mirror of
https://github.com/zed-industries/zed.git
synced 2025-01-29 21:49:33 +00:00
4785520d99
Closes #7645 Release Notes: - Added support for inserting newlines (`\n`) and tabs (`\t`) in editor Regex search replacements ([#7645](https://github.com/zed-industries/zed/issues/7645)).
479 lines
15 KiB
Rust
479 lines
15 KiB
Rust
use aho_corasick::{AhoCorasick, AhoCorasickBuilder};
|
|
use anyhow::{Context, Result};
|
|
use client::proto;
|
|
use itertools::Itertools;
|
|
use language::{char_kind, BufferSnapshot};
|
|
use regex::{Captures, Regex, RegexBuilder};
|
|
use smol::future::yield_now;
|
|
use std::{
|
|
borrow::Cow,
|
|
io::{BufRead, BufReader, Read},
|
|
ops::Range,
|
|
path::Path,
|
|
sync::{Arc, OnceLock},
|
|
};
|
|
use util::paths::PathMatcher;
|
|
|
|
static TEXT_REPLACEMENT_SPECIAL_CHARACTERS_REGEX: OnceLock<Regex> = OnceLock::new();
|
|
|
|
#[derive(Clone, Debug)]
|
|
pub struct SearchInputs {
|
|
query: Arc<str>,
|
|
files_to_include: Vec<PathMatcher>,
|
|
files_to_exclude: Vec<PathMatcher>,
|
|
}
|
|
|
|
impl SearchInputs {
|
|
pub fn as_str(&self) -> &str {
|
|
self.query.as_ref()
|
|
}
|
|
pub fn files_to_include(&self) -> &[PathMatcher] {
|
|
&self.files_to_include
|
|
}
|
|
pub fn files_to_exclude(&self) -> &[PathMatcher] {
|
|
&self.files_to_exclude
|
|
}
|
|
}
|
|
#[derive(Clone, Debug)]
|
|
pub enum SearchQuery {
|
|
Text {
|
|
search: Arc<AhoCorasick>,
|
|
replacement: Option<String>,
|
|
whole_word: bool,
|
|
case_sensitive: bool,
|
|
include_ignored: bool,
|
|
inner: SearchInputs,
|
|
},
|
|
|
|
Regex {
|
|
regex: Regex,
|
|
replacement: Option<String>,
|
|
multiline: bool,
|
|
whole_word: bool,
|
|
case_sensitive: bool,
|
|
include_ignored: bool,
|
|
inner: SearchInputs,
|
|
},
|
|
}
|
|
|
|
impl SearchQuery {
|
|
pub fn text(
|
|
query: impl ToString,
|
|
whole_word: bool,
|
|
case_sensitive: bool,
|
|
include_ignored: bool,
|
|
files_to_include: Vec<PathMatcher>,
|
|
files_to_exclude: Vec<PathMatcher>,
|
|
) -> Result<Self> {
|
|
let query = query.to_string();
|
|
let search = AhoCorasickBuilder::new()
|
|
.ascii_case_insensitive(!case_sensitive)
|
|
.build(&[&query])?;
|
|
let inner = SearchInputs {
|
|
query: query.into(),
|
|
files_to_exclude,
|
|
files_to_include,
|
|
};
|
|
Ok(Self::Text {
|
|
search: Arc::new(search),
|
|
replacement: None,
|
|
whole_word,
|
|
case_sensitive,
|
|
include_ignored,
|
|
inner,
|
|
})
|
|
}
|
|
|
|
pub fn regex(
|
|
query: impl ToString,
|
|
whole_word: bool,
|
|
case_sensitive: bool,
|
|
include_ignored: bool,
|
|
files_to_include: Vec<PathMatcher>,
|
|
files_to_exclude: Vec<PathMatcher>,
|
|
) -> Result<Self> {
|
|
let mut query = query.to_string();
|
|
let initial_query = Arc::from(query.as_str());
|
|
if whole_word {
|
|
let mut word_query = String::new();
|
|
word_query.push_str("\\b");
|
|
word_query.push_str(&query);
|
|
word_query.push_str("\\b");
|
|
query = word_query
|
|
}
|
|
|
|
let multiline = query.contains('\n') || query.contains("\\n");
|
|
let regex = RegexBuilder::new(&query)
|
|
.case_insensitive(!case_sensitive)
|
|
.multi_line(multiline)
|
|
.build()?;
|
|
let inner = SearchInputs {
|
