- From: Denis Ah-Kang <denis@w3.org>
- Date: Tue, 04 Feb 2014 17:32:41 +0400
- To: James Graham <james@hoppipolla.co.uk>, public-test-infra@w3.org
On 02/04/2014 05:07 PM, James Graham wrote: > On 04/02/14 12:58, Denis Ah-Kang wrote: >> Hi, >> >> The guys working on webdriver added a couple of tests [1] a few months >> ago using the python implementation and it seems to work fine except >> that it's not running on wptserve but on a python webserver made by >> Mozilla. The switch shouldn't be too hard according to James. >> >> I was wondering how people feel about starting to write tests based on >> what's already in [1]. We should then be able to automate most of the >> manual tests. >> Any thoughts? > > I am fine with automating the manual tests using webdriver, although we > should enforce the requirement that they can still be run without > webdriver for situations in which it is not available. > > We need to decide on which client language we will support; I am > strongly in favour of using python since that is a requirement for the > tests in any case. +1 > > Obviously webdriver-based tests for other specs should go inside the > path for the spec they are testing and not the webdriver/ path. That > means that this doesn't have any dependence on the webdriver tests > themselves moving to wptserve. There is probably some extra work to do > ensuring that a webdriver test is distinguishable from a normal python > file. And probably some more technical challenges besides. > Can't we go with a naming convention for the webdriver tests that would differentiate them from the other files? Denis
Received on Tuesday, 4 February 2014 13:32:50 UTC