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