W3C home > Mailing lists > Public > public-pointer-events@w3.org > April to June 2015

Re: Automating W3C Test Execution with WebDriver for Pointer Events

From: Rick Byers <rbyers@chromium.org>
Date: Thu, 18 Jun 2015 10:31:47 -0400
Message-ID: <CAFUtAY-7kiG_NU5V_4px4fe-unNjirSr0EjTftpuvWg7Y3vKuQ@mail.gmail.com>
To: David Burns <dburns@mozilla.com>
Cc: James Graham <james@hoppipolla.co.uk>, Arthur Barstow <art.barstow@gmail.com>, Jacob Rossi <Jacob.Rossi@microsoft.com>, Doug Schepers <schepers@w3.org>, "public-browser-tools-testing@w3.org" <public-browser-tools-testing@w3.org>, Pointer Events WG <public-pointer-events@w3.org>, Philippe Le Hegaret <plh@w3.org>, Mike Smith <mike@w3.org>, Scott González <scott.gonzalez@gmail.com>, John Jansen <John.Jansen@microsoft.com>, Dirk Pranke <dpranke@chromium.org>, jsbell@chromium.org
/cc dpranke@ and jsbell@ who are interested in this infrastructure from the
blink side

This sounds great, thank you!

Note that although the WebDriver pointer actions spec is written somewhat
in terms of DOM events, that's not really how we'd want to implement it in
Chrome.  We'd generate the lower level input as seen by the OS so that
we're covering all the code that turns that into (among other things) DOM
events.  As Jacob says, they use a fake OS-level input driver for this.  In
Chrome we don't go quite that low, but instead inject fake input at the
very bottom of the stack in Chrome exactly as if it came from the OS-level
input APIs.

Given that, I don't think it should make any difference to the Web Driver
API design what specific DOM input APIs the user agent supports (i.e.
whether or not Chrome supports pointer events).  A pointer-event-style API
to describe the raw input is fine, it just won't map directly 1:1 with the
events generated (eg. a pointerdown, move and up with source=touch will
trigger sending touch events and possibly a scroll, even when the UA
doesn't support pointer events).

Sound reasonable?

On Thu, Jun 18, 2015 at 9:43 AM, David Burns <dburns@mozilla.com> wrote:

> As James said, A Bugzilla bug[1] would be preferable. Jacob I suggest
> speaking to John Jansen since he is helping move forward the microsoft
> implementation and has already contributed to the WebDriver spec.
> David
> [1]
> https://www.w3.org/Bugs/Public/enter_bug.cgi?comment=&component=WebDriver&product=Browser+Test%2FTools+WG&blocked=20860
> On 18 June 2015 at 12:06, James Graham <james@hoppipolla.co.uk> wrote:
>> On 18/06/15 11:48, Arthur Barstow wrote:
>>> On 6/17/15 7:51 PM, David Burns wrote:
>>>> When the low level actions was written when Chrome had said that they
>>>> were not going to Pointer Events so they are in need of rewrite. It
>>>> also needs a rewrite for other reasons so let me know what changes you
>>>> would like to have and I will try to accomodate them. I will be
>>>> tackling this soon.
>>> Thanks David!
>>> If someone has input, do you prefer a Bugzilla bug [B], a GG Issue [GH],
>>> e-mail to [p-b-t-t], reply to this thread, and/or something else? And if
>>> there is a deadline, please let us know.
>> I think bugzilla is the best place for "this is a bug in the spec" type
>> feedback. For things that are more like topics of discussion I suggest
>> using the list directly. So far GH issues are not really being used.
>> The group is chartered until the end of the year so if you think about
>> deadlines in Process terms, it rather behooves us to have something that
>> looks like it might get published by then. If you want to think about it in
>> more practical terms, there are people implementing right now, so the
>> sooner the better.
Received on Thursday, 18 June 2015 14:32:37 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 18 June 2015 14:32:37 UTC