Deal more appropriately with bare ampersands and pointy brackets; this

module has to deal with "class" HTML-as-deployed as well as XHTML, so we
cannot be as strict as XHTML allows.

This closes SF bug #453059, but uses a different fix than suggested in
the bug comments.

diff --git a/Lib/HTMLParser.py b/Lib/HTMLParser.py
index 39a5d8262ff1bb5cbe4e5f6fd6499e655daef2e7..954ce2647f570a9c9a3cd13e29cc817cf59bf536 100644 (file)
--- a/Lib/HTMLParser.py
+++ b/Lib/HTMLParser.py
@@ -15,7 +15,8 @@ import string
 
 interesting_normal = re.compile('[&<]')
 interesting_cdata = re.compile(r'<(/|\Z)')
-incomplete = re.compile('&([a-zA-Z][-.a-zA-Z0-9]*|#[0-9]*)?')
+incomplete = re.compile('&([a-zA-Z][-.a-zA-Z0-9]*'
+                        '|#([0-9]*|[xX][0-9a-fA-F]*))?')
 
 entityref = re.compile('&([a-zA-Z][-.a-zA-Z0-9]*)[^a-zA-Z0-9]')
 charref = re.compile('&#(?:[0-9]+|[xX][0-9a-fA-F]+)[^0-9a-fA-F]')
@@ -185,11 +186,8 @@ class HTMLParser:
                 elif declopen.match(rawdata, i): # <!
                     k = self.parse_declaration(i)
                 else:
-                    if i < n-1:
-                        raise HTMLParseError(
-                            "invalid '<' construct: %s" % `rawdata[i:i+2]`,
-                            self.getpos())
-                    k = -1
+                    self.handle_data("<")
+                    k = i + 1
                 if k < 0:
                     if end:
                         raise HTMLParseError("EOF in middle of construct",
@@ -203,7 +201,7 @@ class HTMLParser:
                     self.handle_charref(name)
                     k = match.end()
                     if rawdata[k-1] != ';':
-                        k = k-1
+                        k = k - 1
                     i = self.updatepos(i, k)
                     continue
                 match = entityref.match(rawdata, i)
@@ -212,17 +210,19 @@ class HTMLParser:
                     self.handle_entityref(name)
                     k = match.end()
                     if rawdata[k-1] != ';':
-                        k = k-1
+                        k = k - 1
                     i = self.updatepos(i, k)
                     continue
-                if incomplete.match(rawdata, i):
-                    if end:
+                match = incomplete.match(rawdata, i)
+                if match:
+                    rest = rawdata[i:]
+                    if end and rest != "&" and match.group() == rest:
                         raise HTMLParseError(
                             "EOF in middle of entity or char ref",
                             self.getpos())
                     return -1 # incomplete
-                raise HTMLParseError("'&' not part of entity or char ref",
-                                     self.getpos())
+                self.handle_data("&")
+                i = self.updatepos(i, i + 1)
             else:
                 assert 0, "interesting.search() lied"
         # end while

diff --git a/Lib/test/test_htmlparser.py b/Lib/test/test_htmlparser.py
index e0e212c27a5453d9913ce10225c05227ee78ebaa..bb6e0b0fe528dee83653f88b8344cb68633e4eb9 100755 (executable)
--- a/Lib/test/test_htmlparser.py
+++ b/Lib/test/test_htmlparser.py
@@ -1,6 +1,7 @@
 """Tests for HTMLParser.py."""
 
 import HTMLParser
+import pprint
 import sys
 import test_support
 import unittest
@@ -83,9 +84,10 @@ class TestCaseBase(unittest.TestCase):
         for c in self.epilogue:
             parser.feed(c)
         parser.close()
-        self.assert_(parser.get_events() ==
-                     self.initial_events + events + self.final_events,
-                     parser.get_events())
+        events = parser.get_events()
+        self.assertEqual(events,
+                         self.initial_events + events + self.final_events,
+                         "got events:\n" + pprint.pformat(events))
 
     def _run_check_extra(self, source, events):
         self._run_check(source, events, EventCollectorExtra)
@@ -137,6 +139,18 @@ text
     ("data", "\n"),
     ])
 
+    def test_doctype_decl(self):
+        inside = """\
+DOCTYPE html [
+  <!ELEMENT html - O EMPTY>
+  <!ATTLIST html
+      version CDATA #IMPLIED '4.0'>
+  <!-- comment -->
+]"""
+        self._run_check("<!%s>" % inside, [
+            ("decl", inside),
+            ])
+
     def test_bad_nesting(self):
         # Strangely, this *is* supposed to test that overlapping
         # elements are allowed.  HTMLParser is more geared toward
@@ -148,6 +162,16 @@ text
             ("endtag", "b"),
             ])
 
+    def test_bare_ampersands(self):
+        self._run_check("this text & contains & ampersands &", [
+            ("data", "this text & contains & ampersands &"),
+            ])
+
+    def test_bare_pointy_brackets(self):
+        self._run_check("this < text > contains < bare>pointy< brackets", [
+            ("data", "this < text > contains < bare>pointy< brackets"),
+            ])
+
     def test_attr_syntax(self):
         output = [
           ("starttag", "a", [("b", "v"), ("c", "v"), ("d", "v"), ("e", None)])
@@ -199,16 +223,12 @@ text
         self._run_check(["<a b='>'", ">"], output)
 
     def test_starttag_junk_chars(self):
-        self._parse_error("<")
-        self._parse_error("<>")
         self._parse_error("</>")
         self._parse_error("</$>")
         self._parse_error("</")
         self._parse_error("</a")
         self._parse_error("<a<a>")
         self._parse_error("</a<a>")
-        self._parse_error("<$")
-        self._parse_error("<$>")
         self._parse_error("<!")
         self._parse_error("<a $>")
         self._parse_error("<a")