This commit is contained in:
2025-06-12 20:23:52 +02:00
parent 6f591627be
commit 27728fc431
28 changed files with 6730 additions and 0 deletions

View File

@ -0,0 +1,243 @@
use egui::text::{CCursorRange, LayoutJob};
use crate::easy_mark::easy_mark_parser;
/// Highlight easymark, memoizing previous output to save CPU.
///
/// In practice, the highlighter is fast enough not to need any caching.
#[derive(Default)]
pub struct MemoizedHighlighter {
style: egui::Style,
code: String,
output: LayoutJob,
}
impl MemoizedHighlighter {
pub fn highlight(
&mut self,
egui_style: &egui::Style,
code: &str,
cursor: Option<CCursorRange>,
) -> LayoutJob {
if (&self.style, self.code.as_str()) != (egui_style, code) {
self.style = egui_style.clone();
code.clone_into(&mut self.code);
self.output = highlight_easymark(egui_style, code, cursor);
}
self.output.clone()
}
}
pub fn highlight_easymark(
egui_style: &egui::Style,
mut text: &str,
// TODO: hide special characters where cursor isn't
_cursor: Option<CCursorRange>,
) -> LayoutJob {
let mut job = LayoutJob::default();
let mut style = easy_mark_parser::Style::default();
let mut start_of_line = true;
const CODE_INDENT: f32 = 10.0;
while !text.is_empty() {
if start_of_line && text.starts_with("```") {
let astyle = format_from_style(
egui_style,
&easy_mark_parser::Style {
code: true,
..Default::default()
},
);
// Render the initial backticks as spaces
text = &text[3..];
job.append(" ", CODE_INDENT, astyle.clone());
match text.find("\n```") {
Some(n) => {
for line in text[..n + 1].lines() {
job.append(line, CODE_INDENT, astyle.clone());
job.append("\n", 0.0, astyle.clone());
}
// Render the final backticks as spaces
job.append(" ", CODE_INDENT, astyle);
text = &text[n + 4..];
}
None => {
job.append(text, 0.0, astyle.clone());
text = "";
}
};
style = Default::default();
continue;
}
if text.starts_with('`') {
style.code = true;
let end = text[1..]
.find(&['`', '\n'][..])
.map_or_else(|| text.len(), |i| i + 2);
job.append(&text[..end], 0.0, format_from_style(egui_style, &style));
text = &text[end..];
style.code = false;
continue;
}
let skip;
// zero-width space
let _zws = "\u{200b}";
let mut apply_basic_style =
|text: &mut &str,
style: &mut easy_mark_parser::Style,
access: fn(&mut easy_mark_parser::Style) -> &mut bool| {
let skip = if *access(style) {
// Include the character that is ending this style:
job.append(&text[..1], 0.0, format_from_style(egui_style, style));
*text = &text[1..];
0
} else {
1
};
*access(style) ^= true;
skip
};
if text.starts_with('*') {
skip = apply_basic_style(&mut text, &mut style, |style| &mut style.strong);
} else if text.starts_with('/') {
skip = apply_basic_style(&mut text, &mut style, |style| &mut style.italics);
} else if text.starts_with('_') {
skip = apply_basic_style(&mut text, &mut style, |style| &mut style.underline);
} else if text.starts_with('$') {
skip = apply_basic_style(&mut text, &mut style, |style| &mut style.small);
} else if text.starts_with('~') {
skip = apply_basic_style(&mut text, &mut style, |style| &mut style.strikethrough);
} else if text.starts_with('^') {
skip = apply_basic_style(&mut text, &mut style, |style| &mut style.raised);
} else if text.starts_with('\\') && text.len() >= 2 {
skip = 2;
} else if start_of_line && text.starts_with(' ') {
// we don't preview indentation, because it is confusing
skip = 1;
} else if start_of_line && text.starts_with("###### ") {
style.heading = true;
skip = 7;
} else if start_of_line && text.starts_with("##### ") {
style.heading = true;
skip = 6;
} else if start_of_line && text.starts_with("#### ") {
style.heading = true;
skip = 5;
} else if start_of_line && text.starts_with("### ") {
style.heading = true;
skip = 4;
} else if start_of_line && text.starts_with("## ") {
style.heading = true;
skip = 3;
} else if start_of_line && text.starts_with("# ") {
style.heading = true;
skip = 2;
} else if start_of_line && text.starts_with("> ") {
style.quoted = true;
skip = 2;
// we don't preview indentation, because it is confusing
} else if start_of_line && text.trim_start().starts_with("- ") {
job.append("", 0.0, format_from_style(egui_style, &style));
text = &text[2..];
skip = 0;
// we don't preview indentation, because it is confusing
} else {
skip = 0;
}
// Note: we don't preview underline, strikethrough and italics because it confuses things.
