2016-02-04 23:36:06 +00:00
|
|
|
use regex_syntax;
|
2016-04-07 15:46:48 +00:00
|
|
|
use rustc::hir::*;
|
2016-02-24 16:38:57 +00:00
|
|
|
use rustc::lint::*;
|
2017-09-13 13:34:04 +00:00
|
|
|
use rustc::ty;
|
2016-03-31 15:05:43 +00:00
|
|
|
use rustc::middle::const_val::ConstVal;
|
2017-01-13 16:04:56 +00:00
|
|
|
use rustc_const_eval::ConstContext;
|
2017-07-31 10:37:38 +00:00
|
|
|
use rustc::ty::subst::Substs;
|
2016-02-08 22:48:04 +00:00
|
|
|
use std::collections::HashSet;
|
2016-02-24 16:38:57 +00:00
|
|
|
use std::error::Error;
|
2016-02-14 15:55:02 +00:00
|
|
|
use syntax::ast::{LitKind, NodeId};
|
2017-09-05 09:33:04 +00:00
|
|
|
use syntax::codemap::{BytePos, Span};
|
2016-11-23 20:19:03 +00:00
|
|
|
use syntax::symbol::InternedString;
|
2017-09-12 12:26:40 +00:00
|
|
|
use utils::{is_expn_of, match_def_path, match_type, paths, span_help_and_lint, span_lint, opt_def_id};
|
2016-02-04 23:36:06 +00:00
|
|
|
|
2017-09-01 20:43:34 +00:00
|
|
|
/// **What it does:** Checks [regex](https://crates.io/crates/regex) creation
|
|
|
|
/// (with `Regex::new`,`RegexBuilder::new` or `RegexSet::new`) for correct
|
|
|
|
/// regex syntax.
|
2016-07-15 22:25:44 +00:00
|
|
|
///
|
2016-02-04 23:36:06 +00:00
|
|
|
/// **Why is this bad?** This will lead to a runtime panic.
|
|
|
|
///
|
|
|
|
/// **Known problems:** None.
|
|
|
|
///
|
2016-07-15 22:25:44 +00:00
|
|
|
/// **Example:**
|
|
|
|
/// ```rust
|
|
|
|
/// Regex::new("|")
|
|
|
|
/// ```
|
2016-02-04 23:36:06 +00:00
|
|
|
declare_lint! {
|
|
|
|
pub INVALID_REGEX,
|
|
|
|
Deny,
|
2016-08-06 08:18:36 +00:00
|
|
|
"invalid regular expressions"
|
2016-02-04 23:36:06 +00:00
|
|
|
}
|
|
|
|
|
2016-08-06 07:55:04 +00:00
|
|
|
/// **What it does:** Checks for trivial [regex] creation (with `Regex::new`,
|
2016-05-25 22:08:31 +00:00
|
|
|
/// `RegexBuilder::new` or `RegexSet::new`).
|
2016-02-05 22:10:48 +00:00
|
|
|
///
|
2016-07-15 22:25:44 +00:00
|
|
|
/// [regex]: https://crates.io/crates/regex
|
|
|
|
///
|
2016-08-06 07:55:04 +00:00
|
|
|
/// **Why is this bad?** Matching the regex can likely be replaced by `==` or
|
|
|
|
/// `str::starts_with`, `str::ends_with` or `std::contains` or other `str`
|
|
|
|
/// methods.
|
2016-02-05 22:10:48 +00:00
|
|
|
///
|
|
|
|
/// **Known problems:** None.
|
|
|
|
///
|
2016-07-15 22:25:44 +00:00
|
|
|
/// **Example:**
|
|
|
|
/// ```rust
|
|
|
|
/// Regex::new("^foobar")
|
|
|
|
/// ```
|
2016-02-05 22:10:48 +00:00
|
|
|
declare_lint! {
|
|
|
|
pub TRIVIAL_REGEX,
|
|
|
|
Warn,
|
2016-08-06 08:18:36 +00:00
|
|
|
"trivial regular expressions"
|
2016-02-05 22:10:48 +00:00
|
|
|
}
|
|
|
|
|
2016-08-06 07:55:04 +00:00
|
|
|
/// **What it does:** 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).
|
2016-02-07 21:50:54 +00:00
|
|
|
///
|
2016-08-06 07:55:04 +00:00
|
|
|
/// **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.
