1 use crate::consts::{constant, Constant};
2 use crate::utils::{match_def_path, paths, span_lint, span_lint_and_help};
3 use if_chain::if_chain;
4 use rustc_ast::ast::{LitKind, StrStyle};
5 use rustc_data_structures::fx::FxHashSet;
6 use rustc_hir::{BorrowKind, Expr, ExprKind, HirId};
7 use rustc_lint::{LateContext, LateLintPass};
8 use rustc_session::{declare_tool_lint, impl_lint_pass};
9 use rustc_span::source_map::{BytePos, Span};
10 use std::convert::TryFrom;
12 declare_clippy_lint! {
13 /// **What it does:** Checks [regex](https://crates.io/crates/regex) creation
14 /// (with `Regex::new`,`RegexBuilder::new` or `RegexSet::new`) for correct
17 /// **Why is this bad?** This will lead to a runtime panic.
19 /// **Known problems:** None.
27 "invalid regular expressions"
30 declare_clippy_lint! {
31 /// **What it does:** Checks for trivial [regex](https://crates.io/crates/regex)
32 /// creation (with `Regex::new`, `RegexBuilder::new` or `RegexSet::new`).
34 /// **Why is this bad?** Matching the regex can likely be replaced by `==` or
35 /// `str::starts_with`, `str::ends_with` or `std::contains` or other `str`
38 /// **Known problems:** None.
42 /// Regex::new("^foobar")
46 "trivial regular expressions"
49 #[derive(Clone, Default)]
51 spans: FxHashSet<Span>,
55 impl_lint_pass!(Regex => [INVALID_REGEX, TRIVIAL_REGEX]);
57 impl<'tcx> LateLintPass<'tcx> for Regex {
58 fn check_expr(&mut self, cx: &LateContext<'tcx>, expr: &'tcx Expr<'_>) {
60 if let ExprKind::Call(ref fun, ref args) = expr.kind;
61 if let ExprKind::Path(ref qpath) = fun.kind;
63 if let Some(def_id) = cx.qpath_res(qpath, fun.hir_id).opt_def_id();
65 if match_def_path(cx, def_id, &paths::REGEX_NEW) ||
66 match_def_path(cx, def_id, &paths::REGEX_BUILDER_NEW) {
67 check_regex(cx, &args[0], true);
68 } else if match_def_path(cx, def_id, &paths::REGEX_BYTES_NEW) ||
69 match_def_path(cx, def_id, &paths::REGEX_BYTES_BUILDER_NEW) {
70 check_regex(cx, &args[0], false);
71 } else if match_def_path(cx, def_id, &paths::REGEX_SET_NEW) {
72 check_set(cx, &args[0], true);
73 } else if match_def_path(cx, def_id, &paths::REGEX_BYTES_SET_NEW) {
74 check_set(cx, &args[0], false);
81 #[allow(clippy::cast_possible_truncation)] // truncation very unlikely here
83 fn str_span(base: Span, c: regex_syntax::ast::Span, offset: u16) -> Span {
84 let offset = u32::from(offset);
85 let end = base.lo() + BytePos(u32::try_from(c.end.offset).expect("offset too large") + offset);
86 let start = base.lo() + BytePos(u32::try_from(c.start.offset).expect("offset too large") + offset);
87 assert!(start <= end);
88 Span::new(start, end, base.ctxt())
91 fn const_str<'tcx>(cx: &LateContext<'tcx>, e: &'tcx Expr<'_>) -> Option<String> {
92 constant(cx, cx.typeck_results(), e).and_then(|(c, _)| match c {
93 Constant::Str(s) => Some(s),
98 fn is_trivial_regex(s: ®ex_syntax::hir::Hir) -> Option<&'static str> {
99 use regex_syntax::hir::Anchor::{EndText, StartText};
100 use regex_syntax::hir::HirKind::{Alternation, Anchor, Concat, Empty, Literal};
102 let is_literal = |e: &[regex_syntax::hir::Hir]| e.iter().all(|e| matches!(*e.kind(), Literal(_)));
105 Empty | Anchor(_) => Some("the regex is unlikely to be useful as it is"),
106 Literal(_) => Some("consider using `str::contains`"),
107 Alternation(ref exprs) => {
108 if exprs.iter().all(|e| e.kind().is_empty()) {
109 Some("the regex is unlikely to be useful as it is")
114 Concat(ref exprs) => match (exprs[0].kind(), exprs[exprs.len() - 1].kind()) {
115 (&Anchor(StartText), &Anchor(EndText)) if exprs[1..(exprs.len() - 1)].is_empty() => {
116 Some("consider using `str::is_empty`")
118 (&Anchor(StartText), &Anchor(EndText)) if is_literal(&exprs[1..(exprs.len() - 1)]) => {
119 Some("consider using `==` on `str`s")
121 (&Anchor(StartText), &Literal(_)) if is_literal(&exprs[1..]) => Some("consider using `str::starts_with`"),
122 (&Literal(_), &Anchor(EndText)) if is_literal(&exprs[1..(exprs.len() - 1)]) => {
123 Some("consider using `str::ends_with`")
125 _ if is_literal(exprs) => Some("consider using `str::contains`"),
132 fn check_set<'tcx>(cx: &LateContext<'tcx>, expr: &'tcx Expr<'_>, utf8: bool) {
134 if let ExprKind::AddrOf(BorrowKind::Ref, _, ref expr) = expr.kind;
135 if let ExprKind::Array(exprs) = expr.kind;
138 check_regex(cx, expr, utf8);
144 fn check_regex<'tcx>(cx: &LateContext<'tcx>, expr: &'tcx Expr<'_>, utf8: bool) {
145 let mut parser = regex_syntax::ParserBuilder::new()
147 .allow_invalid_utf8(!utf8)
150 if let ExprKind::Lit(ref lit) = expr.kind {
151 if let LitKind::Str(ref r, style) = lit.node {
153 let offset = if let StrStyle::Raw(n) = style { 2 + n } else { 1 };
154 match parser.parse(r) {
156 if let Some(repl) = is_trivial_regex(&r) {
157 span_lint_and_help(cx, TRIVIAL_REGEX, expr.span, "trivial regex", None, repl);
160 Err(regex_syntax::Error::Parse(e)) => {
164 str_span(expr.span, *e.span(), offset),
165 &format!("regex syntax error: {}", e.kind()),
168 Err(regex_syntax::Error::Translate(e)) => {
172 str_span(expr.span, *e.span(), offset),
173 &format!("regex syntax error: {}", e.kind()),
177 span_lint(cx, INVALID_REGEX, expr.span, &format!("regex syntax error: {}", e));
181 } else if let Some(r) = const_str(cx, expr) {
182 match parser.parse(&r) {
184 if let Some(repl) = is_trivial_regex(&r) {
185 span_lint_and_help(cx, TRIVIAL_REGEX, expr.span, "trivial regex", None, repl);
188 Err(regex_syntax::Error::Parse(e)) => {
193 &format!("regex syntax error on position {}: {}", e.span().start.offset, e.kind()),
196 Err(regex_syntax::Error::Translate(e)) => {
201 &format!("regex syntax error on position {}: {}", e.span().start.offset, e.kind()),
205 span_lint(cx, INVALID_REGEX, expr.span, &format!("regex syntax error: {}", e));