From c9e6ad1c35b5fbe4c06c09142b4886d13555803a Mon Sep 17 00:00:00 2001
From: Makoto Kato <m_kato@ga2.so-net.ne.jp>
Date: Wed, 1 Nov 2023 05:46:50 +0000
Subject: [PATCH] Bug 628825 - Don't break line by Break After.
 r=TYLin,jfkthame

Legacy segmenter breaks line by some Break After properties. But UAX#14 rule
doesn't break Break After (BA) at force.

So I would like to remove some legacy rules that isn't ASCII white space from
`nsLineBreaker`.

Differential Revision: https://phabricator.services.mozilla.com/D192174
---
 dom/base/nsLineBreaker.cpp                    | 12 ++---
 dom/base/nsLineBreaker.h                      | 10 ++++
 .../word-break-break-all-ethiopic.html.ini    |  2 -
 .../line-breaking/line-breaking-028.html      | 46 +++++++++++++++++++
 4 files changed, 62 insertions(+), 8 deletions(-)
 delete mode 100644 testing/web-platform/meta/css/css-text/word-break/word-break-break-all-ethiopic.html.ini
 create mode 100644 testing/web-platform/tests/css/css-text/line-breaking/line-breaking-028.html

diff --git a/dom/base/nsLineBreaker.cpp b/dom/base/nsLineBreaker.cpp
index b44c7343d752..ffc224fc278c 100644
--- a/dom/base/nsLineBreaker.cpp
+++ b/dom/base/nsLineBreaker.cpp
@@ -252,7 +252,7 @@ nsresult nsLineBreaker::AppendText(nsAtom* aHyphenationLanguage,
     NS_ASSERTION(!mAfterBreakableSpace && !mBreakHere,
                  "These should not be set");
 
-    while (offset < aLength && !IsSpace(aText[offset])) {
+    while (offset < aLength && !IsSegmentSpace(aText[offset])) {
       mCurrentWord.AppendElement(aText[offset]);
       if (!mCurrentWordMightBeBreakable &&
           !IsNonBreakableChar<char16_t>(aText[offset], mLegacyBehavior)) {
@@ -306,7 +306,7 @@ nsresult nsLineBreaker::AppendText(nsAtom* aHyphenationLanguage,
     offset = aLength;
     while (offset > start) {
       --offset;
-      if (IsSpace(aText[offset])) {
+      if (IsSegmentSpace(aText[offset])) {
         break;
       }
     }
@@ -323,7 +323,7 @@ nsresult nsLineBreaker::AppendText(nsAtom* aHyphenationLanguage,
 
   for (;;) {
     char16_t ch = aText[offset];
-    bool isSpace = IsSpace(ch);
+    bool isSpace = IsSegmentSpace(ch);
     bool isBreakableSpace = isSpace && !(aFlags & BREAK_SUPPRESS_INSIDE);
 
     if (aSink && !noBreaksNeeded) {
@@ -442,7 +442,7 @@ nsresult nsLineBreaker::AppendText(nsAtom* aHyphenationLanguage,
     NS_ASSERTION(!mAfterBreakableSpace && !mBreakHere,
                  "These should not be set");
 
-    while (offset < aLength && !IsSpace(aText[offset])) {
+    while (offset < aLength && !IsSegmentSpace(aText[offset])) {
       mCurrentWord.AppendElement(aText[offset]);
       if (!mCurrentWordMightBeBreakable &&
           !IsNonBreakableChar<uint8_t>(aText[offset], mLegacyBehavior)) {
@@ -486,7 +486,7 @@ nsresult nsLineBreaker::AppendText(nsAtom* aHyphenationLanguage,
     offset = aLength;
     while (offset > start) {
       --offset;
-      if (IsSpace(aText[offset])) {
+      if (IsSegmentSpace(aText[offset])) {
         break;
       }
     }
@@ -496,7 +496,7 @@ nsresult nsLineBreaker::AppendText(nsAtom* aHyphenationLanguage,
 
