11904027177

Committed 19 Nov 2024 12:33AM UTC coverage: 75.477% (+0.3%) from 75.174%

Build # 11904027177

Build Type

push

github

Committed by

web-flow

Commit Message

Move DateTimePattern into pattern module (#5834)

#1317

Also removes `NeoNeverMarker` and fixes #5689

Run Details

258 of 319 new or added lines in 6 files covered. (80.88%)

6967 existing lines in 278 files now uncovered.

54522 of 72237 relevant lines covered (75.48%)

655305.49 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

72.54

/components/segmenter/src/rule_segmenter.rs

// This file is part of ICU4X. For terms of use, please see the file
// called LICENSE at the top level of the ICU4X source tree
// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ).

use crate::complex::ComplexPayloads;
use crate::indices::{Latin1Indices, Utf16Indices};
use crate::provider::*;
use crate::WordType;
use core::str::CharIndices;
use utf8_iter::Utf8CharIndices;

/// A trait allowing for RuleBreakIterator to be generalized to multiple string
/// encoding methods and granularity such as grapheme cluster, word, etc.
pub trait RuleBreakType<'l, 's> {
    /// The iterator over characters.
    type IterAttr: Iterator<Item = (usize, Self::CharType)> + Clone + core::fmt::Debug;

    /// The character type.
    type CharType: Copy + Into<u32> + core::fmt::Debug;

    fn get_current_position_character_len(iter: &RuleBreakIterator<'l, 's, Self>) -> usize;

    fn handle_complex_language(
        iter: &mut RuleBreakIterator<'l, 's, Self>,
        left_codepoint: Self::CharType,
    ) -> Option<usize>;
}

/// Implements the [`Iterator`] trait over the segmenter boundaries of the given string.
///
/// Lifetimes:
///
/// - `'l` = lifetime of the segmenter object from which this iterator was created
/// - `'s` = lifetime of the string being segmented
///
/// The [`Iterator::Item`] is an [`usize`] representing index of a code unit
/// _after_ the boundary (for a boundary at the end of text, this index is the length
/// of the [`str`] or array of code units).
#[derive(Debug)]
pub struct RuleBreakIterator<'l, 's, Y: RuleBreakType<'l, 's> + ?Sized> {
    pub(crate) iter: Y::IterAttr,
    pub(crate) len: usize,
    pub(crate) current_pos_data: Option<(usize, Y::CharType)>,
    pub(crate) result_cache: alloc::vec::Vec<usize>,
    pub(crate) data: &'l RuleBreakDataV2<'l>,
    pub(crate) complex: Option<&'l ComplexPayloads>,
    pub(crate) boundary_property: u8,
    pub(crate) locale_override: Option<&'l RuleBreakDataOverrideV1<'l>>,
}

impl<'l, 's, Y: RuleBreakType<'l, 's> + ?Sized> Iterator for RuleBreakIterator<'l, 's, Y> {
    type Item = usize;

    fn next(&mut self) -> Option<Self::Item> {
        // If we have break point cache by previous run, return this result
        if let Some(&first_result) = self.result_cache.first() {
            let mut i = 0;
            loop {
                if i == first_result {
                    self.result_cache = self.result_cache.iter().skip(1).map(|r| r - i).collect();
                    return self.get_current_position();
                }
                i += Y::get_current_position_character_len(self);
                self.advance_iter();
                if self.is_eof() {
                    self.result_cache.clear();
                    self.boundary_property = self.data.complex_property;
                    return Some(self.len);
                }
            }
        }

        if self.is_eof() {
            self.advance_iter();
            if self.is_eof() && self.len == 0 {
                // Empty string. Since `self.current_pos_data` is always going to be empty,
                // we never read `self.len` except for here, so we can use it to mark that
                // we have already returned the single empty-string breakpoint.
                self.len = 1;
                return Some(0);
            }
            let Some(right_prop) = self.get_current_break_property() else {
                // iterator already reaches to EOT. Reset boundary property for word-like.
                self.boundary_property = 0;
                return None;
            };
            // SOT x anything
            if matches!(
                self.get_break_state_from_table(self.data.sot_property, right_prop),
                BreakState::Break | BreakState::NoMatch
            ) {
                self.boundary_property = 0; // SOT is special type
                return self.get_current_position();
            }
        }

