From c4825a719d3ad79436b3961e52b4fad8428c8bf0 Mon Sep 17 00:00:00 2001
From: Vincent Batoufflet <vincent@batoufflet.info>
Date: Fri, 29 May 2015 13:30:49 +0200
Subject: [PATCH] Add definition lists extension support

---
 block.go      | 82 ++++++++++++++++++++++++++++++++++++++++++++++-----
 block_test.go | 80 +++++++++++++++++++++++++++++++++++++++++++++++++
 html.go       | 27 +++++++++++++----
 markdown.go   |  6 +++-
 4 files changed, 182 insertions(+), 13 deletions(-)

diff --git a/block.go b/block.go
index c9bf2ff..6e61351 100644
--- a/block.go
+++ b/block.go
@@ -166,6 +166,21 @@ func (p *parser) block(out *bytes.Buffer, data []byte) {
 			continue
 		}
 
+		// definition lists:
+		//
+		// Term 1
+		// :   Definition a
+		// :   Definition b
+		//
+		// Term 2
+		// :   Definition c
+		if p.flags&EXTENSION_DEFINITION_LISTS != 0 {
+			if p.dliPrefix(data) > 0 {
+				data = data[p.list(out, data, LIST_TYPE_DEFINITION):]
+				continue
+			}
+		}
+
 		// anything else must look like a normal paragraph
 		// note: this finds underlined headers, too
 		data = data[p.paragraph(out, data):]
@@ -1018,6 +1033,20 @@ func (p *parser) oliPrefix(data []byte) int {
 	return i + 2
 }
 
+// returns definition list item prefix
+func (p *parser) dliPrefix(data []byte) int {
+	i := 0
+
+	// need a : followed by a spaces
+	if data[i] != ':' || data[i+1] != ' ' {
+		return 0
+	}
+	for data[i] == ' ' {
+		i++
+	}
+	return i + 2
+}
+
 // parse ordered or unordered list block
 func (p *parser) list(out *bytes.Buffer, data []byte, flags int) int {
 	i := 0
@@ -1053,7 +1082,19 @@ func (p *parser) listItem(out *bytes.Buffer, data []byte, flags *int) int {
 		i = p.oliPrefix(data)
 	}
 	if i == 0 {
-		return 0
+		i = p.dliPrefix(data)
+		// reset definition term flag
+		if i > 0 {
+			*flags &= ^LIST_TYPE_TERM
+		}
+	}
+	if i == 0 {
+		// if in defnition list, set term flag and continue
+		if *flags&LIST_TYPE_DEFINITION != 0 {
+			*flags |= LIST_TYPE_TERM
+		} else {
+			return 0
+		}
 	}
 
 	// skip leading whitespace on first line
@@ -1063,7 +1104,7 @@ func (p *parser) listItem(out *bytes.Buffer, data []byte, flags *int) int {
 
 	// find the end of the line
 	line := i
-	for data[i-1] != '\n' {
+	for i > 0 && data[i-1] != '\n' {
 		i++
 	}
 
@@ -1107,7 +1148,8 @@ gatherlines:
 		switch {
 		// is this a nested list item?
 		case (p.uliPrefix(chunk) > 0 && !p.isHRule(chunk)) ||
-			p.oliPrefix(chunk) > 0:
+			p.oliPrefix(chunk) > 0 ||
+			p.dliPrefix(chunk) > 0:
 
 			if containsBlankLine {
 				*flags |= LIST_ITEM_CONTAINS_BLOCK
@@ -1138,7 +1180,18 @@ gatherlines:
 		// of this item if it is indented 4 spaces
 		// (regardless of the indentation of the beginning of the item)
 		case containsBlankLine && indent < 4:
-			*flags |= LIST_ITEM_END_OF_LIST
+			if *flags&LIST_TYPE_DEFINITION != 0 && i < len(data)-1 {
+				// is the next item still a part of this list?
+				next := i
+				for data[next] != '\n' {
+					next++
+				}
+				if next < len(data)-2 && data[next+1] != ':' {
+					*flags |= LIST_ITEM_END_OF_LIST
+				}
+			} else {
+				*flags |= LIST_ITEM_END_OF_LIST
+			}
 			break gatherlines
 
 		// a blank line means this should be parsed as a block
@@ -1152,6 +1205,7 @@ gatherlines:
 		if containsBlankLine {
 			containsBlankLine = false
 			raw.WriteByte('\n')
+
 		}
 
 		// add the line into the working buffer without prefix
@@ -1164,8 +1218,8 @@ gatherlines:
 
 	// render the contents of the list item
 	var cooked bytes.Buffer
-	if *flags&LIST_ITEM_CONTAINS_BLOCK != 0 {
-		// intermediate render of block li
+	if *flags&LIST_ITEM_CONTAINS_BLOCK != 0 && *flags&LIST_TYPE_TERM == 0 {
+		// intermediate render of block item, except for definition term
 		if sublist > 0 {
 			p.block(&cooked, rawBytes[:sublist])
 			p.block(&cooked, rawBytes[sublist:])
@@ -1173,7 +1227,7 @@ gatherlines:
 			p.block(&cooked, rawBytes)
 		}
 	} else {
-		// intermediate render of inline li
+		// intermediate render of inline item
 		if sublist > 0 {
 			p.inline(&cooked, rawBytes[:sublist])
 			p.block(&cooked, rawBytes[sublist:])
@@ -1237,6 +1291,13 @@ func (p *parser) paragraph(out *bytes.Buffer, data []byte) int {
 
