Backport of:

------------------------------------------------------------------------
r26444 | srichter | 2004-07-13 11:01:30 -0400 (Tue, 13 Jul 2004) | 5 
lines

While I was working on functional tests, I found the scenario when this
path is chosen, which is when you get the source (for editing for
example) and macro expansion is turned on. It is perfectly fine to pass
an empty dict at this point, since the template itself is not evaluated.

------------------------------------------------------------------------
------------------------------------------------------------------------
r26445 | srichter | 2004-07-13 11:01:55 -0400 (Tue, 13 Jul 2004) | 2 
lines

Converted XXX to TODO.

------------------------------------------------------------------------

pagetemplate.py

@@ -158,8 +158,9 @@ def read(self):
             if not self.expand:
                 return self._text
             try:
-                # XXX not clear how this ever gets called, but the
-                # first arg to pt_render() needs to change if it ever does.
+                # This gets called, if macro expansion is turned on.
+                # Note that an empty dictionary is fine for the context at
+                # this point, since we are not evaluating the template. 
                 return self.pt_render({}, source=1)
             except:
                 return ('%s\n Macro expansion failed\n %s\n-->\n%s' %

tests/test_ptfile.py

@@ -0,0 +1,126 @@
+"""Tests of PageTemplateFile."""
+
+import os
+import tempfile
+import unittest
+
+from zope.pagetemplate.pagetemplatefile import PageTemplateFile
+
+
+class TypeSniffingTestCase(unittest.TestCase):
+
+    TEMPFILENAME = tempfile.mktemp()
+
+    def tearDown(self):
+        if os.path.exists(self.TEMPFILENAME):
+            os.unlink(self.TEMPFILENAME)
+
+    def check_content_type(self, text, expected_type):
+        f = open(self.TEMPFILENAME, "wb")
+        f.write(text)
+        f.close()
+        pt = PageTemplateFile(self.TEMPFILENAME)
+        pt.read()
+        self.assertEqual(pt.content_type, expected_type)
+
+    def test_sniffer_xml_ascii(self):
+        self.check_content_type(
+            "<?xml version='1.0' encoding='ascii'?><doc/>",
+            "text/xml")
+        self.check_content_type(
+            "<?xml\tversion='1.0' encoding='ascii'?><doc/>",
+            "text/xml")
+
+    def test_sniffer_xml_utf8(self):
+        # w/out byte order mark
+        self.check_content_type(
+            "<?xml version='1.0' encoding='utf-8'?><doc/>",
+            "text/xml")
+        self.check_content_type(
+            "<?xml\tversion='1.0' encoding='utf-8'?><doc/>",
+            "text/xml")
+        # with byte order mark
+        self.check_content_type(
+            "\xef\xbb\xbf<?xml version='1.0' encoding='utf-8'?><doc/>",
+            "text/xml")
+        self.check_content_type(
+            "\xef\xbb\xbf<?xml\tversion='1.0' encoding='utf-8'?><doc/>",
+            "text/xml")
+
+    def test_sniffer_xml_utf16_be(self):
+        # w/out byte order mark
+        self.check_content_type(
+            "\0<\0?\0x\0m\0l\0 \0v\0e\0r\0s\0i\0o\0n\0=\0'\01\0.\0000\0'"
+            "\0 \0e\0n\0c\0o\0d\0i\0n\0g\0=\0'\0u\0t\0f\0-\08\0'\0?\0>"
+            "\0<\0d\0o\0c\0/\0>",
+            "text/xml")
+        self.check_content_type(
+            "\0<\0?\0x\0m\0l\0\t\0v\0e\0r\0s\0i\0o\0n\0=\0'\01\0.\0000\0'"
+            "\0 \0e\0n\0c\0o\0d\0i\0n\0g\0=\0'\0u\0t\0f\0-\08\0'\0?\0>"
+            "\0<\0d\0o\0c\0/\0>",
+            "text/xml")
+        # with byte order mark
+        self.check_content_type(
+            "\xfe\xff"
+            "\0<\0?\0x\0m\0l\0 \0v\0e\0r\0s\0i\0o\0n\0=\0'\01\0.\0000\0'"
+            "\0 \0e\0n\0c\0o\0d\0i\0n\0g\0=\0'\0u\0t\0f\0-\08\0'\0?\0>"
+            "\0<\0d\0o\0c\0/\0>",
+            "text/xml")
+        self.check_content_type(
+            "\xfe\xff"
+            "\0<\0?\0x\0m\0l\0\t\0v\0e\0r\0s\0i\0o\0n\0=\0'\01\0.\0000\0'"
+            "\0 \0e\0n\0c\0o\0d\0i\0n\0g\0=\0'\0u\0t\0f\0-\08\0'\0?\0>"
+            "\0<\0d\0o\0c\0/\0>",
+            "text/xml")
+
+    def test_sniffer_xml_utf16_le(self):
+        # w/out byte order mark
+        self.check_content_type(
+            "<\0?\0x\0m\0l\0 \0v\0e\0r\0s\0i\0o\0n\0=\0'\01\0.\0000\0'\0"
+            " \0e\0n\0c\0o\0d\0i\0n\0g\0=\0'\0u\0t\0f\0-\08\0'\0?\0>\0"
+            "<\0d\0o\0c\0/\0>\n",
+            "text/xml")
+        self.check_content_type(
+            "<\0?\0x\0m\0l\0\t\0v\0e\0r\0s\0i\0o\0n\0=\0'\01\0.\0000\0'\0"
+            " \0e\0n\0c\0o\0d\0i\0n\0g\0=\0'\0u\0t\0f\0-\08\0'\0?\0>\0"
+            "<\0d\0o\0c\0/\0>\0",
+            "text/xml")
+        # with byte order mark
+        self.check_content_type(
+            "\xff\xfe"
+            "<\0?\0x\0m\0l\0 \0v\0e\0r\0s\0i\0o\0n\0=\0'\01\0.\0000\0'\0"
+            " \0e\0n\0c\0o\0d\0i\0n\0g\0=\0'\0u\0t\0f\0-\08\0'\0?\0>\0"
+            "<\0d\0o\0c\0/\0>\0",
+            "text/xml")
+        self.check_content_type(
+            "\xff\xfe"
+            "<\0?\0x\0m\0l\0\t\0v\0e\0r\0s\0i\0o\0n\0=\0'\01\0.\0000\0'\0"
+            " \0e\0n\0c\0o\0d\0i\0n\0g\0=\0'\0u\0t\0f\0-\08\0'\0?\0>\0"
+            "<\0d\0o\0c\0/\0>\0",
+            "text/xml")
+
+    HTML_PUBLIC_ID = "-//W3C//DTD HTML 4.01 Transitional//EN"
+    HTML_SYSTEM_ID = "http://www.w3.org/TR/html4/loose.dtd"
+
+    def test_sniffer_html_ascii(self):
+        self.check_content_type(
+            "<!DOCTYPE html [ SYSTEM '%s' ]><html></html>"
+            % self.HTML_SYSTEM_ID,
+            "text/html")
+        self.check_content_type(
+            "<html><head><title>sample document</title></head></html>",
+            "text/html")
+
+    # TODO: This reflects a case that simply isn't handled by the
+    # sniffer; there are many, but it gets it right more often than
+    # before.
+    def donttest_sniffer_xml_simple(self):
+        self.check_content_type("<doc><element/></doc>",
+                                "text/xml")
+
+
+def test_suite():
+    return unittest.makeSuite(TypeSniffingTestCase)
+
+if __name__ == "__main__":
+    unittest.main(defaultTest="test_suite")