230 lines
7.8 KiB
Rust
230 lines
7.8 KiB
Rust
use regex_syntax;
|
|
use rustc::hir::*;
|
|
use rustc::lint::*;
|
|
use rustc::middle::const_val::ConstVal;
|
|
use rustc_const_eval::EvalHint::ExprTypeChecked;
|
|
use rustc_const_eval::eval_const_expr_partial;
|
|
use std::collections::HashSet;
|
|
use std::error::Error;
|
|
use syntax::ast::{LitKind, NodeId};
|
|
use syntax::codemap::{Span, BytePos};
|
|
use syntax::parse::token::InternedString;
|
|
use utils::{is_expn_of, match_def_path, match_type, paths, span_lint, span_help_and_lint};
|
|
|
|
/// **What it does:** This lint checks [regex] creation (with `Regex::new`, `RegexBuilder::new` or
|
|
/// `RegexSet::new`) for correct regex syntax.
|
|
///
|
|
/// **Why is this bad?** This will lead to a runtime panic.
|
|
///
|
|
/// **Known problems:** None.
|
|
///
|
|
/// **Example:** `Regex::new("|")`
|
|
declare_lint! {
|
|
pub INVALID_REGEX,
|
|
Deny,
|
|
"finds invalid regular expressions"
|
|
}
|
|
|
|
/// **What it does:** This lint checks for trivial [regex] creation (with `Regex::new`,
|
|
/// `RegexBuilder::new` or `RegexSet::new`).
|
|
///
|
|
/// **Why is this bad?** This can likely be replaced by `==` or `str::starts_with`,
|
|
/// `str::ends_with` or `std::contains` or other `str` methods.
|
|
///
|
|
/// **Known problems:** None.
|
|
///
|
|
/// **Example:** `Regex::new("^foobar")`
|
|
///
|
|
/// [regex]: https://crates.io/crates/regex
|
|
declare_lint! {
|
|
pub TRIVIAL_REGEX,
|
|
Warn,
|
|
"finds trivial regular expressions"
|
|
}
|
|
|
|
/// **What it does:** This lint checks for usage of `regex!(_)` which as of now is usually slower than `Regex::new(_)` unless called in a loop (which is a bad idea anyway).
|
|
///
|
|
/// **Why is this bad?** Performance, at least for now. The macro version is likely to catch up long-term, but for now the dynamic version is faster.
|
|
///
|
|
/// **Known problems:** None
|
|
///
|
|
/// **Example:** `regex!("foo|bar")`
|
|
declare_lint! {
|
|
pub REGEX_MACRO,
|
|
Warn,
|
|
"finds use of `regex!(_)`, suggests `Regex::new(_)` instead"
|
|
}
|
|
|
|
#[derive(Clone, Default)]
|
|
pub struct RegexPass {
|
|
spans: HashSet<Span>,
|
|
last: Option<NodeId>,
|
|
}
|
|
|
|
impl LintPass for RegexPass {
|
|
fn get_lints(&self) -> LintArray {
|
|
lint_array!(INVALID_REGEX, REGEX_MACRO, TRIVIAL_REGEX)
|
|
}
|
|
}
|
|
|
|
impl LateLintPass for RegexPass {
|
|
fn check_crate(&mut self, _: &LateContext, _: &Crate) {
|
|
self.spans.clear();
|
|
}
|
|
|
|
fn check_block(&mut self, cx: &LateContext, block: &Block) {
|
|
if_let_chain!{[
|
|
self.last.is_none(),
|
|
let Some(ref expr) = block.expr,
|
|
match_type(cx, cx.tcx.expr_ty(expr), &paths::REGEX),
|
|
let Some(span) = is_expn_of(cx, expr.span, "regex"),
|
|
], {
|
|
if !self.spans.contains(&span) {
|
|
span_lint(cx,
|
|
REGEX_MACRO,
|
|
span,
|
|
"`regex!(_)` found. \
|
|
Please use `Regex::new(_)`, which is faster for now.");
|
|
self.spans.insert(span);
|
|
}
|
|
self.last = Some(block.id);
|
|
}}
|
|
}
|
|
|
|
fn check_block_post(&mut self, _: &LateContext, block: &Block) {
|
|
if self.last.map_or(false, |id| block.id == id) {
|
|
self.last = None;
|
|
}
|
|
}
|
|
|
|
fn check_expr(&mut self, cx: &LateContext, expr: &Expr) {
|
|
if_let_chain!{[
|
|
let ExprCall(ref fun, ref args) = expr.node,
|
|
args.len() == 1,
|
|
let Some(def) = cx.tcx.def_map.borrow().get(&fun.id),
|
|
], {
|
|
let def_id = def.def_id();
|
|
if match_def_path(cx, def_id, &paths::REGEX_NEW) {
