urllib2 httplib.BadStatusLine exception while opening a page on an Oracle HTTP Server

ak aktests at gmail.com
Wed Jan 21 20:43:27 CET 2009


On Jan 20, 1:14 am, Steven D'Aprano
<ste... at REMOVE.THIS.cybersource.com.au> wrote:
> On Mon, 19 Jan 2009 13:00:44 -0800, ak wrote:
> > Hi everyone,
>
> > I have a problem with urllib2 on this particular url, hosted on an
> > Oracle HTTP Server
>
> >http://www.orange.sk/eshop/sk/portal/catalog.html?
>
> type=post&subtype=phone&null
>
>
>
> > which gets 302 redirected to
> >https://www.orange.sk/eshop/sk/catalog/post/phones.html, after setting a
> > cookie through the Set-Cookie header field in the 302 reply. This works
> > fin with firefox.
>
> > However, with urllib2 and the following code snippet, it doesn't work
>
> Looking at the BadStatusLine exception raised, the server response line
> is empty. Looking at the source for httpllib suggests to me that the
> server closed the connection early. Perhaps it doesn't like connections
> from urllib2?
>
> I ran a test pretending to be IE using this code:
>
> cookiejar = cookielib.LWPCookieJar()
> opener = urllib2.build_opener(urllib2.HTTPCookieProcessor(cookiejar))
> url = 'http://www.orange.sk/eshop/sk/portal/catalog.html?'\
>     'type=post&subtype=phone&null'
> agent = "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; " \
>     "NeosBrowser; .NET CLR 1.1.4322; .NET CLR 2.0.50727)"
> headers = {'User-Agent': agent}
> req = urllib2.Request(url, data=None, headers=headers)
> try:
>     s=opener.open(req)
> except httplib.BadStatusLine, e:
>     print e, e.line
> else:
>     print "Success"
>
> but it failed. So the problem is not as simple as changing the user-agent
> string.
>
> Other than that, I'm stumped.
>
> --
> Steven

Thanks a lot for confirming this. I also tried with different headers,
including putting *exactly* the same headers as firefox (including
Connection:keep-alive by modifying httplib), it still doesn't work.
The only possible explanation for me is that python's httplib doesn't
handle SSL/TLS 'properly' (not necessarly in the sense of the TLS
spec, but in the sense that every other browser can connect properly
to this website and httplib can't)

If anyone knows an Oracle HTTPS server to confirm this on another
server, it would be nice...



More information about the Python-list mailing list