create a new driver for mechanize.

more informations about mechanize in http://wwwsearch.sourceforge.net/mechanize/

In documentation explain with a little more details what Splinter is and why you should consider use this tool.

Create a django plugin to return django template exceptions.

I heard people coming back to webrat just because capybara does not have a good way to change headers.

we should allow people do do something like:

browser.headers['Content-Type'] = 'application/pdf'

Nowadays splinter tests occur visiting always the same web page.
This approach shades a problem with chrome:

let's say we are logging into a Django website:

browser = Browser('webdriver.chrome')
browser.visit("http://localhost:8000/admin/")

browser.find_by_id("login-form") # returns the login form

browser.fill("username", "johndoe")
browser.fill("password", "123456")
browser.find_by_id("login").first.click()

browser.visit("http://localhost:8000/some-other-page/")
browser.find_by_id("login-form") # should not see the login form, since I already logged in

What happens is that every single time I use browser.visit(), chrome opens a new tab, and all the next calls browser.find_by_something effects the first tab, not the new one.

In order to workaround this behaviour I had to hack splinter to use a javascript call to "location.href".

Below I put an example from my django project (using lettuce):

# -*- coding: utf-8 -*-
import urlparse
import warnings
from lettuce import before, after, world
from splinter.browser import Browser

warnings.simplefilter('ignore')
@before.harvest
def setup_browser(variables):
    world.browser = Browser('webdriver.chrome')

    world._old_visit = world.browser.visit
    _base = 'http://localhost:8000/'
    _redirect = 'window.location.href="%s/";'

    def navigate(path):
        current_path = world.browser.url.replace(_base, '/', 1)
        if path == current_path:
            script = 'window.location.href=window.location.href;'
        else:
            script = _redirect % path.rstrip('/')

        return world.browser.execute_script(script)

    world.browser.visit = navigate
    world._old_visit(_base)
    world.current_path = lambda: urlparse.urlsplit(world.browser.url).path

@after.harvest
def teardown_browser(results):
    world.browser.quit()

like this:

http://docs.python.org/library/warnings.html

Provide methods for mouse events: mouse over, mouse out, and others.

We need to change our markdown text to rst, because it will be easier to maintain the docs and the pypi page.

Support back and advance on history. It is important to note that some websites may block history.back() (from javascript), so I think we can't just do:

browser.execute_script('history.back();')

this feature is out of splinter's scope, hence should be implemented as an external package

splinter should support iframes

It needs a method for reload a page. Like:

browser.reload()

is need create a website for splinter

create a new driver for django test client

more information in http://docs.djangoproject.com/en/1.2/topics/testing/#module-django.test.client

It needs a driver for selenium 2 remote.

With selenium remote is possible use htmlunit driver.

More informations in http://code.google.com/p/selenium/

method for manage select fields

Is need a method for execute script ex:

execute_script("$('body').empty()")

Now find use one method for find by css_selector, xpath, tag, name or id:

browser.find(css_selector='h1')
browser.find(xpath='//h1')
browser.find(tag='h1')
browser.find(name='name')
browser.find(id='firstheader')

Other pattern is use one method for each selector type. Ex:

browser.find_by_css('.my-class')

It's need implements a method for waiting for elements to appear or to disappear on the page.

Enable support to any browser, like googlechrome, opera

It's need create all method, for find more than one element in a page.

all will be provide find by css_selector, xpath, id, tag and name.

To click a link (http://splinter.cobrateam.info/#clicking-links), the docs says that we have in Browser class, the methods: click_link_by_href and click_link_by_text, but it doesn't exists.

To click a link, I have to use browser.find_link_by_text('My Text').first.click()

Is this a problem with the docs, with the source code or with me? :-)

splinter have finders by name, id, css_select, tag and xpath. But check, fill_in, choose get elements by name:

browser.choose("some-radio")

It's need change this, moving these methods from browser class to element class:

checkbox = find_by_name('my-checkbox')
checkbox.check()

Well, as I've saw, those methods just look for an objet using it's name. But, what if we, in example, there're 2 radio options to choose and both have the same name?

The point it's not the reason, but how to deal with that case. I think Splinter must have another options to search controls, like by id, class or even xpath too.

Hi,
I can't install splinter because when I run 'pip install splinter' I get the following error:

Downloading/unpacking splinter
Running setup.py egg_info for package splinter
Traceback (most recent call last):
File "", line 14, in
File "/home/cillian/.virtualenvs/batucada/build/splinter/setup.py", line 4, in
README = open('README.rst').read()
IOError: [Errno 2] No such file or directory: 'README.rst'
Complete output from command python setup.py egg_info:
Traceback (most recent call last):

File "", line 14, in

File "/home/cillian/.virtualenvs/batucada/build/splinter/setup.py", line 4, in

README = open('README.rst').read()

IOError: [Errno 2] No such file or directory: 'README.rst'

Command python setup.py egg_info failed with error code 1

I think it's because the package at http://pypi.python.org/pypi/splinter/0.0.2 does not contain the README.rst file. Looks like a good package by the way.

Provide a drag and drop method.

Splinter should be able to suport JavaScript alert.

is need create some documentation for the splinter

for this let's use sphinx http://sphinx.pocoo.org/

Flask fake web app need down in the end of tests running

Create method for manipulate input files like:

attach_file('Image', '/path/to/image.jpg')

Add support to Internet Explorer using the webdriver Splinter driver.

create a new driver for zombie.js

more informations about zombie.js in http://zombie.labnotes.org/

create a new driver for windmill.

more informations about windmill in http://www.getwindmill.com/

splinter should support multi tabs

New driver for phantomjs: http://dev.umaclan.com/projects/pyphantomjs/

lxml can be declared as a dependence in setup.py, so it would be installed together with splinter when installing with pip.

Some script can return the result of the script. Ex:

evaluate_script('2+2') #returns 4

Method to return a list with all specifieds tags. Ex:

browser.find_all_tags('p')

['Foo', 'Bar', 'Test']

we can create a driver using webkit engine to run actions without a browser, with the same results.

for help us with this, we can use qtwebkit

add this method to driver

Enable support to google chrome browser

create a warning for deprecated method

find in chain for find a element in other element. Ex:

browser.find_by_id('my-id-1').find_by_tag('a')

It would be nice to be able to customize some settings on Firefox profile. Something like this code for Capybara:

class Capybara::Driver::Selenium
  def self.driver
    unless @driver
      profile = Selenium::WebDriver::Firefox::Profile.new
      profile['network.manage-offline-status'] = false
      @driver = Selenium::WebDriver.for :firefox, :profile => profile
      at_exit do
        @driver.quit
      end
    end
    @driver
  end
end

Selenium WebDriver for Python provides a FirefoxProfile class, so it's possible to create a new WebDriver instance with some customizations by hacking FirefoxProfile class. Here is an example, extracted from Rodrigo Manhẽs blog:

from selenium.firefox.firefox_profile import FirefoxProfile
prefs = FirefoxProfile._get_webdriver_prefs()
prefs['network.manage-offline-status'] = 'false'
@staticmethod
def prefs_func():
    return prefs
FirefoxProfile._get_webdriver_prefs = prefs_func

This hack looks ugly, because FirefoxProfile API has a bad design, it is just impossible to create a new FirefoxProfile instance and change some options, as in Capybara :(

Implement a way to execute actions on specific content areas.

Say you have many forms in a page, with similar fields, and want to select a field from a specific form, you could to something like:

in('#form-1').check('foo')

Chrome tests are very unstable, we need to inspect it to understand what to do.

The type method is used to slowly type the values to the text field.

Useful to test fields autocompletion.