_security::GeneralSecurityProfile;
153 let check_non_ascii_idents = cx.builder.lint_level(NON_ASCII_IDENTS).0 != Level::Allow;
154 let check_uncommon_codepoints =
155 cx.builder.lint_level(UNCOMMON_CODEPOINTS).0 != Level::Allow;
156 let check_confusable_idents = cx.builder.lint_level(CONFUSABLE_IDENTS).0 != Level::Allow;
157 let check_mixed_script_confusables =
158 cx.builder.lint_level(MIXED_SCRIPT_CONFUSABLES).0 != Level::Allow;
160 if !check_non_ascii_idents
161 && !check_uncommon_codepoints
162 && !check_confusable_idents
163 && !check_mixed_script_confusables
168 let mut has_non_ascii_idents = false;
169 let symbols = cx.sess.parse_sess.symbol_gallery.symbols.lock();
171 // Sort by `Span` so that error messages make sense with respect to the
172 // order of identifier locations in the code.
173 let mut symbols: Vec<_> = symbols.iter().collect();
174 symbols.sort_by_key(|k| k.1);
176 for (symbol, &sp) in symbols.iter() {
177 let symbol_str = symbol.as_str();
178 if symbol_str.is_ascii() {
181 has_non_ascii_idents = true;
182 cx.struct_span_lint(NON_ASCII_IDENTS, sp, |lint| {
183 lint.build("identifier contains non-ASCII characters").emit()
185 if check_uncommon_codepoints
186 && !symbol_str.chars().all(GeneralSecurityProfile::identifier_allowed)
188 cx.struct_span_lint(UNCOMMON_CODEPOINTS, sp, |lint| {
189 lint.build("identifier contains uncommon Unicode codepoints").emit()
194 if has_non_ascii_idents && check_confusable_idents {
195 let mut skeleton_map: FxHashMap<Symbol, (Symbol, Span, bool)> =
196 FxHashMap::with_capacity_and_hasher(symbols.len(), Default::default());
197 let mut skeleton_buf = String::new();
199 for (&symbol, &sp) in symbols.iter() {
200 use unicode_security::confusable_detection::skeleton;
202 let symbol_str = symbol.as_str();
203 let is_ascii = symbol_str.is_ascii();
205 // Get the skeleton as a `Symbol`.
206 skeleton_buf.clear();
207 skeleton_buf.extend(skeleton(&symbol_str));
208 let skeleton_sym = if *symbol_str == *skeleton_buf {
211 Symbol::intern(&skeleton_buf)
216 .and_modify(|(existing_symbol, existing_span, existing_is_ascii)| {
217 if !*existing_is_ascii || !is_ascii {
218 cx.struct_span_lint(CONFUSABLE_IDENTS, sp, |lint| {
220 "identifier pair considered confusable between `{}` and `{}`",
221 existing_symbol.as_str(),
226 "this is where the previous identifier occurred",
231 if *existing_is_ascii && !is_ascii {
232 *existing_symbol = symbol;
234 *existing_is_ascii = is_ascii;
237 .or_insert((symbol, sp, is_ascii));
241 if has_non_ascii_idents && check_mixed_script_confusables {
242 use unicode_security::is_potential_mixed_script_confusable_char;
243 use unicode_security::mixed_script::AugmentedScriptSet;
246 enum ScriptSetUsage {
247 Suspicious(Vec<char>, Span),
251 let mut script_states: FxHashMap<AugmentedScriptSet, ScriptSetUsage> =
252 FxHashMap::default();
253 let latin_augmented_script_set = AugmentedScriptSet::for_char('A');
254 script_states.insert(latin_augmented_script_set, ScriptSetUsage::Verified);
256 let mut has_suspicous = false;
257 for (symbol, &sp) in symbols.iter() {
258 let symbol_str = symbol.as_str();
259 for ch in symbol_str.chars() {
261 // all ascii characters are covered by exception.
264 if !GeneralSecurityProfile::identifier_allowed(ch) {
265 // this character is covered by `uncommon_codepoints` lint.
268 let augmented_script_set = AugmentedScriptSet::for_char(ch);
270 .entry(augmented_script_set)
271 .and_modify(|existing_state| {
272 if let ScriptSetUsage::Suspicious(ch_list, _) = existing_state {
273 if is_potential_mixed_script_confusable_char(ch) {
276 *existing_state = ScriptSetUsage::Verified;
281 if !is_potential_mixed_script_confusable_char(ch) {
282 ScriptSetUsage::Verified
284 has_suspicous = true;
285 ScriptSetUsage::Suspicious(vec![ch], sp)
292 let verified_augmented_script_sets = script_states
294 .flat_map(|(k, v)| match v {
295 ScriptSetUsage::Verified => Some(*k),
298 .collect::<Vec<_>>();
300 // we're sorting the output here.
301 let mut lint_reports: BTreeMap<(Span, Vec<char>), AugmentedScriptSet> =
304 'outerloop: for (augment_script_set, usage) in script_states {
305 let (mut ch_list, sp) = match usage {
306 ScriptSetUsage::Verified => continue,
307 ScriptSetUsage::Suspicious(ch_list, sp) => (ch_list, sp),
310 if augment_script_set.is_all() {
314 for existing in verified_augmented_script_sets.iter() {
315 if existing.is_all() {
318 let mut intersect = *existing;
319 intersect.intersect_with(augment_script_set);
320 if !intersect.is_empty() && !intersect.is_all() {
325 // We sort primitive chars here and can use unstable sort
326 ch_list.sort_unstable();
328 lint_reports.insert((sp, ch_list), augment_script_set);
331 for ((sp, ch_list), script_set) in lint_reports {
332 cx.struct_span_lint(MIXED_SCRIPT_CONFUSABLES, sp, |lint| {
333 let message = format!(
334 "The usage of Script Group `{}` in this crate consists solely of mixed script confusables",
336 let mut note = "The usage includes ".to_string();
337 for (idx, ch) in ch_list.into_iter().enumerate() {
341 let char_info = format!("'{}' (U+{:04X})", ch, ch as u32);
345 lint.build(&message).note(¬e).note("Please recheck to make sure their usages are indeed what you want.").emit()