// Swallow everything up to the next special character:
let line_end = text[skip..]
.find('\n')
.map_or_else(|| text.len(), |i| (skip + i + 1));
let end = text[skip..]
.find(&['*', '`', '~', '_', '/', '$', '^', '\\', '<', '['][..])
.map_or_else(|| text.len(), |i| (skip + i).max(1));
if line_end <= end {
job.append(
&text[..line_end],
0.0,
format_from_style(egui_style, &style),
);
text = &text[line_end..];
start_of_line = true;
style = Default::default();
} else {
job.append(&text[..end], 0.0, format_from_style(egui_style, &style));
text = &text[end..];
start_of_line = false;
}
}
job
}
fn format_from_style(
egui_style: &egui::Style,
emark_style: &easy_mark_parser::Style,
) -> egui::text::TextFormat {
use egui::{Align, Color32, Stroke, TextStyle};
let color = if emark_style.strong || emark_style.heading {
egui_style.visuals.strong_text_color()
} else if emark_style.quoted {
egui_style.visuals.weak_text_color()
} else {
egui_style.visuals.text_color()
};
let text_style = if emark_style.heading {
TextStyle::Heading
} else if emark_style.code {
TextStyle::Monospace
} else if emark_style.small | emark_style.raised {
TextStyle::Small
} else {
TextStyle::Body
};
let background = if emark_style.code {
egui_style.visuals.code_bg_color
} else {
Color32::TRANSPARENT
};
let underline = if emark_style.underline {
Stroke::new(1.0, color)
} else {
Stroke::NONE
};
let strikethrough = if emark_style.strikethrough {
Stroke::new(1.0, color)
} else {
Stroke::NONE
};
let valign = if emark_style.raised {
Align::TOP
} else {
Align::BOTTOM
};
egui::text::TextFormat {
font_id: text_style.resolve(egui_style),
color,
background,
italics: emark_style.italics,
underline,
strikethrough,
valign,
..Default::default()
}
}

View File

@ -0,0 +1,346 @@
//! A parser for `EasyMark`: a very simple markup language.
//!
//! WARNING: `EasyMark` is subject to change.
//
//! # `EasyMark` design goals:
//! 1. easy to parse
//! 2. easy to learn
//! 3. similar to markdown
#[derive(Copy, Clone, Debug, Eq, PartialEq)]
pub enum Item<'a> {
/// `\n`
// TODO(emilk): add Style here so empty heading still uses up the right amount of space.
Newline,
/// Text
Text(Style, &'a str),
/// title, url
Hyperlink(Style, &'a str, &'a str),
/// leading space before e.g. a [`Self::BulletPoint`].
Indentation(usize),
/// >
QuoteIndent,
/// - a point well made.
BulletPoint,
/// 1. numbered list. The string is the number(s).
