replace tui_markdown with a custom markdown renderer (#3396)

Also, simplify the streaming behavior.

This fixes a number of display issues with streaming markdown, and paves
the way for better markdown features (e.g. customizable styles, syntax
highlighting, markdown-aware wrapping).

Not currently supported:
- footnotes
- tables
- reference-style links
This commit is contained in:
Jeremy Rose
2025-09-10 12:13:53 -07:00
committed by GitHub
parent acb28bf914
commit 8068cc75f8
16 changed files with 2309 additions and 983 deletions

189
codex-rs/Cargo.lock generated
View File

@@ -311,15 +311,6 @@ version = "0.5.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3a8241f3ebb85c056b509d4327ad0358fbbba6ffb340bf388f26350aeda225b1"
[[package]]
name = "bincode"
version = "1.3.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b1f45e9417d87227c7a56d22e471c6206462cba514c7590c09aff4cf6d1ddcad"
dependencies = [
"serde",
]
[[package]]
name = "bit-set"
version = "0.5.3"
@@ -655,7 +646,7 @@ dependencies = [
"tokio-test",
"tokio-util",
"toml",
"toml_edit 0.23.4",
"toml_edit",
"tracing",
"tree-sitter",
"tree-sitter-bash",
@@ -879,6 +870,7 @@ dependencies = [
"path-clean",
"pathdiff",
"pretty_assertions",
"pulldown-cmark",
"rand 0.9.2",
"ratatui",
"regex-lite",
@@ -895,7 +887,6 @@ dependencies = [
"tracing",
"tracing-appender",
"tracing-subscriber",
"tui-markdown",
"unicode-segmentation",
"unicode-width 0.1.14",
"url",
@@ -1763,12 +1754,6 @@ version = "0.3.31"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f90f7dce0722e95104fcb095585910c0977252f286e354b5e3bd38902cd99988"
[[package]]
name = "futures-timer"
version = "3.0.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f288b0a4f20f9a56b5d1da57e2227c661b7b16168e2f72365f57b63326e29b24"
[[package]]
name = "futures-util"
version = "0.3.31"
@@ -1854,12 +1839,6 @@ version = "0.31.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "07e28edb80900c19c28f1072f2e8aeca7fa06b23cd4169cefe1af5aa3260783f"
[[package]]
name = "glob"
version = "0.3.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a8d1add55171497b4705a648c6b583acafb01d58050a51727785f0b2c8e0a2b2"
[[package]]
name = "globset"
version = "0.4.16"
@@ -2567,12 +2546,6 @@ dependencies = [
"libc",
]
[[package]]
name = "linked-hash-map"
version = "0.5.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0717cef1bc8b636c6e1c1bbdefc09e6322da8a9321966e8928ef80d20f7f770f"
[[package]]
name = "linux-raw-sys"
version = "0.4.15"
@@ -3014,28 +2987,6 @@ version = "1.70.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a4895175b425cb1f87721b59f0f286c2092bd4af812243672510e1ac53e2e0ad"
[[package]]
name = "onig"
version = "6.5.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "336b9c63443aceef14bea841b899035ae3abe89b7c486aaf4c5bd8aafedac3f0"
dependencies = [
"bitflags 2.9.1",
"libc",
"once_cell",
"onig_sys",
]
[[package]]
name = "onig_sys"
version = "69.9.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "c7f86c6eef3d6df15f23bcfb6af487cbd2fed4e5581d58d5bf1f5f8b7f6727dc"
dependencies = [
"cc",
"pkg-config",
]
[[package]]
name = "openssl"
version = "0.10.73"
@@ -3361,15 +3312,6 @@ dependencies = [
"yansi",
]
[[package]]
name = "proc-macro-crate"
version = "3.3.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "edce586971a4dfaa28950c6f18ed55e0406c1ab88bbce2c6f6293a7aaba73d35"
dependencies = [
"toml_edit 0.22.27",
]
[[package]]
name = "proc-macro2"
version = "1.0.95"
@@ -3381,9 +3323,9 @@ dependencies = [
[[package]]
name = "pulldown-cmark"
version = "0.13.0"
version = "0.10.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "1e8bbe1a966bd2f362681a44f6edce3c2310ac21e4d5067a6e7ec396297a6ea0"
checksum = "76979bea66e7875e7509c4ec5300112b316af87fa7a252ca91c448b32dfe3993"
dependencies = [
"bitflags 2.9.1",
"getopts",
@@ -3394,9 +3336,9 @@ dependencies = [
[[package]]
name = "pulldown-cmark-escape"
version = "0.11.0"
version = "0.10.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "007d8adb5ddab6f8e3f491ac63566a7d5002cc7ed73901f72057943fa71ae1ae"
checksum = "bd348ff538bc9caeda7ee8cad2d1d48236a1f443c1fa3913c6a02fe0043b1dd3"
[[package]]
name = "pxfm"
@@ -3627,12 +3569,6 @@ version = "0.8.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "2b15c43186be67a4fd63bee50d0303afffcef381492ebe2c5d87f324e1b8815c"
[[package]]
name = "relative-path"
version = "1.9.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ba39f3699c378cd8970968dcbff9c43159ea4cfbd88d43c00b22f2ef10a435d2"
[[package]]
name = "reqwest"
version = "0.12.23"
@@ -3691,51 +3627,12 @@ dependencies = [
"windows-sys 0.52.0",
]
[[package]]
name = "rstest"
version = "0.25.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6fc39292f8613e913f7df8fa892b8944ceb47c247b78e1b1ae2f09e019be789d"
dependencies = [
"futures-timer",
"futures-util",
"rstest_macros",
"rustc_version",
]
[[package]]
name = "rstest_macros"
version = "0.25.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "1f168d99749d307be9de54d23fd226628d99768225ef08f6ffb52e0182a27746"
dependencies = [
"cfg-if",
"glob",
"proc-macro-crate",
"proc-macro2",
"quote",
"regex",
"relative-path",
"rustc_version",
"syn 2.0.104",
"unicode-ident",
]
[[package]]
name = "rustc-demangle"
version = "0.1.25"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "989e6739f80c4ad5b13e0fd7fe89531180375b18520cc8c82080e4dc4035b84f"
[[package]]
name = "rustc_version"
version = "0.4.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "cfcb3a22ef46e85b45de6ee7e79d063319ebb6594faafcf1c225ea92ab6e9b92"
dependencies = [
"semver",
]
[[package]]
name = "rustix"
version = "0.38.44"
@@ -3975,12 +3872,6 @@ dependencies = [
"libc",
]
[[package]]
name = "semver"
version = "1.0.26"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "56e6fa9c48d24d85fb3de5ad847117517440f6beceb7798af16b4a87d616b8d0"
[[package]]
name = "serde"
version = "1.0.219"
@@ -4464,28 +4355,6 @@ dependencies = [
"syn 2.0.104",
]
[[package]]
name = "syntect"
version = "5.2.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "874dcfa363995604333cf947ae9f751ca3af4522c60886774c4963943b4746b1"
dependencies = [
"bincode",
"bitflags 1.3.2",
"flate2",
"fnv",
"once_cell",
"onig",
"plist",
"regex-syntax 0.8.5",
"serde",
"serde_derive",
"serde_json",
"thiserror 1.0.69",
"walkdir",
"yaml-rust",
]
[[package]]
name = "sys-locale"
version = "0.3.2"
@@ -4809,18 +4678,12 @@ dependencies = [
"indexmap 2.10.0",
"serde",
"serde_spanned",
"toml_datetime 0.7.0",
"toml_datetime",
"toml_parser",
"toml_writer",
"winnow",
]
[[package]]
name = "toml_datetime"
version = "0.6.11"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "22cddaf88f4fbc13c51aebbf5f8eceb5c7c5a9da2ac40a13519eb5b0a0e8f11c"
[[package]]
name = "toml_datetime"
version = "0.7.0"
@@ -4830,17 +4693,6 @@ dependencies = [
"serde",
]
[[package]]
name = "toml_edit"
version = "0.22.27"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "41fe8c660ae4257887cf66394862d21dbca4a6ddd26f04a3560410406a2f819a"
dependencies = [
"indexmap 2.10.0",
"toml_datetime 0.6.11",
"winnow",
]
[[package]]
name = "toml_edit"
version = "0.23.4"
@@ -4848,7 +4700,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7211ff1b8f0d3adae1663b7da9ffe396eabe1ca25f0b0bee42b0da29a9ddce93"
dependencies = [
"indexmap 2.10.0",
"toml_datetime 0.7.0",
"toml_datetime",
"toml_parser",
"toml_writer",
"winnow",
@@ -5058,22 +4910,6 @@ dependencies = [
"termcolor",
]
[[package]]
name = "tui-markdown"
version = "0.3.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d10648c25931bfaaf5334ff4e7dc5f3d830e0c50d7b0119b1d5cfe771f540536"
dependencies = [
"ansi-to-tui",
"itertools 0.14.0",
"pretty_assertions",
"pulldown-cmark",
"ratatui",
"rstest",
"syntect",
"tracing",
]
[[package]]
name = "typenum"
version = "1.18.0"
@@ -5855,15 +5691,6 @@ version = "0.13.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ec107c4503ea0b4a98ef47356329af139c0a4f7750e621cf2973cd3385ebcb3d"
[[package]]
name = "yaml-rust"
version = "0.4.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "56c1936c4cc7a1c9ab21a1ebb602eb942ba868cbd44a99cb7cdc5892335e1c85"
dependencies = [
"linked-hash-map",
]
[[package]]
name = "yansi"
version = "1.0.1"

View File

@@ -79,7 +79,7 @@ tokio-stream = "0.1.17"
tracing = { version = "0.1.41", features = ["log"] }
tracing-appender = "0.2.3"
tracing-subscriber = { version = "0.3.19", features = ["env-filter"] }
tui-markdown = "0.3.3"
pulldown-cmark = "0.10"
unicode-segmentation = "1.12.0"
unicode-width = "0.1"
url = "2"

View File

@@ -0,0 +1,15 @@
use std::io::Read;
use std::io::{self};
fn main() {
let mut input = String::new();
if let Err(err) = io::stdin().read_to_string(&mut input) {
eprintln!("failed to read stdin: {err}");
std::process::exit(1);
}
let parser = pulldown_cmark::Parser::new(&input);
for event in parser {
println!("{event:?}");
}
}

View File

@@ -0,0 +1,18 @@
---
source: tui/src/chatwidget/tests.rs
expression: visual
---
> -- Indented code block (4 spaces)
SELECT *
FROM "users"
WHERE "email" LIKE '%@example.com';
```sh
printf 'fenced within fenced\n'
```
{
// comment allowed in jsonc
"path": "C:\\Program Files\\App",
"regex": "^foo.*(bar)?$"
}

View File

@@ -1756,3 +1756,123 @@ fn chatwidget_exec_and_status_layout_vt100_snapshot() {
let visual = vt_lines.join("\n");
assert_snapshot!(visual);
}
// E2E vt100 snapshot for complex markdown with indented and nested fenced code blocks
#[test]
fn chatwidget_markdown_code_blocks_vt100_snapshot() {
let (mut chat, mut rx, _op_rx) = make_chatwidget_manual();
// Simulate a final agent message via streaming deltas instead of a single message
chat.handle_codex_event(Event {
id: "t1".into(),
msg: EventMsg::TaskStarted(TaskStartedEvent {
model_context_window: None,
}),
});
// Build a vt100 visual from the history insertions only (no UI overlay)
let width: u16 = 80;
let height: u16 = 50;
let backend = ratatui::backend::TestBackend::new(width, height);
let mut term = crate::custom_terminal::Terminal::with_options(backend).expect("terminal");
// Place viewport at the last line so that history lines insert above it
term.set_viewport_area(Rect::new(0, height - 1, width, 1));
let mut ansi: Vec<u8> = Vec::new();
// Simulate streaming via AgentMessageDelta in 2-character chunks (no final AgentMessage).
let source: &str = r#"
-- Indented code block (4 spaces)
SELECT *
FROM "users"
WHERE "email" LIKE '%@example.com';
````markdown
```sh
printf 'fenced within fenced\n'
```
````
```jsonc
{
// comment allowed in jsonc
"path": "C:\\Program Files\\App",
"regex": "^foo.*(bar)?$"
}
```
"#;
let mut it = source.chars();
loop {
let mut delta = String::new();
match it.next() {
Some(c) => delta.push(c),
None => break,
}
if let Some(c2) = it.next() {
delta.push(c2);
}
chat.handle_codex_event(Event {
id: "t1".into(),
msg: EventMsg::AgentMessageDelta(AgentMessageDeltaEvent { delta }),
});
// Drive commit ticks and drain emitted history lines into the vt100 buffer.
loop {
chat.on_commit_tick();
let mut inserted_any = false;
while let Ok(app_ev) = rx.try_recv() {
if let AppEvent::InsertHistoryCell(cell) = app_ev {
let lines = cell.display_lines(width);
crate::insert_history::insert_history_lines_to_writer(
&mut term, &mut ansi, lines,
);
inserted_any = true;
}
}
if !inserted_any {
break;
}
}
}
// Finalize the stream without sending a final AgentMessage, to flush any tail.
chat.handle_codex_event(Event {
id: "t1".into(),
msg: EventMsg::TaskComplete(TaskCompleteEvent {
last_agent_message: None,
}),
});
for lines in drain_insert_history(&mut rx) {
crate::insert_history::insert_history_lines_to_writer(&mut term, &mut ansi, lines);
}
let mut parser = vt100::Parser::new(height, width, 0);
parser.process(&ansi);
let mut vt_lines: Vec<String> = (0..height)
.map(|row| {
let mut s = String::with_capacity(width as usize);
for col in 0..width {
if let Some(cell) = parser.screen().cell(row, col) {
if let Some(ch) = cell.contents().chars().next() {
s.push(ch);
} else {
s.push(' ');
}
} else {
s.push(' ');
}
}
s.trim_end().to_string()
})
.collect();
// Compact trailing blank rows for a stable snapshot
while matches!(vt_lines.last(), Some(l) if l.trim().is_empty()) {
vt_lines.pop();
}
let visual = vt_lines.join("\n");
assert_snapshot!(visual);
}