        'a: loop {
            debug_assert!(!self.is_eof());
            let left_codepoint = self.get_current_codepoint()?;
            let left_prop = self.get_break_property(left_codepoint);
            self.advance_iter();

            let Some(right_prop) = self.get_current_break_property() else {
                self.boundary_property = left_prop;
                return Some(self.len);
            };

            // Some segmenter rules doesn't have language-specific rules, we have to use LSTM (or dictionary) segmenter.
            // If property is marked as SA, use it
            if right_prop == self.data.complex_property {
                if left_prop != self.data.complex_property {
                    // break before SA
                    self.boundary_property = left_prop;
                    return self.get_current_position();
                }
                let break_offset = Y::handle_complex_language(self, left_codepoint);
                if break_offset.is_some() {
                    return break_offset;
                }
            }

            match self.get_break_state_from_table(left_prop, right_prop) {
                BreakState::Keep => continue,
                BreakState::Break | BreakState::NoMatch => {
                    self.boundary_property = left_prop;
                    return self.get_current_position();
                }
                BreakState::Index(mut index) | BreakState::Intermediate(mut index) => {
                    // This isn't simple rule set. We need marker to restore iterator to previous position.
                    let mut previous_iter = self.iter.clone();
                    let mut previous_pos_data = self.current_pos_data;
                    let mut previous_left_prop = left_prop;

                    loop {
                        self.advance_iter();

                        let Some(prop) = self.get_current_break_property() else {
                            // Reached EOF. But we are analyzing multiple characters now, so next break may be previous point.
                            self.boundary_property = index;
                            if self.get_break_state_from_table(index, self.data.eot_property)
                                == BreakState::NoMatch
                            {
                                self.boundary_property = previous_left_prop;
                                self.iter = previous_iter;
                                self.current_pos_data = previous_pos_data;
                                return self.get_current_position();
                            }
                            // EOF
                            return Some(self.len);
                        };

                        let previous_break_state_is_cp_prop =
                            index <= self.data.last_codepoint_property;

                        match self.get_break_state_from_table(index, prop) {
                            BreakState::Keep => continue 'a,
                            BreakState::NoMatch => {
                                self.boundary_property = previous_left_prop;
                                self.iter = previous_iter;
                                self.current_pos_data = previous_pos_data;
                                return self.get_current_position();
                            }
                            BreakState::Break => return self.get_current_position(),
                            BreakState::Intermediate(i) => {
                                index = i;
                                if previous_break_state_is_cp_prop {
                                    // Move marker
                                    previous_left_prop = index;
                                }
                                previous_iter = self.iter.clone();
                                previous_pos_data = self.current_pos_data;
                            }
                            BreakState::Index(i) => {
                                index = i;
                                if previous_break_state_is_cp_prop {
                                    // Move marker
                                    previous_iter = self.iter.clone();
                                    previous_pos_data = self.current_pos_data;
                                    previous_left_prop = index;
                                }
                            }
                        }
                    }
                }
            }
        }
    }
}

impl<'l, 's, Y: RuleBreakType<'l, 's> + ?Sized> RuleBreakIterator<'l, 's, Y> {
    pub(crate) fn advance_iter(&mut self) {
        self.current_pos_data = self.iter.next();
    }

    pub(crate) fn is_eof(&self) -> bool {
        self.current_pos_data.is_none()
    }

    pub(crate) fn get_current_break_property(&self) -> Option<u8> {
        self.get_current_codepoint()
            .map(|c| self.get_break_property(c))
    }

    pub(crate) fn get_current_position(&self) -> Option<usize> {
        self.current_pos_data.map(|(pos, _)| pos)
    }

    pub(crate) fn get_current_codepoint(&self) -> Option<Y::CharType> {
        self.current_pos_data.map(|(_, codepoint)| codepoint)
    }

    fn get_break_property(&self, codepoint: Y::CharType) -> u8 {
        // Note: Default value is 0 == UNKNOWN
        if let Some(locale_override) = &self.locale_override {
            let property = locale_override
                .property_table_override
                .get32(codepoint.into());
            if property != 0 {
                return property;
            }
        }
        self.data.property_table.get32(codepoint.into())
    }

    fn get_break_state_from_table(&self, left: u8, right: u8) -> BreakState {
        let idx = left as usize * self.data.property_count as usize + right as usize;
        // We use unwrap_or to fall back to the base case and prevent panics on bad data.
        self.data
            .break_state_table
            .get(idx)
            .unwrap_or(BreakState::Keep)
    }

