mirror of
https://github.com/storytold/deckcraft.git
synced 2026-11-03 00:41:28 -06:00
xmaks82andМагда М.Г.
5fb694d991
* Text: a word doesn't break after ?, !, } or + (URLs wrap as in PowerPoint) Unicode line breaking allows a break after ?, !, } and + inside a word, so a long URL broke after its '?' and after each '+' of a query and left the rest of the line empty. Measured in PowerPoint 2016 for Windows with latinLnBrk off (its default), none of them is a break: the word fills the line and breaks where it no longer fits, as PowerPoint already does after / and \ (break_opportunities). The same synthetic slide now wraps line for line as PowerPoint's PDF does, a Yandex search URL included. * Text tests: the first line doesn't end at the separator * Text tests: check the line is full, not that it avoids ending at a separator * Keep the Unicode break after ?, !, / outside Latin words The PowerPoint no-break rule only holds inside a word: before an ideograph (日本!語) or punctuation the Unicode break stays, as CodeRabbit pointed out. * Only letters and digits continue a word for the no-break rule AL and AI also hold symbols such as & and ×, so a!&b kept no break after !. is_word_char now also requires c.is_alphanumeric() (CodeRabbit). * Test that Latin before CJK and a separator before a space still break --------- Co-authored-by: Магда М.Г. <mail@mgmagda.ru>
319 lines
12 KiB
Rust
319 lines
12 KiB
Rust
use super::style::{Cell, ParaShaped};
|
||
use super::{FontFace, ParaProps};
|
||
|
||
pub(crate) fn measure_str(face: &FontFace, size: f64, s: &str) -> (f64, Vec<(u32, f64)>) {
|
||
let g = deckcraft_fonts::shape(face, s, &[], |c| c);
|
||
let k = size / face.upem.max(1.0);
|
||
let mut x = 0.0;
|
||
let mut out = vec![];
|
||
for gl in g {
|
||
out.push((gl.gid, x + gl.x_offset as f64 * k));
|
||
x += gl.x_advance as f64 * k;
|
||
}
|
||
(x, out)
|
||
}
|
||
|
||
/// Single line spacing as a multiple of the font size. PowerPoint spaces single-spaced lines 1.2
|
||
/// times the size apart in every font (observed on twenty Latin, CJK, Thai and Indic fonts at 18
|
||
/// and 40 pt), whatever the font's own ascent and descent.
|
||
///
|
||
/// Caveat: this was measured on one Windows install of desktop PowerPoint. It may be a Windows or
|
||
/// machine setting; PowerPoint on macOS or the web may space lines by the font's own metrics.
|
||
/// Keep it a single tunable value, and re-measure on the other platform before changing it.
|
||
pub(crate) const LINE_PITCH: f64 = 1.2;
|
||
|
||
/// A line's height above and below the baseline at `size`: [`LINE_PITCH`] × size, split where
|
||
/// the font's Windows ascent and descent put the baseline.
|
||
pub(crate) fn line_metrics(face: &FontFace, size: f64) -> (f64, f64) {
|
||
let h = size * LINE_PITCH;
|
||
(h * face.line_split, h * (1.0 - face.line_split))
|
||
}
|
||
|
||
/// Hangul syllables and jamo: PowerPoint wraps Korean at word boundaries (spaces), not between
|
||
/// syllables.
|
||
fn is_hangul(c: char) -> bool {
|
||
matches!(c, '\u{1100}'..='\u{11FF}' | '\u{3130}'..='\u{318F}' | '\u{A960}'..='\u{A97F}' | '\u{AC00}'..='\u{D7FF}')
|
||
}
|
||
|
||
/// East Asian commas and full stops, the punctuation PowerPoint lets hang past the end of a line
|
||
/// (`hangingPunct`). Closing brackets don't hang: "「。」" overflowing by its "」" moves to the next
|
||
/// line whole.
|
||
fn is_hanging_punct(c: char) -> bool {
|
||
matches!(c, '、' | '。' | '，' | '．' | '､' | '｡')
|
||
}
|
||
|
||
/// Small kana, the prolonged sound mark and iteration marks: PowerPoint keeps them off the start
|
||
/// of a line (strict Japanese line breaking), where UAX #14's default allows a break before them.
