ExpatError: XML or text declaration not at start of entity

[pafcu]

My server is running Wordpress 3.2.1.
The following code:

from wordpress_xmlrpc import Client
wp = Client('<redacted>', '<redacted>', '<redacted>')

results in the exception:

Traceback (most recent call last):
  File "test.py", line 2, in <module>
    wp = Client('<redacted>', '<redacted>', '<redacted>')
  File "/home/pafcu/projects/rsswp/wordpress_xmlrpc/base.py", line 24, in __init__
    self.supported_methods = self.server.mt.supportedMethods()
  File "/usr/lib/python2.7/xmlrpclib.py", line 1224, in __call__
    return self.__send(self.__name, args)
  File "/usr/lib/python2.7/xmlrpclib.py", line 1570, in __request
    verbose=self.__verbose
  File "/usr/lib/python2.7/xmlrpclib.py", line 1264, in request
    return self.single_request(host, handler, request_body, verbose)
  File "/usr/lib/python2.7/xmlrpclib.py", line 1297, in single_request
    return self.parse_response(response)
  File "/usr/lib/python2.7/xmlrpclib.py", line 1462, in parse_response
    p.feed(data)
  File "/usr/lib/python2.7/xmlrpclib.py", line 557, in feed
    self._parser.Parse(data, 0)
xml.parsers.expat.ExpatError: XML or text declaration not at start of entity: line 2, column 0

It seems like the library has trouble with the whitespace at the beginning of the reply from the server.

[maxcutler]

Are you running stock WordPress, or are you using one or more plug-ins that modify XML-RPC functionality?

This exception is actually bubbling up from Python's xmlrpclib module, which handles the XML-RPC protocol spec implementation. A response is not valid XML if it starts with whitespace, so that's why it is failing.

If you can provide more information about your WordPress setup, maybe I can help you find a solution.

[pafcu]

I have deactivated all plugins so it should now act like a stock 3.2.1 install. Is it possible that it is in fact a WordPress bug? Have you successfully tested with version 3.2.1? I'm not sure, but could it also be that xmlrpclib has previously been more lenient with XML errors? I'm running Python 2.7.1.

[maxcutler]

Based on some quick searching, it looks like 2.7 might be at least part of the problem. I primarily use 2.6 or 2.5 on most of my machines, but I can set up 2.7 somewhere and investigate. This is a rough week for me, so it'll probably have to wait until next week. If you want to do some digging and submit a pull request, I'd be happy to integrate it.

[pafcu]

It seems to work OK on Python 2.6.5.

On one hand, this seems to be a bug in WordPress (adding the empty newline in the beginning). On the other hand, python-wordpress-rpc should be able to talk to a WordPress server, even when the server is not standards compliant.

However, it seems like there really isn't any nice solution (except for making a copy of the python xmlrpclib module and removing the offending newline there before passing the data on to the XML parser) so maybe it's just better to wait for a fix to WordPress instead.

[maxcutler]

I just ran a quick test using Python 2.7.1 against a WordPress 3.2.1 instance and didn't encounter the reported exception.

@pafcu, Would you mind setting up a new empty WordPress blog somewhere, pointing your script at that, and see if you still get the exception? If not, that would lead me to believe there's something amiss with this particular blog/site. Any debugging information (raw XML response, more info on your execution and server environments) you can provide would be helpful in diagnosing what is going on.

[pafcu]

I tried it now with a fresh install and funnily enough it works. I have deleted all the plugins in the old install but still experience the same problems there. A diff of the directories show identical contents (except for database details). So seems like it's some Wordpress setting messing things up. Anyway, since it seems like a WordPress bug that only occurs in certain situations it seems unnecessary to handle it in the python-wordpress-xmlrpc library.

[maxcutler]

Alright, I'll close this issue then. I'd still be very curious to figure out what exactly is going on in WordPress to cause such an issue, as that's something that really needs to be patched. But I understand if you don't have time to devote to debugging it and want to just move on.

[evandavey]

I'm getting the same error with Wordpress 3.4.1 and Python 2.7.1.

I am able to use XMLRPC with other programs so suggest there is an issue with python-wordpress-xmlrpc somewhere.

I tried using this https://github.com/maxcutler/wordpress-xmlrpc-dewhitespacer and it attempts to progress further but the plugin seems to degrade performance to the point that requests timeout. The plugin has the same impact on my other software which has no problems parsing the original feed.

UPDATE: I managed to fix this by finding and removing a newline after ?> in my theme. Still, a better solution would be nice.

[jandetlefsen]

yes new lines are the problem in this. this is mostly due to the fact that Wordpress still uses closing PHP tags, which are redundant

[fgm]

FWIW: similar issue with 2.7 talking to a non-Wordpress (Drupal 7) server. The responses also start with a "\n".

[kiwiheretic]

I had the same problem. I managed to capture the xml file on the return path and put the xml file into this pastebin -> http://goo.gl/FGcBmx. Didn't appear to have any leading or trailing newlines.

[StevensUneven]

This is in no way a fix but just a workaround until it gets fixed but what I did to resolve this issue is open the xmlrpclib.py which is usually located in "C:\Python27\Lib" at line 558 I changed self._parser.Parse(data, 0 ) to self._parser.Parse(data.strip(), 0 )

[Strompf]

This is in no way a fix but just a workaround until it gets fixed but what I did to resolve this issue is open the xmlrpclib.py which is usually located in "C:\Python27\Lib" at line 558 I changed self._parser.Parse(data, 0 ) to self._parser.Parse(data.strip(), 0 )

That worked for me! Thank you!