|
query: initial_query,
|
|
files_to_exclude,
|
|
files_to_include,
|
|
};
|
|
Ok(Self::Regex {
|
|
regex,
|
|
replacement: None,
|
|
multiline,
|
|
whole_word,
|
|
case_sensitive,
|
|
include_ignored,
|
|
inner,
|
|
})
|
|
}
|
|
|
|
pub fn from_proto(message: proto::SearchProject) -> Result<Self> {
|
|
if message.regex {
|
|
Self::regex(
|
|
message.query,
|
|
message.whole_word,
|
|
message.case_sensitive,
|
|
message.include_ignored,
|
|
deserialize_path_matches(&message.files_to_include)?,
|
|
deserialize_path_matches(&message.files_to_exclude)?,
|
|
)
|
|
} else {
|
|
Self::text(
|
|
message.query,
|
|
message.whole_word,
|
|
message.case_sensitive,
|
|
message.include_ignored,
|
|
deserialize_path_matches(&message.files_to_include)?,
|
|
deserialize_path_matches(&message.files_to_exclude)?,
|
|
)
|
|
}
|
|
}
|
|
pub fn with_replacement(mut self, new_replacement: String) -> Self {
|
|
match self {
|
|
Self::Text {
|
|
ref mut replacement,
|
|
..
|
|
}
|
|
| Self::Regex {
|
|
ref mut replacement,
|
|
..
|
|
} => {
|
|
*replacement = Some(new_replacement);
|
|
self
|
|
}
|
|
}
|
|
}
|
|
pub fn to_proto(&self, project_id: u64) -> proto::SearchProject {
|
|
proto::SearchProject {
|
|
project_id,
|
|
query: self.as_str().to_string(),
|
|
regex: self.is_regex(),
|
|
whole_word: self.whole_word(),
|
|
case_sensitive: self.case_sensitive(),
|
|
include_ignored: self.include_ignored(),
|
|
files_to_include: self
|
|
.files_to_include()
|
|
.iter()
|
|
.map(|matcher| matcher.to_string())
|
|
.join(","),
|
|
files_to_exclude: self
|
|
.files_to_exclude()
|
|
.iter()
|
|
.map(|matcher| matcher.to_string())
|
|
.join(","),
|
|
}
|
|
}
|
|
|
|
pub fn detect<T: Read>(&self, stream: T) -> Result<bool> {
|
|
if self.as_str().is_empty() {
|
|
return Ok(false);
|
|
}
|
|
|
|
match self {
|
|
Self::Text { search, .. } => {
|
|
let mat = search.stream_find_iter(stream).next();
|
|
match mat {
|
|
Some(Ok(_)) => Ok(true),
|
|
Some(Err(err)) => Err(err.into()),
|
|
None => Ok(false),
|
|
}
|
|
}
|
|
Self::Regex {
|
|
regex, multiline, ..
|
|
} => {
|
|
let mut reader = BufReader::new(stream);
|
|
if *multiline {
|
|
let mut text = String::new();
|
|
if let Err(err) = reader.read_to_string(&mut text) {
|
|
Err(err.into())
|
|
} else {
|
|
Ok(regex.find(&text).is_some())
|
|
}
|
|
} else {
|
|
for line in reader.lines() {
|
|
let line = line?;
|
|
if regex.find(&line).is_some() {
|
|
return Ok(true);
|
|
}
|
|
}
|
|
Ok(false)
|
|
}
|
|
}
|
|
}
|
|
}
|
|
/// Returns the replacement text for this `SearchQuery`.
|
|
pub fn replacement(&self) -> Option<&str> {
|
|
match self {
|
|
SearchQuery::Text { replacement, .. } | SearchQuery::Regex { replacement, .. } => {
|
|
replacement.as_deref()
|
|
}
|
|
}
|
|
}
|
|
/// Replaces search hits if replacement is set. `text` is assumed to be a string that matches this `SearchQuery` exactly, without any leftovers on either side.
|
|
pub fn replacement_for<'a>(&self, text: &'a str) -> Option<Cow<'a, str>> {
|
|
match self {
|
|
SearchQuery::Text { replacement, .. } => replacement.clone().map(Cow::from),
|
|
SearchQuery::Regex {
|
|
regex, replacement, ..
