Fix #534

extension/cjk_test.go

@@ -266,4 +266,50 @@ func TestEastAsianLineBreaks(t *testing.T) {
 		t,
 	)
 
+	// Tests with EastAsianLineBreaksStyleSimple (for Chinese)
+	markdown = goldmark.New(
+		goldmark.WithRendererOptions(
+			html.WithXHTML(),
+			html.WithUnsafe(),
+		),
+		goldmark.WithExtensions(
+			NewCJK(WithEastAsianLineBreaks(EastAsianLineBreaksSimple)),
+		),
+	)
+	no = 20
+	testutil.DoTestCase(
+		markdown,
+		testutil.MarkdownTestCase{
+			No:          no,
+			Description: "中文汉字之间的软回车应被忽略",
+			Markdown:    "被分开成两行\n写的一句话。",
+			Expected:    "<p>被分开成两行写的一句话。</p>",
+		},
+		t,
+	)
+	no = 21
+	testutil.DoTestCase(
+		markdown,
+		testutil.MarkdownTestCase{
+			No:          no,
+			Description: "中文常用标点符号之间的软回车应被忽略（换行符前）",
+			Markdown:    "一，\n二。\n三？\n四！\n五：\n六；\n七。\n八【\n九】\n十『\n九』\n八‘\n七’\n六“\n五”\n四……\n三、\n二",
+			Expected:    "<p>一，二。三？四！五：六；七。八【九】十『九』八‘七’六“五”四……三、二</p>",
+		},
+		t,
+	)
+
+	// 注意：按照中文标点符号规范，中文标点符号与其后的英文字母之间不应该有空格。
+	// 但是目前的实现不支持这种判断，所以还是会有空格。
+	no = 22
+	testutil.DoTestCase(
+		markdown,
+		testutil.MarkdownTestCase{
+			No:          no,
+			Description: "中文与英文混合",
+			Markdown:    "一，\na\n二。\nb",
+			Expected:    "<p>一，\na\n二。\nb</p>",
+		},
+		t,
+	)
 }

util/util_cjk.go

@@ -26,6 +26,12 @@ var cjkSymbolsAndPunctuation = &unicode.RangeTable{
 	},
 }
 
+var unicodeGeneralPunctuation = &unicode.RangeTable{
+	R16: []unicode.Range16{
+		{0x2000, 0x206F, 1},
+	},
+}
+
 var hiragana = &unicode.RangeTable{
 	R16: []unicode.Range16{
 		{0x3040, 0x309F, 1},
@@ -177,14 +183,19 @@ var cjkUnifiedIdeographsExtensionG = &unicode.RangeTable{
 	},
 }
 
-// IsEastAsianWideRune returns trhe if the given rune is an east asian wide character, otherwise false.
+// IsEastAsianWideRune returns true if the given rune is an east asian wide character, otherwise false.
+//
+// Note for halfwidthAndFullwidthForms: some characters in this range are not wide characters,
+// but soft line breaks are ignored for them too.
 func IsEastAsianWideRune(r rune) bool {
 	return unicode.Is(unicode.Hiragana, r) ||
 		unicode.Is(unicode.Katakana, r) ||
 		unicode.Is(unicode.Han, r) ||
 		unicode.Is(unicode.Lm, r) ||
 		unicode.Is(unicode.Hangul, r) ||
-		unicode.Is(cjkSymbolsAndPunctuation, r)
+		unicode.Is(cjkSymbolsAndPunctuation, r) ||
+		unicode.Is(halfwidthAndFullwidthForms, r) ||
+		unicode.Is(unicodeGeneralPunctuation, r)
 }
 
 // IsSpaceDiscardingUnicodeRune returns true if the given rune is space-discarding unicode character, otherwise false.