1 // Copyright 2012-2015 The Rust Project Developers. See the COPYRIGHT
2 // file at the top-level directory of this distribution and at
3 // http://rust-lang.org/COPYRIGHT.
4 //
5 // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
6 // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
7 // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
8 // option. This file may not be copied, modified, or distributed
9 // except according to those terms.
10
11 #[cfg(feature = "bench")]
12 use std::iter;
13 #[cfg(feature = "bench")]
14 use test::{self, Bencher};
15 #[cfg(feature = "bench")]
16 use super::{UnicodeWidthChar, UnicodeWidthStr};
17
18 use std::prelude::v1::*;
19
20 #[cfg(feature = "bench")]
21 #[bench]
cargo(b: &mut Bencher)22 fn cargo(b: &mut Bencher) {
23 let string = iter::repeat('a').take(4096).collect::<String>();
24
25 b.iter(|| {
26 for c in string.chars() {
27 test::black_box(UnicodeWidthChar::width(c));
28 }
29 });
30 }
31
32 #[cfg(feature = "bench")]
33 #[bench]
34 #[allow(deprecated)]
stdlib(b: &mut Bencher)35 fn stdlib(b: &mut Bencher) {
36 let string = iter::repeat('a').take(4096).collect::<String>();
37
38 b.iter(|| {
39 for c in string.chars() {
40 test::black_box(c.width());
41 }
42 });
43 }
44
45 #[cfg(feature = "bench")]
46 #[bench]
simple_if(b: &mut Bencher)47 fn simple_if(b: &mut Bencher) {
48 let string = iter::repeat('a').take(4096).collect::<String>();
49
50 b.iter(|| {
51 for c in string.chars() {
52 test::black_box(simple_width_if(c));
53 }
54 });
55 }
56
57 #[cfg(feature = "bench")]
58 #[bench]
simple_match(b: &mut Bencher)59 fn simple_match(b: &mut Bencher) {
60 let string = iter::repeat('a').take(4096).collect::<String>();
61
62 b.iter(|| {
63 for c in string.chars() {
64 test::black_box(simple_width_match(c));
65 }
66 });
67 }
68
69 #[cfg(feature = "bench")]
70 #[inline]
simple_width_if(c: char) -> Option<usize>71 fn simple_width_if(c: char) -> Option<usize> {
72 let cu = c as u32;
73 if cu < 127 {
74 if cu > 31 {
75 Some(1)
76 } else if cu == 0 {
77 Some(0)
78 } else {
79 None
80 }
81 } else {
82 UnicodeWidthChar::width(c)
83 }
84 }
85
86 #[cfg(feature = "bench")]
87 #[inline]
simple_width_match(c: char) -> Option<usize>88 fn simple_width_match(c: char) -> Option<usize> {
89 match c as u32 {
90 cu if cu == 0 => Some(0),
91 cu if cu < 0x20 => None,
92 cu if cu < 0x7f => Some(1),
93 _ => UnicodeWidthChar::width(c)
94 }
95 }
96 #[cfg(all(feature = "bench", not(feature = "no_std")))]
97 #[bench]
enwik8(b: &mut Bencher)98 fn enwik8(b: &mut Bencher) {
99 // To benchmark, download & unzip `enwik8` from https://data.deepai.org/enwik8.zip
100 let data_path = "bench_data/enwik8";
101 let string = std::fs::read_to_string(data_path).unwrap_or_default();
102 b.iter(|| test::black_box(UnicodeWidthStr::width(string.as_str())));
103 }
104 #[cfg(all(feature = "bench", not(feature = "no_std")))]
105 #[bench]
jawiki(b: &mut Bencher)106 fn jawiki(b: &mut Bencher) {
107 // To benchmark, download & extract `jawiki-20220501-pages-articles-multistream-index.txt` from
108 // https://dumps.wikimedia.org/jawiki/20220501/jawiki-20220501-pages-articles-multistream-index.txt.bz2
109 let data_path = "bench_data/jawiki-20220501-pages-articles-multistream-index.txt";
110 let string = std::fs::read_to_string(data_path).unwrap_or_default();
111 b.iter(|| test::black_box(UnicodeWidthStr::width(string.as_str())));
112 }
113 #[test]
test_str()114 fn test_str() {
115 use super::UnicodeWidthStr;
116
117 assert_eq!(UnicodeWidthStr::width("ｈｅｌｌｏ"), 10);
118 assert_eq!("ｈｅｌｌｏ".width_cjk(), 10);
119 assert_eq!(UnicodeWidthStr::width("\0\0\0\x01\x01"), 0);
120 assert_eq!("\0\0\0\x01\x01".width_cjk(), 0);
121 assert_eq!(UnicodeWidthStr::width(""), 0);
122 assert_eq!("".width_cjk(), 0);
123 assert_eq!(UnicodeWidthStr::width("\u{2081}\u{2082}\u{2083}\u{2084}"), 4);
124 assert_eq!("\u{2081}\u{2082}\u{2083}\u{2084}".width_cjk(), 8);
125 }
126
127 #[test]
test_emoji()128 fn test_emoji() {
129 // Example from the README.
