Commit 0d6d2ffc authored by Florent Xicluna's avatar Florent Xicluna

Issue #2892: preserve iterparse events in case of SyntaxError

parent 20cb029c
...@@ -739,6 +739,7 @@ def iterparse(): ...@@ -739,6 +739,7 @@ def iterparse():
... except ET.ParseError, v: ... except ET.ParseError, v:
... print v ... print v
junk after document element: line 1, column 12 junk after document element: line 1, column 12
end document
""" """
def writefile(): def writefile():
......
...@@ -1210,6 +1210,7 @@ class _IterParseIterator(object): ...@@ -1210,6 +1210,7 @@ class _IterParseIterator(object):
self._close_file = close_source self._close_file = close_source
self._events = [] self._events = []
self._index = 0 self._index = 0
self._error = None
self.root = self._root = None self.root = self._root = None
self._parser = parser self._parser = parser
# wire up the parser for event reporting # wire up the parser for event reporting
...@@ -1255,24 +1256,31 @@ class _IterParseIterator(object): ...@@ -1255,24 +1256,31 @@ class _IterParseIterator(object):
while 1: while 1:
try: try:
item = self._events[self._index] item = self._events[self._index]
self._index += 1
return item
except IndexError: except IndexError:
if self._parser is None: pass
self.root = self._root if self._error:
if self._close_file: e = self._error
self._file.close() self._error = None
raise StopIteration raise e
# load event buffer if self._parser is None:
del self._events[:] self.root = self._root
self._index = 0 if self._close_file:
data = self._file.read(16384) self._file.close()
if data: raise StopIteration
# load event buffer
del self._events[:]
self._index = 0
data = self._file.read(16384)
if data:
try:
self._parser.feed(data) self._parser.feed(data)
else: except SyntaxError as exc:
self._root = self._parser.close() self._error = exc
self._parser = None
else: else:
self._index = self._index + 1 self._root = self._parser.close()
return item self._parser = None
def __iter__(self): def __iter__(self):
return self return self
......
...@@ -2970,6 +2970,7 @@ init_elementtree(void) ...@@ -2970,6 +2970,7 @@ init_elementtree(void)
" self._file = file\n" " self._file = file\n"
" self._events = []\n" " self._events = []\n"
" self._index = 0\n" " self._index = 0\n"
" self._error = None\n"
" self.root = self._root = None\n" " self.root = self._root = None\n"
" b = cElementTree.TreeBuilder()\n" " b = cElementTree.TreeBuilder()\n"
" self._parser = cElementTree.XMLParser(b)\n" " self._parser = cElementTree.XMLParser(b)\n"
...@@ -2978,24 +2979,31 @@ init_elementtree(void) ...@@ -2978,24 +2979,31 @@ init_elementtree(void)
" while 1:\n" " while 1:\n"
" try:\n" " try:\n"
" item = self._events[self._index]\n" " item = self._events[self._index]\n"
" self._index += 1\n"
" return item\n"
" except IndexError:\n" " except IndexError:\n"
" if self._parser is None:\n" " pass\n"
" self.root = self._root\n" " if self._error:\n"
" if self._close_file:\n" " e = self._error\n"
" self._file.close()\n" " self._error = None\n"
" raise StopIteration\n" " raise e\n"
" # load event buffer\n" " if self._parser is None:\n"
" del self._events[:]\n" " self.root = self._root\n"
" self._index = 0\n" " if self._close_file:\n"
" data = self._file.read(16384)\n" " self._file.close()\n"
" if data:\n" " raise StopIteration\n"
" # load event buffer\n"
" del self._events[:]\n"
" self._index = 0\n"
" data = self._file.read(16384)\n"
" if data:\n"
" try:\n"
" self._parser.feed(data)\n" " self._parser.feed(data)\n"
" else:\n" " except SyntaxError as exc:\n"
" self._root = self._parser.close()\n" " self._error = exc\n"
" self._parser = None\n"
" else:\n" " else:\n"
" self._index = self._index + 1\n" " self._root = self._parser.close()\n"
" return item\n" " self._parser = None\n"
" def __iter__(self):\n" " def __iter__(self):\n"
" return self\n" " return self\n"
"cElementTree.iterparse = iterparse\n" "cElementTree.iterparse = iterparse\n"
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment