Handling cookies without urllib2 and cookielib

[Joshua Kugler]

Standard disclaimer: read, googled, read some more. If you have a link,
please free free to point me there.

I'm using HTTPlib to construct some functional tests for a web app we're
writing. We're not using urllib2 because we need support for PUT and
DELETE methods, which urllib2 does not do.

We also need client-side cookie handling. So, I start reading about
cookielib and run across a problem. It's cookie handling is tied quite
tightly to urllib2's request object. httplib has somewhat different
semantics in its request object. So, you can use cookielib with httplib.
And cookie lib has no simple function (that I could find) for passing in a
set-cookie header and getting back a CookieJar object (or even a list of
Cookie objects).

I'm sure I'm not the first to have to deal with httplib and cookies. Anyone
have suggestions or pointers?

j

[Gabriel Genellina]

On 14 dic, 23:44, Joshua Kugler <jkug...@bigfoo t.comwrote:

I'm using HTTPlib to construct some functional tests for a web app we're
writing. We're not using urllib2 because we need support for PUT and
DELETE methods, which urllib2 does not do.

We also need client-side cookie handling. So, I start reading about
cookielib and run across a problem. It's cookie handling is tied quite
tightly to urllib2's request object. httplib has somewhat different
semantics in its request object. So, you can use cookielib with httplib.
And cookie lib has no simple function (that I could find) for passing in a
set-cookie header and getting back a CookieJar object (or even a list of
Cookie objects).

What about correcting the first thing, making urllib2 support HEAD/PUT/
DELETE?

import urllib2

class Request(urllib2 .Request):

def __init__(self, url, data=None, headers={},
origin_req_host =None, unverifiable=Fa lse,
method=None):
urllib2.Request .__init__(self, url, data, headers,
origin_req_host , unverifiable)
self.method = method

def get_method(self ):
if self.method is None:
if self.data is not None:
return "POST"
else:
return "GET"
return self.method

pyf = urllib2.urlopen (Request("http://www.python.org/",
method="HEAD"))
pyprint f.info()
Date: Sun, 16 Dec 2007 00:03:43 GMT
Server: Apache/2.2.3 (Debian) DAV/2 SVN/1.4.2 mod_ssl/2.2.3 OpenSSL/
0.9.8c
Last-Modified: Sat, 15 Dec 2007 16:25:58 GMT
ETag: "60193-3e6a-a24fb180"
Accept-Ranges: bytes
Content-Length: 15978
Connection: close
Content-Type: text/html

pyprint len(f.read())
0

Notes:
a) Instead of urlopen(url,... ) you must use urlopen(Request (url,...))
b) Redirection is not handled correctly in HTTPRedirectHan dler (the
request method should be copied over)
c) I've not verified PUT / DELETE methods
d) I'll try to make a proper patch later

--
Gabriel Genellina

[Joshua Kugler]

Gabriel Genellina wrote:

On 14 dic, 23:44, Joshua Kugler <jkug...@bigfoo t.comwrote:

>I'm using HTTPlib to construct some functional tests for a web app we're
writing. We're not using urllib2 because we need support for PUT and
DELETE methods, which urllib2 does not do.

We also need client-side cookie handling. So, I start reading about
cookielib and run across a problem. It's cookie handling is tied quite
tightly to urllib2's request object. httplib has somewhat different
semantics in its request object. So, you can use cookielib with httplib.
And cookie lib has no simple function (that I could find) for passing in
a set-cookie header and getting back a CookieJar object (or even a list
of Cookie objects).

What about correcting the first thing, making urllib2 support HEAD/PUT/
DELETE?

<SNIP>

We may have to do that, and then hack on the Redirect handler too so it will
properly keep the request method. But that's not our preference, for
obvious reasons. :)

I just find it hard to believe that no one has ever needed to do cookie
handling in a generic way (i.e. input: set-cookie header, output: cookie
objects) before. May have to write my own. Or sublcass/extend cookielib.

j