From 86ae2ed0a644f124475a4aff3b34e229f5b7ec8f Mon Sep 17 00:00:00 2001
From: Leonard Richardson <leonard.richardson@canonical.com>
Date: Sat, 19 Feb 2011 21:16:20 -0500
Subject: Set up an lxml parser that only parses XML, though it's not very
 functional yet.

---
 tests/test_lxml.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

(limited to 'tests/test_lxml.py')

diff --git a/tests/test_lxml.py b/tests/test_lxml.py
index 7e15dcf..7597552 100644
--- a/tests/test_lxml.py
+++ b/tests/test_lxml.py
@@ -131,7 +131,7 @@ class TestLXMLBuilder(SoupTest):
         # But, both lxml and html5lib do their best to parse the
         # contents of a <textarea> as HTML.
         text = '<textarea>Junk like <b> tags and <&<&amp;</textarea>'
-        soup = BeautifulSoup(text)
+        soup = self.soup(text)
         self.assertEquals(len(soup.textarea.contents), 2)
         self.assertEquals(soup.textarea.contents[0], u"Junk like ")
         self.assertEquals(soup.textarea.contents[1].name, 'b')
@@ -141,7 +141,7 @@ class TestLXMLBuilder(SoupTest):
         # The contents of a <script> tag are treated as a literal string,
         # even if that string contains HTML.
         javascript = 'if (i < 2) { alert("<b>foo</b>"); }'
-        soup = BeautifulSoup('<script>%s</script>' % javascript)
+        soup = self.soup('<script>%s</script>' % javascript)
         self.assertEquals(soup.script.string, javascript)
 
     def test_naked_ampersands(self):
@@ -300,7 +300,7 @@ class TestLXMLBuilder(SoupTest):
     def test_entities_converted_on_the_way_out(self):
         text = "<p>&lt;&lt;sacr&eacute;&#32;bleu!&gt;&gt;</p>"
         expected = u"&lt;&lt;sacr\N{LATIN SMALL LETTER E WITH ACUTE} bleu!&gt;&gt;".encode("utf-8")
-        soup = BeautifulSoup(text)
+        soup = self.soup(text)
         str = soup.p.string
         #self.assertEquals(str.encode("utf-8"), expected)
 
-- 
cgit v1.2.3


From 9a936b48fe05666780662c76d5df3b3de7b48074 Mon Sep 17 00:00:00 2001
From: Leonard Richardson <leonard.richardson@canonical.com>
Date: Sat, 19 Feb 2011 21:20:54 -0500
Subject: Oh, good, html5lib correctly handles literals in <textarea> tags.

---
 tests/test_lxml.py | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

(limited to 'tests/test_lxml.py')

diff --git a/tests/test_lxml.py b/tests/test_lxml.py
index 7597552..8670806 100644
--- a/tests/test_lxml.py
+++ b/tests/test_lxml.py
@@ -126,10 +126,9 @@ class TestLXMLBuilder(SoupTest):
 
     def test_literal_in_textarea(self):
         # Anything inside a <textarea> is supposed to be treated as
-        # the literal value of the field, (XXX citation needed).
-        #
-        # But, both lxml and html5lib do their best to parse the
-        # contents of a <textarea> as HTML.
+        # the literal value of the field, (XXX citation
+        # needed). html5lib does this correctly. But, lxml does its
+        # best to parse the contents of a <textarea> as HTML.
         text = '<textarea>Junk like <b> tags and <&<&amp;</textarea>'
         soup = self.soup(text)
         self.assertEquals(len(soup.textarea.contents), 2)
-- 
cgit v1.2.3