    /// Return the status value of break boundary.
    /// If segmenter isn't word, always return WordType::None
    pub fn word_type(&self) -> WordType {
        if !self.result_cache.is_empty() {
            // Dictionary type (CJ and East Asian) is letter.
            return WordType::Letter;
        }
        if self.boundary_property == 0 {
            // break position is SOT / Any
            return WordType::None;
        }
        self.data
            .word_type_table
            .get((self.boundary_property - 1) as usize)
            .unwrap_or(WordType::None)
    }

    /// Return true when break boundary is word-like such as letter/number/CJK
    /// If segmenter isn't word, return false
    pub fn is_word_like(&self) -> bool {
        self.word_type().is_word_like()
    }
}

#[derive(Debug)]
pub struct RuleBreakTypeUtf8;

impl<'s> RuleBreakType<'_, 's> for RuleBreakTypeUtf8 {
    type IterAttr = CharIndices<'s>;
    type CharType = char;

    fn get_current_position_character_len(iter: &RuleBreakIterator<Self>) -> usize {
        iter.get_current_codepoint().map_or(0, |c| c.len_utf8())
    }

    fn handle_complex_language(
        _: &mut RuleBreakIterator<Self>,
        _: Self::CharType,
    ) -> Option<usize> {
        unreachable!()
    }
}

#[derive(Debug)]
pub struct RuleBreakTypePotentiallyIllFormedUtf8;

impl<'s> RuleBreakType<'_, 's> for RuleBreakTypePotentiallyIllFormedUtf8 {
    type IterAttr = Utf8CharIndices<'s>;
    type CharType = char;

    fn get_current_position_character_len(iter: &RuleBreakIterator<Self>) -> usize {
        iter.get_current_codepoint().map_or(0, |c| c.len_utf8())
    }

    fn handle_complex_language(
        _: &mut RuleBreakIterator<Self>,
        _: Self::CharType,
    ) -> Option<usize> {
        unreachable!()
    }
}

#[derive(Debug)]
pub struct RuleBreakTypeLatin1;

impl<'s> RuleBreakType<'_, 's> for RuleBreakTypeLatin1 {
    type IterAttr = Latin1Indices<'s>;
    type CharType = u8;

    fn get_current_position_character_len(_: &RuleBreakIterator<Self>) -> usize {
        unreachable!()
    }

    fn handle_complex_language(
        _: &mut RuleBreakIterator<Self>,
        _: Self::CharType,
    ) -> Option<usize> {
        unreachable!()
    }
}

#[derive(Debug)]
pub struct RuleBreakTypeUtf16;

impl<'s> RuleBreakType<'_, 's> for RuleBreakTypeUtf16 {
    type IterAttr = Utf16Indices<'s>;
    type CharType = u32;

    fn get_current_position_character_len(iter: &RuleBreakIterator<Self>) -> usize {
        match iter.get_current_codepoint() {
            None => 0,
            Some(ch) if ch >= 0x10000 => 2,
            _ => 1,
        }
    }

    fn handle_complex_language(
        _: &mut RuleBreakIterator<Self>,
        _: Self::CharType,
    ) -> Option<usize> {
        unreachable!()
    }
}

