Improve test cases for Intl.Segmenter

Fix containing() w/ n point to surrogate tail.

Bug: v8:6891
Change-Id: I4d7e4083ba409eecaefaf2ec5c67b7226bd7c97c
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2348208
Commit-Queue: Frank Tang <ftang@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/master@{#69363}

src/objects/js-segments.cc

@@ -77,6 +77,9 @@ MaybeHandle<Object> JSSegments::Containing(Isolate* isolate,
     return isolate->factory()->undefined_value();
   }
 
+  // n may point to the surrogate tail- adjust it back to the lead.
+  n = segments->unicode_string().raw()->getChar32Start(n);
+
   icu::BreakIterator* break_iterator = segments->icu_break_iterator().raw();
   // 8. Let startIndex be ! FindBoundary(segmenter, string, n, before).
   int32_t start_index =

test/intl/segmenter/grapheme-containing.js

+// Copyright 2020 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.
+
+// Flags: --harmony-intl-segmenter
+
+const segmenter = new Intl.Segmenter();
+const waving_hand_light_skin_tone = "\uD83D\uDC4B\uD83C\uDFFB";
+const man_light_skin_tone_red_hair =
+    "\uD83D\uDC68\uD83C\uDFFB\u200D\uD83E\uDDB0";
+//                          1                                  2
+//             034566     89012 3      4     56    89          01
+const input = "ซิ่ง Ame\u0301lie台北\uD800\uDCB0 " + waving_hand_light_skin_tone +
+//  2
+//  2345678
+    man_light_skin_tone_red_hair;
+const segments = segmenter.segment(input);
+// Test less than 0
+assertEquals(undefined, segments.containing(-1));
+
+// Test longer than input
+assertEquals(undefined, segments.containing(input.length));
+assertEquals(undefined, segments.containing(input.length + 1));
+
+// Test modifier in Thai
+assertEquals({segment: "ซิ่", index: 0, input}, segments.containing(0));
+assertEquals({segment: "ซิ่", index: 0, input}, segments.containing(1));
+assertEquals({segment: "ซิ่", index: 0, input}, segments.containing(2));
+
+// Test basic Thai
+assertEquals({segment: "ง", index: 3, input}, segments.containing(3));
+
+// Test SPACE
+assertEquals({segment: " ", index: 4, input}, segments.containing(4));
+// Test ASCII
+assertEquals({segment: "A", index: 5, input}, segments.containing(5));
+assertEquals({segment: "m", index: 6, input}, segments.containing(6));
+
+// Test ASCII with modifier
+assertEquals({segment: "e\u0301", index: 7, input}, segments.containing(7));
+assertEquals({segment: "e\u0301", index: 7, input}, segments.containing(8));
+
+// Test ASCII
+assertEquals({segment: "l", index: 9, input}, segments.containing(9));
+assertEquals({segment: "i", index: 10, input}, segments.containing(10));
+assertEquals({segment: "e", index: 11, input}, segments.containing(11));
+
+// Test Han
+assertEquals({segment: "台", index: 12, input}, segments.containing(12));
+assertEquals({segment: "北", index: 13, input}, segments.containing(13));
+
+// Test Surrogate pairs
+assertEquals({segment: "𐂰", index: 14, input}, segments.containing(14));
+assertEquals({segment: "𐂰", index: 14, input}, segments.containing(15));
+
+// Test SPACE
+assertEquals({segment: " ", index: 16, input}, segments.containing(16));
+
+// Test Emoji modifier: U+1F44B U+1F3FB
+const emoji1 = {segment: waving_hand_light_skin_tone, index: 17, input};
+assertEquals(emoji1, segments.containing(17));
+assertEquals(emoji1, segments.containing(18));
+assertEquals(emoji1, segments.containing(19));
+assertEquals(emoji1, segments.containing(20));
+
+// Test Emoji modifiers sequence: U+1F468 U+1F3FB U+200D U+1F9B0
+const emoji2 = {segment: man_light_skin_tone_red_hair, index: 21, input};
+assertEquals(emoji2, segments.containing(21));
+assertEquals(emoji2, segments.containing(22));
+assertEquals(emoji2, segments.containing(23));
+assertEquals(emoji2, segments.containing(24));
+assertEquals(emoji2, segments.containing(25));
+assertEquals(emoji2, segments.containing(26));
+assertEquals(emoji2, segments.containing(27));

test/intl/segmenter/grapheme-nested-next.js

+// Copyright 2020 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.
+
+// Flags: --harmony-intl-segmenter
+
+// Test to ensure the nested calling of the next method won't caused
+// confusion to each other.
+const segmenter = new Intl.Segmenter();
+const input = "ABCD";
+const segments = segmenter.segment(input);
+let result = [];
+for (let v1 of segments) {
+  for (let v2 of segments) {
+    result.push(v1.segment);
+    result.push(v2.segment);
+  }
+  result.push(":");
+}
+assertEquals("AAABACAD:BABBBCBD:CACBCCCD:DADBDCDD:", result.join(""));

