From a4eaafab5a003298c8235f7502fb684da2220ca8 Mon Sep 17 00:00:00 2001
From: Aldo Cortesi <aldo@nullcube.com>
Date: Sun, 6 Feb 2011 16:56:13 +1300
Subject: [PATCH] Further fine-tuning for pretty_xmlish.

---
 libmproxy/utils.py | 16 ++++++++--------
 test/test_utils.py |  5 ++++-
 2 files changed, 12 insertions(+), 9 deletions(-)

diff --git a/libmproxy/utils.py b/libmproxy/utils.py
index 2a8786761..e29910c5a 100644
--- a/libmproxy/utils.py
+++ b/libmproxy/utils.py
@@ -52,13 +52,10 @@ TAG = r"""
         (?P<close>\s*\/)?
         (?P<name>\w+)
         (
-                [a-zA-Z0-9_#:=().%\/]+
-            |
-                "[^\"]*"['\"]*
-            |
-                '[^']*'['\"]*
-            | 
-                \s+
+            [^'"\t >]+ |
+            "[^\"]*"['\"]* |
+            '[^']*'['\"]* |
+            \s+
         )*
         (?P<selfcont>\s*\/\s*)?
         \s*>
@@ -67,7 +64,7 @@ UNI = set(["br", "hr", "img", "input", "area", "link"])
 INDENT = " "*4
 def pretty_xmlish(s):
     """
-        This is a robust, general pretty-printer for XML-ish data. 
+        A robust pretty-printer for XML-ish data.
         Returns a list of lines.
     """
     data, offset, indent, prev = [], 0, 0, None
@@ -87,6 +84,9 @@ def pretty_xmlish(s):
         if not any([i.group("close"), i.group("selfcont"), name in UNI]):
             indent += 1
         prev = name
+    trail = s[offset:]
+    if trail.strip():
+        data.append(s[offset:])
     return data
 
 
diff --git a/test/test_utils.py b/test/test_utils.py
index 1ec4f2f50..5cf81e2ef 100644
--- a/test/test_utils.py
+++ b/test/test_utils.py
@@ -242,6 +242,9 @@ class upretty_xmlish(libpry.AutoTree):
         assert f(r"<foo a='b\"'>")
         assert f(r'<a b=(a.b) href="foo">')
         assert f('<td width=25%>')
+        assert f('<form name="search" action="/search.php" method="get" accept-charset="utf-8" class="search">')
+        assert f('<img src="gif" width="125" height="16" alt=&quot;&quot; />')
+
 
     def test_all(self):
         def isbalanced(ret):
@@ -278,7 +281,7 @@ class upretty_xmlish(libpry.AutoTree):
         isbalanced(ret)
 
         s = "gobbledygook"
-        print utils.pretty_xmlish(s)
+        assert utils.pretty_xmlish(s) == ["gobbledygook"]