View File

@@ -97,7 +97,17 @@ pub fn insert_history_lines_to_writer<B, W>(
for line in wrapped {
queue!(writer, Print("\r\n")).ok();
write_spans(writer, &line).ok();
// Merge line-level style into each span so that ANSI colors reflect
// line styles (e.g., blockquotes with green fg).
let merged_spans: Vec<Span> = line
.spans
.iter()
.map(|s| Span {
style: s.style.patch(line.style),
content: s.content.clone(),
})
.collect();
write_spans(writer, merged_spans.iter()).ok();
}
queue!(writer, ResetScrollRegion).ok();
@@ -264,6 +274,10 @@ where
#[cfg(test)]
mod tests {
use super::*;
use crate::markdown_render::render_markdown_text;
use ratatui::layout::Rect;
use ratatui::style::Color;
use vt100::Parser;
#[test]
fn writes_bold_then_regular_spans() {
@@ -292,4 +306,240 @@ mod tests {
String::from_utf8(expected).unwrap()
);
}
#[test]
fn vt100_blockquote_line_emits_green_fg() {
// Set up a small off-screen terminal
let width: u16 = 40;
let height: u16 = 10;
let backend = ratatui::backend::TestBackend::new(width, height);
let mut term = crate::custom_terminal::Terminal::with_options(backend).expect("terminal");
// Place viewport on the last line so history inserts scroll upward
let viewport = Rect::new(0, height - 1, width, 1);
term.set_viewport_area(viewport);
// Build a blockquote-like line: apply line-level green style and prefix "> "
let mut line: Line<'static> = Line::from(vec!["> ".into(), "Hello world".into()]);
line = line.style(Color::Green);
let mut ansi: Vec<u8> = Vec::new();
insert_history_lines_to_writer(&mut term, &mut ansi, vec![line]);
// Parse ANSI using vt100 and assert at least one non-default fg color appears
let mut parser = Parser::new(height, width, 0);
parser.process(&ansi);
let mut saw_colored = false;
'outer: for row in 0..height {
for col in 0..width {
if let Some(cell) = parser.screen().cell(row, col)
&& cell.has_contents()
&& cell.fgcolor() != vt100::Color::Default
{
saw_colored = true;
break 'outer;
}
}
}
assert!(
saw_colored,
"expected at least one colored cell in vt100 output"
);
}
#[test]
fn vt100_blockquote_wrap_preserves_color_on_all_wrapped_lines() {
// Force wrapping by using a narrow viewport width and a long blockquote line.
let width: u16 = 20;
let height: u16 = 8;
let backend = ratatui::backend::TestBackend::new(width, height);
let mut term = crate::custom_terminal::Terminal::with_options(backend).expect("terminal");
// Viewport is the last line so history goes directly above it.
let viewport = Rect::new(0, height - 1, width, 1);
term.set_viewport_area(viewport);
// Create a long blockquote with a distinct prefix and enough text to wrap.
let mut line: Line<'static> = Line::from(vec![
"> ".into(),
"This is a long quoted line that should wrap".into(),
]);
line = line.style(Color::Green);
let mut ansi: Vec<u8> = Vec::new();
insert_history_lines_to_writer(&mut term, &mut ansi, vec![line]);
// Parse and inspect the final screen buffer.
let mut parser = Parser::new(height, width, 0);
parser.process(&ansi);
let screen = parser.screen();
// Collect rows that are non-empty; these should correspond to our wrapped lines.
let mut non_empty_rows: Vec<u16> = Vec::new();
for row in 0..height {
let mut any = false;
for col in 0..width {
if let Some(cell) = screen.cell(row, col)
&& cell.has_contents()
&& cell.contents() != "\0"
&& cell.contents() != " "
{
any = true;
break;
}
}
if any {
non_empty_rows.push(row);
}
}
// Expect at least two rows due to wrapping.
assert!(
non_empty_rows.len() >= 2,
"expected wrapped output to span >=2 rows, got {non_empty_rows:?}",
);
// For each non-empty row, ensure all non-space cells are using a non-default fg color.
for row in non_empty_rows {
for col in 0..width {
if let Some(cell) = screen.cell(row, col) {
let contents = cell.contents();
if !contents.is_empty() && contents != " " {
assert!(
cell.fgcolor() != vt100::Color::Default,
"expected non-default fg on row {row} col {col}, got {:?}",
cell.fgcolor()
);
}
}
}
}
}
#[test]
fn vt100_colored_prefix_then_plain_text_resets_color() {
let width: u16 = 40;
let height: u16 = 6;
let backend = ratatui::backend::TestBackend::new(width, height);
let mut term = crate::custom_terminal::Terminal::with_options(backend).expect("terminal");
let viewport = Rect::new(0, height - 1, width, 1);
term.set_viewport_area(viewport);
// First span colored, rest plain.
let line: Line<'static> = Line::from(vec![
Span::styled("1. ", ratatui::style::Style::default().fg(Color::LightBlue)),
Span::raw("Hello world"),
]);
let mut ansi: Vec<u8> = Vec::new();
insert_history_lines_to_writer(&mut term, &mut ansi, vec![line]);
let mut parser = Parser::new(height, width, 0);
parser.process(&ansi);
let screen = parser.screen();
// Find the first non-empty row; verify first three cells are colored, following cells default.
'rows: for row in 0..height {
let mut has_text = false;
for col in 0..width {
if let Some(cell) = screen.cell(row, col)
&& cell.has_contents()
&& cell.contents() != " "
{
has_text = true;
break;
}
}
if !has_text {
continue;
}
// Expect "1. Hello world" starting at col 0.
for col in 0..3 {
let cell = screen.cell(row, col).unwrap();
assert!(
cell.fgcolor() != vt100::Color::Default,
"expected colored prefix at col {col}, got {:?}",
cell.fgcolor()
);
}
for col in 3..(3 + "Hello world".len() as u16) {
let cell = screen.cell(row, col).unwrap();
assert_eq!(
cell.fgcolor(),
vt100::Color::Default,
"expected default color for plain text at col {col}, got {:?}",
cell.fgcolor()
);
}
break 'rows;
}
}
#[test]
fn vt100_deep_nested_mixed_list_third_level_marker_is_colored() {
// Markdown with five levels (ordered → unordered → ordered → unordered → unordered).
let md = "1. First\n - Second level\n 1. Third level (ordered)\n - Fourth level (bullet)\n - Fifth level to test indent consistency\n";
let text = render_markdown_text(md);
let lines: Vec<Line<'static>> = text.lines.clone();
let width: u16 = 60;
let height: u16 = 12;
let backend = ratatui::backend::TestBackend::new(width, height);
let mut term = crate::custom_terminal::Terminal::with_options(backend).expect("terminal");
let viewport = ratatui::layout::Rect::new(0, height - 1, width, 1);
term.set_viewport_area(viewport);
let mut ansi: Vec<u8> = Vec::new();
insert_history_lines_to_writer(&mut term, &mut ansi, lines);
let mut parser = Parser::new(height, width, 0);
parser.process(&ansi);
let screen = parser.screen();
// Reconstruct screen rows as strings to locate the 3rd level line.
let mut rows: Vec<String> = Vec::with_capacity(height as usize);
for row in 0..height {
let mut s = String::with_capacity(width as usize);
for col in 0..width {
if let Some(cell) = screen.cell(row, col) {
if let Some(ch) = cell.contents().chars().next() {
s.push(ch);
} else {
s.push(' ');
}
} else {
s.push(' ');
}
}
rows.push(s.trim_end().to_string());
}
let needle = "1. Third level (ordered)";
let row_idx = rows
.iter()
.position(|r| r.contains(needle))
.unwrap_or_else(|| {
panic!("expected to find row containing {needle:?}, have rows: {rows:?}")
});
let col_start = rows[row_idx].find(needle).unwrap() as u16; // column where '1' starts
// Verify that the numeric marker ("1.") at the third level is colored
// (non-default fg) and the content after the following space resets to default.
for c in [col_start, col_start + 1] {
let cell = screen.cell(row_idx as u16, c).unwrap();
assert!(
cell.fgcolor() != vt100::Color::Default,
"expected colored 3rd-level marker at row {row_idx} col {c}, got {:?}",
cell.fgcolor()
);
}
let content_col = col_start + 3; // skip '1', '.', and the space
if let Some(cell) = screen.cell(row_idx as u16, content_col) {
assert_eq!(
cell.fgcolor(),
vt100::Color::Default,
"expected default color for 3rd-level content at row {row_idx} col {content_col}, got {:?}",
cell.fgcolor()
);
}
}
}

View File

@@ -45,6 +45,7 @@ pub mod insert_history;
mod key_hint;
pub mod live_wrap;
mod markdown;
mod markdown_render;
mod markdown_stream;
pub mod onboarding;
mod pager_overlay;