|
2016-02-07 21:50:54 +00:00
|
|
|
///
|
2016-08-06 07:55:04 +00:00
|
|
|
/// **Known problems:** None.
|
2016-02-07 21:50:54 +00:00
|
|
|
///
|
2016-07-15 22:25:44 +00:00
|
|
|
/// **Example:**
|
|
|
|
/// ```rust
|
|
|
|
/// regex!("foo|bar")
|
|
|
|
/// ```
|
2016-02-07 21:50:54 +00:00
|
|
|
declare_lint! {
|
|
|
|
pub REGEX_MACRO,
|
2016-02-08 22:48:04 +00:00
|
|
|
Warn,
|
2016-08-06 08:18:36 +00:00
|
|
|
"use of `regex!(_)` instead of `Regex::new(_)`"
|
2016-02-07 21:50:54 +00:00
|
|
|
}
|
|
|
|
|
2016-02-14 15:55:02 +00:00
|
|
|
#[derive(Clone, Default)]
|
2016-06-10 14:17:20 +00:00
|
|
|
pub struct Pass {
|
2016-02-14 15:55:02 +00:00
|
|
|
spans: HashSet<Span>,
|
2016-02-24 16:38:57 +00:00
|
|
|
last: Option<NodeId>,
|
2016-02-14 15:55:02 +00:00
|
|
|
}
|
2016-02-04 23:36:06 +00:00
|
|
|
|
2016-06-10 14:17:20 +00:00
|
|
|
impl LintPass for Pass {
|
2016-02-04 23:36:06 +00:00
|
|
|
fn get_lints(&self) -> LintArray {
|
2016-02-07 21:50:54 +00:00
|
|
|
lint_array!(INVALID_REGEX, REGEX_MACRO, TRIVIAL_REGEX)
|
2016-02-04 23:36:06 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-12-07 12:13:40 +00:00
|
|
|
impl<'a, 'tcx> LateLintPass<'a, 'tcx> for Pass {
|
|
|
|
fn check_crate(&mut self, _: &LateContext<'a, 'tcx>, _: &'tcx Crate) {
|
2016-02-14 15:55:02 +00:00
|
|
|
self.spans.clear();
|
2016-02-07 21:50:54 +00:00
|
|
|
}
|
|
|
|
|
2016-12-07 12:13:40 +00:00
|
|
|
fn check_block(&mut self, cx: &LateContext<'a, 'tcx>, block: &'tcx Block) {
|
2016-02-14 15:55:02 +00:00
|
|
|
if_let_chain!{[
|
|
|
|
self.last.is_none(),
|
|
|
|
let Some(ref expr) = block.expr,
|
2017-01-13 16:04:56 +00:00
|
|
|
match_type(cx, cx.tables.expr_ty(expr), &paths::REGEX),
|
2017-03-31 22:14:04 +00:00
|
|
|
let Some(span) = is_expn_of(expr.span, "regex"),
|
2016-02-14 15:55:02 +00:00
|
|
|
], {
|
|
|
|
if !self.spans.contains(&span) {
|
|
|
|
span_lint(cx,
|
|
|
|
REGEX_MACRO,
|
|
|
|
span,
|
|
|
|
"`regex!(_)` found. \
|
|
|
|
Please use `Regex::new(_)`, which is faster for now.");
|
2016-05-25 15:15:19 +00:00
|
|
|
self.spans.insert(span);
|
2016-02-14 15:55:02 +00:00
|
|
|
}
|
|
|
|
self.last = Some(block.id);
|
|
|
|
}}
|
|
|
|
}
|
2016-02-24 16:38:57 +00:00
|
|
|
|
2016-12-07 12:13:40 +00:00
|
|
|
fn check_block_post(&mut self, _: &LateContext<'a, 'tcx>, block: &'tcx Block) {
|
2016-02-14 15:55:02 +00:00
|
|
|
if self.last.map_or(false, |id| block.id == id) {
|
2016-02-24 16:38:57 +00:00
|
|
|
self.last = None;
|
2016-02-14 15:55:02 +00:00
|
|
|
}
|
|
|
|
}
|
2016-02-07 21:50:54 +00:00
|
|
|
|
2016-12-07 12:13:40 +00:00
|
|
|
fn check_expr(&mut self, cx: &LateContext<'a, 'tcx>, expr: &'tcx Expr) {
|
2016-02-04 23:36:06 +00:00
|
|
|
if_let_chain!{[
|
|
|
|
let ExprCall(ref fun, ref args) = expr.node,