test/intl/segmenter/grapheme-next-after-containing.js

+// Copyright 2020 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.
+
+// Flags: --harmony-intl-segmenter
+
+// Test to ensure the calling of containing() won't impact the calling of
+// the next() method.
+const segmenter = new Intl.Segmenter();
+const man_light_skin_tone_red_hair =
+    "\uD83D\uDC68\uD83C\uDFFB\u200D\uD83E\uDDB0";
+const input = "ABCD" + man_light_skin_tone_red_hair;
+const segments = segmenter.segment(input);
+for (let i = 0; i < input.length; i++) {
+  let idx = i < 4 ? i : 4;
+  let expectation = i < 4 ? input[i] : man_light_skin_tone_red_hair;
+  assertEquals({segment: expectation, index: idx, input},
+      segments.containing(i));
+  let result = [];
+  for (let v of segments) {
+    result.push(v.segment);
+    result.push(":");
+    // Ensure the value n passing into segments.containing(n) will not impact
+    // the result of next().
+    assertEquals({segment: expectation, index: idx, input},
+        segments.containing(i));
+  }
+  assertEquals("A:B:C:D:" + man_light_skin_tone_red_hair + ":",
+      result.join(""));
+}

test/intl/segmenter/word-containing.js

+// Copyright 2020 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.
+
+// Flags: --harmony-intl-segmenter
+
+const segmenter = new Intl.Segmenter(undefined, {granularity: 'word'});
+const waving_hand_light_skin_tone = "\uD83D\uDC4B\uD83C\uDFFB";
+const man_light_skin_tone_red_hair =
+    "\uD83D\uDC68\uD83C\uDFFB\u200D\uD83E\uDDB0";
+//                          1                                  2
+//             034566     89012 3      4     56    89          01
+const input = "ซิ่ง Ame\u0301lie台北\uD800\uDCB0 " + waving_hand_light_skin_tone +
+//  2
+//  2345678
+    man_light_skin_tone_red_hair;
+const segments = segmenter.segment(input);
+// Test less than 0
+assertEquals(undefined, segments.containing(-1));
+
+// Test longer than input
+assertEquals(undefined, segments.containing(input.length));
+assertEquals(undefined, segments.containing(input.length + 1));
+
+// Test modifier in Thai
+const thai = "ซิ่ง";
+let pos = 0;
+for (let i = pos; i < pos + thai.length; i++) {
+  assertEquals({segment: "ซิ่ง", index: pos, input, isWordLike: true},
+      segments.containing(i));
+}
+pos += thai.length;
+
+// Test SPACE
+assertEquals({segment: " ", index: pos, input, isWordLike: false},
+    segments.containing(pos));
+pos++;
+
+// Test Latin with modifier
+const latin_with_modifier = "Ame\u0301lie";
+for (let i = pos; i < pos + latin_with_modifier.length; i++) {
+  assertEquals(
+      {segment: latin_with_modifier, index: pos, input, isWordLike: true},
+      segments.containing(i));
+}
+pos += latin_with_modifier.length;
+
+// Test Han
+const taipei = "台北";
+for (let i = pos; i < pos + taipei.length; i++) {
+  assertEquals({segment: taipei, index: pos, input, isWordLike: true},
+      segments.containing(i));
+}
+pos += taipei.length;
+
+// Test Surrogate pair
+const surrogate = "\uD800\uDCB0";
+for (let i = pos; i < pos + surrogate.length; i++) {
+  assertEquals({segment: surrogate, index: pos, input, isWordLike: true},
+      segments.containing(14));
+}
+pos += surrogate.length;
+
+// Test SPACE
+assertEquals({segment: " ", index: pos, input, isWordLike: false},
+    segments.containing(pos));
+pos++;
+
+// Test Emoji modifier: U+1F44B U+1F3FB
+for (let i = pos; i < pos + waving_hand_light_skin_tone.length; i++) {
+  assertEquals({segment: waving_hand_light_skin_tone, index: pos, input,
+    isWordLike: false},
+    segments.containing(i));
+}
+pos += waving_hand_light_skin_tone.length;
+
+// Test Emoji modifiers sequence: U+1F468 U+1F3FB U+200D U+1F9B0
+for (let i = pos; i < pos + man_light_skin_tone_red_hair.length; i++) {
+  assertEquals({segment: man_light_skin_tone_red_hair, index: pos, input,
+    isWordLike: false},
+    segments.containing(i));
+}
+pos += man_light_skin_tone_red_hair.length;