Document the meaning of the scroll wheel event in PPAPI.

Document the meaning of the scroll wheel event in PPAPI.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=75025

[brettw, 2011-02-15 03:31:27]

After some experimenting, I think the current meanings of the scroll wheel
events is best, so I just documented the current meanings of the scroll events.
We can add an additional "hardware tick counts" later if we deem it necessary
(the structure is padded and the rest is guaranteed 0, so a plugin could check
these later values against 0 to see if they're present.

[Peter Kasting, 2011-02-15 20:03:59]

http://codereview.chromium.org/6528007/diff/1/ppapi/c/pp_input_event.h
File ppapi/c/pp_input_event.h (right):

http://codereview.chromium.org/6528007/diff/1/ppapi/c/pp_input_event.h#newcod...
ppapi/c/pp_input_event.h:229: * event. This value will not take into account
acceleration for faster
This would be ideal, but it's inaccurate.  On Mac (the only platform that
currently accelerates AFAIK), these values are actually accelerated, not
unaccelerated.

[brettw, 2011-02-15 20:17:59]

I removed the acceleration stuff from the tick event.

[Peter Kasting, 2011-02-15 20:43:23]

On 2011/02/15 20:17:59, brettw wrote:
> I removed the acceleration stuff from the tick event.

I don't know that we want to do that -- we probably want to say explicitly that
this is accelerated on continuous devices and unaccelerated on discrete devices.

Or we want to just fix WebInputEventFactory.mm to always send these as
unaccelerated values; maybe since Avi wrote his comments Apple has fixed the
bugs in the kCGScrollWheelEventDeltaAxis* values.

[Avi (use Gerrit), 2011-02-15 20:56:46]

This is a tough call. No, Apple hasn't fixed any bugs. Do we declare what it is
or what it should be? :(

[brettw, 2011-02-15 21:06:40]

On Tue, Feb 15, 2011 at 12:56 PM,  <avi@chromium.org> wrote:
> This is a tough call. No, Apple hasn't fixed any bugs. Do we declare what it
> is
> or what it should be? :(

Given that nobody has time to work on it, that the issues seem
complicated and subtle, and that the current behavior is reasonable
for most uses, I'm inclined to document how it is.

Brett

[Peter Kasting, 2011-02-15 21:07:48]

On 2011/02/15 21:06:40, brettw wrote:
> On Tue, Feb 15, 2011 at 12:56 PM,  <mailto:avi@chromium.org> wrote:
> > This is a tough call. No, Apple hasn't fixed any bugs. Do we declare what it
> > is
> > or what it should be? :(
> 
> Given that nobody has time to work on it, that the issues seem
> complicated and subtle, and that the current behavior is reasonable
> for most uses, I'm inclined to document how it is.

Yes, if there's no way we can make the tick value always be unaccelerated, we
should document what happens today.

[brettw, 2011-02-15 21:34:14]

On Tue, Feb 15, 2011 at 1:07 PM,  <pkasting@chromium.org> wrote:
> On 2011/02/15 21:06:40, brettw wrote:
>>
>> On Tue, Feb 15, 2011 at 12:56 PM,  <mailto:avi@chromium.org> wrote:
>> > This is a tough call. No, Apple hasn't fixed any bugs. Do we declare
>> > what it
>> > is
>> > or what it should be? :(
>
>> Given that nobody has time to work on it, that the issues seem
>> complicated and subtle, and that the current behavior is reasonable
>> for most uses, I'm inclined to document how it is.
>
> Yes, if there's no way we can make the tick value always be unaccelerated,
> we
> should document what happens today.

I added back acceleration as being system-specific. I don't think
we're in a good position to say definitively what it means. I have an
Apple mouse on my Linux box. I seem to be getting discrete mouse
events. I suspect Windows trackpad scrolls will be unaccelerated, but
Mac ones will be.

Brett

[Peter Kasting, 2011-02-15 21:46:41]

LGTM