1	// This file is part of ICU4X. For terms of use, please see the file
2	// called LICENSE at the top level of the ICU4X source tree
3	// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ).
4
5	use crate::complex::ComplexPayloads;
6	use crate::indices::{Latin1Indices, Utf16Indices};
7	use crate::provider::*;
8	use crate::WordType;
9	use core::str::CharIndices;
10	use utf8_iter::Utf8CharIndices;
11
12	/// A trait allowing for RuleBreakIterator to be generalized to multiple string
13	/// encoding methods and granularity such as grapheme cluster, word, etc.
14	pub trait RuleBreakType<'l, 's> {
15	/// The iterator over characters.
16	type IterAttr: Iterator<Item = (usize, Self::CharType)> + Clone + core::fmt::Debug;
17
18	/// The character type.
19	type CharType: Copy + Into<u32> + core::fmt::Debug;
20
21	fn get_current_position_character_len(iter: &RuleBreakIterator<'l, 's, Self>) -> usize;
22
23	fn handle_complex_language(
24	iter: &mut RuleBreakIterator<'l, 's, Self>,
25	left_codepoint: Self::CharType,
26	) -> Option<usize>;
27	}
28
29	/// Implements the [`Iterator`] trait over the segmenter boundaries of the given string.
30	///
31	/// Lifetimes:
32	///
33	/// - `'l` = lifetime of the segmenter object from which this iterator was created
34	/// - `'s` = lifetime of the string being segmented
35	///
36	/// The [`Iterator::Item`] is an [`usize`] representing index of a code unit
37	/// _after_ the boundary (for a boundary at the end of text, this index is the length
38	/// of the [`str`] or array of code units).
39	#[derive(Debug)]
40	pub struct RuleBreakIterator<'l, 's, Y: RuleBreakType<'l, 's> + ?Sized> {
41	pub(crate) iter: Y::IterAttr,
42	pub(crate) len: usize,
43	pub(crate) current_pos_data: Option<(usize, Y::CharType)>,
44	pub(crate) result_cache: alloc::vec::Vec<usize>,
45	pub(crate) data: &'l RuleBreakDataV2<'l>,
46	pub(crate) complex: Option<&'l ComplexPayloads>,
47	pub(crate) boundary_property: u8,
48	pub(crate) locale_override: Option<&'l RuleBreakDataOverrideV1<'l>>,
49	}
50
51	impl<'l, 's, Y: RuleBreakType<'l, 's> + ?Sized> Iterator for RuleBreakIterator<'l, 's, Y> {
52	type Item = usize;
53
54	fn next(&mut self) -> Option<Self::Item> {	91,092✔
55	// If we have break point cache by previous run, return this result
56	if let Some(&first_result) = self.result_cache.first() {	91,092✔
57	let mut i = 0;	2✔
58	loop {	2✔
59	if i == first_result {	6✔
60	self.result_cache = self.result_cache.iter().skip(1).map(\|r\| r - i).collect();	×
UNCOV 61	return self.get_current_position();	×
62	}
63	i += Y::get_current_position_character_len(self);	6✔
64	self.advance_iter();	6✔
65	if self.is_eof() {	6✔
66	self.result_cache.clear();	2✔
67	self.boundary_property = self.data.complex_property;	2✔
68	return Some(self.len);	2✔
69	}
70	}
71	}
72
73	if self.is_eof() {	91,090✔
74	self.advance_iter();	7,760✔
75	if self.is_eof() && self.len == 0 {	7,760✔
76	// Empty string. Since `self.current_pos_data` is always going to be empty,
77	// we never read `self.len` except for here, so we can use it to mark that
78	// we have already returned the single empty-string breakpoint.
79	self.len = 1;	3✔
80	return Some(0);	3✔
81	}
82	let Some(right_prop) = self.get_current_break_property() else {	7,757✔
83	// iterator already reaches to EOT. Reset boundary property for word-like.
84	self.boundary_property = 0;	3,863✔
85	return None;	3,863✔
86	};
87	// SOT x anything
88	if matches!(	3,894✔
89	self.get_break_state_from_table(self.data.sot_property, right_prop),	3,894✔
90	BreakState::Break \| BreakState::NoMatch
91	) {
92	self.boundary_property = 0; // SOT is special type	3,894✔
93	return self.get_current_position();	3,894✔
94	}
95	}
96
97	'a: loop {
98	debug_assert!(!self.is_eof());	188,531✔
99	let left_codepoint = self.get_current_codepoint()?;	279,623✔
100	let left_prop = self.get_break_property(left_codepoint);	188,531✔
101	self.advance_iter();	188,531✔
