Mirrors/servo
1
0
Fork 0
mirror of https://github.com/servo/servo.git synced 2025-08-06 06:00:15 +01:00
Code Issues Projects Releases Packages Wiki Activity

Use byte indices instead of char indices for text runs

Browse source 
Replace character indices with UTF-8 byte offsets throughout the code dealing
with text shaping and breaking.  This eliminates a lot of complexity when
converting from one to the other, and interoperates better with the rest of
the Rust ecosystem.
This commit is contained in:
Matt Brubeck 2016-04-27 11:22:02 -07:00
parent dba878dfb2
commit 659305fe0a
15 changed files with 259 additions and 437 deletions
Download patch file Download diff file Expand all files Collapse all files

81
components/gfx/text/text_run.rs
View file

@ -11,11 +11,11 @@ use std::cell::Cell;
use std::cmp::{Ordering, max};
use std::slice::Iter;
use std::sync::Arc;
use text::glyph::{CharIndex, GlyphStore};
use text::glyph::{ByteIndex, GlyphStore};
use webrender_traits;
thread_local! {
static INDEX_OF_FIRST_GLYPH_RUN_CACHE: Cell<Option<(*const TextRun, CharIndex, usize)>> =
static INDEX_OF_FIRST_GLYPH_RUN_CACHE: Cell<Option<(*const TextRun, ByteIndex, usize)>> =
Cell::new(None)
}
@ -51,19 +51,19 @@ impl Drop for TextRun {
pub struct GlyphRun {
/// The glyphs.
pub glyph_store: Arc<GlyphStore>,
/// The range of characters in the containing run.
pub range: Range<CharIndex>,
/// The byte range of characters in the containing run.
pub range: Range<ByteIndex>,
}
pub struct NaturalWordSliceIterator<'a> {
glyphs: &'a [GlyphRun],
index: usize,
range: Range<CharIndex>,
range: Range<ByteIndex>,
reverse: bool,
}
impl GlyphRun {
fn compare(&self, key: &CharIndex) -> Ordering {
fn compare(&self, key: &ByteIndex) -> Ordering {
if *key < self.range.begin() {
Ordering::Greater
} else if *key >= self.range.end() {
@ -79,16 +79,16 @@ impl GlyphRun {
pub struct TextRunSlice<'a> {
/// The glyph store that the glyphs in this slice belong to.
pub glyphs: &'a GlyphStore,
/// The character index that this slice begins at, relative to the start of the *text run*.
pub offset: CharIndex,
/// The byte index that this slice begins at, relative to the start of the *text run*.
pub offset: ByteIndex,
/// The range that these glyphs encompass, relative to the start of the *glyph store*.
pub range: Range<CharIndex>,
pub range: Range<ByteIndex>,
}
impl<'a> TextRunSlice<'a> {
/// Returns the range that these glyphs encompass, relative to the start of the *text run*.
#[inline]
pub fn text_run_range(&self) -> Range<CharIndex> {
pub fn text_run_range(&self) -> Range<ByteIndex> {
let mut range = self.range;
range.shift_by(self.offset);
range
@ -116,15 +116,15 @@ impl<'a> Iterator for NaturalWordSliceIterator<'a> {
self.index += 1;
}
let mut char_range = self.range.intersect(&slice_glyphs.range);
let mut byte_range = self.range.intersect(&slice_glyphs.range);
let slice_range_begin = slice_glyphs.range.begin();
char_range.shift_by(-slice_range_begin);
byte_range.shift_by(-slice_range_begin);
if !char_range.is_empty() {
if !byte_range.is_empty() {
Some(TextRunSlice {
glyphs: &*slice_glyphs.glyph_store,
offset: slice_range_begin,
range: char_range,
range: byte_range,
})
} else {
None
@ -133,9 +133,10 @@ impl<'a> Iterator for NaturalWordSliceIterator<'a> {
}
pub struct CharacterSliceIterator<'a> {
text: &'a str,
glyph_run: Option<&'a GlyphRun>,
glyph_run_iter: Iter<'a, GlyphRun>,
range: Range<CharIndex>,
range: Range<ByteIndex>,
}
impl<'a> Iterator for CharacterSliceIterator<'a> {
@ -150,8 +151,13 @@ impl<'a> Iterator for CharacterSliceIterator<'a> {
};
debug_assert!(!self.range.is_empty());
let index_to_return = self.range.begin();
self.range.adjust_by(CharIndex(1), CharIndex(-1));
let byte_start = self.range.begin();
let byte_len = match self.text[byte_start.to_usize()..].chars().next() {
Some(ch) => ByteIndex(ch.len_utf8() as isize),
None => unreachable!() // XXX refactor?
};
self.range.adjust_by(byte_len, -byte_len);
if self.range.is_empty() {
// We're done.
self.glyph_run = None
@ -160,11 +166,11 @@ impl<'a> Iterator for CharacterSliceIterator<'a> {
self.glyph_run = self.glyph_run_iter.next();
}
let index_within_glyph_run = index_to_return - glyph_run.range.begin();
let index_within_glyph_run = byte_start - glyph_run.range.begin();
Some(TextRunSlice {
glyphs: &*glyph_run.glyph_store,
offset: glyph_run.range.begin(),
range: Range::new(index_within_glyph_run, CharIndex(1)),
range: Range::new(index_within_glyph_run, byte_len),
})
}
}
@ -187,9 +193,9 @@ impl<'a> TextRun {
-> Vec<GlyphRun> {
// TODO(Issue #230): do a better job. See Gecko's LineBreaker.
let mut glyphs = vec!();
let (mut byte_i, mut char_i) = (0, CharIndex(0));
let mut byte_i = 0;
let mut cur_slice_is_whitespace = false;
let (mut byte_last_boundary, mut char_last_boundary) = (0, CharIndex(0));