|
||
fn is_cj_nonstarter(c: char) -> bool {
|
||
matches!(
|
||
c,
|
||
'ぁ' | 'ぃ'
|
||
| 'ぅ'
|
||
| 'ぇ'
|
||
| 'ぉ'
|
||
| 'っ'
|
||
| 'ゃ'
|
||
| 'ゅ'
|
||
| 'ょ'
|
||
| 'ゎ'
|
||
| 'ゕ'
|
||
| 'ゖ'
|
||
| 'ァ'
|
||
| 'ィ'
|
||
| 'ゥ'
|
||
| 'ェ'
|
||
| 'ォ'
|
||
| 'ッ'
|
||
| 'ャ'
|
||
| 'ュ'
|
||
| 'ョ'
|
||
| 'ヮ'
|
||
| 'ヵ'
|
||
| 'ヶ'
|
||
| 'ー'
|
||
| '々'
|
||
| '〻'
|
||
| 'ゝ'
|
||
| 'ゞ'
|
||
| 'ヽ'
|
||
| 'ヾ'
|
||
| '\u{31F0}'..='\u{31FF}'
|
||
| 'ｧ'..='ｰ'
|
||
)
|
||
}
|
||
|
||
/// A character that continues a Latin-style word: a letter or digit that Unicode does not break
|
||
/// between, as opposed to CJK ideographs, kana, spaces, punctuation and symbols such as `&` or `×`
|
||
/// (which share the AL and AI line-break classes with letters).
|
||
fn is_word_char(c: char) -> bool {
|
||
use unicode_linebreak::BreakClass::{Alphabetic, Ambiguous, HebrewLetter, Numeric};
|
||
matches!(unicode_linebreak::break_property(c as u32), Alphabetic | Ambiguous | HebrewLetter | Numeric) && c.is_alphanumeric()
|
||
}
|
||
|
||
/// Break opportunities: allowed[i] = a line may start at char i. Unicode line breaking, then
|
||
/// PowerPoint's differences: without `latinLnBrk` a word doesn't break after `/`, `\`, `?`, `!`, `}`
|
||
/// or `+` inside a word (measured in PowerPoint 2016 for Windows; before an ideograph or a space
|
||
/// the Unicode break stays), and Korean breaks only between words.
|
||
pub(crate) fn break_opportunities(text: &str, latin_break: bool) -> (Vec<bool>, Vec<bool>) {
|
||
let n = text.chars().count();
|
||
let mut allowed = vec![false; n + 1];
|
||
let mut mandatory = vec![false; n + 1];
|
||
let mut byte_to_char = std::collections::HashMap::new();
|
||
for (ci, (b, _)) in text.char_indices().enumerate() {
|
||
byte_to_char.insert(b, ci);
|
||
}
|
||
byte_to_char.insert(text.len(), n);
|
||
for (b, op) in unicode_linebreak::linebreaks(text) {
|
||
if let Some(&ci) = byte_to_char.get(&b) {
|
||
if let Some(a) = allowed.get_mut(ci) {
|
||
*a = true;
|
||
}
|
||
if op == unicode_linebreak::BreakOpportunity::Mandatory
|
||
&& let Some(m) = mandatory.get_mut(ci)
|
||
{
|
||
*m = true;
|
||
}
|
||
}
|
||
}
|
||
let chars: Vec<char> = text.chars().collect();
|
||
for (i, w) in chars.windows(2).enumerate() {
|
||
let (prev, next) = (w[0], w[1]);
|
||
// PowerPoint never breaks after a zero-width space, nor before an em dash (only after it).
|
||
let keep = if (is_hangul(prev) && is_hangul(next)) || prev == '\u{200B}' || next == '\u{2014}' || is_cj_nonstarter(next) {
|
||
false
|
||
} else {
|
||
latin_break || !matches!(prev, '/' | '\\' | '?' | '!' | '}' | '+') || !is_word_char(next)
|
||
};
|
||
if !keep
|
||
&& !mandatory.get(i + 1).copied().unwrap_or(false)
|
||
&& let Some(a) = allowed.get_mut(i + 1)
|
||
{
|
||
*a = false;
|
||
}
|
||
}
|
||
(allowed, mandatory)
|
||
}
|
||
|
||
pub(crate) struct RawLine {
|
||
pub(crate) start: usize,
|
||
pub(crate) end: usize,
|
||
/// End including trailing spaces / the break char consumed.
