64da43ef36
This is a reland of c9c3ec4c14
Original change's description:
> [intl] Clean up intl_segmenter flag
>
> Intl.Segmenter shipped in m87 and launched.
>
> Bug: v8:11225
> Change-Id: I4213e261e1aea717c1281f19785a8c29ff1bbd8b
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2570461
> Commit-Queue: Frank Tang <ftang@chromium.org>
> Reviewed-by: Shu-yu Guo <syg@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#71653}
Bug: v8:11225, v8:11240
Change-Id: Ibded9038671862d90206d328f8a12db51c40e63c
Cq-Include-Trybots: luci.v8.try:v8_linux64_gc_stress_custom_snapshot_dbg_ng,v8_linux_arm64_gc_stress_dbg_ng,v8_linux_gc_stress_dbg_ng,v8_mac64_gc_stress_dbg_ng
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2579043
Commit-Queue: Frank Tang <ftang@chromium.org>
Reviewed-by: Shu-yu Guo <syg@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71691}
43 lines
2.7 KiB
JavaScript
43 lines
2.7 KiB
JavaScript
// Copyright 2018 the V8 project authors. All rights reserved.
|
||
// Use of this source code is governed by a BSD-style license that can be
|
||
// found in the LICENSE file.
|
||
|
||
const seg = new Intl.Segmenter([], {granularity: "word"})
|
||
for (const text of [
|
||
"Hello world!", // English
|
||
" Hello world! ", // English with space before/after
|
||
" Hello world? Foo bar!", // English
|
||
"Jedovatou mambu objevila žena v zahrádkářské kolonii.", // Czech
|
||
"Việt Nam: Nhất thể hóa sẽ khác Trung Quốc?", // Vietnamese
|
||
"Σοβαρές ενστάσεις Κομισιόν για τον προϋπολογισμό της Ιταλίας", // Greek
|
||
"Решение Индии о покупке российских С-400 расценили как вызов США", // Russian
|
||
"הרופא שהציל נשים והנערה ששועבדה ע", // Hebrew,
|
||
"ترامب للملك سلمان: أنا جاد للغاية.. عليك دفع المزيد", // Arabic
|
||
"भारत की एस 400 मिसाइल के मुकाबले पाक की थाड, जानें कौन कितना ताकतवर", // Hindi
|
||
"ரெட் அலர்ட் எச்சரிக்கை; புதுச்சேரியில் நாளை அரசு விடுமுறை!", // Tamil
|
||
"'ఉత్తర్వులు అందే వరకు ఓటర్ల తుది జాబితాను వెబ్సైట్లో పెట్టవద్దు'", // Telugu
|
||
"台北》抹黑柯P失敗？朱學恒酸：姚文智氣pupu嗆大老闆", // Chinese
|
||
"วัดไทรตีระฆังเบาลงช่วงเข้าพรรษา เจ้าอาวาสเผยคนร้องเรียนรับผลกรรมแล้ว", // Thai
|
||
"九州北部の一部が暴風域に入りました(日直予報士 2018年10月06日) - 日本気象協会 tenki.jp", // Japanese
|
||
"법원 “다스 지분 처분권·수익권 모두 MB가 보유”", // Korean
|
||
]) {
|
||
const segments = seg.segment(text);
|
||
let results = [];
|
||
var pos = -1;
|
||
for (let s of segments) {
|
||
assertEquals(["segment", "index", "input", "isWordLike"], Object.keys(s));
|
||
assertEquals(typeof s.isWordLike, "boolean");
|
||
assertEquals(typeof s.index, "number");
|
||
assertEquals(typeof s.segment, "string");
|
||
assertEquals(typeof s.input, "string");
|
||
assertEquals(text, s.input);
|
||
assertEquals(text.substring(s.index, s.index + s.segment.length),
|
||
s.segment);
|
||
assertTrue(pos < s.index);
|
||
pos = s.index;
|
||
results.push(s.segment);
|
||
}
|
||
assertTrue(pos < text.length);
|
||
assertEquals(text, results.join(""));
|
||
}
|