View File

@@ -1,8 +1,6 @@
use crate::citation_regex::CITATION_REGEX;
use codex_core::config::Config;
use codex_core::config_types::UriBasedFileOpener;
use ratatui::text::Line;
use std::borrow::Cow;
use std::path::Path;
pub(crate) fn append_markdown(
@@ -19,238 +17,13 @@ fn append_markdown_with_opener_and_cwd(
file_opener: UriBasedFileOpener,
cwd: &Path,
) {
// Historically, we fed the entire `markdown_source` into the renderer in
// one pass. However, fenced code blocks sometimes lost leading whitespace
// when formatted by the markdown renderer/highlighter. To preserve code
// block content exactly, split the source into "text" and "code" segments:
// - Render non-code text through `tui_markdown` (with citation rewrite).
// - Render code block content verbatim as plain lines without additional
// formatting, preserving leading spaces.
for seg in split_text_and_fences(markdown_source) {
match seg {
Segment::Text(s) => {
let processed = rewrite_file_citations(&s, file_opener, cwd);
let rendered = tui_markdown::from_str(&processed);
crate::render::line_utils::push_owned_lines(&rendered.lines, lines);
}
Segment::Code { content, .. } => {
// Emit the code content exactly as-is, line by line.
// We don't attempt syntax highlighting to avoid whitespace bugs.
for line in content.split_inclusive('\n') {
// split_inclusive keeps the trailing \n; we want lines without it.
let line = if let Some(stripped) = line.strip_suffix('\n') {
stripped
} else {
line
};
let owned_line: Line<'static> = line.to_string().into();
lines.push(owned_line);
}
}
}
}
}
/// Rewrites file citations in `src` into markdown hyperlinks using the
/// provided `scheme` (`vscode`, `cursor`, etc.). The resulting URI follows the
/// format expected by VS Code-compatible file openers:
///
/// ```text
/// <scheme>://file<ABS_PATH>:<LINE>
/// ```
fn rewrite_file_citations<'a>(
src: &'a str,
file_opener: UriBasedFileOpener,
cwd: &Path,
) -> Cow<'a, str> {
// Map enum values to the corresponding URI scheme strings.
let scheme: &str = match file_opener.get_scheme() {
Some(scheme) => scheme,
None => return Cow::Borrowed(src),
};
CITATION_REGEX.replace_all(src, |caps: &regex_lite::Captures<'_>| {
let file = &caps[1];
let start_line = &caps[2];
// Resolve the path against `cwd` when it is relative.
let absolute_path = {
let p = Path::new(file);
let absolute_path = if p.is_absolute() {
path_clean::clean(p)
} else {
path_clean::clean(cwd.join(p))
};
// VS Code expects forward slashes even on Windows because URIs use
// `/` as the path separator.
absolute_path.to_string_lossy().replace('\\', "/")
};
// Render as a normal markdown link so the downstream renderer emits
// the hyperlink escape sequence (when supported by the terminal).
//
// In practice, sometimes multiple citations for the same file, but with a
// different line number, are shown sequentially, so we:
// - include the line number in the label to disambiguate them
// - add a space after the link to make it easier to read
format!("[{file}:{start_line}]({scheme}://file{absolute_path}:{start_line}) ")
})
}
// use shared helper from `line_utils`
// Minimal code block splitting.
// - Recognizes fenced blocks opened by ``` or ~~~ (allowing leading whitespace).
// The opening fence may include a language string which we ignore.
// The closing fence must be on its own line (ignoring surrounding whitespace).
// - Additionally recognizes indented code blocks that begin after a blank line
// with a line starting with at least 4 spaces or a tab, and continue for
// consecutive lines that are blank or also indented by >= 4 spaces or a tab.
enum Segment {
Text(String),
Code {
_lang: Option<String>,
content: String,
},
}
fn split_text_and_fences(src: &str) -> Vec<Segment> {
let mut segments = Vec::new();
let mut curr_text = String::new();
#[derive(Copy, Clone, PartialEq)]
enum CodeMode {
None,
Fenced,
Indented,
}
let mut code_mode = CodeMode::None;
let mut fence_token = "";
let mut code_lang: Option<String> = None;
let mut code_content = String::new();
// We intentionally do not require a preceding blank line for indented code blocks,
// since streamed model output often omits it. This favors preserving indentation.
for line in src.split_inclusive('\n') {
let line_no_nl = line.strip_suffix('\n');
let trimmed_start = match line_no_nl {
Some(l) => l.trim_start(),
None => line.trim_start(),
};
if code_mode == CodeMode::None {
let open = if trimmed_start.starts_with("```") {
Some("```")
} else if trimmed_start.starts_with("~~~") {
Some("~~~")
} else {
None
};
if let Some(tok) = open {
// Flush pending text segment.
if !curr_text.is_empty() {
segments.push(Segment::Text(curr_text.clone()));
curr_text.clear();
}
fence_token = tok;
// Capture language after the token on this line (before newline).
let after = &trimmed_start[tok.len()..];
let lang = after.trim();
code_lang = if lang.is_empty() {
None
} else {
Some(lang.to_string())
};
code_mode = CodeMode::Fenced;
code_content.clear();
// Do not include the opening fence line in output.
continue;
}
// Check for start of an indented code block: only after a blank line
// (or at the beginning), and the line must start with >=4 spaces or a tab.
let raw_line = match line_no_nl {
Some(l) => l,
None => line,
};
let leading_spaces = raw_line.chars().take_while(|c| *c == ' ').count();
let starts_with_tab = raw_line.starts_with('\t');
// Consider any line that begins with >=4 spaces or a tab to start an
// indented code block. This favors preserving indentation even when a
// preceding blank line is omitted (common in streamed model output).
let starts_indented_code = (leading_spaces >= 4) || starts_with_tab;
if starts_indented_code {
// Flush pending text and begin an indented code block.
if !curr_text.is_empty() {
segments.push(Segment::Text(curr_text.clone()));
curr_text.clear();
}
code_mode = CodeMode::Indented;
code_content.clear();
code_content.push_str(line);
// Inside code now; do not treat this line as normal text.
continue;
}
// Normal text line.
curr_text.push_str(line);
} else {
match code_mode {
CodeMode::Fenced => {
// inside fenced code: check for closing fence on its own line
let trimmed = match line_no_nl {
Some(l) => l.trim(),
None => line.trim(),
};
if trimmed == fence_token {
// End code block: emit segment without fences
segments.push(Segment::Code {
_lang: code_lang.take(),
content: code_content.clone(),
});
code_content.clear();
code_mode = CodeMode::None;
fence_token = "";
continue;
}
// Accumulate code content exactly as-is.
code_content.push_str(line);
}
CodeMode::Indented => {
// Continue while the line is blank, or starts with >=4 spaces, or a tab.
let raw_line = match line_no_nl {
Some(l) => l,
None => line,
};
let is_blank = raw_line.trim().is_empty();
let leading_spaces = raw_line.chars().take_while(|c| *c == ' ').count();
let starts_with_tab = raw_line.starts_with('\t');
if is_blank || leading_spaces >= 4 || starts_with_tab {
code_content.push_str(line);
} else {
// Close the indented code block and reprocess this line as normal text.
segments.push(Segment::Code {
_lang: None,
content: code_content.clone(),
});
code_content.clear();
code_mode = CodeMode::None;
// Now handle current line as text.
curr_text.push_str(line);
}
}
CodeMode::None => unreachable!(),
}
}
}
if code_mode != CodeMode::None {
// Unterminated code fence: treat accumulated content as a code segment.
segments.push(Segment::Code {
_lang: code_lang.take(),
content: code_content.clone(),
});
} else if !curr_text.is_empty() {
segments.push(Segment::Text(curr_text.clone()));
}
segments
// Render via pulldown-cmark and rewrite citations during traversal (outside code blocks).
let rendered = crate::markdown_render::render_markdown_text_with_citations(
markdown_source,
file_opener.get_scheme(),
cwd,
);
crate::render::line_utils::push_owned_lines(&rendered.lines, lines);
}
#[cfg(test)]
@@ -258,88 +31,6 @@ mod tests {
use super::*;
use pretty_assertions::assert_eq;
#[test]
fn citation_is_rewritten_with_absolute_path() {
let markdown = "See 【F:/src/main.rs†L42-L50】 for details.";
let cwd = Path::new("/workspace");
let result = rewrite_file_citations(markdown, UriBasedFileOpener::VsCode, cwd);
assert_eq!(
"See [/src/main.rs:42](vscode://file/src/main.rs:42) for details.",
result
);
}
#[test]
fn citation_is_rewritten_with_relative_path() {
let markdown = "Refer to 【F:lib/mod.rs†L5】 here.";
let cwd = Path::new("/home/user/project");
let result = rewrite_file_citations(markdown, UriBasedFileOpener::Windsurf, cwd);
assert_eq!(
"Refer to [lib/mod.rs:5](windsurf://file/home/user/project/lib/mod.rs:5) here.",
result
);
}
#[test]
fn citation_followed_by_space_so_they_do_not_run_together() {
let markdown = "References on lines 【F:src/foo.rs†L24】【F:src/foo.rs†L42】";
let cwd = Path::new("/home/user/project");
let result = rewrite_file_citations(markdown, UriBasedFileOpener::VsCode, cwd);
assert_eq!(
"References on lines [src/foo.rs:24](vscode://file/home/user/project/src/foo.rs:24) [src/foo.rs:42](vscode://file/home/user/project/src/foo.rs:42) ",
result
);
}
#[test]
fn citation_unchanged_without_file_opener() {
let markdown = "Look at 【F:file.rs†L1】.";
let cwd = Path::new("/");
let unchanged = rewrite_file_citations(markdown, UriBasedFileOpener::VsCode, cwd);
// The helper itself always rewrites this test validates behaviour of
// append_markdown when `file_opener` is None.
let mut out = Vec::new();
append_markdown_with_opener_and_cwd(markdown, &mut out, UriBasedFileOpener::None, cwd);
// Convert lines back to string for comparison.
let rendered: String = out
.iter()
.flat_map(|l| l.spans.iter())
.map(|s| s.content.clone())
.collect::<Vec<_>>()
.join("");
assert_eq!(markdown, rendered);
// Ensure helper rewrites.
assert_ne!(markdown, unchanged);
}
#[test]
fn fenced_code_blocks_preserve_leading_whitespace() {
let src = "```\n indented\n\t\twith tabs\n four spaces\n```\n";
let cwd = Path::new("/");
let mut out = Vec::new();
append_markdown_with_opener_and_cwd(src, &mut out, UriBasedFileOpener::None, cwd);
let rendered: Vec<String> = out
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(
rendered,
vec![
" indented".to_string(),
"\t\twith tabs".to_string(),
" four spaces".to_string()
]
);
}
#[test]
fn citations_not_rewritten_inside_code_blocks() {
let src = "Before 【F:/x.rs†L1】\n```\nInside 【F:/x.rs†L2】\n```\nAfter 【F:/x.rs†L3】\n";
@@ -355,19 +46,31 @@ mod tests {
.collect::<String>()
})
.collect();
// Expect first and last lines rewritten, middle line unchanged.
assert!(rendered[0].contains("vscode://file"));
assert_eq!(rendered[1], "Inside 【F:/x.rs†L2】");
assert!(matches!(rendered.last(), Some(s) if s.contains("vscode://file")));
// Expect a line containing the inside text unchanged.
assert!(rendered.iter().any(|s| s.contains("Inside 【F:/x.rs†L2】")));
// And first/last sections rewritten.
assert!(
rendered
.first()
.map(|s| s.contains("vscode://file"))
.unwrap_or(false)
);
assert!(
rendered
.last()
.map(|s| s.contains("vscode://file"))
.unwrap_or(false)
);
}
#[test]
fn indented_code_blocks_preserve_leading_whitespace() {
let src = "Before\n code 1\n\tcode with tab\n code 2\nAfter\n";
// Basic sanity: indented code with surrounding blank lines should produce the indented line.
let src = "Before\n\n code 1\n\nAfter\n";
let cwd = Path::new("/");
let mut out = Vec::new();
append_markdown_with_opener_and_cwd(src, &mut out, UriBasedFileOpener::None, cwd);
let rendered: Vec<String> = out
let lines: Vec<String> = out
.iter()
.map(|l| {
l.spans
@@ -376,16 +79,7 @@ mod tests {
.collect::<String>()
})
.collect();
assert_eq!(
rendered,
vec![
"Before".to_string(),
" code 1".to_string(),
"\tcode with tab".to_string(),
" code 2".to_string(),
"After".to_string()
]
);
assert_eq!(lines, vec!["Before", "", " code 1", "", "After"]);
}
#[test]
@@ -403,11 +97,17 @@ mod tests {
.collect::<String>()
})
.collect();
// Expect first and last lines rewritten, and the indented code line present
// unchanged (citations inside not rewritten). We do not assert on blank
// separator lines since the markdown renderer may normalize them.
assert!(rendered.iter().any(|s| s.contains("vscode://file")));
assert!(rendered.iter().any(|s| s == " Inside 【F:/x.rs†L2】"));
assert!(
rendered
.iter()
.any(|s| s.contains("Start") && s.contains("vscode://file"))
);
assert!(
rendered
.iter()
.any(|s| s.contains("End") && s.contains("vscode://file"))
);
assert!(rendered.iter().any(|s| s.contains("Inside 【F:/x.rs†L2】")));
}
#[test]
@@ -435,27 +135,6 @@ mod tests {
);
}
#[test]
fn tui_markdown_splits_ordered_marker_and_text() {
// With marker and content on the same line, tui_markdown keeps it as one line
// even in the surrounding section context.
let rendered = tui_markdown::from_str("Loose vs. tight list items:\n1. Tight item\n");
let lines: Vec<String> = rendered
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert!(
lines.iter().any(|w| w == "1. Tight item"),
"expected single line '1. Tight item' in context: {lines:?}"
);
}
#[test]
fn append_markdown_matches_tui_markdown_for_ordered_item() {
use codex_core::config_types::UriBasedFileOpener;
@@ -480,72 +159,6 @@ mod tests {
assert_eq!(lines, vec!["1. Tight item".to_string()]);
}
#[test]
fn tui_markdown_shape_for_loose_tight_section() {
// Use the exact source from the session deltas used in tests.
let source = r#"
Loose vs. tight list items:
1. Tight item
2. Another tight item
3.
Loose item
"#;
let rendered = tui_markdown::from_str(source);
let lines: Vec<String> = rendered
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
// Join into a single string and assert the exact shape we observe
// from tui_markdown in this larger context (marker and content split).
let joined = {
let mut s = String::new();
for (i, l) in lines.iter().enumerate() {
s.push_str(l);
if i + 1 < lines.len() {
s.push('\n');
}
}
s
};
let expected = r#"Loose vs. tight list items:
1.
Tight item
2.
Another tight item
3.
Loose item"#;
assert_eq!(
joined, expected,
"unexpected tui_markdown shape: {joined:?}"
);
}
#[test]
fn split_text_and_fences_keeps_ordered_list_line_as_text() {
// No fences here; expect a single Text segment containing the full input.
let src = "Loose vs. tight list items:\n1. Tight item\n";
let segs = super::split_text_and_fences(src);
assert_eq!(
segs.len(),
1,
"expected single text segment, got {}",
segs.len()
);
match &segs[0] {
super::Segment::Text(s) => assert_eq!(s, src),
_ => panic!("expected Text segment for non-fence input"),
}
}
#[test]
fn append_markdown_keeps_ordered_list_line_unsplit_in_context() {
use codex_core::config_types::UriBasedFileOpener;