102
103	let Some(right_prop) = self.get_current_break_property() else {	188,531✔
104	self.boundary_property = left_prop;	3,665✔
105	return Some(self.len);	3,665✔
106	};
107
108	// Some segmenter rules doesn't have language-specific rules, we have to use LSTM (or dictionary) segmenter.
109	// If property is marked as SA, use it
110	if right_prop == self.data.complex_property {	184,866✔
111	if left_prop != self.data.complex_property {	2✔
112	// break before SA
113	self.boundary_property = left_prop;	×
UNCOV 114	return self.get_current_position();	×
115	}
116	let break_offset = Y::handle_complex_language(self, left_codepoint);	2✔
117	if break_offset.is_some() {	2✔
118	return break_offset;	2✔
119	}
120	}
121
122	match self.get_break_state_from_table(left_prop, right_prop) {	184,864✔
123	BreakState::Keep => continue,
124	BreakState::Break \| BreakState::NoMatch => {
125	self.boundary_property = left_prop;	71,983✔
126	return self.get_current_position();	71,983✔
127	}
128	BreakState::Index(mut index) \| BreakState::Intermediate(mut index) => {	15,290✔
129	// This isn't simple rule set. We need marker to restore iterator to previous position.
130	let mut previous_iter = self.iter.clone();	15,290✔
131	let mut previous_pos_data = self.current_pos_data;	15,290✔
132	let mut previous_left_prop = left_prop;	15,290✔
133
134	loop {	15,290✔
135	self.advance_iter();	22,836✔
136
137	let Some(prop) = self.get_current_break_property() else {	22,810✔
138	// Reached EOF. But we are analyzing multiple characters now, so next break may be previous point.
139	self.boundary_property = index;	226✔
140	if self.get_break_state_from_table(index, self.data.eot_property)	226✔
141	== BreakState::NoMatch
142	{
143	self.boundary_property = previous_left_prop;	16✔
144	self.iter = previous_iter;	16✔
145	self.current_pos_data = previous_pos_data;	16✔
146	return self.get_current_position();	16✔
147	}
148	// EOF
149	return Some(self.len);	210✔
150	};
151
152	let previous_break_state_is_cp_prop =
153	index <= self.data.last_codepoint_property;	22,603✔
154
155	match self.get_break_state_from_table(index, prop) {	22,573✔
156	BreakState::Keep => continue 'a,
157	BreakState::NoMatch => {
158	self.boundary_property = previous_left_prop;	962✔
159	self.iter = previous_iter;	962✔
160	self.current_pos_data = previous_pos_data;	962✔
161	return self.get_current_position();	962✔
162	}
163	BreakState::Break => return self.get_current_position(),	6,470✔
164	BreakState::Intermediate(i) => {	638✔
165	index = i;	638✔
166	if previous_break_state_is_cp_prop {	638✔
167	// Move marker
168	previous_left_prop = index;	94✔
169	}
170	previous_iter = self.iter.clone();	638✔
171	previous_pos_data = self.current_pos_data;	638✔
172	}	638✔
173	BreakState::Index(i) => {	6,908✔
174	index = i;	6,908✔
175	if previous_break_state_is_cp_prop {	8,602✔
176	// Move marker
177	previous_iter = self.iter.clone();	1,694✔
178	previous_pos_data = self.current_pos_data;	1,694✔
179	previous_left_prop = index;	1,694✔
180	}
181	}
182	}
183	}
184	}	15,268✔
185	}
186	}
187	}	91,070✔
188	}
189
190	impl<'l, 's, Y: RuleBreakType<'l, 's> + ?Sized> RuleBreakIterator<'l, 's, Y> {
191	pub(crate) fn advance_iter(&mut self) {	242,132✔
192	self.current_pos_data = self.iter.next();	242,132✔
193	}	242,132✔
194
195	pub(crate) fn is_eof(&self) -> bool {	322,444✔
196	self.current_pos_data.is_none()	322,444✔
197	}	322,444✔
198
199	pub(crate) fn get_current_break_property(&self) -> Option<u8> {	242,245✔
200	self.get_current_codepoint()	242,245✔
201	.map(\|c\| self.get_break_property(c))	225,936✔
202	}	242,245✔
203
204	pub(crate) fn get_current_position(&self) -> Option<usize> {	92,444✔
205	self.current_pos_data.map(\|(pos, _)\| pos)	184,871✔
206	}	92,444✔
207
208	pub(crate) fn get_current_codepoint(&self) -> Option<Y::CharType> {	441,006✔
209	self.current_pos_data.map(\|(_, codepoint)\| codepoint)	865,683✔
210	}	441,006✔
211