let mut byte_last_boundary = 0;
for ch in text.chars() {
// Slices alternate between whitespace and non-whitespace,
@ -225,14 +231,13 @@ impl<'a> TextRun {
glyphs.push(GlyphRun {
glyph_store: font.shape_text(slice, &options),
range: Range::new(char_last_boundary, char_i - char_last_boundary),
range: Range::new(ByteIndex(byte_last_boundary as isize),
ByteIndex((byte_i - byte_last_boundary) as isize)),
});
byte_last_boundary = byte_i;
char_last_boundary = char_i;
}
byte_i = byte_i + ch.len_utf8();
char_i = char_i + CharIndex(1);
}
// Create a glyph store for the final slice if it's nonempty.
@ -248,7 +253,8 @@ impl<'a> TextRun {
glyphs.push(GlyphRun {
glyph_store: font.shape_text(slice, &options),
range: Range::new(char_last_boundary, char_i - char_last_boundary),
range: Range::new(ByteIndex(byte_last_boundary as isize),
ByteIndex((byte_i - byte_last_boundary) as isize)),
});
}
@ -263,7 +269,7 @@ impl<'a> TextRun {
self.font_metrics.descent
}
pub fn advance_for_range(&self, range: &Range<CharIndex>) -> Au {
pub fn advance_for_range(&self, range: &Range<ByteIndex>) -> Au {
if range.is_empty() {
return Au(0)
}
@ -272,24 +278,24 @@ impl<'a> TextRun {
// TODO(Issue #98): using inter-char and inter-word spacing settings when measuring text
self.natural_word_slices_in_range(range)
.fold(Au(0), |advance, slice| {
advance + slice.glyphs.advance_for_char_range(&slice.range)
advance + slice.glyphs.advance_for_byte_range(&slice.range)
})
}
pub fn metrics_for_range(&self, range: &Range<CharIndex>) -> RunMetrics {
pub fn metrics_for_range(&self, range: &Range<ByteIndex>) -> RunMetrics {
RunMetrics::new(self.advance_for_range(range),
self.font_metrics.ascent,
self.font_metrics.descent)
}
pub fn metrics_for_slice(&self, glyphs: &GlyphStore, slice_range: &Range<CharIndex>)
pub fn metrics_for_slice(&self, glyphs: &GlyphStore, slice_range: &Range<ByteIndex>)
-> RunMetrics {
RunMetrics::new(glyphs.advance_for_char_range(slice_range),
RunMetrics::new(glyphs.advance_for_byte_range(slice_range),
self.font_metrics.ascent,
self.font_metrics.descent)
}
pub fn min_width_for_range(&self, range: &Range<CharIndex>) -> Au {
pub fn min_width_for_range(&self, range: &Range<ByteIndex>) -> Au {
debug!("iterating outer range {:?}", range);
self.natural_word_slices_in_range(range).fold(Au(0), |max_piece_width, slice| {
debug!("iterated on {:?}[{:?}]", slice.offset, slice.range);
@ -297,8 +303,8 @@ impl<'a> TextRun {
})
}
/// Returns the index of the first glyph run containing the given character index.
fn index_of_first_glyph_run_containing(&self, index: CharIndex) -> Option<usize> {
/// Returns the index of the first glyph run containing the given byte index.
fn index_of_first_glyph_run_containing(&self, index: ByteIndex) -> Option<usize> {
let self_ptr = self as *const TextRun;
INDEX_OF_FIRST_GLYPH_RUN_CACHE.with(|index_of_first_glyph_run_cache| {
if let Some((last_text_run, last_index, last_result)) =
@ -319,7 +325,7 @@ impl<'a> TextRun {
/// Returns an iterator that will iterate over all slices of glyphs that represent natural
/// words in the given range.
pub fn natural_word_slices_in_range(&'a self, range: &Range<CharIndex>)
pub fn natural_word_slices_in_range(&'a self, range: &Range<ByteIndex>)
-> NaturalWordSliceIterator<'a> {
let index = match self.index_of_first_glyph_run_containing(range.begin()) {
None => self.glyphs.len(),
@ -335,13 +341,13 @@ impl<'a> TextRun {
/// Returns an iterator that over natural word slices in visual order (left to right or
/// right to left, depending on the bidirectional embedding level).
pub fn natural_word_slices_in_visual_order(&'a self, range: &Range<CharIndex>)
pub fn natural_word_slices_in_visual_order(&'a self, range: &Range<ByteIndex>)
-> NaturalWordSliceIterator<'a> {
// Iterate in reverse order if bidi level is RTL.
let reverse = self.bidi_level % 2 == 1;
let index = if reverse {
match self.index_of_first_glyph_run_containing(range.end() - CharIndex(1)) {
match self.index_of_first_glyph_run_containing(range.end() - ByteIndex(1)) {
Some(i) => i + 1, // In reverse mode, index points one past the next element.
None => 0
}
@ -361,7 +367,7 @@ impl<'a> TextRun {
/// Returns an iterator that will iterate over all slices of glyphs that represent individual
/// characters in the given range.
pub fn character_slices_in_range(&'a self, range: &Range<CharIndex>)
pub fn character_slices_in_range(&'a self, range: &Range<ByteIndex>)
-> CharacterSliceIterator<'a> {
let index = match self.index_of_first_glyph_run_containing(range.begin()) {
None => self.glyphs.len(),
@ -370,6 +376,7 @@ impl<'a> TextRun {
let mut glyph_run_iter = self.glyphs[index..].iter();
let first_glyph_run = glyph_run_iter.next();
CharacterSliceIterator {
text: &self.text,
glyph_run: first_glyph_run,
glyph_run_iter: glyph_run_iter,
range: *range,