|
|
} => {
|
|
if let Some(replacement) = replacement {
|
|
let replacement = TEXT_REPLACEMENT_SPECIAL_CHARACTERS_REGEX
|
|
.get_or_init(|| Regex::new(r"\\\\|\\n|\\t").unwrap())
|
|
.replace_all(replacement, |c: &Captures| {
|
|
match c.get(0).unwrap().as_str() {
|
|
r"\\" => "\\",
|
|
r"\n" => "\n",
|
|
r"\t" => "\t",
|
|
x => unreachable!("Unexpected escape sequence: {}", x),
|
|
}
|
|
});
|
|
Some(regex.replace(text, replacement))
|
|
} else {
|
|
None
|
|
}
|
|
}
|
|
}
|
|
}
|
|
pub async fn search(
|
|
&self,
|
|
buffer: &BufferSnapshot,
|
|
subrange: Option<Range<usize>>,
|
|
) -> Vec<Range<usize>> {
|
|
const YIELD_INTERVAL: usize = 20000;
|
|
|
|
if self.as_str().is_empty() {
|
|
return Default::default();
|
|
}
|
|
|
|
let range_offset = subrange.as_ref().map(|r| r.start).unwrap_or(0);
|
|
let rope = if let Some(range) = subrange {
|
|
buffer.as_rope().slice(range)
|
|
} else {
|
|
buffer.as_rope().clone()
|
|
};
|
|
|
|
let mut matches = Vec::new();
|
|
match self {
|
|
Self::Text {
|
|
search, whole_word, ..
|
|
} => {
|
|
for (ix, mat) in search
|
|
.stream_find_iter(rope.bytes_in_range(0..rope.len()))
|
|
.enumerate()
|
|
{
|
|
if (ix + 1) % YIELD_INTERVAL == 0 {
|
|
yield_now().await;
|
|
}
|
|
|
|
let mat = mat.unwrap();
|
|
if *whole_word {
|
|
let scope = buffer.language_scope_at(range_offset + mat.start());
|
|
let kind = |c| char_kind(&scope, c);
|
|
|
|
let prev_kind = rope.reversed_chars_at(mat.start()).next().map(kind);
|
|
let start_kind = kind(rope.chars_at(mat.start()).next().unwrap());
|
|
let end_kind = kind(rope.reversed_chars_at(mat.end()).next().unwrap());
|
|
let next_kind = rope.chars_at(mat.end()).next().map(kind);
|
|
if Some(start_kind) == prev_kind || Some(end_kind) == next_kind {
|
|
continue;
|
|
}
|
|
}
|
|
matches.push(mat.start()..mat.end())
|
|
}
|
|
}
|
|
|
|
Self::Regex {
|
|
regex, multiline, ..
|
|
} => {
|
|
if *multiline {
|
|
let text = rope.to_string();
|
|
for (ix, mat) in regex.find_iter(&text).enumerate() {
|
|
if (ix + 1) % YIELD_INTERVAL == 0 {
|
|
yield_now().await;
|
|
}
|
|
|
|
matches.push(mat.start()..mat.end());
|
|
}
|
|
} else {
|
|
let mut line = String::new();
|
|
let mut line_offset = 0;
|
|
for (chunk_ix, chunk) in rope.chunks().chain(["\n"]).enumerate() {
|
|
if (chunk_ix + 1) % YIELD_INTERVAL == 0 {
|
|
yield_now().await;
|
|
}
|
|
|
|
for (newline_ix, text) in chunk.split('\n').enumerate() {
|
|
if newline_ix > 0 {
|
|
for mat in regex.find_iter(&line) {
|
|
let start = line_offset + mat.start();
|
|
let end = line_offset + mat.end();
|
|
matches.push(start..end);
|
|
}
|
|
|
|
line_offset += line.len() + 1;
|
|
line.clear();
|
|
}
|
|
line.push_str(text);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
matches
|
|
}
|
|
|
|
pub fn is_empty(&self) -> bool {
|
|
self.as_str().is_empty()
|
|
}
|
|
|
|
pub fn as_str(&self) -> &str {
|
|
self.as_inner().as_str()
|
|
}
|
|
|
|
pub fn whole_word(&self) -> bool {
|
|
match self {
|
|
Self::Text { whole_word, .. } => *whole_word,
|
|
Self::Regex { whole_word, .. } => *whole_word,
|
|
}
|
|
}
|
|
|
|
pub fn case_sensitive(&self) -> bool {
|
|
match self {
|
|
Self::Text { case_sensitive, .. } => *case_sensitive,
|
|
Self::Regex { case_sensitive, .. } => *case_sensitive,
|
|
}
|
|
}
|
|
|
|
pub fn include_ignored(&self) -> bool {
|
|
match self {
|
|
Self::Text {
|
|
include_ignored, ..
