From b05bfd6e8b379025359eff8f18987d1c262ecda9 Mon Sep 17 00:00:00 2001
From: Alex Kocharin <alex@kocharin.ru>
Date: Tue, 10 Mar 2015 18:23:37 +0300
Subject: [PATCH] Punycode domain names only in known schemas

In urls like `skype:XXX` it prevents XXX to be parsed as domain
name and encoded with punycode.
---
 lib/index.js                            | 32 +++++++++++++++++++------
 test/fixtures/markdown-it/normalize.txt | 16 +++++++++++++
 2 files changed, 41 insertions(+), 7 deletions(-)

diff --git a/lib/index.js b/lib/index.js
index 6a0310e..7e1d237 100644
--- a/lib/index.js
+++ b/lib/index.js
@@ -21,7 +21,7 @@ var config = {
 };
 
 
-var BAD_PROTOCOLS   = [ 'vbscript', 'javascript', 'file' ];
+var BAD_PROTOCOLS    = [ 'vbscript', 'javascript', 'file' ];
 
 function validateLink(url) {
   // url should be normalized at this point, and existing entities are decoded
@@ -34,13 +34,23 @@ function validateLink(url) {
   return true;
 }
 
+var RECODE_HOSTNAME_FOR = [ 'http:', 'https:', 'mailto:' ];
+
 function normalizeLink(url) {
   var parsed = mdurl.parse(url, true);
 
   if (parsed.hostname) {
-    try {
-      parsed.hostname = punycode.toASCII(parsed.hostname);
-    } catch(er) {}
+    // Encode hostnames in urls like:
+    // `http://host/`, `https://host/`, `mailto:user@host`, `//host/`
+    //
+    // We don't encode unknown schemas, because it's likely that we encode
+    // something we shouldn't (e.g. `skype:name` treated as `skype:host`)
+    //
+    if (!parsed.protocol || RECODE_HOSTNAME_FOR.indexOf(parsed.protocol) >= 0) {
+      try {
+        parsed.hostname = punycode.toASCII(parsed.hostname);
+      } catch(er) {}
+    }
   }
 
   return mdurl.encode(mdurl.format(parsed));
@@ -50,9 +60,17 @@ function normalizeLinkText(url) {
   var parsed = mdurl.parse(url, true);
 
   if (parsed.hostname) {
-    try {
-      parsed.hostname = punycode.toUnicode(parsed.hostname);
-    } catch(er) {}
+    // Encode hostnames in urls like:
+    // `http://host/`, `https://host/`, `mailto:user@host`, `//host/`
+    //
+    // We don't encode unknown schemas, because it's likely that we encode
+    // something we shouldn't (e.g. `skype:name` treated as `skype:host`)
+    //
+    if (!parsed.protocol || RECODE_HOSTNAME_FOR.indexOf(parsed.protocol) >= 0) {
+      try {
+        parsed.hostname = punycode.toUnicode(parsed.hostname);
+      } catch(er) {}
+    }
   }
 
   return mdurl.decode(mdurl.format(parsed));
diff --git a/test/fixtures/markdown-it/normalize.txt b/test/fixtures/markdown-it/normalize.txt
index 71e1ad8..c59165d 100644
--- a/test/fixtures/markdown-it/normalize.txt
+++ b/test/fixtures/markdown-it/normalize.txt
@@ -43,6 +43,22 @@ Invalid punycode (non-ascii):
 <p><a href="http://xn--xn---emd.com/">http://xn–γ.com/</a></p>
 .
 
+Two slashes should start a domain:
+
+.
+[](//☃.net/)
+.
+<p><a href="//xn--n3h.net/"></a></p>
+.
+
+Don't encode domains in unknown schemas:
+
+.
+[](skype:γγγ)
+.
+<p><a href="skype:%CE%B3%CE%B3%CE%B3"></a></p>
+.
+
 Should auto-add protocol to autolinks:
 
 .