From d2d9dc0fc8210104ad2fbcbbda2df964671b3fb8 Mon Sep 17 00:00:00 2001
From: Niklas Fasching <niklas.fasching@gmail.com>
Date: Sun, 2 Dec 2018 18:24:51 +0100
Subject: [PATCH] Add auto links: Try to render unmarked urls as links

---
 org/inline.go             | 34 ++++++++++++++++++++++++++++++++--
 org/org.go                |  4 +++-
 org/testdata/example.html |  1 +
 org/testdata/example.org  |  1 +
 4 files changed, 37 insertions(+), 3 deletions(-)

diff --git a/org/inline.go b/org/inline.go
index 2b78cd4..62bf0d3 100644
--- a/org/inline.go
+++ b/org/inline.go
@@ -24,8 +24,12 @@ type RegularLink struct {
 	Protocol    string
 	Description []Node
 	URL         string
+	AutoLink    bool
 }
 
+var validURLCharacters = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-._~:/?#[]@!$&'()*+,;="
+var autolinkProtocols = regexp.MustCompile(`(https?|ftp|file)`)
+
 var redundantSpaces = regexp.MustCompile("[ \t]+")
 var subScriptSuperScriptRegexp = regexp.MustCompile(`([_^])\{(.*?)\}`)
 var footnoteRegexp = regexp.MustCompile(`\[fn:([\w-]+?)(:(.*?))?\]`)
@@ -33,7 +37,7 @@ var footnoteRegexp = regexp.MustCompile(`\[fn:([\w-]+?)(:(.*?))?\]`)
 func (d *Document) parseInline(input string) (nodes []Node) {
 	previous, current := 0, 0
 	for current < len(input) {
-		consumed, node := 0, (Node)(nil)
+		rewind, consumed, node := 0, 0, (Node)(nil)
 		switch input[current] {
 		case '^':
 			consumed, node = d.parseSubOrSuperScript(input, current)
@@ -45,6 +49,9 @@ func (d *Document) parseInline(input string) (nodes []Node) {
 			consumed, node = d.parseRegularLinkOrFootnoteReference(input, current)
 		case '\\':
 			consumed, node = d.parseExplicitLineBreak(input, current)
+		case ':':
+			rewind, consumed, node = d.parseAutoLink(input, current)
+			current -= rewind
 		}
 		if consumed != 0 {
 			if current > previous {
@@ -125,6 +132,29 @@ func (d *Document) parseFootnoteReference(input string, start int) (int, Node) {
 	return 0, nil
 }
 
+func (d *Document) parseAutoLink(input string, start int) (int, int, Node) {
+	if len(input[start:]) < 3 || input[start+1] != '/' || input[start+2] != '/' {
+		return 0, 0, nil
+	}
+	protocolStart, protocol := start-1, ""
+	for ; protocolStart > 0 && unicode.IsLetter(rune(input[protocolStart])); protocolStart-- {
+	}
+	if m := autolinkProtocols.FindStringSubmatch(input[protocolStart:start]); m != nil {
+		protocol = m[1]
+	} else {
+		return 0, 0, nil
+	}
+	end := start
+	for ; end < len(input) && strings.ContainsRune(validURLCharacters, rune(input[end])); end++ {
+	}
+	path := input[start:end]
+	if path == "://" {
+		return 0, 0, nil
+	}
+	link := RegularLink{protocol, []Node{Text{protocol + path}}, protocol + path, true}
+	return len(protocol), len(path + protocol), link
+}
+
 func (d *Document) parseRegularLink(input string, start int) (int, Node) {
 	if len(input[start:]) == 0 || input[start+1] != '[' {
 		return 0, nil
@@ -147,7 +177,7 @@ func (d *Document) parseRegularLink(input string, start int) (int, Node) {
 	if len(parts) == 2 {
 		protocol = parts[0]
 	}
-	return consumed, RegularLink{protocol, description, link}
+	return consumed, RegularLink{protocol, description, link, false}
 }
 
 func (d *Document) parseEmphasis(input string, start int) (int, Node) {
diff --git a/org/org.go b/org/org.go
index 866881c..203981c 100644
--- a/org/org.go
+++ b/org/org.go
@@ -246,7 +246,9 @@ func (w *OrgWriter) writeRegularLink(l RegularLink) {
 	descriptionWriter := w.emptyClone()
 	descriptionWriter.writeNodes(l.Description...)
 	description := descriptionWriter.String()
-	if l.URL != description {
+	if l.AutoLink {
+		w.WriteString(l.URL)
+	} else if l.URL != description {
 		w.WriteString(fmt.Sprintf("[[%s][%s]]", l.URL, description))
 	} else {
 		w.WriteString(fmt.Sprintf("[[%s]]", l.URL))
diff --git a/org/testdata/example.html b/org/testdata/example.html
index 9786b14..4d6af74 100644
--- a/org/testdata/example.html
+++ b/org/testdata/example.html
@@ -58,6 +58,7 @@
 <li>regular link <a href="https://example.com">https://example.com</a> link without description</li>
 <li>regular link <a href="https://example.com">example.com</a> link with description</li>
 <li>regular link to a file (image) <img src="my-img.png" alt="file:my-img.png" title="file:my-img.png" /></li>
+<li>auto link, i.e. not inside <code class="verbatim">\[[square brackets]\]</code> <a href="https://www.example.com">https://www.example.com</a></li>
 </ol>
 </li>
 </ul>
diff --git a/org/testdata/example.org b/org/testdata/example.org
index 789cdac..258964e 100644
--- a/org/testdata/example.org
+++ b/org/testdata/example.org
@@ -50,6 +50,7 @@ this one is cheating a little as tags are ALWAYS printed right aligned to a give
   1. regular link [[https://example.com]] link without description
   2. regular link [[https://example.com][example.com]] link with description
   3. regular link to a file (image) [[file:my-img.png]]
+  4. auto link, i.e. not inside =\[[square brackets]\]= https://www.example.com
 ** Footnotes
 - normal footnote reference [fn:1]
 - further references to the same footnote should not [fn:1] render duplicates in the footnote list