View File

@@ -0,0 +1,566 @@
use crate::citation_regex::CITATION_REGEX;
use pulldown_cmark::CodeBlockKind;
use pulldown_cmark::CowStr;
use pulldown_cmark::Event;
use pulldown_cmark::HeadingLevel;
use pulldown_cmark::Options;
use pulldown_cmark::Parser;
use pulldown_cmark::Tag;
use pulldown_cmark::TagEnd;
use ratatui::style::Style;
use ratatui::style::Stylize;
use ratatui::text::Line;
use ratatui::text::Span;
use ratatui::text::Text;
use std::borrow::Cow;
use std::path::Path;
#[derive(Clone, Debug)]
struct IndentContext {
prefix: Vec<Span<'static>>,
marker: Option<Vec<Span<'static>>>,
is_list: bool,
}
impl IndentContext {
fn new(prefix: Vec<Span<'static>>, marker: Option<Vec<Span<'static>>>, is_list: bool) -> Self {
Self {
prefix,
marker,
is_list,
}
}
}
#[allow(dead_code)]
pub(crate) fn render_markdown_text(input: &str) -> Text<'static> {
let mut options = Options::empty();
options.insert(Options::ENABLE_STRIKETHROUGH);
let parser = Parser::new_ext(input, options);
let mut w = Writer::new(parser, None, None);
w.run();
w.text
}
pub(crate) fn render_markdown_text_with_citations(
input: &str,
scheme: Option<&str>,
cwd: &Path,
) -> Text<'static> {
let mut options = Options::empty();
options.insert(Options::ENABLE_STRIKETHROUGH);
let parser = Parser::new_ext(input, options);
let mut w = Writer::new(
parser,
scheme.map(|s| s.to_string()),
Some(cwd.to_path_buf()),
);
w.run();
w.text
}
struct Writer<'a, I>
where
I: Iterator<Item = Event<'a>>,
{
iter: I,
text: Text<'static>,
inline_styles: Vec<Style>,
indent_stack: Vec<IndentContext>,
list_indices: Vec<Option<u64>>,
link: Option<String>,
needs_newline: bool,
pending_marker_line: bool,
in_paragraph: bool,
scheme: Option<String>,
cwd: Option<std::path::PathBuf>,
in_code_block: bool,
}
impl<'a, I> Writer<'a, I>
where
I: Iterator<Item = Event<'a>>,
{
fn new(iter: I, scheme: Option<String>, cwd: Option<std::path::PathBuf>) -> Self {
Self {
iter,
text: Text::default(),
inline_styles: Vec::new(),
indent_stack: Vec::new(),
list_indices: Vec::new(),
link: None,
needs_newline: false,
pending_marker_line: false,
in_paragraph: false,
scheme,
cwd,
in_code_block: false,
}
}
fn run(&mut self) {
while let Some(ev) = self.iter.next() {
self.handle_event(ev);
}
}
fn handle_event(&mut self, event: Event<'a>) {
match event {
Event::Start(tag) => self.start_tag(tag),
Event::End(tag) => self.end_tag(tag),
Event::Text(text) => self.text(text),
Event::Code(code) => self.code(code),
Event::SoftBreak => self.soft_break(),
Event::HardBreak => self.hard_break(),
Event::Rule => {
if !self.text.lines.is_empty() {
self.push_blank_line();
}
self.push_line(Line::from("———"));
self.needs_newline = true;
}
Event::Html(html) => self.html(html, false),
Event::InlineHtml(html) => self.html(html, true),
Event::FootnoteReference(_) => {}
Event::TaskListMarker(_) => {}
}
}
fn start_tag(&mut self, tag: Tag<'a>) {
match tag {
Tag::Paragraph => self.start_paragraph(),
Tag::Heading { level, .. } => self.start_heading(level),
Tag::BlockQuote => self.start_blockquote(),
Tag::CodeBlock(kind) => {
let indent = match kind {
CodeBlockKind::Fenced(_) => None,
CodeBlockKind::Indented => Some(Span::from(" ".repeat(4))),
};
let lang = match kind {
CodeBlockKind::Fenced(lang) => Some(lang.to_string()),
CodeBlockKind::Indented => None,
};
self.start_codeblock(lang, indent)
}
Tag::List(start) => self.start_list(start),
Tag::Item => self.start_item(),
Tag::Emphasis => self.push_inline_style(Style::new().italic()),
Tag::Strong => self.push_inline_style(Style::new().bold()),
Tag::Strikethrough => self.push_inline_style(Style::new().crossed_out()),
Tag::Link { dest_url, .. } => self.push_link(dest_url.to_string()),
Tag::HtmlBlock
| Tag::FootnoteDefinition(_)
| Tag::Table(_)
| Tag::TableHead
| Tag::TableRow
| Tag::TableCell
| Tag::Image { .. }
| Tag::MetadataBlock(_) => {}
}
}
fn end_tag(&mut self, tag: TagEnd) {
match tag {
TagEnd::Paragraph => self.end_paragraph(),
TagEnd::Heading(_) => self.end_heading(),
TagEnd::BlockQuote => self.end_blockquote(),
TagEnd::CodeBlock => self.end_codeblock(),
TagEnd::List(_) => self.end_list(),
TagEnd::Item => {
self.indent_stack.pop();
self.pending_marker_line = false;
}
TagEnd::Emphasis | TagEnd::Strong | TagEnd::Strikethrough => self.pop_inline_style(),
TagEnd::Link => self.pop_link(),
TagEnd::HtmlBlock
| TagEnd::FootnoteDefinition
| TagEnd::Table
| TagEnd::TableHead
| TagEnd::TableRow
| TagEnd::TableCell
| TagEnd::Image
| TagEnd::MetadataBlock(_) => {}
}
}
fn start_paragraph(&mut self) {
if self.needs_newline {
self.push_blank_line();
}
self.push_line(Line::default());
self.needs_newline = false;
self.in_paragraph = true;
}
fn end_paragraph(&mut self) {
self.needs_newline = true;
self.in_paragraph = false;
self.pending_marker_line = false;
}
fn start_heading(&mut self, level: HeadingLevel) {
if self.needs_newline {
self.push_line(Line::default());
self.needs_newline = false;
}
let heading_style = match level {
HeadingLevel::H1 => Style::new().bold().underlined(),
HeadingLevel::H2 => Style::new().bold(),
HeadingLevel::H3 => Style::new().bold().italic(),
HeadingLevel::H4 => Style::new().italic(),
HeadingLevel::H5 => Style::new().italic(),
HeadingLevel::H6 => Style::new().italic(),
};
let content = format!("{} ", "#".repeat(level as usize));
self.push_line(Line::from(vec![Span::styled(content, heading_style)]));
self.push_inline_style(heading_style);
self.needs_newline = false;
}
fn end_heading(&mut self) {
self.needs_newline = true;
self.pop_inline_style();
}
fn start_blockquote(&mut self) {
if self.needs_newline {
self.push_blank_line();
self.needs_newline = false;
}
self.indent_stack
.push(IndentContext::new(vec![Span::from("> ")], None, false));
}
fn end_blockquote(&mut self) {
self.indent_stack.pop();
self.needs_newline = true;
}
fn text(&mut self, text: CowStr<'a>) {
if self.pending_marker_line {
self.push_line(Line::default());
}
self.pending_marker_line = false;
if self.in_code_block
&& !self.needs_newline
&& self
.text
.lines
.last()
.map(|line| !line.spans.is_empty())
.unwrap_or(false)
{
self.push_line(Line::default());
}
for (i, line) in text.lines().enumerate() {
if self.needs_newline {
self.push_line(Line::default());
self.needs_newline = false;
}
if i > 0 {
self.push_line(Line::default());
}
let mut content = line.to_string();
if !self.in_code_block
&& let (Some(scheme), Some(cwd)) = (&self.scheme, &self.cwd)
{
let cow = rewrite_file_citations_with_scheme(&content, Some(scheme.as_str()), cwd);
if let std::borrow::Cow::Owned(s) = cow {
content = s;
}
}
let span = Span::styled(
content,
self.inline_styles.last().copied().unwrap_or_default(),
);
self.push_span(span);
}
self.needs_newline = false;
}
fn code(&mut self, code: CowStr<'a>) {
if self.pending_marker_line {
self.push_line(Line::default());
self.pending_marker_line = false;
}
let span = Span::from(code.into_string()).dim();
self.push_span(span);
}
fn html(&mut self, html: CowStr<'a>, inline: bool) {
self.pending_marker_line = false;
for (i, line) in html.lines().enumerate() {
if self.needs_newline {
self.push_line(Line::default());
self.needs_newline = false;
}
if i > 0 {
self.push_line(Line::default());
}
let style = self.inline_styles.last().copied().unwrap_or_default();
self.push_span(Span::styled(line.to_string(), style));
}
self.needs_newline = !inline;
}
fn hard_break(&mut self) {
self.push_line(Line::default());
}
fn soft_break(&mut self) {
self.push_line(Line::default());
}
fn start_list(&mut self, index: Option<u64>) {
if self.list_indices.is_empty() && self.needs_newline {
self.push_line(Line::default());
}
self.list_indices.push(index);
}
fn end_list(&mut self) {
self.list_indices.pop();
self.needs_newline = true;
}
fn start_item(&mut self) {
self.pending_marker_line = true;
let depth = self.list_indices.len();
let is_ordered = self
.list_indices
.last()
.map(|index| index.is_some())
.unwrap_or(false);
let width = depth * 4 - 3;
let marker = if let Some(last_index) = self.list_indices.last_mut() {
match last_index {
None => Some(vec![Span::from(" ".repeat(width - 1) + "- ")]),
Some(index) => {
*index += 1;
Some(vec![format!("{:width$}. ", *index - 1).light_blue()])
}
}
} else {
None
};
let indent_prefix = if depth == 0 {
Vec::new()
} else {
let indent_len = if is_ordered { width + 2 } else { width + 1 };
vec![Span::from(" ".repeat(indent_len))]
};
self.indent_stack
.push(IndentContext::new(indent_prefix, marker, true));
self.needs_newline = false;
}
fn start_codeblock(&mut self, _lang: Option<String>, indent: Option<Span<'static>>) {
if !self.text.lines.is_empty() {
self.push_blank_line();
}
self.in_code_block = true;
self.indent_stack.push(IndentContext::new(
vec![indent.unwrap_or_default()],
None,
false,
));
// let opener = match lang {
// Some(l) if !l.is_empty() => format!("```{l}"),
// _ => "```".to_string(),
// };
// self.push_line(opener.into());
self.needs_newline = true;
}
fn end_codeblock(&mut self) {
// self.push_line("```".into());
self.needs_newline = true;
self.in_code_block = false;
self.indent_stack.pop();
}
fn push_inline_style(&mut self, style: Style) {
let current = self.inline_styles.last().copied().unwrap_or_default();
let merged = current.patch(style);
self.inline_styles.push(merged);
}
fn pop_inline_style(&mut self) {
self.inline_styles.pop();
}
fn push_link(&mut self, dest_url: String) {
self.link = Some(dest_url);
}
fn pop_link(&mut self) {
if let Some(link) = self.link.take() {
self.push_span(" (".into());
self.push_span(link.cyan().underlined());
self.push_span(")".into());
}
}
fn push_line(&mut self, line: Line<'static>) {
let mut line = line;
let was_pending = self.pending_marker_line;
let mut spans = self.current_prefix_spans();
spans.append(&mut line.spans);
let blockquote_active = self
.indent_stack
.iter()
.any(|ctx| ctx.prefix.iter().any(|s| s.content.contains('>')));
let style = if blockquote_active {
Style::new().green()
} else {
line.style
};
self.text.lines.push(Line::from_iter(spans).style(style));
if was_pending {
self.pending_marker_line = false;
}
}
fn push_span(&mut self, span: Span<'static>) {
if let Some(last) = self.text.lines.last_mut() {
last.push_span(span);
} else {
self.push_line(Line::from(vec![span]));
}
}
fn push_blank_line(&mut self) {
if self.indent_stack.iter().all(|ctx| ctx.is_list) {
self.text.lines.push(Line::default());
} else {
self.push_line(Line::default());
}
}
fn current_prefix_spans(&self) -> Vec<Span<'static>> {
let mut prefix: Vec<Span<'static>> = Vec::new();
let last_marker_index = if self.pending_marker_line {
self.indent_stack
.iter()
.enumerate()
.rev()
.find_map(|(i, ctx)| if ctx.marker.is_some() { Some(i) } else { None })
} else {
None
};
let last_list_index = self.indent_stack.iter().rposition(|ctx| ctx.is_list);
for (i, ctx) in self.indent_stack.iter().enumerate() {
if self.pending_marker_line {
if Some(i) == last_marker_index
&& let Some(marker) = &ctx.marker
{
prefix.extend(marker.iter().cloned());
continue;
}
if ctx.is_list && last_marker_index.is_some_and(|idx| idx > i) {
continue;
}
} else if ctx.is_list && Some(i) != last_list_index {
continue;
}
prefix.extend(ctx.prefix.iter().cloned());
}
prefix
}
}
pub(crate) fn rewrite_file_citations_with_scheme<'a>(
src: &'a str,
scheme_opt: Option<&str>,
cwd: &Path,
) -> Cow<'a, str> {
let scheme: &str = match scheme_opt {
Some(s) => s,
None => return Cow::Borrowed(src),
};
CITATION_REGEX.replace_all(src, |caps: &regex_lite::Captures<'_>| {
let file = &caps[1];
let start_line = &caps[2];
// Resolve the path against `cwd` when it is relative.
let absolute_path = {
let p = Path::new(file);
let absolute_path = if p.is_absolute() {
path_clean::clean(p)
} else {
path_clean::clean(cwd.join(p))
};
// VS Code expects forward slashes even on Windows because URIs use
// `/` as the path separator.
absolute_path.to_string_lossy().replace('\\', "/")
};
// Render as a normal markdown link so the downstream renderer emits
// the hyperlink escape sequence (when supported by the terminal).
//
// In practice, sometimes multiple citations for the same file, but with a
// different line number, are shown sequentially, so we:
// - include the line number in the label to disambiguate them
// - add a space after the link to make it easier to read
format!("[{file}:{start_line}]({scheme}://file{absolute_path}:{start_line}) ")
})
}
#[cfg(test)]
mod markdown_render_tests {
include!("markdown_render_tests.rs");
}
#[cfg(test)]
mod tests {
use super::*;
use pretty_assertions::assert_eq;
#[test]
fn citation_is_rewritten_with_absolute_path() {
let markdown = "See 【F:/src/main.rs†L42-L50】 for details.";
let cwd = Path::new("/workspace");
let result = rewrite_file_citations_with_scheme(markdown, Some("vscode"), cwd);
assert_eq!(
"See [/src/main.rs:42](vscode://file/src/main.rs:42) for details.",
result
);
}
#[test]
fn citation_followed_by_space_so_they_do_not_run_together() {
let markdown = "References on lines 【F:src/foo.rs†L24】【F:src/foo.rs†L42】";
let cwd = Path::new("/home/user/project");
let result = rewrite_file_citations_with_scheme(markdown, Some("vscode"), cwd);
assert_eq!(
"References on lines [src/foo.rs:24](vscode://file/home/user/project/src/foo.rs:24) [src/foo.rs:42](vscode://file/home/user/project/src/foo.rs:42) ",
result
);
}
#[test]
fn citation_unchanged_without_file_opener() {
let markdown = "Look at 【F:file.rs†L1】.";
let cwd = Path::new("/");
let unchanged = rewrite_file_citations_with_scheme(markdown, Some("vscode"), cwd);
// The helper itself always rewrites this test validates behaviour of
// append_markdown when `file_opener` is None.
let rendered = render_markdown_text_with_citations(markdown, None, cwd);
// Convert lines back to string for comparison.
let rendered: String = rendered
.lines
.iter()
.flat_map(|l| l.spans.iter())
.map(|s| s.content.clone())
.collect::<Vec<_>>()
.join("");
assert_eq!(markdown, rendered);
// Ensure helper rewrites.
assert_ne!(markdown, unchanged);
}
}

