Boris Ozegovic <[EMAIL PROTECTED]> writes: > Hi > > I have HTTP client which accepts cookies. If client allready has cookie, > but that cookie has expired, server sends him new cookie, and in response > object Set-Cookie: header everything is fine, but if I reload request, > client sends expired cookie, and not the new one. In cookiejar there is > only new and valid cookie, and if I use regular browser everything is fine. > The code is following: > > urlopen = urllib2.urlopen > Request = urllib2.Request > cj = cookielib.LWPCookieJar() > COOKIEFILE = 'cookies.lwp' > > if os.path.isfile(COOKIEFILE): > # if we have a cookie file already saved > # then load the cookies into the Cookie Jar > cj.load(COOKIEFILE) > > opener = urllib2.build_opener(urllib2.HTTPCookieProcessor(cj)) > urllib2.install_opener(opener) > url = "http://localhost:8000/files/index.html"; > params = {'question':question} > data = urllib.urlencode(params) > Request(url, data) > try: > response = urlopen(Request) > etc. > > Only if I create new request object the new cookie is send, but I don't > want to create new object. And idea?
You seem to suggest in your first paragraph that you call urlopen more than once, but in the code you show, you call it only once. This is the way HTTP cookies work: you can't send back a cookie before you've received it. So, you need two HTTP requests: one to get the cookie, and a second to send it back again. However, some websites arrange things so that a single attempt to load a URL results in multiple HTTP requests. One common way to do that is by using an "HTTP Refresh" -- look at the top of the HTML to see if there's a META tag something like this: <meta http-equiv="refresh" content="1" /> This means that the browser should reload the page after a pause of one second. Often, the reloaded page then does not contain another meta refresh tag, so that the page does not reload every second forever (the server knows not to include the tag because it received the cookie). This tag can be supported by appropriate Python code. So, two HTTP requests needn't necessarily mean two urlopen calls. This library handles refresh redirections, and supports essentially the same interface as urllib2 and cookielib (the union of the two interfaces): http://wwwsearch.sourceforge.net/mechanize/ John -- http://mail.python.org/mailman/listinfo/python-list
