- From: Tobie Langel <tobie@w3.org>
- Date: Mon, 18 Mar 2013 11:18:25 +0100
- To: Robin Berjon <robin@w3.org>
- Cc: public-test-infra@w3.org
On Monday, March 18, 2013 at 11:03 AM, Robin Berjon wrote: > On 15/03/2013 17:24 , Tobie Langel wrote: > > I wrote down[1] a very early and very high level plan of how to > > architect a WebDriver-based framework. Note this framework would > > fallback to running the tests in a more regular manner should > > WebDriver not be supported. In which case, ref tests wouldn't be > > run. > > A few questions: > > • Is Mozilla really the only native implementation? I thought that > Chromium had native support too (but I couldn't confirm or infirm this > from the Web). No. I got lazy. Also, this is a wiki. :) > • Beyond the architecture to get WD tests to run, is there something we > need to agree upon for how these tests are written, how the drivers > knows what to do, etc.? I ask because I've been thinking about WD tests > for the Editing APIs and might be tempted to give those a stab at some > point. So, right now, I've only been thinking only about running testharness.js and ref tests within that context, and imagining what a small subset of a11y tests could be. Getting a PoC up for that is next on my todo list. To write raw WD tests, we'll probably need to agree and specify a single language binding to the WD JSON protocol. My understanding is there is a related, non-normative deliverable within the Browser Testing and Tools WG. We might want to look into that a bit more closely at some point. --tobie
Received on Monday, 18 March 2013 10:18:34 UTC