lxml.etree.parse method don't close the file

[Rodrigo F. S.]

in the method parse it open the file but don't close: (looking at link http://codespeak.net/lxml/api/xml.etree.ElementTree-pysrc.html#)

def parse(self, source, parser=None):
    if not hasattr(source, "read"):
        source = open(source, "rb")
    if not parser:
        parser = XMLTreeBuilder()
    while 1:
        data = source.read(32768)
        if not data:
            break
            parser.feed(data)
    self._root = parser.close()
    return self._root

and it should be:

def parse(self, source, parser=None):
    if not hasattr(source, "read"):
        source = open(source, "rb")
    if not parser:
        parser = XMLTreeBuilder()
    while 1:
        data = source.read(32768)
        if not data:
            break
            parser.feed(data)
    source.close() # add this line
    self._root = parser.close()
    return self._root

I got this bug because my program is reading the xml file into a timer, and after some minutes it got the 'too many files open' error into the console

of course, I did a working arround opening the file out of lxml and passing the file handler to the parse method, and doing a file_handler.close() after the parse..

In [3]: ## -- PLEASE PROVIDE THE FOLLOWING INFORMATION: --
In [4]: import sys
In [5]: from lxml import etree
In [6]:
In [7]: print("%-20s: %s" % ('Python', sys.version_info))
Python : (2, 6, 5, 'final', 0)
In [8]: print("%-20s: %s" % ('lxml.etree', etree.LXML_VERSION))
lxml.etree : (2, 2, 4, 0)
In [9]: print("%-20s: %s" % ('libxml used', etree.LIBXML_VERSION))
libxml used : (2, 7, 6)
In [10]: print("%-20s: %s" % ('libxml compiled', etree.LIBXML_COMPILED_VERSION))
libxml compiled : (2, 7, 6)
In [11]: print("%-20s: %s" % ('libxslt used', etree.LIBXSLT_VERSION))
libxslt used : (1, 1, 26)
In [12]: print("%-20s: %s" % ('libxslt compiled', etree.LIBXSLT_COMPILED_VERSION))
libxslt compiled : (1, 1, 26)
In [13]: ## -----

[scoder (scoder)]

Rodrigo F. S., 09.02.2011 13:39:
> in the method parse it open the file but don't close: (looking at link
> http://codespeak.net/lxml/api/xml.etree.ElementTree-pysrc.html#)

Note that this source file is part of ElementTree, not lxml.

> I got this bug because my program is reading the xml file into a timer,
> and after some minutes it got the 'too many files open' error into the
> console
>
> of course, I did a working arround opening the file out of lxml and
> passing the file handler to the parse method, and doing a
> file_handler.close() after the parse..

> lxml.etree : (2, 2, 4, 0)

If you are referring to lxml.etree, this was likely fixed in lxml 2.3.
Please upgrade.

Stefan