Make XMLSyntaxError have normal SyntaxError metadata

3 files changed, 37 insertions, 6 deletions

diff --git a/src/lxml/parser.pxi b/src/lxml/parser.pxi
index 46eb99cc..639e5acf 100644
--- a/src/lxml/parser.pxi
+++ b/src/lxml/parser.pxi
@@ -15,10 +15,20 @@ class ParseError(LxmlSyntaxError):
 
     For compatibility with ElementTree 1.3 and later.
     """
-    def __init__(self, message, code, line, column):
+    def __init__(self, message, code, line, column, filename = None):
         super(_ParseError, self).__init__(message)
-        self.position = (line, column)
+        self.lineno, self.offset = (line, column - 1)
         self.code = code
+        self.filename = filename
+
+    @property
+    def position(self):
+        return self.lineno, self.offset + 1
+
+    @position.setter
+    def position(self, new_pos):
+        self.lineno, column = new_pos
+        self.offset = column - 1
 
 cdef object _ParseError = ParseError
 
@@ -630,9 +640,10 @@ cdef int _raiseParseError(xmlparser.xmlParserCtxt* ctxt, filename,
         column = ctxt.lastError.int2
         if ctxt.lastError.line > 0:
             message = u"line %d: %s" % (line, message)
-        raise XMLSyntaxError(message, code, line, column)
+        raise XMLSyntaxError(message, code, line, column, filename)
     else:
-        raise XMLSyntaxError(None, xmlerror.XML_ERR_INTERNAL_ERROR, 0, 0)
+        raise XMLSyntaxError(None, xmlerror.XML_ERR_INTERNAL_ERROR, 0, 0,
+                             filename)
 
 cdef xmlDoc* _handleParseResult(_ParserContext context,
                                 xmlparser.xmlParserCtxt* c_ctxt,
@@ -1341,7 +1352,8 @@ cdef class _FeedParser(_BaseParser):
         """
         if not self._feed_parser_running:
             raise XMLSyntaxError(u"no element found",
-                                 xmlerror.XML_ERR_INTERNAL_ERROR, 0, 0)
+                                 xmlerror.XML_ERR_INTERNAL_ERROR, 0, 0,
+                                 self._filename)
 
         context = self._getPushParserContext()
         pctxt = context._c_ctxt
diff --git a/src/lxml/tests/test_errors.py b/src/lxml/tests/test_errors.py
index 068e5e3f..784847d5 100644
--- a/src/lxml/tests/test_errors.py
+++ b/src/lxml/tests/test_errors.py
@@ -49,6 +49,24 @@ class ErrorTestCase(HelperTestCase):
         finally:
             sys.settrace(trace_func)
 
+    def test_xmlsyntaxerror_has_info(self):
+        broken_xml_name = 'test_broken.xml'
+        broken_xml_path = os.path.join(this_dir, broken_xml_name)
+        fail_msg = 'test_broken.xml should raise an etree.XMLSyntaxError'
+        try:
+            etree.parse(broken_xml_path)
+        except etree.XMLSyntaxError as e:
+            # invariant
+            self.assertEqual(e.position, (e.lineno, e.offset + 1), 'position and lineno/offset out of sync')
+            # SyntaxError info derived from file & contents
+            self.assertTrue(e.filename.endswith(broken_xml_name), 'filename must be preserved')
+            self.assertEqual(e.lineno, 1)
+            self.assertEqual(e.offset, 10)
+        except Exception as e:
+            self.fail('{}, not {}'.format(fail_msg, type(e)))
+        else:
+            self.fail('test_broken.xml should raise an etree.XMLSyntaxError')
+
 
 def test_suite():
     suite = unittest.TestSuite()
diff --git a/src/lxml/xmlerror.pxi b/src/lxml/xmlerror.pxi
index 12f84c9b..31154685 100644
--- a/src/lxml/xmlerror.pxi
+++ b/src/lxml/xmlerror.pxi
@@ -231,12 +231,13 @@ cdef class _BaseErrorLog:
             message = default_message
         line = self._first_error.line
         column = self._first_error.column
+        filename = self._first_error.filename
         if line > 0:
             if column > 0:
                 message = u"%s, line %d, column %d" % (message, line, column)
             else:
                 message = u"%s, line %d" % (message, line)
-        return exctype(message, code, line, column)
+        return exctype(message, code, line, column, filename)
 
     @cython.final
     cdef _buildExceptionMessage(self, default_message):