212	fn get_break_property(&self, codepoint: Y::CharType) -> u8 {	423,287✔
213	// Note: Default value is 0 == UNKNOWN
214	if let Some(locale_override) = &self.locale_override {	423,287✔
215	let property = locale_override	47,648✔
216	.property_table_override
217	.get32(codepoint.into());	23,824✔
218	if property != 0 {	23,824✔
219	return property;	1,739✔
220	}
221	}
222	self.data.property_table.get32(codepoint.into())	421,548✔
223	}	423,287✔
224
225	fn get_break_state_from_table(&self, left: u8, right: u8) -> BreakState {	211,353✔
226	let idx = left as usize * self.data.property_count as usize + right as usize;	211,353✔
227	// We use unwrap_or to fall back to the base case and prevent panics on bad data.
228	self.data	422,706✔
229	.break_state_table
230	.get(idx)
231	.unwrap_or(BreakState::Keep)	211,353✔
232	}	211,353✔
233
234	/// Return the status value of break boundary.
235	/// If segmenter isn't word, always return WordType::None
236	pub fn word_type(&self) -> WordType {	×
UNCOV 237	if !self.result_cache.is_empty() {	×
238	// Dictionary type (CJ and East Asian) is letter.
239	return WordType::Letter;	×
240	}
UNCOV 241	if self.boundary_property == 0 {	×
242	// break position is SOT / Any
UNCOV 243	return WordType::None;	×
244	}
UNCOV 245	self.data	×
246	.word_type_table
UNCOV 247	.get((self.boundary_property - 1) as usize)	×
UNCOV 248	.unwrap_or(WordType::None)	×
249	}	×
250
251	/// Return true when break boundary is word-like such as letter/number/CJK
252	/// If segmenter isn't word, return false
253	pub fn is_word_like(&self) -> bool {
254	self.word_type().is_word_like()
255	}
256	}
257
258	#[derive(Debug)]	×
259	pub struct RuleBreakTypeUtf8;
260
261	impl<'s> RuleBreakType<'_, 's> for RuleBreakTypeUtf8 {
262	type IterAttr = CharIndices<'s>;
263	type CharType = char;
264
UNCOV 265	fn get_current_position_character_len(iter: &RuleBreakIterator<Self>) -> usize {	×
UNCOV 266	iter.get_current_codepoint().map_or(0, \|c\| c.len_utf8())	×
UNCOV 267	}	×
268
UNCOV 269	fn handle_complex_language(	×
270	_: &mut RuleBreakIterator<Self>,
271	_: Self::CharType,
272	) -> Option<usize> {
UNCOV 273	unreachable!()	×
274	}
275	}
276
277	#[derive(Debug)]	×
278	pub struct RuleBreakTypePotentiallyIllFormedUtf8;
279
280	impl<'s> RuleBreakType<'_, 's> for RuleBreakTypePotentiallyIllFormedUtf8 {
281	type IterAttr = Utf8CharIndices<'s>;
282	type CharType = char;
283
UNCOV 284	fn get_current_position_character_len(iter: &RuleBreakIterator<Self>) -> usize {	×
UNCOV 285	iter.get_current_codepoint().map_or(0, \|c\| c.len_utf8())	×
UNCOV 286	}	×
287
UNCOV 288	fn handle_complex_language(	×
289	_: &mut RuleBreakIterator<Self>,
290	_: Self::CharType,
291	) -> Option<usize> {
UNCOV 292	unreachable!()	×
293	}
294	}
295
UNCOV 296	#[derive(Debug)]	×
297	pub struct RuleBreakTypeLatin1;
298
299	impl<'s> RuleBreakType<'_, 's> for RuleBreakTypeLatin1 {
300	type IterAttr = Latin1Indices<'s>;
301	type CharType = u8;
302
UNCOV 303	fn get_current_position_character_len(_: &RuleBreakIterator<Self>) -> usize {	×
UNCOV 304	unreachable!()	×
305	}
306
UNCOV 307	fn handle_complex_language(	×
308	_: &mut RuleBreakIterator<Self>,
309	_: Self::CharType,
310	) -> Option<usize> {
UNCOV 311	unreachable!()	×
312	}
313	}
314
315	#[derive(Debug)]	×
316	pub struct RuleBreakTypeUtf16;
317
318	impl<'s> RuleBreakType<'_, 's> for RuleBreakTypeUtf16 {
319	type IterAttr = Utf16Indices<'s>;
320	type CharType = u32;
321
UNCOV 322	fn get_current_position_character_len(iter: &RuleBreakIterator<Self>) -> usize {	×
UNCOV 323	match iter.get_current_codepoint() {	×
UNCOV 324	None => 0,	×
325	Some(ch) if ch >= 0x10000 => 2,	×
UNCOV 326	_ => 1,	×
327	}
UNCOV 328	}	×
329
UNCOV 330	fn handle_complex_language(	×
331	_: &mut RuleBreakIterator<Self>,
332	_: Self::CharType,
333	) -> Option<usize> {
UNCOV 334	unreachable!()	×
335	}
336	}

zbraniecki / icu4x / 11904027177

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous