Spawning multiple browsers from Selenium RC utilizing Python

Question

I've been trying to develop an automated test case solution using Selenium RC and Python and after lengthy testing I've hit a pretty hard block in the road, so to speak.

I have three files: unit.py, case1.py, and case1m.py

unit.py configures instances of case1m.py with a browser and a port, then runs the test by sending the case1m instance through unittest.main().

The case1.py file is a vanilla case that is generated from Selenium IDE; when run from the command line, it executes the test case and exits with OK. I used this file to help debug the failing points of the other two files.

Here is the source for all three files:


unit.py:

import unittest
from case1m import case1m

browser = "*chrome"
port = 4444

a = case1m()
a.setBrowser("*chrome",4444)
unittest.main(a)

case1m.py - handles browser/port arguments and runs selenium test cases:

from selenium import selenium
import unittest, time, re

class case1m(unittest.TestCase):
    def setBrowser(self,b,p):
        print "entered setBrowser"
        self.browser = b
        self.port = p
        print "leaving setBrowser"
        self.setUp()

    def setUp(self):
        print self.browser,", ",self.port
        self.verificationErrors = []
        self.selenium = selenium("localhost", self.browser, self.port, "http://megagate-ffcdcb.xl_net.internal/")
        self.selenium.start()
        print "end setUp"
        self.runTest()

    def runTest(self):
        print "entered runTest"
        sel = self.selenium
        sel.open("/seltest/")
        try: self.failUnless(sel.is_text_present("BODY"))
        except AssertionError, e: self.verificationErrors.append(str(e))
        print "leaving runTest"
        self.tearDown()     

    def tearDown(self):
        print "entered tearDown"
        self.selenium.stop()
        self.assertEqual([], self.verificationErrors)
        print "leaving tearDown"

case1.py:

from selenium import selenium
import unittest, time, re

class case1(unittest.TestCase):
    def setUp(self):
        print "entered setUp"
        self.verificationErrors = []
        self.selenium = selenium("localhost", 4444, "*chrome", "http://megagate-ffcdcb.xl_net.internal/")
        self.selenium.start()

    def runTest(self):
        sel = self.selenium
        sel.open("/seltest/")
        try: self.failUnless(sel.is_text_present("BODY"))
        except AssertionError, e: self.verificationErrors.append(str(e))

    def tearDown(self):
        self.selenium.stop()
        self.assertEqual([], self.verificationErrors)

if __name__ == '__main__':
    unittest.main()

The first problem I ran into was passing the browser and port values to an instance of the case1m class. I tried using __init__ to collect them as arguments, but apparently sub-classing the TestCase class and then adding an __init__ override causes problems; the setUp(), runTest() and tearDown() methods no longer triggered automatically as they do in the case1 class.

So instead, I overrode and inserted a setBrowser() method to collect the values and create the browser and port variables within the class instance. This again causes the same issue as before, so I resorted to inserting method calls into setUp(), runTest() and tearDown(). When executed, it runs until it tries the do_command() method in the selenium instance.

Here is the error:

Traceback (most recent call last):
File "C:\sel-test\unit.py", line 13, in
a.setBrowser("*chrome",4444)
File "C:\sel-test\case1m.py", line 10, in setBrowser
self.setUp()
File "C:\sel-test\case1m.py", line 16, in setUp
self.selenium.start()
File "C:\Python26\lib\selenium.py", line 190, in start
result = self.get_string("getNewBrowserSession", [self.browserStartCommand, self.browserURL, self.extensionJs])
File "C:\Python26\lib\selenium.py", line 225, in get_string
result = self.do_command(verb, args)
File "C:\Python26\lib\selenium.py", line 213, in do_command
conn.request("POST", "/selenium-server/driver/", body, headers)
File "C:\Python26\lib\httplib.py", line 910, in request
self._send_request(method, url, body, headers)
File "C:\Python26\lib\httplib.py", line 947, in _send_request
self.endheaders()
File "C:\Python26\lib\httplib.py", line 904, in endheaders
self._send_output()
File "C:\Python26\lib\httplib.py", line 776, in _send_output
self.send(msg)
File "C:\Python26\lib\httplib.py", line 735, in send
self.connect()
File "C:\Python26\lib\httplib.py", line 716, in connect
self.timeout)
File "C:\Python26\lib\socket.py", line 500, in create_connection
for res in getaddrinfo(host, port, 0, SOCK_STREAM):
socket.gaierror: [Errno 10109] getaddrinfo failed

My questions is: why does the unit.py/case1m.py combination result in socket.gaierror when the case1.py file will run without error? From what I can see, the selenium class should be receiving the exact same information by the time it reaches self.do_command(). The only difference is that case1.py is being run directly from the commandline, while case1m.py is being run as an imported module.

Answer 1

Looking at the 2 code snippets side by side, I think you have inverted the browser and port arguments. This is probably the source of your error.

case1.py (runs fine):

self.selenium = selenium("localhost", 4444, "*chrome", "http://megagate-ffcdcb.xl_net.internal/")

case1m.py (socket error):

self.selenium = selenium("localhost", self.browser, self.port, "http://megagate-ffcdcb.xl_net.internal/")