NumberedPoint(&'a str),
/// ---
Separator,
/// language, code
CodeBlock(&'a str, &'a str),
}
#[derive(Copy, Clone, Debug, Default, Eq, PartialEq)]
pub struct Style {
/// # heading (large text)
pub heading: bool,
/// > quoted (slightly dimmer color or other font style)
pub quoted: bool,
/// `code` (monospace, some other color)
pub code: bool,
/// self.strong* (emphasized, e.g. bold)
pub strong: bool,
/// _underline_
pub underline: bool,
/// ~strikethrough~
pub strikethrough: bool,
/// /italics/
pub italics: bool,
/// $small$
pub small: bool,
/// ^raised^
pub raised: bool,
}
/// Parser for the `EasyMark` markup language.
pub struct Parser<'a> {
/// The remainder of the input text
s: &'a str,
/// Are we at the start of a line?
start_of_line: bool,
/// Current self.style. Reset after a newline.
style: Style,
}
impl<'a> Parser<'a> {
pub fn new(s: &'a str) -> Self {
Self {
s,
start_of_line: true,
style: Style::default(),
}
}
/// `1. `, `42. ` etc.
fn numbered_list(&mut self) -> Option<Item<'a>> {
let n_digits = self.s.chars().take_while(|c| c.is_ascii_digit()).count();
if n_digits > 0 && self.s.chars().skip(n_digits).take(2).eq(". ".chars()) {
let number = &self.s[..n_digits];
self.s = &self.s[(n_digits + 2)..];
self.start_of_line = false;
return Some(Item::NumberedPoint(number));
}
None
}
// ```{language}\n{code}```
fn code_block(&mut self) -> Option<Item<'a>> {
if let Some(language_start) = self.s.strip_prefix("```") {
if let Some(newline) = language_start.find('\n') {
let language = &language_start[..newline];
let code_start = &language_start[newline + 1..];
if let Some(end) = code_start.find("\n```") {
let code = &code_start[..end].trim();
self.s = &code_start[end + 4..];
self.start_of_line = false;
return Some(Item::CodeBlock(language, code));
} else {
self.s = "";
return Some(Item::CodeBlock(language, code_start));
}
}
}
None
}
// `code`
fn inline_code(&mut self) -> Option<Item<'a>> {
if let Some(rest) = self.s.strip_prefix('`') {
self.s = rest;
self.start_of_line = false;
self.style.code = true;
let rest_of_line = &self.s[..self.s.find('\n').unwrap_or(self.s.len())];
if let Some(end) = rest_of_line.find('`') {
let item = Item::Text(self.style, &self.s[..end]);
self.s = &self.s[end + 1..];
self.style.code = false;
return Some(item);
} else {
let end = rest_of_line.len();
let item = Item::Text(self.style, rest_of_line);
self.s = &self.s[end..];
self.style.code = false;
return Some(item);
}
}
None
}
/// `<url>` or `[link](url)`
fn url(&mut self) -> Option<Item<'a>> {
if self.s.starts_with('<') {
let this_line = &self.s[..self.s.find('\n').unwrap_or(self.s.len())];
if let Some(url_end) = this_line.find('>') {
let url = &self.s[1..url_end];
self.s = &self.s[url_end + 1..];
self.start_of_line = false;
return Some(Item::Hyperlink(self.style, url, url));
}
}
// [text](url)
if self.s.starts_with('[') {
let this_line = &self.s[..self.s.find('\n').unwrap_or(self.s.len())];
if let Some(bracket_end) = this_line.find(']') {
let text = &this_line[1..bracket_end];
if this_line[bracket_end + 1..].starts_with('(') {
if let Some(parens_end) = this_line[bracket_end + 2..].find(')') {
let parens_end = bracket_end + 2 + parens_end;
let url = &self.s[bracket_end + 2..parens_end];
self.s = &self.s[parens_end + 1..];
self.start_of_line = false;
return Some(Item::Hyperlink(self.style, text, url));
}
}
}
}
None
}
}
impl<'a> Iterator for Parser<'a> {
type Item = Item<'a>;
fn next(&mut self) -> Option<Self::Item> {
loop {
if self.s.is_empty() {
return None;
}
// \n
if self.s.starts_with('\n') {
self.s = &self.s[1..];
self.start_of_line = true;
