diff --git a/scripts/unicode.py b/scripts/unicode.py index 580b3c0..617f8cd 100755 --- a/scripts/unicode.py +++ b/scripts/unicode.py @@ -280,7 +280,8 @@ def emit_break_module(f, break_table, break_cats, name): f.write((" %sC_" % Name[0]) + cat + ",\n") f.write(""" } - fn bsearch_range_value_table(c: char, r: &'static [(char, char, %sCat)]) -> %sCat { + fn bsearch_range_value_table(c: char, r: &'static [(char, char, %sCat)]) -> (u32, u32, %sCat) { + use core; use core::cmp::Ordering::{Equal, Less, Greater}; match r.binary_search_by(|&(lo, hi, _)| { if lo <= c && c <= hi { Equal } @@ -288,14 +289,20 @@ def emit_break_module(f, break_table, break_cats, name): else { Greater } }) { Ok(idx) => { - let (_, _, cat) = r[idx]; - cat + let (lower, upper, cat) = r[idx]; + (lower as u32, upper as u32, cat) + } + Err(idx) => { + ( + if idx > 0 { r[idx-1].1 as u32 + 1 } else { 0 }, + r.get(idx).map(|c|c.0 as u32 - 1).unwrap_or(core::u32::MAX), + %sC_Any, + ) } - Err(_) => %sC_Any } } - pub fn %s_category(c: char) -> %sCat { + pub fn %s_category(c: char) -> (u32, u32, %sCat) { bsearch_range_value_table(c, %s_cat_table) } diff --git a/src/grapheme.rs b/src/grapheme.rs index cde6526..b66536e 100644 --- a/src/grapheme.rs +++ b/src/grapheme.rs @@ -178,6 +178,8 @@ pub struct GraphemeCursor { // Set if a call to `prev_boundary` or `next_boundary` was suspended due // to needing more input. resuming: bool, + // Cached grapheme category and associated scalar value range. + grapheme_cat_cache: (u32, u32, GraphemeCat), } /// An error return indicating that not enough content was available in the @@ -276,9 +278,20 @@ impl GraphemeCursor { pre_context_offset: None, ris_count: None, resuming: false, + grapheme_cat_cache: (0, 0, GraphemeCat::GC_Control), } } + fn grapheme_category(&mut self, ch: char) -> GraphemeCat { + use tables::grapheme as gr; + // If this char isn't within the cached range, update the cache to the + // range that includes it. + if (ch as u32) < self.grapheme_cat_cache.0 || (ch as u32) > self.grapheme_cat_cache.1 { + self.grapheme_cat_cache = gr::grapheme_category(ch); + } + self.grapheme_cat_cache.2 + } + // Not sure I'm gonna keep this, the advantage over new() seems thin. /// Set the cursor to a new location in the same string. @@ -349,7 +362,7 @@ impl GraphemeCursor { self.pre_context_offset = None; if self.is_extended && chunk_start + chunk.len() == self.offset { let ch = chunk.chars().rev().next().unwrap(); - if gr::grapheme_category(ch) == gr::GC_Prepend { + if self.grapheme_category(ch) == gr::GC_Prepend { self.decide(false); // GB9b return; } @@ -359,7 +372,7 @@ impl GraphemeCursor { GraphemeState::Emoji => self.handle_emoji(chunk, chunk_start), _ => if self.cat_before.is_none() && self.offset == chunk.len() + chunk_start { let ch = chunk.chars().rev().next().unwrap(); - self.cat_before = Some(gr::grapheme_category(ch)); + self.cat_before = Some(self.grapheme_category(ch)); }, } } @@ -393,7 +406,7 @@ impl GraphemeCursor { use tables::grapheme as gr; let mut ris_count = self.ris_count.unwrap_or(0); for ch in chunk.chars().rev() { - if gr::grapheme_category(ch) != gr::GC_Regional_Indicator { + if self.grapheme_category(ch) != gr::GC_Regional_Indicator { self.ris_count = Some(ris_count); self.decide((ris_count % 2) == 0); return; @@ -413,13 +426,13 @@ impl GraphemeCursor { use tables::grapheme as gr; let mut iter = chunk.chars().rev(); if let Some(ch) = iter.next() { - if gr::grapheme_category(ch) != gr::GC_ZWJ { + if self.grapheme_category(ch) != gr::GC_ZWJ { self.decide(true); return; } } for ch in iter { - match gr::grapheme_category(ch) { + match self.grapheme_category(ch) { gr::GC_Extend => (), gr::GC_Extended_Pictographic => { self.decide(false); @@ -481,7 +494,7 @@ impl GraphemeCursor { let offset_in_chunk = self.offset - chunk_start; if self.cat_after.is_none() { let ch = chunk[offset_in_chunk..].chars().next().unwrap(); - self.cat_after = Some(gr::grapheme_category(ch)); + self.cat_after = Some(self.grapheme_category(ch)); } if self.offset == chunk_start { let mut need_pre_context = true; @@ -497,7 +510,7 @@ impl GraphemeCursor { } if self.cat_before.is_none() { let ch = chunk[..offset_in_chunk].chars().rev().next().unwrap(); - self.cat_before = Some(gr::grapheme_category(ch)); + self.cat_before = Some(self.grapheme_category(ch)); } match check_pair(self.cat_before.unwrap(), self.cat_after.unwrap()) { PairResult::NotBreak => return self.decision(false), @@ -553,7 +566,6 @@ impl GraphemeCursor { /// assert_eq!(cursor.next_boundary(&s[2..4], 2), Ok(None)); /// ``` pub fn next_boundary(&mut self, chunk: &str, chunk_start: usize) -> Result, GraphemeIncomplete> { - use tables::grapheme as gr; if self.offset == self.len { return Ok(None); } @@ -562,14 +574,14 @@ impl GraphemeCursor { loop { if self.resuming { if self.cat_after.is_none() { - self.cat_after = Some(gr::grapheme_category(ch)); + self.cat_after = Some(self.grapheme_category(ch)); } } else { self.offset += ch.len_utf8(); self.state = GraphemeState::Unknown; self.cat_before = self.cat_after.take(); if self.cat_before.is_none() { - self.cat_before = Some(gr::grapheme_category(ch)); + self.cat_before = Some(self.grapheme_category(ch)); } if self.cat_before.unwrap() == GraphemeCat::GC_Regional_Indicator { self.ris_count = self.ris_count.map(|c| c + 1); @@ -578,7 +590,7 @@ impl GraphemeCursor { } if let Some(next_ch) = iter.next() { ch = next_ch; - self.cat_after = Some(gr::grapheme_category(ch)); + self.cat_after = Some(self.grapheme_category(ch)); } else if self.offset == self.len { self.decide(true); } else { @@ -629,7 +641,6 @@ impl GraphemeCursor { /// assert_eq!(cursor.prev_boundary(&s[0..2], 0), Ok(None)); /// ``` pub fn prev_boundary(&mut self, chunk: &str, chunk_start: usize) -> Result, GraphemeIncomplete> { - use tables::grapheme as gr; if self.offset == 0 { return Ok(None); } @@ -644,7 +655,7 @@ impl GraphemeCursor { return Err(GraphemeIncomplete::PrevChunk); } if self.resuming { - self.cat_before = Some(gr::grapheme_category(ch)); + self.cat_before = Some(self.grapheme_category(ch)); } else { self.offset -= ch.len_utf8(); self.cat_after = self.cat_before.take(); @@ -654,12 +665,12 @@ impl GraphemeCursor { } if let Some(prev_ch) = iter.next() { ch = prev_ch; - self.cat_before = Some(gr::grapheme_category(ch)); + self.cat_before = Some(self.grapheme_category(ch)); } else if self.offset == 0 { self.decide(true); } else { self.resuming = true; - self.cat_after = Some(gr::grapheme_category(ch)); + self.cat_after = Some(self.grapheme_category(ch)); return Err(GraphemeIncomplete::PrevChunk); } } diff --git a/src/sentence.rs b/src/sentence.rs index c16c927..48cafba 100644 --- a/src/sentence.rs +++ b/src/sentence.rs @@ -115,7 +115,7 @@ mod fwd { for next_char in ahead.chars() { //( ¬(OLetter | Upper | Lower | ParaSep | SATerm) )* Lower - match se::sentence_category(next_char) { + match se::sentence_category(next_char).2 { se::SC_Lower => return true, se::SC_OLetter | se::SC_Upper | @@ -182,7 +182,7 @@ mod fwd { let position_before = self.pos; let state_before = self.state.clone(); - let next_cat = se::sentence_category(next_char); + let next_cat = se::sentence_category(next_char).2; self.pos += next_char.len_utf8(); self.state = self.state.next(next_cat); diff --git a/src/tables.rs b/src/tables.rs index bfd7290..8e98c9a 100644 --- a/src/tables.rs +++ b/src/tables.rs @@ -345,7 +345,8 @@ pub mod grapheme { GC_ZWJ, } - fn bsearch_range_value_table(c: char, r: &'static [(char, char, GraphemeCat)]) -> GraphemeCat { + fn bsearch_range_value_table(c: char, r: &'static [(char, char, GraphemeCat)]) -> (u32, u32, GraphemeCat) { + use core; use core::cmp::Ordering::{Equal, Less, Greater}; match r.binary_search_by(|&(lo, hi, _)| { if lo <= c && c <= hi { Equal } @@ -353,14 +354,20 @@ pub mod grapheme { else { Greater } }) { Ok(idx) => { - let (_, _, cat) = r[idx]; - cat + let (lower, upper, cat) = r[idx]; + (lower as u32, upper as u32, cat) + } + Err(idx) => { + ( + if idx > 0 { r[idx-1].1 as u32 + 1 } else { 0 }, + r.get(idx).map(|c|c.0 as u32 - 1).unwrap_or(core::u32::MAX), + GC_Any, + ) } - Err(_) => GC_Any } } - pub fn grapheme_category(c: char) -> GraphemeCat { + pub fn grapheme_category(c: char) -> (u32, u32, GraphemeCat) { bsearch_range_value_table(c, grapheme_cat_table) } @@ -980,7 +987,8 @@ pub mod word { WC_ZWJ, } - fn bsearch_range_value_table(c: char, r: &'static [(char, char, WordCat)]) -> WordCat { + fn bsearch_range_value_table(c: char, r: &'static [(char, char, WordCat)]) -> (u32, u32, WordCat) { + use core; use core::cmp::Ordering::{Equal, Less, Greater}; match r.binary_search_by(|&(lo, hi, _)| { if lo <= c && c <= hi { Equal } @@ -988,14 +996,20 @@ pub mod word { else { Greater } }) { Ok(idx) => { - let (_, _, cat) = r[idx]; - cat + let (lower, upper, cat) = r[idx]; + (lower as u32, upper as u32, cat) + } + Err(idx) => { + ( + if idx > 0 { r[idx-1].1 as u32 + 1 } else { 0 }, + r.get(idx).map(|c|c.0 as u32 - 1).unwrap_or(core::u32::MAX), + WC_Any, + ) } - Err(_) => WC_Any } } - pub fn word_category(c: char) -> WordCat { + pub fn word_category(c: char) -> (u32, u32, WordCat) { bsearch_range_value_table(c, word_cat_table) } @@ -1439,7 +1453,8 @@ pub mod emoji { EC_Extended_Pictographic, } - fn bsearch_range_value_table(c: char, r: &'static [(char, char, EmojiCat)]) -> EmojiCat { + fn bsearch_range_value_table(c: char, r: &'static [(char, char, EmojiCat)]) -> (u32, u32, EmojiCat) { + use core; use core::cmp::Ordering::{Equal, Less, Greater}; match r.binary_search_by(|&(lo, hi, _)| { if lo <= c && c <= hi { Equal } @@ -1447,14 +1462,20 @@ pub mod emoji { else { Greater } }) { Ok(idx) => { - let (_, _, cat) = r[idx]; - cat + let (lower, upper, cat) = r[idx]; + (lower as u32, upper as u32, cat) + } + Err(idx) => { + ( + if idx > 0 { r[idx-1].1 as u32 + 1 } else { 0 }, + r.get(idx).map(|c|c.0 as u32 - 