recipe-224043.py 文件源码

def getpage(self, url_pair):
        # Incoming argument name is a (URL, fragment) pair.
        # The page may have been cached in the name_table variable.
        url, fragment = url_pair
        if self.name_table.has_key(url):
            return self.name_table[url]

        scheme, path = urllib.splittype(url)
        if scheme in ('mailto', 'news', 'javascript', 'telnet'):
            self.note(1, " Not checking %s URL" % scheme)
            return None
        isint = self.inroots(url)

        # Ensure that openpage gets the URL pair to
        # print out its error message and record the error pair
        # correctly.
        if not isint:
            if not self.checkext:
                self.note(1, " Not checking ext link")
                return None
            f = self.openpage(url_pair)
            if f:
                self.safeclose(f)
            return None
        text, nurl = self.readhtml(url_pair)

        if nurl != url:
            self.note(1, " Redirected to %s", nurl)
            url = nurl
        if text:
            return Page(text, url, maxpage=self.maxpage, checker=self)

    # These next three functions take (URL, fragment) pairs as
    # arguments, so that openpage() receives the appropriate tuple to
    # record error messages.