From a6c1c021f31ea688a7f8e8d904f9db3a18f11876 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ren=C3=A9=20=27Necoro=27=20Neumann?= <necoro@necoro.eu>
Date: Sun, 10 May 2020 22:49:18 +0200
Subject: [PATCH] Do not add spaces in front of certain punctuation marks like
 '.', '?' or ')'.

But '(' would still get its space.
---
 html2text.go      | 29 ++++++++++++++++++++++++-----
 html2text_test.go |  8 ++++++++
 2 files changed, 32 insertions(+), 5 deletions(-)

diff --git a/html2text.go b/html2text.go
index 12da0d4..1fe39fb 100644
--- a/html2text.go
+++ b/html2text.go
@@ -72,8 +72,8 @@ func FromHTMLNode(doc *html.Node, o ...Options) (string, error) {
 	}
 
 	ctx := textifyTraverseContext{
-		buf:     bytes.Buffer{},
-		options: options,
+		buf:         bytes.Buffer{},
+		options:     options,
 		citationMap: map[string]int{},
 	}
 	if err := ctx.traverse(doc); err != nil {
@@ -431,6 +431,25 @@ func (ctx *textifyTraverseContext) traverseChildren(node *html.Node) error {
 	return nil
 }
 
+// Tests r for being a character where no space should be inserted in front of.
+func punctNoSpaceBefore(r rune) bool {
+	switch r {
+	case '.', ',', ';', '!', '?', ')', ']', '>':
+		return true
+	default:
+		return false
+	}
+}
+
+// Tests r for being a character where no space should be inserted after.
+func punctNoSpaceAfter(r rune) bool {
+	switch r {
+	case '(', '[', '<':
+		return true
+	default:
+		return false
+	}
+}
 func (ctx *textifyTraverseContext) emit(data string) error {
 	if data == "" {
 		return nil
@@ -441,14 +460,14 @@ func (ctx *textifyTraverseContext) emit(data string) error {
 	)
 	for _, line := range lines {
 		runes := []rune(line)
-		startsWithSpace := unicode.IsSpace(runes[0])
-		if !startsWithSpace && !ctx.endsWithSpace && !strings.HasPrefix(data, ".") {
+		startsWithSpace := unicode.IsSpace(runes[0]) || punctNoSpaceBefore(runes[0])
+		if !startsWithSpace && !ctx.endsWithSpace {
 			if err = ctx.buf.WriteByte(' '); err != nil {
 				return err
 			}
 			ctx.lineLength++
 		}
-		ctx.endsWithSpace = unicode.IsSpace(runes[len(runes)-1])
+		ctx.endsWithSpace = unicode.IsSpace(runes[len(runes)-1]) || punctNoSpaceAfter(runes[len(runes)-1])
 		for _, c := range line {
 			if _, err = ctx.buf.WriteString(string(c)); err != nil {
 				return err
diff --git a/html2text_test.go b/html2text_test.go
index 71558a5..92f2575 100644
--- a/html2text_test.go
+++ b/html2text_test.go
@@ -542,6 +542,14 @@ func TestCitationStyleLinks(t *testing.T) {
 			`<a href="">Link</a>`,
 			"Link",
 		},
+		{
+			`<a href="http://example1.com/">Link1</a><a href="http://example2.com/">Link2</a>`,
+			"Link1 [1] Link2 [2]\n\n[1] http://example1.com/\n[2] http://example2.com/",
+		},
+		{
+			`<a href="http://example1.com/">Link1</a> (<a href="http://example2.com/">Link2</a>)`,
+			"Link1 [1] (Link2 [2])\n\n[1] http://example1.com/\n[2] http://example2.com/",
+		},
 		{
 			`<a href="http://example1.com/">Link1</a>? <a href="http://example2.com/">Link2</a>!`,
 			"Link1 [1]? Link2 [2]!\n\n[1] http://example1.com/\n[2] http://example2.com/",