1).unwrap_or(core::u32::MAX), + EC_Any, + ) } - Err(_) => EC_Any } } - pub fn emoji_category(c: char) -> EmojiCat { + pub fn emoji_category(c: char) -> (u32, u32, EmojiCat) { bsearch_range_value_table(c, emoji_cat_table) } @@ -1535,7 +1556,8 @@ pub mod sentence { SC_Upper, } - fn bsearch_range_value_table(c: char, r: &'static [(char, char, SentenceCat)]) -> SentenceCat { + fn bsearch_range_value_table(c: char, r: &'static [(char, char, SentenceCat)]) -> (u32, u32, SentenceCat) { + use core; use core::cmp::Ordering::{Equal, Less, Greater}; match r.binary_search_by(|&(lo, hi, _)| { if lo <= c && c <= hi { Equal } @@ -1543,14 +1565,20 @@ pub mod sentence { else { Greater } }) { Ok(idx) => { - let (_, _, cat) = r[idx]; - cat + let (lower, upper, cat) = r[idx]; + (lower as u32, upper as u32, cat) + } + Err(idx) => { + ( + if idx > 0 { r[idx-1].1 as u32 + 1 } else { 0 }, + r.get(idx).map(|c|c.0 as u32 - 1).unwrap_or(core::u32::MAX), + SC_Any, + ) } - Err(_) => SC_Any } } - pub fn sentence_category(c: char) -> SentenceCat { + pub fn sentence_category(c: char) -> (u32, u32, SentenceCat) { bsearch_range_value_table(c, sentence_cat_table) } diff --git a/src/word.rs b/src/word.rs index 6e9c049..5cf111a 100644 --- a/src/word.rs +++ b/src/word.rs @@ -125,7 +125,7 @@ enum RegionalState { fn is_emoji(ch: char) -> bool { use tables::emoji; - emoji::emoji_category(ch) == emoji::EmojiCat::EC_Extended_Pictographic + emoji::emoji_category(ch).2 == emoji::EmojiCat::EC_Extended_Pictographic } impl<'a> Iterator for UWordBounds<'a> { @@ -164,7 +164,7 @@ impl<'a> Iterator for UWordBounds<'a> { prev_zwj = cat == wd::WC_ZWJ; // if there's a category cached, grab it cat = match self.cat { - None => wd::word_category(ch), + None => wd::word_category(ch).2, _ => self.cat.take().unwrap() }; take_cat = true; @@ -391,7 +391,7 @@ impl<'a> DoubleEndedIterator for UWordBounds<'a> { // if there's a category cached, grab it cat = match self.catb { - None => wd::word_category(ch), + None => wd::word_category(ch).2, _ => self.catb.take().unwrap() }; take_cat = true; @@ -533,7 +533,7 @@ impl<'a> DoubleEndedIterator for UWordBounds<'a> { if regional_state == RegionalState::Unknown { let count = self.string[..previdx] .chars().rev() - .map(|c| wd::word_category(c)) + .map(|c| wd::word_category(c).2) .filter(|&c| ! (c == wd::WC_ZWJ || c == wd::WC_Extend || c == wd::WC_Format)) .take_while(|&c| c == wd::WC_Regional_Indicator) .count(); @@ -624,7 +624,7 @@ impl<'a> UWordBounds<'a> { let nidx = idx + self.string[idx..].chars().next().unwrap().len_utf8(); if nidx < self.string.len() { let nch = self.string[nidx..].chars().next().unwrap(); - Some(wd::word_category(nch)) + Some(wd::word_category(nch).2) } else { None } @@ -635,7 +635,7 @@ impl<'a> UWordBounds<'a> { use tables::word as wd; if idx > 0 { let nch = self.string[..idx].chars().next_back().unwrap(); - Some(wd::word_category(nch)) + Some(wd::word_category(nch).2) } else { None } pFad - Phonifier reborn Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved. Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages. Alternative Proxies:Alternative ProxypFad ProxypFad v3 ProxypFad v4 Proxy
Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.
Alternative Proxies:
Alternative Proxy
pFad Proxy
pFad v3 Proxy
pFad v4 Proxy