Yehowshua/ZIm
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Add new terminal hyperlink tests (#28525)

Browse source 
Part of #28238

This PR refactors `FindHyperlink` handling and associated code in
`terminal.rs` into its own file for improved testability, and adds
tests.

Release Notes:

- N/A
This commit is contained in:
Dave Waggoner 2025-06-06 05:08:20 -07:00 • committed by GitHub
parent 709523bf36
commit 8837e5564d
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
4 changed files with 1232 additions and 260 deletions
Download patch file Download diff file Expand all files Collapse all files

1
crates/terminal/Cargo.toml
View file

@ -39,3 +39,4 @@ windows.workspace = true
[dev-dependencies]
rand.workspace = true
url.workspace = true

273
crates/terminal/src/terminal.rs
View file

@ -3,6 +3,7 @@ pub mod mappings;
pub use alacritty_terminal;
mod pty_info;
mod terminal_hyperlinks;
pub mod terminal_settings;
use alacritty_terminal::{
@ -39,11 +40,11 @@ use mappings::mouse::{
use collections::{HashMap, VecDeque};
use futures::StreamExt;
use pty_info::PtyProcessInfo;
use regex::Regex;
use serde::{Deserialize, Serialize};
use settings::Settings;
use smol::channel::{Receiver, Sender};
use task::{HideStrategy, Shell, TaskId};
use terminal_hyperlinks::RegexSearches;
use terminal_settings::{AlternateScroll, CursorShape, TerminalSettings};
use theme::{ActiveTheme, Theme};
use util::{paths::home_dir, truncate_and_trailoff};
@ -52,10 +53,10 @@ use std::{
borrow::Cow,
cmp::{self, min},
fmt::Display,
ops::{Deref, Index, RangeInclusive},
ops::{Deref, RangeInclusive},
path::PathBuf,
process::ExitStatus,
sync::{Arc, LazyLock},
sync::Arc,
time::Duration,
};
use thiserror::Error;
@ -93,7 +94,6 @@ actions!(
const SCROLL_MULTIPLIER: f32 = 4.;
#[cfg(not(target_os = "macos"))]
const SCROLL_MULTIPLIER: f32 = 1.;
const MAX_SEARCH_LINES: usize = 100;
const DEBUG_TERMINAL_WIDTH: Pixels = px(500.);
const DEBUG_TERMINAL_HEIGHT: Pixels = px(30.);
const DEBUG_CELL_WIDTH: Pixels = px(5.);
@ -314,25 +314,6 @@ impl Display for TerminalError {
// https://github.com/alacritty/alacritty/blob/cb3a79dbf6472740daca8440d5166c1d4af5029e/extra/man/alacritty.5.scd?plain=1#L207-L213
const DEFAULT_SCROLL_HISTORY_LINES: usize = 10_000;
pub const MAX_SCROLL_HISTORY_LINES: usize = 100_000;
const URL_REGEX: &str = r#"(ipfs:|ipns:|magnet:|mailto:|gemini://|gopher://|https://|http://|news:|file://|git://|ssh:|ftp://)[^\u{0000}-\u{001F}\u{007F}-\u{009F}<>"\s{-}\^⟨⟩`]+"#;
// Optional suffix matches MSBuild diagnostic suffixes for path parsing in PathLikeWithPosition
// https://learn.microsoft.com/en-us/visualstudio/msbuild/msbuild-diagnostic-format-for-tasks
const WORD_REGEX: &str =
r#"[\$\+\w.\[\]:/\\@\-~()]+(?:\((?:\d+|\d+,\d+)\))|[\$\+\w.\[\]:/\\@\-~()]+"#;
const PYTHON_FILE_LINE_REGEX: &str = r#"File "(?P<file>[^"]+)", line (?P<line>\d+)"#;
static PYTHON_FILE_LINE_MATCHER: LazyLock<Regex> =
LazyLock::new(|| Regex::new(PYTHON_FILE_LINE_REGEX).unwrap());
fn python_extract_path_and_line(input: &str) -> Option<(&str, u32)> {
if let Some(captures) = PYTHON_FILE_LINE_MATCHER.captures(input) {
let path_part = captures.name("file")?.as_str();
let line_number: u32 = captures.name("line")?.as_str().parse().ok()?;
return Some((path_part, line_number));
}
None
}
pub struct TerminalBuilder {
terminal: Terminal,
@ -497,9 +478,7 @@ impl TerminalBuilder {
next_link_id: 0,
selection_phase: SelectionPhase::Ended,
// hovered_word: false,
url_regex: RegexSearch::new(URL_REGEX).unwrap(),
word_regex: RegexSearch::new(WORD_REGEX).unwrap(),
python_file_line_regex: RegexSearch::new(PYTHON_FILE_LINE_REGEX).unwrap(),
hyperlink_regex_searches: RegexSearches::new(),
vi_mode_enabled: false,
is_ssh_terminal,
python_venv_directory,
@ -657,9 +636,7 @@ pub struct Terminal {
scroll_px: Pixels,
next_link_id: usize,
selection_phase: SelectionPhase,
url_regex: RegexSearch,
word_regex: RegexSearch,
python_file_line_regex: RegexSearch,
hyperlink_regex_searches: RegexSearches,
task: Option<TaskState>,
vi_mode_enabled: bool,
is_ssh_terminal: bool,
@ -926,122 +903,14 @@ impl Terminal {
)
.grid_clamp(term, Boundary::Grid);
let link = term.grid().index(point).hyperlink();
let found_word = if link.is_some() {
let mut min_index = point;
loop {
let new_min_index = min_index.sub(term, Boundary::Cursor, 1);
if new_min_index == min_index
|| term.grid().index(new_min_index).hyperlink() != link
{
break;
} else {
min_index = new_min_index
}
}
let mut max_index = point;
loop {
let new_max_index = max_index.add(term, Boundary::Cursor, 1);
if new_max_index == max_index
|| term.grid().index(new_max_index).hyperlink() != link
{
break;
} else {
max_index = new_max_index
}
}
let url = link.unwrap().uri().to_owned();
let url_match = min_index..=max_index;
Some((url, true, url_match))
} else if let Some(url_match) = regex_match_at(term, point, &mut self.url_regex) {
let url = term.bounds_to_string(*url_match.start(), *url_match.end());
Some((url, true, url_match))
} else if let Some(python_match) =
regex_match_at(term, point, &mut self.python_file_line_regex)
{
let matching_line =
term.bounds_to_string(*python_match.start(), *python_match.end());
python_extract_path_and_line(&matching_line).map(|(file_path, line_number)| {
(format!("{file_path}:{line_number}"), false, python_match)
})
} else if let Some(word_match) = regex_match_at(term, point, &mut self.word_regex) {
let file_path = term.bounds_to_string(*word_match.start(), *word_match.end());
let (sanitized_match, sanitized_word) = 'sanitize: {
let mut word_match = word_match;
let mut file_path = file_path;
if is_path_surrounded_by_common_symbols(&file_path) {
word_match = Match::new(
word_match.start().add(term, Boundary::Grid, 1),
word_match.end().sub(term, Boundary::Grid, 1),
);
file_path = file_path[1..file_path.len() - 1].to_owned();
}
while file_path.ends_with(':') {
file_path.pop();
word_match = Match::new(
*word_match.start(),
word_match.end().sub(term, Boundary::Grid, 1),
);
}
let mut colon_count = 0;
for c in file_path.chars() {
if c == ':' {
colon_count += 1;
}
}
// strip trailing comment after colon in case of
// file/at/path.rs:row:column:description or error message
// so that the file path is `file/at/path.rs:row:column`
if colon_count > 2 {
let last_index = file_path.rfind(':').unwrap();
let prev_is_digit = last_index > 0
&& file_path
.chars()
.nth(last_index - 1)
.map_or(false, |c| c.is_ascii_digit());
let next_is_digit = last_index < file_path.len() - 1
&& file_path
.chars()
.nth(last_index + 1)
.map_or(true, |c| c.is_ascii_digit());
if prev_is_digit && !next_is_digit {
let stripped_len = file_path.len() - last_index;
word_match = Match::new(
*word_match.start(),
word_match.end().sub(term, Boundary::Grid, stripped_len),
);
file_path = file_path[0..last_index].to_owned();
}
}
break 'sanitize (word_match, file_path);
};
Some((sanitized_word, false, sanitized_match))
} else {
None
};
match found_word {
match terminal_hyperlinks::find_from_grid_point(
term,
point,
&mut self.hyperlink_regex_searches,
) {
Some((maybe_url_or_path, is_url, url_match)) => {
let target = if is_url {
// Treat "file://" URLs like file paths to ensure
// that line numbers at the end of the path are
// handled correctly
if let Some(path) = maybe_url_or_path.strip_prefix("file://") {
MaybeNavigationTarget::PathLike(PathLikeTarget {
maybe_path: path.to_string(),
terminal_dir: self.working_directory(),
})
} else {
MaybeNavigationTarget::Url(maybe_url_or_path.clone())
}
MaybeNavigationTarget::Url(maybe_url_or_path.clone())
} else {
MaybeNavigationTarget::PathLike(PathLikeTarget {
maybe_path: maybe_url_or_path.clone(),
@ -1954,14 +1823,6 @@ pub fn row_to_string(row: &Row<Cell>) -> String {
.collect::<String>()
}
fn is_path_surrounded_by_common_symbols(path: &str) -> bool {
// Avoid detecting `[]` or `()` strings as paths, surrounded by common symbols
path.len() > 2
// The rest of the brackets and various quotes cannot be matched by the [`WORD_REGEX`] hence not checked for.
&& (path.starts_with('[') && path.ends_with(']')
|| path.starts_with('(') && path.ends_with(')'))
}
const TASK_DELIMITER: &str = "⏵ ";
fn task_summary(task: &TaskState, error_code: Option<i32>) -> (bool, String, String) {
let escaped_full_label = task.full_label.replace("\r\n", "\r").replace('\n', "\r");
@ -2031,30 +1892,6 @@ impl Drop for Terminal {
impl EventEmitter<Event> for Terminal {}
/// Based on alacritty/src/display/hint.rs > regex_match_at
/// Retrieve the match, if the specified point is inside the content matching the regex.
fn regex_match_at<T>(term: &Term<T>, point: AlacPoint, regex: &mut RegexSearch) -> Option<Match> {
visible_regex_match_iter(term, regex).find(|rm| rm.contains(&point))
}
/// Copied from alacritty/src/display/hint.rs:
/// Iterate over all visible regex matches.
pub fn visible_regex_match_iter<'a, T>(
term: &'a Term<T>,
regex: &'a mut RegexSearch,
) -> impl Iterator<Item = Match> + 'a {
let viewport_start = Line(-(term.grid().display_offset() as i32));
let viewport_end = viewport_start + term.bottommost_line();
let mut start = term.line_search_left(AlacPoint::new(viewport_start, Column(0)));
let mut end = term.line_search_right(AlacPoint::new(viewport_end, Column(0)));
start.line = start.line.max(viewport_start - MAX_SEARCH_LINES);
end.line = end.line.min(viewport_end + MAX_SEARCH_LINES);
RegexIter::new(start, end, AlacDirection::Right, term, regex)
.skip_while(move |rm| rm.end().line < viewport_start)
.take_while(move |rm| rm.start().line <= viewport_end)
}
fn make_selection(range: &RangeInclusive<AlacPoint>) -> Selection {
let mut selection = Selection::new(SelectionType::Simple, *range.start(), AlacDirection::Left);
selection.update(*range.end(), AlacDirection::Right);
@ -2177,8 +2014,7 @@ mod tests {
use rand::{Rng, distributions::Alphanumeric, rngs::ThreadRng, thread_rng};
use crate::{
IndexedCell, TerminalBounds, TerminalContent, content_index_for_mouse,
python_extract_path_and_line, rgb_for_index,
IndexedCell, TerminalBounds, TerminalContent, content_index_for_mouse, rgb_for_index,
};
#[test]
@ -2312,87 +2148,4 @@ mod tests {
..Default::default()
}
}
fn re_test(re: &str, hay: &str, expected: Vec<&str>) {
let results: Vec<_> = regex::Regex::new(re)
.unwrap()
.find_iter(hay)
.map(|m| m.as_str())
.collect();
assert_eq!(results, expected);
}
#[test]
fn test_url_regex() {
re_test(
crate::URL_REGEX,
"test http://example.com test mailto:bob@example.com train",
vec!["http://example.com", "mailto:bob@example.com"],
);
}
#[test]
fn test_word_regex() {
re_test(
crate::WORD_REGEX,
"hello, world! \"What\" is this?",
vec!["hello", "world", "What", "is", "this"],
);
}
#[test]
fn test_word_regex_with_linenum() {
// filename(line) and filename(line,col) as used in MSBuild output
// should be considered a single "word", even though comma is
// usually a word separator
re_test(
crate::WORD_REGEX,
"a Main.cs(20) b",
vec!["a", "Main.cs(20)", "b"],
);
re_test(
crate::WORD_REGEX,
"Main.cs(20,5) Error desc",
vec!["Main.cs(20,5)", "Error", "desc"],
);
// filename:line:col is a popular format for unix tools
re_test(
crate::WORD_REGEX,
"a Main.cs:20:5 b",
vec!["a", "Main.cs:20:5", "b"],
);
// Some tools output "filename:line:col:message", which currently isn't
// handled correctly, but might be in the future
re_test(
crate::WORD_REGEX,
"Main.cs:20:5:Error desc",
vec!["Main.cs:20:5:Error", "desc"],
);
}
#[test]
fn test_python_file_line_regex() {
re_test(
crate::PYTHON_FILE_LINE_REGEX,
"hay File \"/zed/bad_py.py\", line 8 stack",
vec!["File \"/zed/bad_py.py\", line 8"],
);
re_test(crate::PYTHON_FILE_LINE_REGEX, "unrelated", vec![]);
}
#[test]
fn test_python_file_line() {
let inputs: Vec<(&str, Option<(&str, u32)>)> = vec![
(
"File \"/zed/bad_py.py\", line 8",
Some(("/zed/bad_py.py", 8u32)),
),
("File \"path/to/zed/bad_py.py\"", None),
("unrelated", None),
("", None),
];
let actual = inputs
.iter()
.map(|input| python_extract_path_and_line(input.0))
.collect::<Vec<_>>();
let expected = inputs.iter().map(|(_, output)| *output).collect::<Vec<_>>();
assert_eq!(actual, expected);
}
}

1217
crates/terminal/src/terminal_hyperlinks.rs Normal file
View file

File diff suppressed because it is too large Load diff