View File

@@ -0,0 +1,995 @@
use pretty_assertions::assert_eq;
use ratatui::style::Stylize;
use ratatui::text::Line;
use ratatui::text::Span;
use ratatui::text::Text;
use crate::markdown_render::render_markdown_text;
use insta::assert_snapshot;
#[test]
fn empty() {
assert_eq!(render_markdown_text(""), Text::default());
}
#[test]
fn paragraph_single() {
assert_eq!(
render_markdown_text("Hello, world!"),
Text::from("Hello, world!")
);
}
#[test]
fn paragraph_soft_break() {
assert_eq!(
render_markdown_text("Hello\nWorld"),
Text::from_iter(["Hello", "World"])
);
}
#[test]
fn paragraph_multiple() {
assert_eq!(
render_markdown_text("Paragraph 1\n\nParagraph 2"),
Text::from_iter(["Paragraph 1", "", "Paragraph 2"])
);
}
#[test]
fn headings() {
let md = "# Heading 1\n## Heading 2\n### Heading 3\n#### Heading 4\n##### Heading 5\n###### Heading 6\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["# ".bold().underlined(), "Heading 1".bold().underlined()]),
Line::default(),
Line::from_iter(["## ".bold(), "Heading 2".bold()]),
Line::default(),
Line::from_iter(["### ".bold().italic(), "Heading 3".bold().italic()]),
Line::default(),
Line::from_iter(["#### ".italic(), "Heading 4".italic()]),
Line::default(),
Line::from_iter(["##### ".italic(), "Heading 5".italic()]),
Line::default(),
Line::from_iter(["###### ".italic(), "Heading 6".italic()]),
]);
assert_eq!(text, expected);
}
#[test]
fn blockquote_single() {
let text = render_markdown_text("> Blockquote");
let expected = Text::from(Line::from_iter(["> ", "Blockquote"]).green());
assert_eq!(text, expected);
}
#[test]
fn blockquote_soft_break() {
// Soft break via lazy continuation should render as a new line in blockquotes.
let text = render_markdown_text("> This is a blockquote\nwith a soft break\n");
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(
lines,
vec![
"> This is a blockquote".to_string(),
"> with a soft break".to_string()
]
);
}
#[test]
fn blockquote_multiple_with_break() {
let text = render_markdown_text("> Blockquote 1\n\n> Blockquote 2\n");
let expected = Text::from_iter([
Line::from_iter(["> ", "Blockquote 1"]).green(),
Line::default(),
Line::from_iter(["> ", "Blockquote 2"]).green(),
]);
assert_eq!(text, expected);
}
#[test]
fn blockquote_three_paragraphs_short_lines() {
let md = "> one\n>\n> two\n>\n> three\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["> ", "one"]).green(),
Line::from_iter(["> "]).green(),
Line::from_iter(["> ", "two"]).green(),
Line::from_iter(["> "]).green(),
Line::from_iter(["> ", "three"]).green(),
]);
assert_eq!(text, expected);
}
#[test]
fn blockquote_nested_two_levels() {
let md = "> Level 1\n>> Level 2\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["> ", "Level 1"]).green(),
Line::from_iter(["> "]).green(),
Line::from_iter(["> ", "> ", "Level 2"]).green(),
]);
assert_eq!(text, expected);
}
#[test]
fn blockquote_with_list_items() {
let md = "> - item 1\n> - item 2\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["> ", "- ", "item 1"]).green(),
Line::from_iter(["> ", "- ", "item 2"]).green(),
]);
assert_eq!(text, expected);
}
#[test]
fn blockquote_with_ordered_list() {
let md = "> 1. first\n> 2. second\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(vec![
Span::from("> "),
"1. ".light_blue(),
Span::from("first"),
])
.green(),
Line::from_iter(vec![
Span::from("> "),
"2. ".light_blue(),
Span::from("second"),
])
.green(),
]);
assert_eq!(text, expected);
}
#[test]
fn blockquote_list_then_nested_blockquote() {
let md = "> - parent\n> > child\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["> ", "- ", "parent"]).green(),
Line::from_iter(["> ", " ", "> ", "child"]).green(),
]);
assert_eq!(text, expected);
}
#[test]
fn list_item_with_inline_blockquote_on_same_line() {
let md = "1. > quoted\n";
let text = render_markdown_text(md);
let mut lines = text.lines.iter();
let first = lines.next().expect("one line");
// Expect content to include the ordered marker, a space, "> ", and the text
let s: String = first.spans.iter().map(|sp| sp.content.clone()).collect();
assert_eq!(s, "1. > quoted");
}
#[test]
fn blockquote_surrounded_by_blank_lines() {
let md = "foo\n\n> bar\n\nbaz\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(
lines,
vec![
"foo".to_string(),
"".to_string(),
"> bar".to_string(),
"".to_string(),
"baz".to_string(),
]
);
}
#[test]
fn blockquote_in_ordered_list_on_next_line() {
// Blockquote begins on a new line within an ordered list item; it should
// render inline on the same marker line.
let md = "1.\n > quoted\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(lines, vec!["1. > quoted".to_string()]);
}
#[test]
fn blockquote_in_unordered_list_on_next_line() {
// Blockquote begins on a new line within an unordered list item; it should
// render inline on the same marker line.
let md = "-\n > quoted\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(lines, vec!["- > quoted".to_string()]);
}
#[test]
fn blockquote_two_paragraphs_inside_ordered_list_has_blank_line() {
// Two blockquote paragraphs inside a list item should be separated by a blank line.
let md = "1.\n > para 1\n >\n > para 2\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(
lines,
vec![
"1. > para 1".to_string(),
" > ".to_string(),
" > para 2".to_string(),
],
"expected blockquote content to stay aligned after list marker"
);
}
#[test]
fn blockquote_inside_nested_list() {
let md = "1. A\n - B\n > inner\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(lines, vec!["1. A", " - B", " > inner"]);
}
#[test]
fn list_item_text_then_blockquote() {
let md = "1. before\n > quoted\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(lines, vec!["1. before", " > quoted"]);
}
#[test]
fn list_item_blockquote_then_text() {
let md = "1.\n > quoted\n after\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(lines, vec!["1. > quoted", " > after"]);
}
#[test]
fn list_item_text_blockquote_text() {
let md = "1. before\n > quoted\n after\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(lines, vec!["1. before", " > quoted", " > after"]);
}
#[test]
fn blockquote_with_heading_and_paragraph() {
let md = "> # Heading\n> paragraph text\n";
let text = render_markdown_text(md);
// Validate on content shape; styling is handled elsewhere
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(
lines,
vec![
"> # Heading".to_string(),
"> ".to_string(),
"> paragraph text".to_string(),
]
);
}
#[test]
fn blockquote_heading_inherits_heading_style() {
let text = render_markdown_text("> # test header\n> in blockquote\n");
assert_eq!(
text.lines,
[
Line::from_iter([
"> ".into(),
"# ".bold().underlined(),
"test header".bold().underlined(),
])
.green(),
Line::from_iter(["> "]).green(),
Line::from_iter(["> ", "in blockquote"]).green(),
]
);
}
#[test]
fn blockquote_with_code_block() {
let md = "> ```\n> code\n> ```\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(lines, vec!["> code".to_string()]);
}
#[test]
fn blockquote_with_multiline_code_block() {
let md = "> ```\n> first\n> second\n> ```\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(lines, vec!["> first", "> second"]);
}
#[test]
fn nested_blockquote_with_inline_and_fenced_code() {
/*
let md = \"> Nested quote with code:\n\
> > Inner quote and `inline code`\n\
> >\n\
> > ```\n\
> > # fenced code inside a quote\n\
> > echo \"hello from a quote\"\n\
> > ```\n";
*/
let md = r#"> Nested quote with code:
> > Inner quote and `inline code`
> >
> > ```
> > # fenced code inside a quote
> > echo "hello from a quote"
> > ```
"#;
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(
lines,
vec![
"> Nested quote with code:".to_string(),
"> ".to_string(),
"> > Inner quote and inline code".to_string(),
"> > ".to_string(),
"> > # fenced code inside a quote".to_string(),
"> > echo \"hello from a quote\"".to_string(),
]
);
}
#[test]
fn list_unordered_single() {
let text = render_markdown_text("- List item 1\n");
let expected = Text::from_iter([Line::from_iter(["- ", "List item 1"])]);
assert_eq!(text, expected);
}
#[test]
fn list_unordered_multiple() {
let text = render_markdown_text("- List item 1\n- List item 2\n");
let expected = Text::from_iter([
Line::from_iter(["- ", "List item 1"]),
Line::from_iter(["- ", "List item 2"]),
]);
assert_eq!(text, expected);
}
#[test]
fn list_ordered() {
let text = render_markdown_text("1. List item 1\n2. List item 2\n");
let expected = Text::from_iter([
Line::from_iter(["1. ".light_blue(), "List item 1".into()]),
Line::from_iter(["2. ".light_blue(), "List item 2".into()]),
]);
assert_eq!(text, expected);
}
#[test]
fn list_nested() {
let text = render_markdown_text("- List item 1\n - Nested list item 1\n");
let expected = Text::from_iter([
Line::from_iter(["- ", "List item 1"]),
Line::from_iter([" - ", "Nested list item 1"]),
]);
assert_eq!(text, expected);
}
#[test]
fn list_ordered_custom_start() {
let text = render_markdown_text("3. First\n4. Second\n");
let expected = Text::from_iter([
Line::from_iter(["3. ".light_blue(), "First".into()]),
Line::from_iter(["4. ".light_blue(), "Second".into()]),
]);
assert_eq!(text, expected);
}
#[test]
fn nested_unordered_in_ordered() {
let md = "1. Outer\n - Inner A\n - Inner B\n2. Next\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["1. ".light_blue(), "Outer".into()]),
Line::from_iter([" - ", "Inner A"]),
Line::from_iter([" - ", "Inner B"]),
Line::from_iter(["2. ".light_blue(), "Next".into()]),
]);
assert_eq!(text, expected);
}
#[test]
fn nested_ordered_in_unordered() {
let md = "- Outer\n 1. One\n 2. Two\n- Last\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["- ", "Outer"]),
Line::from_iter([" 1. ".light_blue(), "One".into()]),
Line::from_iter([" 2. ".light_blue(), "Two".into()]),
Line::from_iter(["- ", "Last"]),
]);
assert_eq!(text, expected);
}
#[test]
fn loose_list_item_multiple_paragraphs() {
let md = "1. First paragraph\n\n Second paragraph of same item\n\n2. Next item\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["1. ".light_blue(), "First paragraph".into()]),
Line::default(),
Line::from_iter([" ", "Second paragraph of same item"]),
Line::from_iter(["2. ".light_blue(), "Next item".into()]),
]);
assert_eq!(text, expected);
}
#[test]
fn tight_item_with_soft_break() {
let md = "- item line1\n item line2\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["- ", "item line1"]),
Line::from_iter([" ", "item line2"]),
]);
assert_eq!(text, expected);
}
#[test]
fn deeply_nested_mixed_three_levels() {
let md = "1. A\n - B\n 1. C\n2. D\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["1. ".light_blue(), "A".into()]),
Line::from_iter([" - ", "B"]),
Line::from_iter([" 1. ".light_blue(), "C".into()]),
Line::from_iter(["2. ".light_blue(), "D".into()]),
]);
assert_eq!(text, expected);
}
#[test]
fn loose_items_due_to_blank_line_between_items() {
let md = "1. First\n\n2. Second\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["1. ".light_blue(), "First".into()]),
Line::from_iter(["2. ".light_blue(), "Second".into()]),
]);
assert_eq!(text, expected);
}
#[test]
fn mixed_tight_then_loose_in_one_list() {
let md = "1. Tight\n\n2.\n Loose\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["1. ".light_blue(), "Tight".into()]),
Line::from_iter(["2. ".light_blue(), "Loose".into()]),
]);
assert_eq!(text, expected);
}
#[test]
fn ordered_item_with_indented_continuation_is_tight() {