|
2016-12-01 21:31:56 +00:00
|
|
|
let ExprPath(ref qpath) = fun.node,
|
2016-05-25 15:15:19 +00:00
|
|
|
args.len() == 1,
|
2017-09-12 12:26:40 +00:00
|
|
|
let Some(def_id) = opt_def_id(cx.tables.qpath_def(qpath, fun.hir_id)),
|
2016-02-04 23:36:06 +00:00
|
|
|
], {
|
2017-01-13 16:04:56 +00:00
|
|
|
if match_def_path(cx.tcx, def_id, &paths::REGEX_NEW) ||
|
|
|
|
match_def_path(cx.tcx, def_id, &paths::REGEX_BUILDER_NEW) {
|
2016-05-25 15:15:19 +00:00
|
|
|
check_regex(cx, &args[0], true);
|
2017-01-13 16:04:56 +00:00
|
|
|
} else if match_def_path(cx.tcx, def_id, &paths::REGEX_BYTES_NEW) ||
|
|
|
|
match_def_path(cx.tcx, def_id, &paths::REGEX_BYTES_BUILDER_NEW) {
|
2016-05-25 19:36:51 +00:00
|
|
|
check_regex(cx, &args[0], false);
|
2017-01-13 16:04:56 +00:00
|
|
|
} else if match_def_path(cx.tcx, def_id, &paths::REGEX_SET_NEW) {
|
2016-05-25 15:15:19 +00:00
|
|
|
check_set(cx, &args[0], true);
|
2017-01-13 16:04:56 +00:00
|
|
|
} else if match_def_path(cx.tcx, def_id, &paths::REGEX_BYTES_SET_NEW) {
|
2016-05-25 15:15:19 +00:00
|
|
|
check_set(cx, &args[0], false);
|
2016-02-05 15:48:35 +00:00
|
|
|
}
|
2016-02-04 23:36:06 +00:00
|
|
|
}}
|
|
|
|
}
|
|
|
|
}
|
2016-02-05 15:48:35 +00:00
|
|
|
|
|
|
|
#[allow(cast_possible_truncation)]
|
|
|
|
fn str_span(base: Span, s: &str, c: usize) -> Span {
|
2016-05-07 22:56:23 +00:00
|
|
|
let mut si = s.char_indices().skip(c);
|
|
|
|
|
2016-06-05 23:42:39 +00:00
|
|
|
match (si.next(), si.next()) {
|
2016-05-07 22:56:23 +00:00
|
|
|
(Some((l, _)), Some((h, _))) => {
|
2017-09-03 21:15:15 +00:00
|
|
|
Span::new(base.lo() + BytePos(l as u32), base.lo() + BytePos(h as u32), base.ctxt())
|
2016-12-20 17:21:30 +00:00
|
|
|
},
|
2016-05-07 22:56:23 +00:00
|
|
|
_ => base,
|
2016-02-29 11:19:32 +00:00
|
|
|
}
|
2016-02-05 15:48:35 +00:00
|
|
|
}
|
|
|
|
|
2017-09-13 13:34:04 +00:00
|
|
|
fn const_str<'a, 'tcx>(cx: &LateContext<'a, 'tcx>, e: &'tcx Expr) -> Option<InternedString> {
|
2017-07-31 10:37:38 +00:00
|
|
|
let parent_item = cx.tcx.hir.get_parent(e.id);
|
|
|
|
let parent_def_id = cx.tcx.hir.local_def_id(parent_item);
|
|
|
|
let substs = Substs::identity_for_item(cx.tcx, parent_def_id);
|
|
|
|
match ConstContext::new(cx.tcx, cx.param_env.and(substs), cx.tables).eval(e) {
|
2017-09-13 13:34:04 +00:00
|
|
|
Ok(&ty::Const { val: ConstVal::Str(r), .. }) => Some(r),
|
2016-02-24 16:38:57 +00:00
|
|
|
_ => None,
|
2016-02-05 15:48:35 +00:00
|
|
|
}
|
|
|
|
}
|
2016-02-05 22:10:48 +00:00
|
|
|
|
2016-02-06 17:06:39 +00:00
|
|
|
fn is_trivial_regex(s: ®ex_syntax::Expr) -> Option<&'static str> {
|
|
|
|
use regex_syntax::Expr;
|
2016-02-05 22:10:48 +00:00
|
|
|
|
2016-02-06 17:06:39 +00:00
|
|
|
match *s {
|
|
|
|