130 use super::UnicodeWidthStr;
131
132 assert_eq!(UnicodeWidthStr::width("��"), 2); // Woman
133 assert_eq!(UnicodeWidthStr::width("��"), 2); // Microscope
134 assert_eq!(UnicodeWidthStr::width("����"), 4); // Woman scientist
135 }
136
137 #[test]
test_char()138 fn test_char() {
139 use super::UnicodeWidthChar;
140 #[cfg(feature = "no_std")]
141 use core::option::Option::{Some, None};
142
143 assert_eq!(UnicodeWidthChar::width('ｈ'), Some(2));
144 assert_eq!('ｈ'.width_cjk(), Some(2));
145 assert_eq!(UnicodeWidthChar::width('\x00'), Some(0));
146 assert_eq!('\x00'.width_cjk(), Some(0));
147 assert_eq!(UnicodeWidthChar::width('\x01'), None);
148 assert_eq!('\x01'.width_cjk(), None);
149 assert_eq!(UnicodeWidthChar::width('\u{2081}'), Some(1));
150 assert_eq!('\u{2081}'.width_cjk(), Some(2));
151 }
152
153 #[test]
test_char2()154 fn test_char2() {
155 use super::UnicodeWidthChar;
156 #[cfg(feature = "no_std")]
157 use core::option::Option::{Some, None};
158
159 assert_eq!(UnicodeWidthChar::width('\x00'),Some(0));
160 assert_eq!('\x00'.width_cjk(),Some(0));
161
162 assert_eq!(UnicodeWidthChar::width('\x0A'),None);
163 assert_eq!('\x0A'.width_cjk(),None);
164
165 assert_eq!(UnicodeWidthChar::width('w'),Some(1));
166 assert_eq!('w'.width_cjk(),Some(1));
167
168 assert_eq!(UnicodeWidthChar::width('ｈ'),Some(2));
169 assert_eq!('ｈ'.width_cjk(),Some(2));
170
171 assert_eq!(UnicodeWidthChar::width('\u{AD}'),Some(1));
172 assert_eq!('\u{AD}'.width_cjk(),Some(1));
173
174 assert_eq!(UnicodeWidthChar::width('\u{1160}'),Some(0));
175 assert_eq!('\u{1160}'.width_cjk(),Some(0));
176
177 assert_eq!(UnicodeWidthChar::width('\u{a1}'),Some(1));
178 assert_eq!('\u{a1}'.width_cjk(),Some(2));
179
180 assert_eq!(UnicodeWidthChar::width('\u{300}'),Some(0));
181 assert_eq!('\u{300}'.width_cjk(),Some(0));
182 }
183
184 #[test]
unicode_12()185 fn unicode_12() {
186 use super::UnicodeWidthChar;
187 #[cfg(feature = "no_std")]
188 use core::option::Option::{Some, None};
189
190 assert_eq!(UnicodeWidthChar::width('\u{1F971}'), Some(2));
191 }
192