 		// did we find a blank line marking the end of the paragraph?
 		if n := p.isEmpty(current); n > 0 {
+			// did this blank line followed by a definition list item?
+			if p.flags&EXTENSION_DEFINITION_LISTS != 0 {
+				if i < len(data)-1 && data[i+1] == ':' {
+					return p.list(out, data[prev:], LIST_TYPE_DEFINITION)
+				}
+			}
+
 			p.renderParagraph(out, data[:i])
 			return i + n
 		}
@@ -1295,6 +1356,13 @@ func (p *parser) paragraph(out *bytes.Buffer, data []byte) int {
 			return i
 		}
 
+		// if there's a definition list item, prev line is a definition term
+		if p.flags&EXTENSION_DEFINITION_LISTS != 0 {
+			if p.dliPrefix(current) != 0 {
+				return p.list(out, data[prev:], LIST_TYPE_DEFINITION)
+			}
+		}
+
 		// if there's a list after this, paragraph is over
 		if p.flags&EXTENSION_NO_EMPTY_LINE_BEFORE_BLOCK != 0 {
 			if p.uliPrefix(current) != 0 ||
diff --git a/block_test.go b/block_test.go
index ddcf0ff..4819c6b 100644
--- a/block_test.go
+++ b/block_test.go
@@ -801,6 +801,86 @@ func TestOrderedList(t *testing.T) {
 	doTestsBlock(t, tests, 0)
 }
 
+func TestDefinitionList(t *testing.T) {
+	var tests = []string{
+		"Term 1\n:   Definition a\n",
+		"<dl>\n<dt>Term 1</dt>\n<dd>Definition a</dd>\n</dl>\n",
+
+		"Term 1\n:   Definition a \n",
+		"<dl>\n<dt>Term 1</dt>\n<dd>Definition a</dd>\n</dl>\n",
+
+		"Term 1\n:   Definition a\n:   Definition b\n",
+		"<dl>\n<dt>Term 1</dt>\n<dd>Definition a</dd>\n<dd>Definition b</dd>\n</dl>\n",
+
+		"Term 1\n:   Definition a\n\nTerm 2\n:   Definition b\n",
+		"<dl>\n" +
+			"<dt>Term 1</dt>\n" +
+			"<dd>Definition a</dd>\n" +
+			"<dt>Term 2</dt>\n" +
+			"<dd>Definition b</dd>\n" +
+			"</dl>\n",
+
+		"Term 1\n:   Definition a\n:   Definition b\n\nTerm 2\n:   Definition c\n",
+		"<dl>\n" +
+			"<dt>Term 1</dt>\n" +
+			"<dd>Definition a</dd>\n" +
+			"<dd>Definition b</dd>\n" +
+			"<dt>Term 2</dt>\n" +
+			"<dd>Definition c</dd>\n" +
+			"</dl>\n",
+
+		"Term 1\n\n:   Definition a\n\nTerm 2\n\n:   Definition b\n",
+		"<dl>\n" +
+			"<dt>Term 1</dt>\n" +
+			"<dd><p>Definition a</p></dd>\n" +
+			"<dt>Term 2</dt>\n" +
+			"<dd><p>Definition b</p></dd>\n" +
+			"</dl>\n",
+
+		"Term 1\n\n:   Definition a\n\n:   Definition b\n\nTerm 2\n\n:   Definition c\n",
+		"<dl>\n" +
+			"<dt>Term 1</dt>\n" +
+			"<dd><p>Definition a</p></dd>\n" +
+			"<dd><p>Definition b</p></dd>\n" +
+			"<dt>Term 2</dt>\n" +
+			"<dd><p>Definition c</p></dd>\n" +
+			"</dl>\n",
+
+		"Term 1\n:   Definition a\nNext line\n",
+		"<dl>\n<dt>Term 1</dt>\n<dd>Definition a\nNext line</dd>\n</dl>\n",
+
+		"Term 1\n:   Definition a\n  Next line\n",
+		"<dl>\n<dt>Term 1</dt>\n<dd>Definition a\nNext line</dd>\n</dl>\n",
+
+		"Term 1\n:   Definition a \n  Next line \n",
+		"<dl>\n<dt>Term 1</dt>\n<dd>Definition a\nNext line</dd>\n</dl>\n",
+
+		"Term 1\n:   Definition a\nNext line\n\nTerm 2\n:   Definition b",
+		"<dl>\n" +
+			"<dt>Term 1</dt>\n" +
+			"<dd>Definition a\nNext line</dd>\n" +
+			"<dt>Term 2</dt>\n" +
+			"<dd>Definition b</dd>\n" +
+			"</dl>\n",
+
+		"Term 1\n: Definition a\n",
+		"<dl>\n<dt>Term 1</dt>\n<dd>Definition a</dd>\n</dl>\n",
+
+		"Term 1\n:Definition a\n",
+		"<p>Term 1\n:Definition a</p>\n",
+
+		"Term 1\n\n:   Definition a\n\nTerm 2\n\n:   Definition b\n\nText 1",
+		"<dl>\n" +
+			"<dt>Term 1</dt>\n" +
+			"<dd><p>Definition a</p></dd>\n" +
+			"<dt>Term 2</dt>\n" +
+			"<dd><p>Definition b</p></dd>\n" +
+			"</dl>\n" +
+			"\n<p>Text 1</p>\n",
+	}
+	doTestsBlock(t, tests, EXTENSION_DEFINITION_LISTS)
+}
+
 func TestPreformattedHtml(t *testing.T) {
 	var tests = []string{
 		"<div></div>\n",
diff --git a/html.go b/html.go
index 4c31cac..264aae5 100644
--- a/html.go
+++ b/html.go
@@ -375,7 +375,9 @@ func (options *Html) List(out *bytes.Buffer, text func() bool, flags int) {
 	marker := out.Len()
 	doubleSpace(out)
 
