rust/clippy_lints/src/regex.rs

211 lines
8.0 KiB
Rust
Raw Normal View History

use clippy_utils::consts::{constant, Constant};
use clippy_utils::diagnostics::{span_lint, span_lint_and_help};
use clippy_utils::{match_def_path, paths};
2018-11-27 14:14:15 -06:00
use if_chain::if_chain;
2020-02-29 21:23:33 -06:00
use rustc_ast::ast::{LitKind, StrStyle};
use rustc_hir::{BorrowKind, Expr, ExprKind};
2020-01-12 00:08:41 -06:00
use rustc_lint::{LateContext, LateLintPass};
2021-09-28 11:40:06 -05:00
use rustc_session::{declare_lint_pass, declare_tool_lint};
use rustc_span::source_map::{BytePos, Span};
use std::convert::TryFrom;
2016-02-04 17:36:06 -06:00
2018-03-28 08:24:26 -05:00
declare_clippy_lint! {
/// ### What it does
/// Checks [regex](https://crates.io/crates/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.
///
/// ### Example
2019-03-05 16:23:50 -06:00
/// ```ignore
/// Regex::new("|")
/// ```
Added `clippy::version` attribute to all normal lints So, some context for this, well, more a story. I'm not used to scripting, I've never really scripted anything, even if it's a valuable skill. I just never really needed it. Now, `@flip1995` correctly suggested using a script for this in `rust-clippy#7813`... And I decided to write a script using nushell because why not? This was a mistake... I spend way more time on this than I would like to admit. It has definitely been more than 4 hours. It shouldn't take that long, but me being new to scripting and nushell just wasn't a good mixture... Anyway, here is the script that creates another script which adds the versions. Fun... Just execute this on the `gh-pages` branch and the resulting `replacer.sh` in `clippy_lints` and it should all work. ```nu mv v0.0.212 rust-1.00.0; mv beta rust-1.57.0; mv master rust-1.58.0; let paths = (open ./rust-1.58.0/lints.json | select id id_span | flatten | select id path); let versions = ( ls | where name =~ "rust-" | select name | format {name}/lints.json | each { open $it | select id | insert version $it | str substring "5,11" version} | group-by id | rotate counter-clockwise id version | update version {get version | first 1} | flatten | select id version); $paths | each { |row| let version = ($versions | where id == ($row.id) | format {version}) let idu = ($row.id | str upcase) $"sed -i '0,/($idu),/{s/pub ($idu),/#[clippy::version = "($version)"]\n pub ($idu),/}' ($row.path)" } | str collect ";" | str find-replace --all '1.00.0' 'pre 1.29.0' | save "replacer.sh"; ``` And this still has some problems, but at this point I just want to be done -.-
2021-10-21 14:06:26 -05:00
#[clippy::version = "pre 1.29.0"]
2016-02-04 17:36:06 -06:00
pub INVALID_REGEX,
2018-03-28 08:24:26 -05:00
correctness,
"invalid regular expressions"
2016-02-04 17:36:06 -06:00
}
2018-03-28 08:24:26 -05:00
declare_clippy_lint! {
/// ### What it does
/// Checks for trivial [regex](https://crates.io/crates/regex)
/// creation (with `Regex::new`, `RegexBuilder::new`, or `RegexSet::new`).
///
/// ### 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.
///
/// ### Known problems
/// If the same regex is going to be applied to multiple
/// inputs, the precomputations done by `Regex` construction can give
/// significantly better performance than any of the `str`-based methods.
