1 use clippy_utils::diagnostics::span_lint_and_sugg;
2 use clippy_utils::is_lint_allowed;
3 use clippy_utils::source::snippet;
4 use rustc_ast::ast::LitKind;
5 use rustc_errors::Applicability;
6 use rustc_hir::{Expr, ExprKind, HirId};
7 use rustc_lint::{LateContext, LateLintPass};
8 use rustc_session::{declare_lint_pass, declare_tool_lint};
9 use rustc_span::source_map::Span;
10 use unicode_normalization::UnicodeNormalization;
12 declare_clippy_lint! {
13 /// **What it does:** Checks for invisible Unicode characters in the code.
15 /// **Why is this bad?** Having an invisible character in the code makes for all
16 /// sorts of April fools, but otherwise is very much frowned upon.
18 /// **Known problems:** None.
20 /// **Example:** You don't see it, but there may be a zero-width space or soft hyphen
21 /// somewhere in this text.
22 pub INVISIBLE_CHARACTERS,
24 "using an invisible character in a string literal, which is confusing"
27 declare_clippy_lint! {
28 /// **What it does:** Checks for non-ASCII characters in string literals.
30 /// **Why is this bad?** Yeah, we know, the 90's called and wanted their charset
31 /// back. Even so, there still are editors and other programs out there that
32 /// don't work well with Unicode. So if the code is meant to be used
33 /// internationally, on multiple operating systems, or has other portability
34 /// requirements, activating this lint could be useful.
36 /// **Known problems:** None.
40 /// let x = String::from("€");
42 /// Could be written as:
44 /// let x = String::from("\u{20ac}");
46 pub NON_ASCII_LITERAL,
48 "using any literal non-ASCII chars in a string literal instead of using the `\\u` escape"
51 declare_clippy_lint! {
52 /// **What it does:** Checks for string literals that contain Unicode in a form
53 /// that is not equal to its
54 /// [NFC-recomposition](http://www.unicode.org/reports/tr15/#Norm_Forms).
56 /// **Why is this bad?** If such a string is compared to another, the results
57 /// may be surprising.
59 /// **Known problems** None.
61 /// **Example:** You may not see it, but "à"" and "à"" aren't the same string. The
62 /// former when escaped is actually `"a\u{300}"` while the latter is `"\u{e0}"`.
65 "using a Unicode literal not in NFC normal form (see [Unicode tr15](http://www.unicode.org/reports/tr15/) for further information)"
68 declare_lint_pass!(Unicode => [INVISIBLE_CHARACTERS, NON_ASCII_LITERAL, UNICODE_NOT_NFC]);
70 impl LateLintPass<'_> for Unicode {
71 fn check_expr(&mut self, cx: &LateContext<'_>, expr: &'_ Expr<'_>) {
72 if let ExprKind::Lit(ref lit) = expr.kind {
73 if let LitKind::Str(_, _) = lit.node {
74 check_str(cx, lit.span, expr.hir_id);
80 fn escape<T: Iterator<Item = char>>(s: T) -> String {
81 let mut result = String::new();
84 for d in c.escape_unicode() {
94 fn check_str(cx: &LateContext<'_>, span: Span, id: HirId) {
95 let string = snippet(cx, span, "");
96 if string.chars().any(|c| ['\u{200B}', '\u{ad}', '\u{2060}'].contains(&c)) {
101 "invisible character detected",
102 "consider replacing the string with",
104 .replace("\u{200B}", "\\u{200B}")
105 .replace("\u{ad}", "\\u{AD}")
106 .replace("\u{2060}", "\\u{2060}"),
107 Applicability::MachineApplicable,
110 if string.chars().any(|c| c as u32 > 0x7F) {
115 "literal non-ASCII character detected",
116 "consider replacing the string with",
117 if is_lint_allowed(cx, UNICODE_NOT_NFC, id) {
118 escape(string.chars())
122 Applicability::MachineApplicable,
125 if is_lint_allowed(cx, NON_ASCII_LITERAL, id) && string.chars().zip(string.nfc()).any(|(a, b)| a != b) {
130 "non-NFC Unicode sequence detected",
131 "consider replacing the string with",
132 string.nfc().collect::<String>(),
133 Applicability::MachineApplicable,