let md = "1. Foo\n Bar\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["1. ".light_blue(), "Foo".into()]),
Line::from_iter([" ", "Bar"]),
]);
assert_eq!(text, expected);
}
#[test]
fn inline_code() {
let text = render_markdown_text("Example of `Inline code`");
let expected = Line::from_iter(["Example of ".into(), "Inline code".dim()]).into();
assert_eq!(text, expected);
}
#[test]
fn strong() {
assert_eq!(
render_markdown_text("**Strong**"),
Text::from(Line::from("Strong".bold()))
);
}
#[test]
fn emphasis() {
assert_eq!(
render_markdown_text("*Emphasis*"),
Text::from(Line::from("Emphasis".italic()))
);
}
#[test]
fn strikethrough() {
assert_eq!(
render_markdown_text("~~Strikethrough~~"),
Text::from(Line::from("Strikethrough".crossed_out()))
);
}
#[test]
fn strong_emphasis() {
let text = render_markdown_text("**Strong *emphasis***");
let expected = Text::from(Line::from_iter([
"Strong ".bold(),
"emphasis".bold().italic(),
]));
assert_eq!(text, expected);
}
#[test]
fn link() {
let text = render_markdown_text("[Link](https://example.com)");
let expected = Text::from(Line::from_iter([
"Link".into(),
" (".into(),
"https://example.com".cyan().underlined(),
")".into(),
]));
assert_eq!(text, expected);
}
#[test]
fn code_block_unhighlighted() {
let text = render_markdown_text("```rust\nfn main() {}\n```\n");
let expected = Text::from_iter([Line::from_iter(["", "fn main() {}"])]);
assert_eq!(text, expected);
}
#[test]
fn code_block_multiple_lines_root() {
let md = "```\nfirst\nsecond\n```\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["", "first"]),
Line::from_iter(["", "second"]),
]);
assert_eq!(text, expected);
}
#[test]
fn code_block_indented() {
let md = " function greet() {\n console.log(\"Hi\");\n }\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter([" ", "function greet() {"]),
Line::from_iter([" ", " console.log(\"Hi\");"]),
Line::from_iter([" ", "}"]),
]);
assert_eq!(text, expected);
}
#[test]
fn horizontal_rule_renders_em_dashes() {
let md = "Before\n\n---\n\nAfter\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(lines, vec!["Before", "", "———", "", "After"]);
}
#[test]
fn code_block_with_inner_triple_backticks_outer_four() {
let md = r#"````text
Here is a code block that shows another fenced block:
```md
# Inside fence
- bullet
- `inline code`
```
````
"#;
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(
lines,
vec![
"Here is a code block that shows another fenced block:".to_string(),
String::new(),
"```md".to_string(),
"# Inside fence".to_string(),
"- bullet".to_string(),
"- `inline code`".to_string(),
"```".to_string(),
]
);
}
#[test]
fn code_block_inside_unordered_list_item_is_indented() {
let md = "- Item\n\n ```\n code line\n ```\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(lines, vec!["- Item", "", " code line"]);
}
#[test]
fn code_block_multiple_lines_inside_unordered_list() {
let md = "- Item\n\n ```\n first\n second\n ```\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(lines, vec!["- Item", "", " first", " second"]);
}
#[test]
fn code_block_inside_unordered_list_item_multiple_lines() {
let md = "- Item\n\n ```\n first\n second\n ```\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(lines, vec!["- Item", "", " first", " second"]);
}
#[test]
fn markdown_render_complex_snapshot() {
let md = r#"# H1: Markdown Streaming Test
Intro paragraph with bold **text**, italic *text*, and inline code `x=1`.
Combined bold-italic ***both*** and escaped asterisks \*literal\*.
Auto-link: <https://example.com> and reference link [ref][r1].
Link with title: [hover me](https://example.com "Example") and mailto <mailto:test@example.com>.
Image: ![alt text](https://example.com/img.png "Title")
> Blockquote level 1
>> Blockquote level 2 with `inline code`
- Unordered list item 1
- Nested bullet with italics _inner_
- Unordered list item 2 with ~~strikethrough~~
1. Ordered item one
2. Ordered item two with sublist:
1) Alt-numbered subitem
- [ ] Task: unchecked
- [x] Task: checked with link [home](https://example.org)
---
Table below (alignment test):
| Left | Center | Right |
|:-----|:------:|------:|
| a | b | c |
Inline HTML: <sup>sup</sup> and <sub>sub</sub>.
HTML block:
<div style="border:1px solid #ccc;padding:2px">inline block</div>
Escapes: \_underscores\_, backslash \\, ticks ``code with `backtick` inside``.
Emoji shortcodes: :sparkles: :tada: (if supported).
Hard break test (line ends with two spaces)
Next line should be close to previous.
Footnote reference here[^1] and another[^longnote].
Horizontal rule with asterisks:
***
Fenced code block (JSON):
```json
{ "a": 1, "b": [true, false] }
```
Fenced code with tildes and triple backticks inside:
~~~markdown
To close ``` you need tildes.
~~~
Indented code block:
for i in range(3): print(i)
Definition-like list:
Term
: Definition with `code`.
Character entities: &amp; &lt; &gt; &quot; &#39;
[^1]: This is the first footnote.
[^longnote]: A longer footnote with a link to [Rust](https://www.rust-lang.org/).
Escaped pipe in text: a \| b \| c.
URL with parentheses: [link](https://example.com/path_(with)_parens).
[r1]: https://example.com/ref "Reference link title"
"#;
let text = render_markdown_text(md);
// Convert to plain text lines for snapshot (ignore styles)
let rendered = text
.lines
.iter()
.map(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
})
.collect::<Vec<_>>()
.join("\n");
assert_snapshot!(rendered);
}
#[test]
fn ordered_item_with_code_block_and_nested_bullet() {
let md = "1. **item 1**\n\n2. **item 2**\n ```\n code\n ```\n - `PROCESS_START` (a `OnceLock<Instant>`) keeps the start time for the entire process.\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|line| {
line.spans
.iter()
.map(|span| span.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(
lines,
vec![
"1. item 1".to_string(),
"2. item 2".to_string(),
String::new(),
" code".to_string(),
" - PROCESS_START (a OnceLock<Instant>) keeps the start time for the entire process.".to_string(),
]
);
}
#[test]
fn nested_five_levels_mixed_lists() {
let md = "1. First\n - Second level\n 1. Third level (ordered)\n - Fourth level (bullet)\n - Fifth level to test indent consistency\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["1. ".light_blue(), "First".into()]),
Line::from_iter([" - ", "Second level"]),
Line::from_iter([" 1. ".light_blue(), "Third level (ordered)".into()]),
Line::from_iter([" - ", "Fourth level (bullet)"]),
Line::from_iter([
" - ",
"Fifth level to test indent consistency",
]),
]);
assert_eq!(text, expected);
}
#[test]
fn html_inline_is_verbatim() {
let md = "Hello <span>world</span>!";
let text = render_markdown_text(md);
let expected: Text = Line::from_iter(["Hello ", "<span>", "world", "</span>", "!"]).into();
assert_eq!(text, expected);
}
#[test]
fn html_block_is_verbatim_multiline() {
let md = "<div>\n <span>hi</span>\n</div>\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["<div>"]),
Line::from_iter([" <span>hi</span>"]),
Line::from_iter(["</div>"]),
]);
assert_eq!(text, expected);
}
#[test]
fn html_in_tight_ordered_item_soft_breaks_with_space() {
let md = "1. Foo\n <i>Bar</i>\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["1. ".light_blue(), "Foo".into()]),
Line::from_iter([" ", "<i>", "Bar", "</i>"]),
]);
assert_eq!(text, expected);
}
#[test]
fn html_continuation_paragraph_in_unordered_item_indented() {
let md = "- Item\n\n <em>continued</em>\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["- ", "Item"]),
Line::default(),
Line::from_iter([" ", "<em>", "continued", "</em>"]),
]);
assert_eq!(text, expected);
}
#[test]
fn unordered_item_continuation_paragraph_is_indented() {
let md = "- Intro\n\n Continuation paragraph line 1\n Continuation paragraph line 2\n";
let text = render_markdown_text(md);
let lines: Vec<String> = text
.lines
.iter()
.map(|line| {
line.spans
.iter()
.map(|span| span.content.clone())
.collect::<String>()
})
.collect();
assert_eq!(
lines,
vec![
"- Intro".to_string(),
String::new(),
" Continuation paragraph line 1".to_string(),
" Continuation paragraph line 2".to_string(),
]
);
}
#[test]
fn ordered_item_continuation_paragraph_is_indented() {
let md = "1. Intro\n\n More details about intro\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["1. ".light_blue(), "Intro".into()]),
Line::default(),
Line::from_iter([" ", "More details about intro"]),
]);
assert_eq!(text, expected);
}
#[test]
fn nested_item_continuation_paragraph_is_indented() {
let md = "1. A\n - B\n\n Continuation for B\n2. C\n";
let text = render_markdown_text(md);
let expected = Text::from_iter([
Line::from_iter(["1. ".light_blue(), "A".into()]),
Line::from_iter([" - ", "B"]),
Line::default(),
Line::from_iter([" ", "Continuation for B"]),
Line::from_iter(["2. ".light_blue(), "C".into()]),
]);
assert_eq!(text, expected);
}

View File

@@ -4,8 +4,6 @@ use codex_core::config::Config;
use ratatui::text::Line;
use crate::markdown;
use crate::render::markdown_utils::is_inside_unclosed_fence;
use crate::render::markdown_utils::strip_empty_fenced_code_blocks;
/// Newline-gated accumulator that renders markdown and commits only fully
/// completed logical lines.
@@ -42,6 +40,7 @@ impl MarkdownStreamCollector {
}
pub fn push_delta(&mut self, delta: &str) {
tracing::trace!("push_delta: {delta:?}");
self.buffer.push_str(delta);
}
@@ -49,14 +48,15 @@ impl MarkdownStreamCollector {
/// since the last commit. When the buffer does not end with a newline, the
/// final rendered line is considered incomplete and is not emitted.
pub fn commit_complete_lines(&mut self, config: &Config) -> Vec<Line<'static>> {
// In non-test builds, unwrap an outer ```markdown fence during commit as well,
// so fence markers never appear in streamed history.
let source = unwrap_markdown_language_fence_if_enabled(self.buffer.clone());
let source = strip_empty_fenced_code_blocks(&source);
let source = self.buffer.clone();
let last_newline_idx = source.rfind('\n');
let source = if let Some(last_newline_idx) = last_newline_idx {
source[..=last_newline_idx].to_string()
} else {
return Vec::new();
};
let mut rendered: Vec<Line<'static>> = Vec::new();
markdown::append_markdown(&source, &mut rendered, config);
let mut complete_line_count = rendered.len();
if complete_line_count > 0
&& crate::render::line_utils::is_blank_line_spaces_only(
@@ -65,87 +65,12 @@ impl MarkdownStreamCollector {
{
complete_line_count -= 1;
}
// Heuristic: if the buffer ends with a double newline and the last non-blank
// rendered line looks like a list bullet with inline content (e.g., "- item"),
// defer committing that line. Subsequent context (e.g., another list item)
// can cause the renderer to split the bullet marker and text into separate
// logical lines ("- " then "item"), which would otherwise duplicate content.
if self.buffer.ends_with("\n\n") && complete_line_count > 0 {
let last = &rendered[complete_line_count - 1];
let mut text = String::new();
for s in &last.spans {
text.push_str(&s.content);
}
if text.starts_with("- ") && text.trim() != "-" {
complete_line_count = complete_line_count.saturating_sub(1);
}
}
if !self.buffer.ends_with('\n') {
complete_line_count = complete_line_count.saturating_sub(1);
// If we're inside an unclosed fenced code block, also drop the
// last rendered line to avoid committing a partial code line.
if is_inside_unclosed_fence(&source) {
complete_line_count = complete_line_count.saturating_sub(1);
}