|
||
next: usize,
|
||
}
|
||
|
||
/// Wrapping geometry of one paragraph: line widths, start offsets, wrap flag and tab settings.
|
||
pub(crate) struct BreakParams<'a> {
|
||
pub(crate) first_w: f64,
|
||
pub(crate) rest_w: f64,
|
||
pub(crate) wrap: bool,
|
||
pub(crate) tab_stops: &'a [Tab],
|
||
pub(crate) default_tab: f64,
|
||
pub(crate) x0_first: f64,
|
||
pub(crate) x0_rest: f64,
|
||
}
|
||
|
||
pub(crate) fn break_lines(p: &ParaShaped, params: &BreakParams) -> Vec<RawLine> {
|
||
let BreakParams { first_w, rest_w, wrap, tab_stops, default_tab, x0_first, x0_rest } = *params;
|
||
let text: String = p.cells.iter().map(|c| c.ch).collect();
|
||
let (allowed, _) = break_opportunities(&text, p.props.latin_line_break.unwrap_or(false));
|
||
let hang = p.props.hanging_punct.unwrap_or(false);
|
||
let n = p.cells.len();
|
||
let mut lines = vec![];
|
||
let mut start = 0;
|
||
if n == 0 {
|
||
return vec![RawLine { start: 0, end: 0, next: 0 }];
|
||
}
|
||
while start < n {
|
||
let avail = if lines.is_empty() { first_w } else { rest_w };
|
||
let x0 = if lines.is_empty() { x0_first } else { x0_rest };
|
||
let mut x = 0.0;
|
||
let mut last_break: Option<usize> = None;
|
||
let mut i = start;
|
||
let mut end = n;
|
||
let mut next = n;
|
||
while i < n {
|
||
let c = p.cells.get(i).map(|c| c.ch).unwrap_or(' ');
|
||
if c == '\u{b}' || c == '\n' {
|
||
end = i;
|
||
next = i + 1;
|
||
break;
|
||
}
|
||
let adv = if c == '\t' {
|
||
tab_advance(x0 + x, tab_stops, default_tab, tab_segment(&p.cells, i))
|
||
} else {
|
||
p.cells.get(i).map(|c| c.adv).unwrap_or(0.0)
|
||
};
|
||
let breakable = |j: usize| p.cells.get(j).is_none_or(|c| c.breakable(j));
|
||
if i > start && allowed.get(i).copied().unwrap_or(false) && breakable(i) {
|
||
last_break = Some(i);
|
||
}
|
||
// Closing East Asian punctuation hangs past the end instead of wrapping
|
||
// (`hangingPunct`); the line breaks after it.
|
||
if wrap && hang && x + adv > avail + 0.01 && x <= avail + 0.01 && is_hanging_punct(c) && i > start {
|
||
end = i + 1;
|
||
next = i + 1;
|
||
break;
|
||
}
|
||
// A tab that would run past the end moves to the next line (and to its first stop there).
|
||
if wrap && c == '\t' && x + adv > avail + 0.01 && i > start {
|
||
end = i;
|
||
next = i;
|
||
while end > start && p.cells.get(end - 1).is_some_and(|c| c.ch == ' ') {
|
||
end -= 1;
|
||
}
|
||
break;
|
||
}
|
||
if wrap && x + adv > avail + 0.01 && !c.is_whitespace() && i > start {
|
||
// Break at the last opportunity, or force mid-word (between clusters).
|
||
let b = last_break
|
||
.filter(|b| *b > start)
|
||
.unwrap_or_else(|| (start + 1..=i).rev().find(|&j| breakable(j)).or_else(|| (i + 1..n).find(|&j| breakable(j))).unwrap_or(n));
|
||
end = b;
|
||
next = b;
|
||
// Trailing spaces belong to the line but don't count.