///
/// ### Example
2019-03-05 16:23:50 -06:00
/// ```ignore
/// Regex::new("^foobar")
/// ```
Added `clippy::version` attribute to all normal lints So, some context for this, well, more a story. I'm not used to scripting, I've never really scripted anything, even if it's a valuable skill. I just never really needed it. Now, `@flip1995` correctly suggested using a script for this in `rust-clippy#7813`... And I decided to write a script using nushell because why not? This was a mistake... I spend way more time on this than I would like to admit. It has definitely been more than 4 hours. It shouldn't take that long, but me being new to scripting and nushell just wasn't a good mixture... Anyway, here is the script that creates another script which adds the versions. Fun... Just execute this on the `gh-pages` branch and the resulting `replacer.sh` in `clippy_lints` and it should all work. ```nu mv v0.0.212 rust-1.00.0; mv beta rust-1.57.0; mv master rust-1.58.0; let paths = (open ./rust-1.58.0/lints.json | select id id_span | flatten | select id path); let versions = ( ls | where name =~ "rust-" | select name | format {name}/lints.json | each { open $it | select id | insert version $it | str substring "5,11" version} | group-by id | rotate counter-clockwise id version | update version {get version | first 1} | flatten | select id version); $paths | each { |row| let version = ($versions | where id == ($row.id) | format {version}) let idu = ($row.id | str upcase) $"sed -i '0,/($idu),/{s/pub ($idu),/#[clippy::version = "($version)"]\n pub ($idu),/}' ($row.path)" } | str collect ";" | str find-replace --all '1.00.0' 'pre 1.29.0' | save "replacer.sh"; ``` And this still has some problems, but at this point I just want to be done -.-
2021-10-21 14:06:26 -05:00
#[clippy::version = "pre 1.29.0"]
2016-02-05 16:10:48 -06:00
pub TRIVIAL_REGEX,
nursery,
"trivial regular expressions"
2016-02-05 16:10:48 -06:00
}
2021-09-28 11:40:06 -05:00
declare_lint_pass!(Regex => [INVALID_REGEX, TRIVIAL_REGEX]);
2016-02-04 17:36:06 -06:00
impl<'tcx> LateLintPass<'tcx> for Regex {
fn check_expr(&mut self, cx: &LateContext<'tcx>, expr: &'tcx Expr<'_>) {
if_chain! {
if let ExprKind::Call(fun, args) = expr.kind;
2019-09-27 10:16:06 -05:00
if let ExprKind::Path(ref qpath) = fun.kind;
if args.len() == 1;
if let Some(def_id) = cx.qpath_res(qpath, fun.hir_id).opt_def_id();
then {
2019-05-17 16:53:54 -05:00
if match_def_path(cx, def_id, &paths::REGEX_NEW) ||
match_def_path(cx, def_id, &paths::REGEX_BUILDER_NEW) {
check_regex(cx, &args[0], true);
2019-05-17 16:53:54 -05:00
} else if match_def_path(cx, def_id, &paths::REGEX_BYTES_NEW) ||
match_def_path(cx, def_id, &paths::REGEX_BYTES_BUILDER_NEW) {
check_regex(cx, &args[0], false);
2019-05-17 16:53:54 -05:00
} else if match_def_path(cx, def_id, &paths::REGEX_SET_NEW) {
check_set(cx, &args[0], true);
2019-05-17 16:53:54 -05:00
} else if match_def_path(cx, def_id, &paths::REGEX_BYTES_SET_NEW) {
check_set(cx, &args[0], false);
}
2016-02-05 09:48:35 -06:00
}
}
2016-02-04 17:36:06 -06:00
}
}
2016-02-05 09:48:35 -06:00
#[allow(clippy::cast_possible_truncation)] // truncation very unlikely here
#[must_use]
2018-04-19 01:30:07 -05:00
fn str_span(base: Span, c: regex_syntax::ast::Span, offset: u16) -> Span {
let offset = u32::from(offset);
let end = base.lo() + BytePos(u32::try_from(c.end.offset).expect("offset too large") + offset);
let start = base.lo() + BytePos(u32::try_from(c.start.offset).expect("offset too large") + offset);
2018-03-13 09:02:40 -05:00
assert!(start <= end);
2021-04-18 07:27:04 -05:00
Span::new(start, end, base.ctxt(), base.parent())
2016-02-05 09:48:35 -06:00
}
fn const_str<'tcx>(cx: &LateContext<'tcx>, e: &'tcx Expr<'_>) -> Option<String> {
2020-07-17 03:47:04 -05:00
constant(cx, cx.typeck_results(), e).and_then(|(c, _)| match c {
2018-03-13 05:38:11 -05:00
Constant::Str(s) => Some(s),
2016-02-24 10:38:57 -06:00
_ => None,
2018-03-13 05:38:11 -05:00
})
2016-02-05 09:48:35 -06:00
}
2016-02-05 16:10:48 -06:00
2018-03-13 09:02:40 -05:00
fn is_trivial_regex(s: &regex_syntax::hir::Hir) -> Option<&'static str> {
2020-02-21 02:39:38 -06:00
