From 9fc8c9d8660c52c8390ba80e50556577e179b984 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Vytautas=20=C5=A0altenis?= <vytas@rtfb.lt>
Date: Sat, 25 Jan 2014 21:42:34 +0200
Subject: [PATCH] Fix bug with overzealous autolink processing

When the source Markdown contains an anchor tag with URL as link text
(i.e. <a href=...>http://foo.bar</a>), autolink converts that link text
into another anchor tag, which is nonsense. Detect this situation with
regexp and early exit autolink processing.
---
 inline.go      | 19 +++++++++++++++++++
 inline_test.go | 15 +++++++++++++++
 2 files changed, 34 insertions(+)

diff --git a/inline.go b/inline.go
index 0348dbf..d83404e 100644
--- a/inline.go
+++ b/inline.go
@@ -15,9 +15,14 @@ package blackfriday
 
 import (
 	"bytes"
+	"regexp"
 	"strconv"
 )
 
+var (
+	anchorRe = regexp.MustCompile(`^(<a\shref="` + urlRe + `"(\stitle="[^"<>]+")?\s?>` + urlRe + `<\/a>)`)
+)
+
 // Functions to parse text within a block
 // Each function returns the number of chars taken care of
 // data is the complete block being rendered
@@ -618,6 +623,20 @@ func autoLink(p *parser, out *bytes.Buffer, data []byte, offset int) int {
 		return 0
 	}
 
+	// Now a more expensive check to see if we're not inside an anchor element
+	anchorStart := offset
+	offsetFromAnchor := 0
+	for anchorStart > 0 && data[anchorStart] != '<' {
+		anchorStart--
+		offsetFromAnchor++
+	}
+
+	anchorStr := anchorRe.Find(data[anchorStart:])
+	if anchorStr != nil {
+		out.Write(anchorStr[offsetFromAnchor:])
+		return len(anchorStr) - offsetFromAnchor
+	}
+
 	// scan backward for a word boundary
 	rewind := 0
 	for offset-rewind > 0 && rewind <= 7 && isletter(data[offset-rewind-1]) {
diff --git a/inline_test.go b/inline_test.go
index 57bef81..7cb3cdf 100644
--- a/inline_test.go
+++ b/inline_test.go
@@ -674,6 +674,21 @@ func TestAutoLink(t *testing.T) {
 		"even a > can be escaped <http://new.com?q=\\>&etc>\n",
 		"<p>even a &gt; can be escaped <a href=\"http://new.com?q=&gt;&amp;etc\">" +
 			"http://new.com?q=&gt;&amp;etc</a></p>\n",
+
+		"<a href=\"http://fancy.com\">http://fancy.com</a>\n",
+		"<p><a href=\"http://fancy.com\">http://fancy.com</a></p>\n",
+
+		"<a href=\"http://fancy.com\">This is a link</a>\n",
+		"<p><a href=\"http://fancy.com\">This is a link</a></p>\n",
+
+		"<a href=\"http://www.fancy.com/A_B.pdf\">http://www.fancy.com/A_B.pdf</a>\n",
+		"<p><a href=\"http://www.fancy.com/A_B.pdf\">http://www.fancy.com/A_B.pdf</a></p>\n",
+
+		"(<a href=\"http://www.fancy.com/A_B\">http://www.fancy.com/A_B</a> (\n",
+		"<p>(<a href=\"http://www.fancy.com/A_B\">http://www.fancy.com/A_B</a> (</p>\n",
+
+		"(<a href=\"http://www.fancy.com/A_B\">http://www.fancy.com/A_B</a> (part two: <a href=\"http://www.fancy.com/A_B\">http://www.fancy.com/A_B</a>)).\n",
+		"<p>(<a href=\"http://www.fancy.com/A_B\">http://www.fancy.com/A_B</a> (part two: <a href=\"http://www.fancy.com/A_B\">http://www.fancy.com/A_B</a>)).</p>\n",
 	}
 	doTestsInline(t, tests)
 }