|
||
while end > start && p.cells.get(end - 1).is_some_and(|c| c.ch == ' ') {
|
||
end -= 1;
|
||
}
|
||
break;
|
||
}
|
||
x += adv;
|
||
i += 1;
|
||
}
|
||
if i >= n {
|
||
end = n;
|
||
next = n;
|
||
}
|
||
lines.push(RawLine { start, end, next });
|
||
if next <= start {
|
||
// No progress: force one char.
|
||
let last = lines.len() - 1;
|
||
let to = (start + 1..n).find(|&j| p.cells.get(j).is_none_or(|c| c.breakable(j))).unwrap_or(n);
|
||
if let Some(l) = lines.get_mut(last) {
|
||
l.end = to;
|
||
l.next = to;
|
||
}
|
||
start = to;
|
||
} else {
|
||
start = next;
|
||
}
|
||
if next == n && p.cells.last().is_some_and(|c| c.ch == '\u{b}') && start >= n {
|
||
// A trailing line break opens an empty last line.
|
||
lines.push(RawLine { start: n, end: n, next: n });
|
||
}
|
||
}
|
||
lines
|
||
}
|
||
|
||
/// How a tab stop aligns the text after it (`a:tab/@algn`).
|
||
#[derive(Clone, Copy, Debug, PartialEq)]
|
||
enum TabAlign {
|
||
Left,
|
||
Center,
|
||
Right,
|
||
/// The first "." of the text sits at the stop (text without one ends there).
|
||
Decimal,
|
||
}
|
||
|
||
#[derive(Clone, Copy, Debug, PartialEq)]
|
||
pub(crate) struct Tab {
|
||
pos: f64,
|
||
align: TabAlign,
|
||
}
|
||
|
||
pub(crate) fn tabs_of(props: &ParaProps) -> Vec<Tab> {
|
||
props
|
||
.tabs
|
||
.as_ref()
|
||
.map(|t| {
|
||
t.iter()
|
||
.map(|s| Tab {
|
||
pos: s.pos,
|
||
align: match s.align.as_str() {
|
||
"ctr" => TabAlign::Center,
|
||
"r" => TabAlign::Right,
|
||
"dec" => TabAlign::Decimal,
|
||
_ => TabAlign::Left,
|
||
},
|
||
})
|
||
.collect()
|
||
})
|
||
.unwrap_or_default()
|
||
}
|
||
|
||
/// Width of the text after the tab at cell `i` up to the next tab or line break, and the width of
|
||
/// that text before its first ".".
|
||
pub(crate) fn tab_segment(cells: &[Cell], i: usize) -> (f64, f64) {
|
||
let mut w = 0.0;
|
||
let mut dec = None;
|
||
for c in cells.iter().skip(i + 1) {
|
||
if matches!(c.ch, '\t'..='\u{b}') {
|
||
break;
|
||
}
|
||
if c.ch == '.' && dec.is_none() {
|
||
dec = Some(w);
|
||
}
|
||
w += c.adv;
|
||
}
|
||
(w, dec.unwrap_or(w))
|
||
}
|
||
|
||
/// Advance of a tab at `x` (from the start edge) to the next stop, placing the text after it
|
||
/// (`seg`: its width, and its width before the decimal point) as the stop aligns it but never
|
||
/// moving back, as PowerPoint does; past the last stop, to the next default stop.
|
||
pub(crate) fn tab_advance(x: f64, stops: &[Tab], default_tab: f64, seg: (f64, f64)) -> f64 {
|
||
for s in stops {
|
||
if s.pos > x + 0.5 {
|
||
let to = s.pos - x;
|
||
return match s.align {
|
||
TabAlign::Left => to,
|
||
TabAlign::Right => (to - seg.0).max(0.0),
|
||
TabAlign::Center => (to - seg.0 / 2.0).max(0.0),
|
||
TabAlign::Decimal => (to - seg.1).max(0.0),
|
||
};
|
||
}
|
||
}
|
||
let d = if default_tab > 1.0 { default_tab } else { 72.0 };
|
||
let next = ((x / d).floor() + 1.0) * d;
|
||
(next - x).max(1.0)
|
||
}
|