self.style = Style::default();
return Some(Item::Newline);
}
// Ignore line break (continue on the same line)
if self.s.starts_with("\\\n") && self.s.len() >= 2 {
self.s = &self.s[2..];
self.start_of_line = false;
continue;
}
// \ escape (to show e.g. a backtick)
if self.s.starts_with('\\') && self.s.len() >= 2 {
let text = &self.s[1..2];
self.s = &self.s[2..];
self.start_of_line = false;
return Some(Item::Text(self.style, text));
}
if self.start_of_line {
// leading space (indentation)
if self.s.starts_with(' ') {
let length = self.s.find(|c| c != ' ').unwrap_or(self.s.len());
self.s = &self.s[length..];
self.start_of_line = true; // indentation doesn't count
return Some(Item::Indentation(length));
}
// # Heading
if let Some(after) = self.s.strip_prefix("# ") {
self.s = after;
self.start_of_line = false;
self.style.heading = true;
continue;
}
// > quote
if let Some(after) = self.s.strip_prefix("> ") {
self.s = after;
self.start_of_line = true; // quote indentation doesn't count
self.style.quoted = true;
return Some(Item::QuoteIndent);
}
// - bullet point
if self.s.starts_with("- ") {
self.s = &self.s[2..];
self.start_of_line = false;
return Some(Item::BulletPoint);
}
// `1. `, `42. ` etc.
if let Some(item) = self.numbered_list() {
return Some(item);
}
// --- separator
if let Some(after) = self.s.strip_prefix("---") {
self.s = after.trim_start_matches('-'); // remove extra dashes
self.s = self.s.strip_prefix('\n').unwrap_or(self.s); // remove trailing newline
self.start_of_line = false;
return Some(Item::Separator);
}
// ```{language}\n{code}```
if let Some(item) = self.code_block() {
return Some(item);
}
}
// `code`
if let Some(item) = self.inline_code() {
return Some(item);
}
if let Some(rest) = self.s.strip_prefix('*') {
self.s = rest;
self.start_of_line = false;
self.style.strong = !self.style.strong;
continue;
}
if let Some(rest) = self.s.strip_prefix('_') {
self.s = rest;
self.start_of_line = false;
self.style.underline = !self.style.underline;
continue;
}
if let Some(rest) = self.s.strip_prefix('~') {
self.s = rest;
self.start_of_line = false;
self.style.strikethrough = !self.style.strikethrough;
continue;
}
if let Some(rest) = self.s.strip_prefix('/') {
self.s = rest;
self.start_of_line = false;
self.style.italics = !self.style.italics;
continue;
}
if let Some(rest) = self.s.strip_prefix('$') {
self.s = rest;
self.start_of_line = false;
self.style.small = !self.style.small;
continue;
}
if let Some(rest) = self.s.strip_prefix('^') {
self.s = rest;
self.start_of_line = false;
self.style.raised = !self.style.raised;
continue;
}
// `<url>` or `[link](url)`
if let Some(item) = self.url() {
return Some(item);
}
// Swallow everything up to the next special character:
let end = self
.s
.find(&['*', '`', '~', '_', '/', '$', '^', '\\', '<', '[', '\n'][..])
.map_or_else(|| self.s.len(), |special| special.max(1));
let item = Item::Text(self.style, &self.s[..end]);
self.s = &self.s[end..];
self.start_of_line = false;
return Some(item);
}
}
}
#[test]
fn test_easy_mark_parser() {
let items: Vec<_> = Parser::new("~strikethrough `code`~").collect();
assert_eq!(
items,
vec![
Item::Text(
Style {
strikethrough: true,
..Default::default()
},
"strikethrough "
),
Item::Text(
Style {
code: true,
strikethrough: true,
..Default::default()
},
"code"
),
]
);
}

7
src/easy_mark/mod.rs Normal file
View File

@ -0,0 +1,7 @@
//! Experimental markup language
mod easy_mark_highlighter;
pub mod easy_mark_parser;
pub use easy_mark_highlighter::{MemoizedHighlighter, highlight_easymark};
pub use easy_mark_parser as parser;