// If the next (incomplete) line appears to begin a list item,
// also defer the previous completed line because the renderer may
// retroactively treat it as part of the list (e.g., ordered list item 1).
if let Some(last_nl) = source.rfind('\n') {
let tail = &source[last_nl + 1..];
if starts_with_list_marker(tail) {
complete_line_count = complete_line_count.saturating_sub(1);
}
}
}
// Conservatively withhold trailing list-like lines (unordered or ordered)
// because streaming mid-item can cause the renderer to later split or
// restructure them (e.g., duplicating content or separating the marker).
// Only defers lines at the end of the out slice so previously committed
// lines remain stable.
if complete_line_count > self.committed_line_count {
let mut safe_count = complete_line_count;
while safe_count > self.committed_line_count {
let l = &rendered[safe_count - 1];
let mut text = String::new();
for s in &l.spans {
text.push_str(&s.content);
}
let listish = is_potentially_volatile_list_line(&text);
if listish {
safe_count -= 1;
continue;
}
break;
}
complete_line_count = safe_count;
}
if self.committed_line_count >= complete_line_count {
return Vec::new();
}
let out_slice = &rendered[self.committed_line_count..complete_line_count];
// Strong correctness: while a fenced code block is open (no closing fence yet),
// do not emit any new lines from inside it. Wait until the fence closes to emit
// the entire block together. This avoids stray backticks and misformatted content.
if is_inside_unclosed_fence(&source) {
return Vec::new();
}
// Additional conservative hold-back: if exactly one short, plain word
// line would be emitted, defer it. This avoids committing a lone word
// that might become the first ordered-list item once the next delta
// arrives (e.g., next line starts with "2 " or "2. ").
if out_slice.len() == 1 {
let mut s = String::new();
for sp in &out_slice[0].spans {
s.push_str(&sp.content);
}
if is_short_plain_word(&s) {
return Vec::new();
}
}
let out = out_slice.to_vec();
self.committed_line_count = complete_line_count;
@@ -157,12 +82,19 @@ impl MarkdownStreamCollector {
/// for rendering. Optionally unwraps ```markdown language fences in
/// non-test builds.
pub fn finalize_and_drain(&mut self, config: &Config) -> Vec<Line<'static>> {
let mut source: String = self.buffer.clone();
let raw_buffer = self.buffer.clone();
let mut source: String = raw_buffer.clone();
if !source.ends_with('\n') {
source.push('\n');
}
let source = unwrap_markdown_language_fence_if_enabled(source);
let source = strip_empty_fenced_code_blocks(&source);
tracing::debug!(
raw_len = raw_buffer.len(),
source_len = source.len(),
"markdown finalize (raw length: {}, rendered length: {})",
raw_buffer.len(),
source.len()
);
tracing::trace!("markdown finalize (raw source):\n---\n{source}\n---");
let mut rendered: Vec<Line<'static>> = Vec::new();
markdown::append_markdown(&source, &mut rendered, config);
@@ -179,122 +111,6 @@ impl MarkdownStreamCollector {
}
}
#[inline]
fn is_potentially_volatile_list_line(text: &str) -> bool {
let t = text.trim_end();
if t == "-" || t == "*" || t == "- " || t == "* " {
return true;
}
if t.starts_with("- ") || t.starts_with("* ") {
return true;
}
// ordered list like "1. " or "23. "
let mut it = t.chars().peekable();
let mut saw_digit = false;
while let Some(&ch) = it.peek() {
if ch.is_ascii_digit() {
saw_digit = true;
it.next();
continue;
}
break;
}
if saw_digit && it.peek() == Some(&'.') {
// consume '.'
it.next();
if it.peek() == Some(&' ') {
return true;
}
}
false
}
#[inline]
fn starts_with_list_marker(text: &str) -> bool {
let t = text.trim_start();
if t.starts_with("- ") || t.starts_with("* ") || t.starts_with("-\t") || t.starts_with("*\t") {
return true;
}
// ordered list marker like "1 ", "1. ", "23 ", "23. "
let mut it = t.chars().peekable();
let mut saw_digit = false;
while let Some(&ch) = it.peek() {
if ch.is_ascii_digit() {
saw_digit = true;
it.next();
} else {
break;
}
}
if !saw_digit {
return false;
}
match it.peek() {
Some('.') => {
it.next();
matches!(it.peek(), Some(' '))
}
Some(' ') => true,
_ => false,
}
}
#[inline]
fn is_short_plain_word(s: &str) -> bool {
let t = s.trim();
if t.is_empty() || t.len() > 5 {
return false;
}
t.chars().all(|c| c.is_alphanumeric())
}
/// fence helpers are provided by `crate::render::markdown_utils`
#[cfg(test)]
fn unwrap_markdown_language_fence_if_enabled(s: String) -> String {
// In tests, keep content exactly as provided to simplify assertions.
s
}
#[cfg(not(test))]
fn unwrap_markdown_language_fence_if_enabled(s: String) -> String {
// Best-effort unwrap of a single outer fenced markdown block.
// Recognizes common forms like ```markdown, ```md (any case), optional
// surrounding whitespace, and flexible trailing newlines/CRLF.
// If the block is not recognized, return the input unchanged.
let lines = s.lines().collect::<Vec<_>>();
if lines.len() < 2 {
return s;
}
// Identify opening fence and language.
let open = lines.first().map(|l| l.trim_start()).unwrap_or("");
if !open.starts_with("```") {
return s;
}
let lang = open.trim_start_matches("```").trim();
let is_markdown_lang = lang.eq_ignore_ascii_case("markdown") || lang.eq_ignore_ascii_case("md");
if !is_markdown_lang {
return s;
}
// Find the last non-empty line and ensure it is a closing fence.
let mut last_idx = lines.len() - 1;
while last_idx > 0 && lines[last_idx].trim().is_empty() {
last_idx -= 1;
}
if lines[last_idx].trim() != "```" {
return s;
}
// Reconstruct the inner content between the fences.
let mut out = String::new();
for l in lines.iter().take(last_idx).skip(1) {
out.push_str(l);
out.push('\n');
}
out
}
pub(crate) struct StepResult {
pub history: Vec<Line<'static>>, // lines to insert into history this step
}
@@ -373,6 +189,7 @@ mod tests {
use super::*;
use codex_core::config::Config;
use codex_core::config::ConfigOverrides;
use ratatui::style::Color;
fn test_config() -> Config {
let overrides = ConfigOverrides {
@@ -406,6 +223,125 @@ mod tests {
assert_eq!(out.len(), 1);
}
#[test]
fn e2e_stream_blockquote_simple_is_green() {
let cfg = test_config();
let out = super::simulate_stream_markdown_for_tests(&["> Hello\n"], true, &cfg);
assert_eq!(out.len(), 1);
let l = &out[0];
assert_eq!(
l.style.fg,
Some(Color::Green),
"expected blockquote line fg green, got {:?}",
l.style.fg
);
}
#[test]
fn e2e_stream_blockquote_nested_is_green() {
let cfg = test_config();
let out =
super::simulate_stream_markdown_for_tests(&["> Level 1\n>> Level 2\n"], true, &cfg);
// Filter out any blank lines that may be inserted at paragraph starts.
let non_blank: Vec<_> = out
.into_iter()
.filter(|l| {
let s = l
.spans
.iter()
.map(|sp| sp.content.clone())
.collect::<Vec<_>>()
.join("");
let t = s.trim();
// Ignore quote-only blank lines like ">" inserted at paragraph boundaries.
!(t.is_empty() || t == ">")
})
.collect();
assert_eq!(non_blank.len(), 2);
assert_eq!(non_blank[0].style.fg, Some(Color::Green));
assert_eq!(non_blank[1].style.fg, Some(Color::Green));
}
#[test]
fn e2e_stream_blockquote_with_list_items_is_green() {
let cfg = test_config();
let out =
super::simulate_stream_markdown_for_tests(&["> - item 1\n> - item 2\n"], true, &cfg);
assert_eq!(out.len(), 2);
assert_eq!(out[0].style.fg, Some(Color::Green));
assert_eq!(out[1].style.fg, Some(Color::Green));
}
#[test]
fn e2e_stream_nested_mixed_lists_ordered_marker_is_light_blue() {
let cfg = test_config();
let md = [
"1. First\n",
" - Second level\n",
" 1. Third level (ordered)\n",
" - Fourth level (bullet)\n",
" - Fifth level to test indent consistency\n",
];
let out = super::simulate_stream_markdown_for_tests(&md, true, &cfg);
// Find the line that contains the third-level ordered text
let find_idx = out.iter().position(|l| {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<String>()
.contains("Third level (ordered)")
});
let idx = find_idx.expect("expected third-level ordered line");
let line = &out[idx];
// Expect at least one span on this line to be styled light blue
let has_light_blue = line
.spans
.iter()
.any(|s| s.style.fg == Some(ratatui::style::Color::LightBlue));
assert!(
has_light_blue,
"expected an ordered-list marker span with light blue fg on: {line:?}"
);
}
#[test]
fn e2e_stream_blockquote_wrap_preserves_green_style() {
let cfg = test_config();
let long = "> This is a very long quoted line that should wrap across multiple columns to verify style preservation.";
let out = super::simulate_stream_markdown_for_tests(&[long, "\n"], true, &cfg);
// Wrap to a narrow width to force multiple output lines.
let wrapped = crate::wrapping::word_wrap_lines(
out.iter().collect::<Vec<_>>(),
crate::wrapping::RtOptions::new(24),
);
// Filter out purely blank lines
let non_blank: Vec<_> = wrapped
.into_iter()
.filter(|l| {
let s = l
.spans
.iter()
.map(|sp| sp.content.clone())
.collect::<Vec<_>>()
.join("");
!s.trim().is_empty()
})
.collect();
assert!(
non_blank.len() >= 2,
"expected wrapped blockquote to span multiple lines"
);
for (i, l) in non_blank.iter().enumerate() {
assert_eq!(
l.style.fg,
Some(Color::Green),
"wrapped line {} should preserve green style, got {:?}",
i,
l.style.fg
);
}
}
#[test]
fn heading_starts_on_new_line_when_following_paragraph() {
let cfg = test_config();
@@ -490,7 +426,7 @@ mod tests {
.collect();
assert_eq!(
s1,
vec!["Sounds good!", ""],
vec!["Sounds good!"],
"expected paragraph followed by blank separator before heading chunk"
);
@@ -509,7 +445,7 @@ mod tests {
.collect();
assert_eq!(
s2,
vec!["## Adding Bird subcommand"],
vec!["", "## Adding Bird subcommand"],
"expected the heading line only on the final commit"
);
@@ -531,18 +467,6 @@ mod tests {
vec!["Hello."],
"unexpected markdown lines: {rendered_strings:?}"
);
let line_to_string = |l: &ratatui::text::Line<'_>| -> String {
l.spans
.iter()
.map(|s| s.content.clone())
.collect::<Vec<_>>()
.join("")
};
assert_eq!(line_to_string(&out1[0]), "Sounds good!");
assert_eq!(line_to_string(&out1[1]), "");
assert_eq!(line_to_string(&out2[0]), "## Adding Bird subcommand");
}
fn lines_to_plain_strings(lines: &[ratatui::text::Line<'_>]) -> Vec<String> {
@@ -560,35 +484,11 @@ mod tests {
#[test]
fn lists_and_fences_commit_without_duplication() {
let cfg = test_config();
// List case
let deltas = vec!["- a\n- ", "b\n- c\n"];
let streamed = simulate_stream_markdown_for_tests(&deltas, true, &cfg);
let streamed_str = lines_to_plain_strings(&streamed);
let mut rendered_all: Vec<ratatui::text::Line<'static>> = Vec::new();
crate::markdown::append_markdown("- a\n- b\n- c\n", &mut rendered_all, &cfg);
let rendered_all_str = lines_to_plain_strings(&rendered_all);
assert_eq!(
streamed_str, rendered_all_str,
"list streaming should equal full render without duplication"
);
assert_streamed_equals_full(&["- a\n- ", "b\n- c\n"]);
// Fenced code case: stream in small chunks
let deltas2 = vec!["```", "\nco", "de 1\ncode 2\n", "```\n"];
let streamed2 = simulate_stream_markdown_for_tests(&deltas2, true, &cfg);
let streamed2_str = lines_to_plain_strings(&streamed2);
let mut rendered_all2: Vec<ratatui::text::Line<'static>> = Vec::new();
crate::markdown::append_markdown("```\ncode 1\ncode 2\n```\n", &mut rendered_all2, &cfg);
let rendered_all2_str = lines_to_plain_strings(&rendered_all2);
assert_eq!(
streamed2_str, rendered_all2_str,
"fence streaming should equal full render without duplication"
);