use regex_syntax::hir::Anchor::{EndText, StartText};
use regex_syntax::hir::HirKind::{Alternation, Anchor, Concat, Empty, Literal};
2018-03-13 09:02:40 -05:00
let is_literal = |e: &[regex_syntax::hir::Hir]| e.iter().all(|e| matches!(*e.kind(), Literal(_)));
2018-03-13 09:02:40 -05:00
match *s.kind() {
2018-11-27 14:14:15 -06:00
Empty | Anchor(_) => Some("the regex is unlikely to be useful as it is"),
2018-03-13 09:02:40 -05:00
Literal(_) => Some("consider using `str::contains`"),
2018-11-27 14:14:15 -06:00
Alternation(ref exprs) => {
if exprs.iter().all(|e| e.kind().is_empty()) {
Some("the regex is unlikely to be useful as it is")
} else {
None
}
2018-03-13 09:02:40 -05:00
},
Concat(ref exprs) => match (exprs[0].kind(), exprs[exprs.len() - 1].kind()) {
2018-11-27 14:14:15 -06:00
(&Anchor(StartText), &Anchor(EndText)) if exprs[1..(exprs.len() - 1)].is_empty() => {
Some("consider using `str::is_empty`")
},
(&Anchor(StartText), &Anchor(EndText)) if is_literal(&exprs[1..(exprs.len() - 1)]) => {
Some("consider using `==` on `str`s")
},
2018-03-13 09:02:40 -05:00
(&Anchor(StartText), &Literal(_)) if is_literal(&exprs[1..]) => Some("consider using `str::starts_with`"),
2018-11-27 14:14:15 -06:00
(&Literal(_), &Anchor(EndText)) if is_literal(&exprs[1..(exprs.len() - 1)]) => {
Some("consider using `str::ends_with`")
},
2018-03-13 09:02:40 -05:00
_ if is_literal(exprs) => Some("consider using `str::contains`"),
2017-09-05 04:33:04 -05:00
_ => None,
2016-12-20 11:21:30 -06:00
},
2016-02-06 11:06:39 -06:00
_ => None,
2016-02-05 16:10:48 -06:00
}
}
2016-05-25 10:15:19 -05:00
fn check_set<'tcx>(cx: &LateContext<'tcx>, expr: &'tcx Expr<'_>, utf8: bool) {
if_chain! {
if let ExprKind::AddrOf(BorrowKind::Ref, _, expr) = expr.kind;
2019-12-27 01:12:26 -06:00
if let ExprKind::Array(exprs) = expr.kind;
then {
for expr in exprs {
check_regex(cx, expr, utf8);
}
2016-05-25 10:15:19 -05:00
}
}
2016-05-25 10:15:19 -05:00
}
fn check_regex<'tcx>(cx: &LateContext<'tcx>, expr: &'tcx Expr<'_>, utf8: bool) {
2018-04-07 15:18:51 -05:00
let mut parser = regex_syntax::ParserBuilder::new()
.unicode(true)
2018-04-07 15:18:51 -05:00
.allow_invalid_utf8(!utf8)
.build();
2016-05-25 10:15:19 -05:00
2019-09-27 10:16:06 -05:00
if let ExprKind::Lit(ref lit) = expr.kind {
if let LitKind::Str(ref r, style) = lit.node {
let r = &r.as_str();
2018-03-13 09:02:40 -05:00
let offset = if let StrStyle::Raw(n) = style { 2 + n } else { 1 };
match parser.parse(r) {
2018-11-27 14:14:15 -06:00
Ok(r) => {
if let Some(repl) = is_trivial_regex(&r) {
span_lint_and_help(cx, TRIVIAL_REGEX, expr.span, "trivial regex", None, repl);
2018-11-27 14:14:15 -06:00
}
2018-03-13 09:02:40 -05:00
},
Err(regex_syntax::Error::Parse(e)) => {
span_lint(
cx,
INVALID_REGEX,
str_span(expr.span, *e.span(), offset),
&format!("regex syntax error: {}", e.kind()),
);
},
Err(regex_syntax::Error::Translate(e)) => {
span_lint(
cx,
INVALID_REGEX,
str_span(expr.span, *e.span(), offset),
&format!("regex syntax error: {}", e.kind()),
2017-09-05 04:33:04 -05:00
);
2016-12-20 11:21:30 -06:00
},
2016-05-25 10:15:19 -05:00
Err(e) => {
2018-11-27 14:14:15 -06:00
span_lint(cx, INVALID_REGEX, expr.span, &format!("regex syntax error: {}", e));
2016-12-20 11:21:30 -06:00
},
2016-05-25 10:15:19 -05:00
}
}
} else if let Some(r) = const_str(cx, expr) {
2018-03-13 09:02:40 -05:00
match parser.parse(&r) {
2018-11-27 14:14:15 -06:00
Ok(r) => {
if let Some(repl) = is_trivial_regex(&r) {
span_lint_and_help(cx, TRIVIAL_REGEX, expr.span, "trivial regex", None, repl);
2018-11-27 14:14:15 -06:00
}
2018-03-13 09:02:40 -05:00
},
Err(regex_syntax::Error::Parse(e)) => {
span_lint(
cx,
INVALID_REGEX,
expr.span,
&format!("regex syntax error on position {}: {}", e.span().start.offset, e.kind()),
);
},
Err(regex_syntax::Error::Translate(e)) => {
span_lint(
cx,
INVALID_REGEX,
expr.span,
&format!("regex syntax error on position {}: {}", e.span().start.offset, e.kind()),
2017-09-05 04:33:04 -05:00
);
2016-12-20 11:21:30 -06:00
},
2016-05-25 10:15:19 -05:00
Err(e) => {
2018-11-27 14:14:15 -06:00
span_lint(cx, INVALID_REGEX, expr.span, &format!("regex syntax error: {}", e));
2016-12-20 11:21:30 -06:00
},
2016-05-25 10:15:19 -05:00
}
}
}