|
|
check_regex(cx, &args[0], true);
|
|
} else if match_def_path(cx, def_id, &paths::REGEX_BYTES_NEW) {
|
|
check_regex(cx, &args[0], false);
|
|
} else if match_def_path(cx, def_id, &paths::REGEX_BUILDER_NEW) {
|
|
check_regex(cx, &args[0], true);
|
|
} else if match_def_path(cx, def_id, &paths::REGEX_BYTES_BUILDER_NEW) {
|
|
check_regex(cx, &args[0], false);
|
|
} else if match_def_path(cx, def_id, &paths::REGEX_SET_NEW) {
|
|
check_set(cx, &args[0], true);
|
|
} else if match_def_path(cx, def_id, &paths::REGEX_BYTES_SET_NEW) {
|
|
check_set(cx, &args[0], false);
|
|
}
|
|
}}
|
|
}
|
|
}
|
|
|
|
#[allow(cast_possible_truncation)]
|
|
fn str_span(base: Span, s: &str, c: usize) -> Span {
|
|
let mut si = s.char_indices().skip(c);
|
|
|
|
match (si.next(), si.next()) {
|
|
(Some((l, _)), Some((h, _))) => {
|
|
Span {
|
|
lo: base.lo + BytePos(l as u32),
|
|
hi: base.lo + BytePos(h as u32),
|
|
..base
|
|
}
|
|
}
|
|
_ => base,
|
|
}
|
|
}
|
|
|
|
fn const_str(cx: &LateContext, e: &Expr) -> Option<InternedString> {
|
|
match eval_const_expr_partial(cx.tcx, e, ExprTypeChecked, None) {
|
|
Ok(ConstVal::Str(r)) => Some(r),
|
|
_ => None,
|
|
}
|
|
}
|
|
|
|
fn is_trivial_regex(s: ®ex_syntax::Expr) -> Option<&'static str> {
|
|
use regex_syntax::Expr;
|
|
|
|
match *s {
|
|
Expr::Empty | Expr::StartText | Expr::EndText => Some("the regex is unlikely to be useful as it is"),
|
|
Expr::Literal { .. } => Some("consider using `str::contains`"),
|
|
Expr::Concat(ref exprs) => {
|
|
match exprs.len() {
|
|
2 => {
|
|
match (&exprs[0], &exprs[1]) {
|
|
(&Expr::StartText, &Expr::EndText) => Some("consider using `str::is_empty`"),
|
|
(&Expr::StartText, &Expr::Literal { .. }) => Some("consider using `str::starts_with`"),
|
|
(&Expr::Literal { .. }, &Expr::EndText) => Some("consider using `str::ends_with`"),
|
|
_ => None,
|
|
}
|
|
}
|
|
3 => {
|
|
if let (&Expr::StartText, &Expr::Literal {..}, &Expr::EndText) = (&exprs[0], &exprs[1], &exprs[2]) {
|
|
Some("consider using `==` on `str`s")
|
|
} else {
|
|
None
|
|
}
|
|
}
|
|
_ => None,
|
|
}
|
|
}
|
|
_ => None,
|
|
}
|
|
}
|
|
|
|
fn check_set(cx: &LateContext, expr: &Expr, utf8: bool) {
|
|
if_let_chain! {[
|
|
let ExprAddrOf(_, ref expr) = expr.node,
|
|
let ExprVec(ref exprs) = expr.node,
|
|
], {
|
|
for expr in exprs {
|
|
check_regex(cx, expr, utf8);
|
|
}
|
|
}}
|
|
}
|
|
|
|
fn check_regex(cx: &LateContext, expr: &Expr, utf8: bool) {
|
|
let builder = regex_syntax::ExprBuilder::new().unicode(utf8);
|
|
|
|
if let ExprLit(ref lit) = expr.node {
|
|
if let LitKind::Str(ref r, _) = lit.node {
|
|
match builder.parse(r) {
|
|
Ok(r) => {
|
|
if let Some(repl) = is_trivial_regex(&r) {
|
|
span_help_and_lint(cx, TRIVIAL_REGEX, expr.span,
|
|
"trivial regex",
|
|
&format!("consider using {}", repl));
|
|
}
|
|
}
|
|
Err(e) => {
|
|
span_lint(cx,
|
|
INVALID_REGEX,
|
|
str_span(expr.span, r, e.position()),
|
|
&format!("regex syntax error: {}",
|
|
e.description()));
|
|
}
|
|
}
|
|
}
|
|
} else if let Some(r) = const_str(cx, expr) {
|
|
match builder.parse(&r) {
|
|
Ok(r) => {
|
|
if let Some(repl) = is_trivial_regex(&r) {
|
|
span_help_and_lint(cx, TRIVIAL_REGEX, expr.span,
|
|
"trivial regex",
|
|
&format!("consider using {}", repl));
|
|
}
|
|
}
|
|
Err(e) => {
|
|
span_lint(cx,
|
|
INVALID_REGEX,
|
|
expr.span,
|
|
&format!("regex syntax error on position {}: {}",
|
|
e.position(),
|
|
e.description()));
|
|
}
|
|
}
|
|
}
|
|
}
|