|
|
} => *include_ignored,
|
|
Self::Regex {
|
|
include_ignored, ..
|
|
} => *include_ignored,
|
|
}
|
|
}
|
|
|
|
pub fn is_regex(&self) -> bool {
|
|
matches!(self, Self::Regex { .. })
|
|
}
|
|
|
|
pub fn files_to_include(&self) -> &[PathMatcher] {
|
|
self.as_inner().files_to_include()
|
|
}
|
|
|
|
pub fn files_to_exclude(&self) -> &[PathMatcher] {
|
|
self.as_inner().files_to_exclude()
|
|
}
|
|
|
|
pub fn file_matches(&self, file_path: Option<&Path>) -> bool {
|
|
match file_path {
|
|
Some(file_path) => {
|
|
let mut path = file_path.to_path_buf();
|
|
loop {
|
|
if self
|
|
.files_to_exclude()
|
|
.iter()
|
|
.any(|exclude_glob| exclude_glob.is_match(&path))
|
|
{
|
|
return false;
|
|
} else if self.files_to_include().is_empty()
|
|
|| self
|
|
.files_to_include()
|
|
.iter()
|
|
.any(|include_glob| include_glob.is_match(&path))
|
|
{
|
|
return true;
|
|
} else if !path.pop() {
|
|
return false;
|
|
}
|
|
}
|
|
}
|
|
None => self.files_to_include().is_empty(),
|
|
}
|
|
}
|
|
pub fn as_inner(&self) -> &SearchInputs {
|
|
match self {
|
|
Self::Regex { inner, .. } | Self::Text { inner, .. } => inner,
|
|
}
|
|
}
|
|
}
|
|
|
|
fn deserialize_path_matches(glob_set: &str) -> anyhow::Result<Vec<PathMatcher>> {
|
|
glob_set
|
|
.split(',')
|
|
.map(str::trim)
|
|
.filter(|glob_str| !glob_str.is_empty())
|
|
.map(|glob_str| {
|
|
PathMatcher::new(glob_str)
|
|
.with_context(|| format!("deserializing path match glob {glob_str}"))
|
|
})
|
|
.collect()
|
|
}
|
|
|
|
#[cfg(test)]
|
|
mod tests {
|
|
use super::*;
|
|
|
|
#[test]
|
|
fn path_matcher_creation_for_valid_paths() {
|
|
for valid_path in [
|
|
"file",
|
|
"Cargo.toml",
|
|
".DS_Store",
|
|
"~/dir/another_dir/",
|
|
"./dir/file",
|
|
"dir/[a-z].txt",
|
|
"../dir/filé",
|
|
] {
|
|
let path_matcher = PathMatcher::new(valid_path).unwrap_or_else(|e| {
|
|
panic!("Valid path {valid_path} should be accepted, but got: {e}")
|
|
});
|
|
assert!(
|
|
path_matcher.is_match(valid_path),
|
|
"Path matcher for valid path {valid_path} should match itself"
|
|
)
|
|
}
|
|
}
|
|
|
|
#[test]
|
|
fn path_matcher_creation_for_globs() {
|
|
for invalid_glob in ["dir/[].txt", "dir/[a-z.txt", "dir/{file"] {
|
|
match PathMatcher::new(invalid_glob) {
|
|
Ok(_) => panic!("Invalid glob {invalid_glob} should not be accepted"),
|
|
Err(_expected) => {}
|
|
}
|
|
}
|
|
|
|
for valid_glob in [
|
|
"dir/?ile",
|
|
"dir/*.txt",
|
|
"dir/**/file",
|
|
"dir/[a-z].txt",
|
|
"{dir,file}",
|
|
] {
|
|
match PathMatcher::new(valid_glob) {
|
|
Ok(_expected) => {}
|
|
Err(e) => panic!("Valid glob {valid_glob} should be accepted, but got: {e}"),
|
|
}
|
|
}
|
|
}
|
|
}
|