-	if flags&LIST_TYPE_ORDERED != 0 {
+	if flags&LIST_TYPE_DEFINITION != 0 {
+		out.WriteString("<dl>")
+	} else if flags&LIST_TYPE_ORDERED != 0 {
 		out.WriteString("<ol>")
 	} else {
 		out.WriteString("<ul>")
@@ -384,7 +386,9 @@ func (options *Html) List(out *bytes.Buffer, text func() bool, flags int) {
 		out.Truncate(marker)
 		return
 	}
-	if flags&LIST_TYPE_ORDERED != 0 {
+	if flags&LIST_TYPE_DEFINITION != 0 {
+		out.WriteString("</dl>\n")
+	} else if flags&LIST_TYPE_ORDERED != 0 {
 		out.WriteString("</ol>\n")
 	} else {
 		out.WriteString("</ul>\n")
@@ -392,12 +396,25 @@ func (options *Html) List(out *bytes.Buffer, text func() bool, flags int) {
 }
 
 func (options *Html) ListItem(out *bytes.Buffer, text []byte, flags int) {
-	if flags&LIST_ITEM_CONTAINS_BLOCK != 0 || flags&LIST_ITEM_BEGINNING_OF_LIST != 0 {
+	if (flags&LIST_ITEM_CONTAINS_BLOCK != 0 && flags&LIST_TYPE_DEFINITION == 0) ||
+		flags&LIST_ITEM_BEGINNING_OF_LIST != 0 {
 		doubleSpace(out)
 	}
-	out.WriteString("<li>")
+	if flags&LIST_TYPE_TERM != 0 {
+		out.WriteString("<dt>")
+	} else if flags&LIST_TYPE_DEFINITION != 0 {
+		out.WriteString("<dd>")
+	} else {
+		out.WriteString("<li>")
+	}
 	out.Write(text)
-	out.WriteString("</li>\n")
+	if flags&LIST_TYPE_TERM != 0 {
+		out.WriteString("</dt>\n")
+	} else if flags&LIST_TYPE_DEFINITION != 0 {
+		out.WriteString("</dd>\n")
+	} else {
+		out.WriteString("</li>\n")
+	}
 }
 
 func (options *Html) Paragraph(out *bytes.Buffer, text func() bool) {
diff --git a/markdown.go b/markdown.go
index fcb8ac2..3ca79d9 100644
--- a/markdown.go
+++ b/markdown.go
@@ -44,6 +44,7 @@ const (
 	EXTENSION_TITLEBLOCK                             // Titleblock ala pandoc
 	EXTENSION_AUTO_HEADER_IDS                        // Create the header ID from the text
 	EXTENSION_BACKSLASH_LINE_BREAK                   // translate trailing backslashes into line breaks
+	EXTENSION_DEFINITION_LISTS                       // render definition lists
 
 	commonHtmlFlags = 0 |
 		HTML_USE_XHTML |
@@ -59,7 +60,8 @@ const (
 		EXTENSION_STRIKETHROUGH |
 		EXTENSION_SPACE_HEADERS |
 		EXTENSION_HEADER_IDS |
-		EXTENSION_BACKSLASH_LINE_BREAK
+		EXTENSION_BACKSLASH_LINE_BREAK |
+		EXTENSION_DEFINITION_LISTS
 )
 
 // These are the possible flag values for the link renderer.
@@ -76,6 +78,8 @@ const (
 // These are mostly of interest if you are writing a new output format.
 const (
 	LIST_TYPE_ORDERED = 1 << iota
+	LIST_TYPE_DEFINITION
+	LIST_TYPE_TERM
 	LIST_ITEM_CONTAINS_BLOCK
 	LIST_ITEM_BEGINNING_OF_LIST
 	LIST_ITEM_END_OF_LIST