   for (;;) {
     uint8_t ch = aText[offset];
-    bool isSpace = IsSpace(ch);
+    bool isSpace = IsSegmentSpace(ch);
     bool isBreakableSpace = isSpace && !(aFlags & BREAK_SUPPRESS_INSIDE);
 
     if (aSink) {
diff --git a/dom/base/nsLineBreaker.h b/dom/base/nsLineBreaker.h
index 6701cc756931..6e409cea6dda 100644
--- a/dom/base/nsLineBreaker.h
+++ b/dom/base/nsLineBreaker.h
@@ -242,6 +242,16 @@ class nsLineBreaker {
                              const char16_t* aTextStart,
                              const char16_t* aTextLimit, uint8_t* aBreakState);
 
+  inline constexpr bool IsSegmentSpace(char16_t u) const {
+    if (mLegacyBehavior) {
+      return nsLineBreaker::IsSpace(u);
+    }
+
+    return u == 0x0020 ||  // SPACE u
+           u == 0x0009 ||  // CHARACTER TABULATION
+           u == 0x000D;    // CARRIAGE RETURN
+  }
+
   AutoTArray<char16_t, 100> mCurrentWord;
   // All the items that contribute to mCurrentWord
   AutoTArray<TextItem, 2> mTextItems;
diff --git a/testing/web-platform/meta/css/css-text/word-break/word-break-break-all-ethiopic.html.ini b/testing/web-platform/meta/css/css-text/word-break/word-break-break-all-ethiopic.html.ini
deleted file mode 100644
index 4cd08ee4c345..000000000000
--- a/testing/web-platform/meta/css/css-text/word-break/word-break-break-all-ethiopic.html.ini
+++ /dev/null
@@ -1,2 +0,0 @@
-[word-break-break-all-ethiopic.html]
-  expected: FAIL
diff --git a/testing/web-platform/tests/css/css-text/line-breaking/line-breaking-028.html b/testing/web-platform/tests/css/css-text/line-breaking/line-breaking-028.html
new file mode 100644
index 000000000000..c82d4762a597
--- /dev/null
+++ b/testing/web-platform/tests/css/css-text/line-breaking/line-breaking-028.html
@@ -0,0 +1,46 @@
+<!doctype html>
+<html>
+<meta charset="utf-8">
+<title>CSS Text — line breaking around Break After and Exclamation</title>
+<meta name=assert content="When ‘white-space’ allows wrapping, line breaking behavior defined for BA and EX line-breaking classes in [UAX14] must be honored.">
+<link rel=help href="https://www.w3.org/TR/css-text-3/#line-breaking">
+<link rel=help href="https://bugzilla.mozilla.org/show_bug.cgi?id=628825">
+<link rel=author title="Makoto Kato" href="mailto:m_kato@ga2.so-net.ne.jp">
+<script src="/resources/testharness.js"></script>
+<script src="/resources/testharnessreport.js"></script>
+<style>
+.test > div {
+  font-family: monospace;
+  font-size: 25px;
+  width: 1ch;
+  line-height: 30px;
+}
+</style>
+<body>
+
+<div class="test">
+  <div id="nonbreakable1">X&#x1361;!</div>
+  <div id="nonbreakable2">X&#x1680;!</div>
+  <div id="nonbreakable3">X&#x2009;!</div>
+  <div id="nonbreakable4">X&#x205F;!</div>
+</div>
+
+<script>
+test(function() {
+  assert_true(document.getElementById('nonbreakable1').offsetHeight <= 35);
+}, "U+0x1361 (BA) and U+0x0021 (EX)");
+test(function() {
+  assert_true(document.getElementById('nonbreakable2').offsetHeight <= 35);
+}, "U+0x1680 (BA) and U+0x0021 (EX)");
+test(function() {
+  assert_true(document.getElementById('nonbreakable3').offsetHeight <= 35);
+}, "U+0x2009 (BA) and U+0x0021 (EX)");
+test(function() {
+  assert_true(document.getElementById('nonbreakable4').offsetHeight <= 35);
+}, "U+0x205F (BA) and U+0x0021 (EX)");
+</script>
+
+<div id='log'></div>
+
+</body>
+</html>