Expr::Empty | Expr::StartText | Expr::EndText => Some("the regex is unlikely to be useful as it is"),
|
2016-04-14 18:14:03 +00:00
|
|
|
Expr::Literal { .. } => Some("consider using `str::contains`"),
|
2017-09-05 09:33:04 +00:00
|
|
|
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`"),
|
2016-02-06 17:06:39 +00:00
|
|
|
_ => None,
|
2017-09-05 09:33:04 +00:00
|
|
|
},
|
|
|
|
3 => if let (&Expr::StartText, &Expr::Literal { .. }, &Expr::EndText) = (&exprs[0], &exprs[1], &exprs[2]) {
|
|
|
|
Some("consider using `==` on `str`s")
|
|
|
|
} else {
|
|
|
|
None
|
|
|
|
},
|
|
|
|
_ => None,
|
2016-12-20 17:21:30 +00:00
|
|
|
},
|
2016-02-06 17:06:39 +00:00
|
|
|
_ => None,
|
2016-02-05 22:10:48 +00:00
|
|
|
}
|
|
|
|
}
|
2016-05-25 15:15:19 +00:00
|
|
|
|
2017-09-13 13:34:04 +00:00
|
|
|
fn check_set<'a, 'tcx>(cx: &LateContext<'a, 'tcx>, expr: &'tcx Expr, utf8: bool) {
|
2016-05-25 15:15:19 +00:00
|
|
|
if_let_chain! {[
|
|
|
|
let ExprAddrOf(_, ref expr) = expr.node,
|
2016-09-30 13:35:24 +00:00
|
|
|
let ExprArray(ref exprs) = expr.node,
|
2016-05-25 15:15:19 +00:00
|
|
|
], {
|
|
|
|
for expr in exprs {
|
|
|
|
check_regex(cx, expr, utf8);
|
|
|
|
}
|
|
|
|
}}
|
|
|
|
}
|
|
|
|
|
2017-09-13 13:34:04 +00:00
|
|
|
fn check_regex<'a, 'tcx>(cx: &LateContext<'a, 'tcx>, expr: &'tcx Expr, utf8: bool) {
|
2016-05-25 15:15:19 +00:00
|
|
|
let builder = regex_syntax::ExprBuilder::new().unicode(utf8);
|
|
|
|
|
|
|
|
if let ExprLit(ref lit) = expr.node {
|
|
|
|
if let LitKind::Str(ref r, _) = lit.node {
|
2017-03-30 08:21:13 +00:00
|
|
|
let r = &r.as_str();
|
2016-05-25 15:15:19 +00:00
|
|
|
match builder.parse(r) {
|
2017-09-05 09:33:04 +00:00
|
|
|
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),
|
|
|
|
);
|
2016-12-20 17:21:30 +00:00
|
|
|
},
|
2016-05-25 15:15:19 +00:00
|
|
|
Err(e) => {
|
2017-08-09 07:30:56 +00:00
|
|
|
span_lint(
|
|
|
|
cx,
|
|
|
|
INVALID_REGEX,
|
|
|
|
str_span(expr.span, r, e.position()),
|
|
|
|
&format!("regex syntax error: {}", e.description()),
|
|
|
|
);
|
2016-12-20 17:21:30 +00:00
|
|
|
},
|
2016-05-25 15:15:19 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
} else if let Some(r) = const_str(cx, expr) {
|
|
|
|
match builder.parse(&r) {
|
2017-09-05 09:33:04 +00:00
|
|
|
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),
|
|
|
|
);
|
2016-12-20 17:21:30 +00:00
|
|
|
},
|
2016-05-25 15:15:19 +00:00
|
|
|
Err(e) => {
|
2017-08-09 07:30:56 +00:00
|
|
|
span_lint(
|
|
|
|
cx,
|
|
|
|
INVALID_REGEX,
|
|
|
|
expr.span,
|
|
|
|
&format!("regex syntax error on position {}: {}", e.position(), e.description()),
|
|
|
|
);
|
2016-12-20 17:21:30 +00:00
|
|
|
},
|
2016-05-25 15:15:19 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|