assert_streamed_equals_full(&["```", "\nco", "de 1\ncode 2\n", "```\n"]);
}
#[test]
@@ -622,6 +522,56 @@ mod tests {
);
}
#[test]
fn e2e_stream_deep_nested_third_level_marker_is_light_blue() {
let cfg = test_config();
let md = "1. First\n - Second level\n 1. Third level (ordered)\n - Fourth level (bullet)\n - Fifth level to test indent consistency\n";
let streamed = super::simulate_stream_markdown_for_tests(&[md], true, &cfg);
let streamed_strs = lines_to_plain_strings(&streamed);
// Locate the third-level line in the streamed output; avoid relying on exact indent.
let target_suffix = "1. Third level (ordered)";
let mut found = None;
for line in &streamed {
let s: String = line.spans.iter().map(|sp| sp.content.clone()).collect();
if s.contains(target_suffix) {
found = Some(line.clone());
break;
}
}
let line = found.unwrap_or_else(|| {
panic!("expected to find the third-level ordered list line; got: {streamed_strs:?}")
});
// The marker (including indent and "1.") is expected to be in the first span
// and colored LightBlue; following content should be default color.
assert!(
!line.spans.is_empty(),
"expected non-empty spans for the third-level line"
);
let marker_span = &line.spans[0];
assert_eq!(
marker_span.style.fg,
Some(Color::LightBlue),
"expected LightBlue 3rd-level ordered marker, got {:?}",
marker_span.style.fg
);
// Find the first non-empty non-space content span and verify it is default color.
let mut content_fg = None;
for sp in &line.spans[1..] {
let t = sp.content.trim();
if !t.is_empty() {
content_fg = Some(sp.style.fg);
break;
}
}
assert_eq!(
content_fg.flatten(),
None,
"expected default color for 3rd-level content, got {content_fg:?}"
);
}
#[test]
fn empty_fenced_block_is_dropped_and_separator_preserved_before_heading() {
let cfg = test_config();
@@ -768,16 +718,12 @@ mod tests {
let expected = vec![
"Loose vs. tight list items:".to_string(),
"".to_string(),
"1. ".to_string(),
"Tight item".to_string(),
"2. ".to_string(),
"Another tight item".to_string(),
"3. ".to_string(),
"Loose item with its own paragraph.".to_string(),
"1. Tight item".to_string(),
"2. Another tight item".to_string(),
"3. Loose item with its own paragraph.".to_string(),
"".to_string(),
"This paragraph belongs to the same list item.".to_string(),
"4. ".to_string(),
"Second loose item with a nested list after a blank line.".to_string(),
" This paragraph belongs to the same list item.".to_string(),
"4. Second loose item with a nested list after a blank line.".to_string(),
" - Nested bullet under a loose item".to_string(),
" - Another nested bullet".to_string(),
];
@@ -788,63 +734,39 @@ mod tests {
}
// Targeted tests derived from fuzz findings. Each asserts streamed == full render.
#[test]
fn fuzz_class_bare_dash_then_task_item() {
fn assert_streamed_equals_full(deltas: &[&str]) {
let cfg = test_config();
// Case similar to: ["two\n", "- \n* [x] done "]
let deltas = vec!["two\n", "- \n* [x] done \n"];
let streamed = simulate_stream_markdown_for_tests(&deltas, true, &cfg);
let streamed = simulate_stream_markdown_for_tests(deltas, true, &cfg);
let streamed_strs = lines_to_plain_strings(&streamed);
let full: String = deltas.iter().copied().collect();
let mut rendered: Vec<ratatui::text::Line<'static>> = Vec::new();
crate::markdown::append_markdown(&full, &mut rendered, &cfg);
let rendered_strs = lines_to_plain_strings(&rendered);
assert_eq!(streamed_strs, rendered_strs);
assert_eq!(streamed_strs, rendered_strs, "full:\n---\n{full}\n---");
}
#[test]
fn fuzz_class_bullet_duplication_variant_1() {
let cfg = test_config();
// Case similar to: ["aph.\n- let one\n- bull", "et two\n\n second paragraph "]
let deltas = vec!["aph.\n- let one\n- bull", "et two\n\n second paragraph \n"];
let streamed = simulate_stream_markdown_for_tests(&deltas, true, &cfg);
let streamed_strs = lines_to_plain_strings(&streamed);
let full: String = deltas.iter().copied().collect();
let mut rendered: Vec<ratatui::text::Line<'static>> = Vec::new();
crate::markdown::append_markdown(&full, &mut rendered, &cfg);
let rendered_strs = lines_to_plain_strings(&rendered);
assert_eq!(streamed_strs, rendered_strs);
assert_streamed_equals_full(&[
"aph.\n- let one\n- bull",
"et two\n\n second paragraph \n",
]);
}
#[test]
fn fuzz_class_bullet_duplication_variant_2() {
let cfg = test_config();
// Case similar to: ["- e\n c", "e\n- bullet two\n\n second paragraph in bullet two\n"]
let deltas = vec![
assert_streamed_equals_full(&[
"- e\n c",
"e\n- bullet two\n\n second paragraph in bullet two\n",
];
let streamed = simulate_stream_markdown_for_tests(&deltas, true, &cfg);
let streamed_strs = lines_to_plain_strings(&streamed);
let full: String = deltas.iter().copied().collect();
let mut rendered: Vec<ratatui::text::Line<'static>> = Vec::new();
crate::markdown::append_markdown(&full, &mut rendered, &cfg);
let rendered_strs = lines_to_plain_strings(&rendered);
assert_eq!(streamed_strs, rendered_strs);
]);
}
#[test]
fn fuzz_class_ordered_list_split_weirdness() {
let cfg = test_config();
// Case similar to: ["one\n2", " two\n- \n* [x] d"]
let deltas = vec!["one\n2", " two\n- \n* [x] d\n"];
let streamed = simulate_stream_markdown_for_tests(&deltas, true, &cfg);
let streamed_strs = lines_to_plain_strings(&streamed);
let full: String = deltas.iter().copied().collect();
let mut rendered: Vec<ratatui::text::Line<'static>> = Vec::new();
crate::markdown::append_markdown(&full, &mut rendered, &cfg);
let rendered_strs = lines_to_plain_strings(&rendered);
assert_eq!(streamed_strs, rendered_strs);
fn streaming_html_block_then_text_matches_full() {
assert_streamed_equals_full(&[
"HTML block:\n",
"<div>inline block</div>\n",
"more stuff\n",
]);
}
}

View File

@@ -1,72 +0,0 @@
/// Returns true if the provided text contains an unclosed fenced code block
/// (opened by ``` or ~~~, closed by a matching fence on its own line).
pub fn is_inside_unclosed_fence(s: &str) -> bool {
let mut open = false;
for line in s.lines() {
let t = line.trim_start();
if t.starts_with("```") || t.starts_with("~~~") {
if !open {
open = true;
} else {
// closing fence on same pattern toggles off
open = false;
}
}
}
open
}
/// Remove fenced code blocks that contain no content (whitespace-only) to avoid
/// streaming empty code blocks like ```lang\n``` or ```\n```.
pub fn strip_empty_fenced_code_blocks(s: &str) -> String {
// Only remove complete fenced blocks that contain no non-whitespace content.
// Leave all other content unchanged to avoid affecting partial streams.
let lines: Vec<&str> = s.lines().collect();
let mut out = String::with_capacity(s.len());
let mut i = 0usize;
while i < lines.len() {
let line = lines[i];
let trimmed_start = line.trim_start();
let fence_token = if trimmed_start.starts_with("```") {
"```"
} else if trimmed_start.starts_with("~~~") {
"~~~"
} else {
""
};
if !fence_token.is_empty() {
// Find a matching closing fence on its own line.
let mut j = i + 1;
let mut has_content = false;
let mut found_close = false;
while j < lines.len() {
let l = lines[j];
if l.trim() == fence_token {
found_close = true;
break;
}
if !l.trim().is_empty() {
has_content = true;
}
j += 1;
}
if found_close && !has_content {
// Drop i..=j and insert at most a single blank separator line.
if !out.ends_with('\n') {
out.push('\n');
}
i = j + 1;
continue;
}
// Not an empty fenced block; emit as-is.
out.push_str(line);
out.push('\n');
i += 1;
} else {
out.push_str(line);
out.push('\n');
i += 1;
}
}
out
}

View File

@@ -1,3 +1,2 @@
pub mod highlight;
pub mod line_utils;
pub mod markdown_utils;

View File

@@ -0,0 +1,62 @@
---
source: tui/src/markdown_render_tests.rs
expression: rendered
---
# H1: Markdown Streaming Test
Intro paragraph with bold text, italic text, and inline code x=1.
Combined bold-italic both and escaped asterisks *literal*.
Auto-link: https://example.com (https://example.com) and reference link [ref][r1].
Link with title: hover me (https://example.com) and mailto mailto:test@example.com (mailto:test@example.com).
Image: alt text
> Blockquote level 1
>
> > Blockquote level 2 with inline code
- Unordered list item 1
- Nested bullet with italics inner
- Unordered list item 2 with strikethrough
1. Ordered item one
2. Ordered item two with sublist:
1. Alt-numbered subitem
- [ ] Task: unchecked
- [x] Task: checked with link home (https://example.org)
———
Table below (alignment test):
| Left | Center | Right |
|:-----|:------:|------:|
| a | b | c |
Inline HTML: <sup>sup</sup> and <sub>sub</sub>.
HTML block:
<div style="border:1px solid #ccc;padding:2px">inline block</div>
Escapes: \_underscores\_, backslash \\, ticks ``code with `backtick` inside``.
Emoji shortcodes: :sparkles: :tada: (if supported).
Hard break test (line ends with two spaces)
Next line should be close to previous.
Footnote reference here[^1] and another[^longnote].
Horizontal rule with asterisks:
***
Fenced code block (JSON):
```json
{ "a": 1, "b": [true, false] }
```
Fenced code with tildes and triple backticks inside:
~~~markdown
To close ``` you need tildes.
~~~
Indented code block:
for i in range(3): print(i)
Definition-like list:
Term
: Definition with `code`.
Character entities: &amp; &lt; &gt; &quot; &#39;
[^1]: This is the first footnote.
[^longnote]: A longer footnote with a link to [Rust](https://www.rust-lang.org/).
Escaped pipe in text: a \| b \| c.
URL with parentheses: [link](https://example.com/path_(with)_parens).
[r1]: https://example.com/ref "Reference link title"

View File

@@ -380,16 +380,12 @@ mod tests {
let expected = vec![
"Loose vs. tight list items:".to_string(),
"".to_string(),
"1. ".to_string(),
"Tight item".to_string(),
"2. ".to_string(),
"Another tight item".to_string(),
"3. ".to_string(),
"Loose item with its own paragraph.".to_string(),
"1. Tight item".to_string(),
"2. Another tight item".to_string(),
"3. Loose item with its own paragraph.".to_string(),
"".to_string(),
"This paragraph belongs to the same list item.".to_string(),
"4. ".to_string(),
"Second loose item with a nested list after a blank line.".to_string(),
" This paragraph belongs to the same list item.".to_string(),
"4. Second loose item with a nested list after a blank line.".to_string(),
" - Nested bullet under a loose item".to_string(),
" - Another nested bullet".to_string(),
];

View File

@@ -187,10 +187,17 @@ where
// Build first wrapped line with initial indent.
let mut first_line = rt_opts.initial_indent.clone();
first_line.style = first_line.style.patch(line.style);
{
let mut sliced = slice_line_spans(line, &span_bounds, first_line_range);
let sliced = slice_line_spans(line, &span_bounds, first_line_range);
let mut spans = first_line.spans;
spans.append(&mut sliced.spans);
spans.append(
&mut sliced
.spans
.into_iter()
.map(|s| s.patch_style(line.style))
.collect(),
);
first_line.spans = spans;
out.push(first_line);
}
@@ -209,10 +216,17 @@ where
continue;
}
let mut subsequent_line = rt_opts.subsequent_indent.clone();
subsequent_line.style = subsequent_line.style.patch(line.style);
let offset_range = (r.start + base)..(r.end + base);
let mut sliced = slice_line_spans(line, &span_bounds, &offset_range);
let sliced = slice_line_spans(line, &span_bounds, &offset_range);
let mut spans = subsequent_line.spans;
spans.append(&mut sliced.spans);
spans.append(
&mut sliced
.spans
.into_iter()
.map(|s| s.patch_style(line.style))
.collect(),
);
subsequent_line.spans = spans;
out.push(subsequent_line);
}