Use same parameters for jpeg decoding quality on all platforms.

Use same parameters for jpeg decoding quality on all platforms.

This patch will change android to use same decoding parameters as
other platforms, thus changing back to dct=int and dither=fs since
the faster settings caused visible artifacts that did not justify
the small performance gain.

BUG=385515
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=192409

[Olle Liljenzin, 2015-03-20 13:43:33]

ollel@opera.com changed reviewers:
+ aelias@chromium.org, kbr@chromium.org, noel@chromium.org,
pkasting@chromium.org

[Olle Liljenzin, 2015-03-20 13:48:58]

Created a new review that replaces:
https://codereview.chromium.org/339143002/

[aelias_OOO_until_Jul13, 2015-03-20 18:33:20]

aelias@chromium.org changed reviewers:
+ chrishtr@chromium.org

[aelias_OOO_until_Jul13, 2015-03-20 18:33:39]

lgtm, adding chrishtr@ for OWNERS.

[Ken Russell (switch to Gerrit), 2015-03-20 18:49:28]

owners LGTM

[Peter Kasting, 2015-03-20 19:13:47]

LGTM

[chrishtr, 2015-03-20 19:16:51]

lgtm

[Noel Gordon, 2015-03-21 05:49:26]

https://codereview.chromium.org/1023193002/diff/1/Source/config.h
File Source/config.h (left):

https://codereview.chromium.org/1023193002/diff/1/Source/config.h#oldcode89
Source/config.h:89: #define WTF_USE_LOW_QUALITY_IMAGE_INTERPOLATION 1
Seems unrelated to JPEG.  Can I assume this interpolation quality flag is not
used anywhere?

[Noel Gordon, 2015-03-22 13:11:27]

lgtm

> https://codereview.chromium.org/1023193002/diff/1/Source/config.h
> File Source/config.h (left):
> 
> https://codereview.chromium.org/1023193002/diff/1/Source/config.h#oldcode89
> Source/config.h:89: #define WTF_USE_LOW_QUALITY_IMAGE_INTERPOLATION 1
> Seems unrelated to JPEG.  Can I assume this interpolation quality flag is not
> used anywhere?

Seems so according to cs.chromium.org

[Noel Gordon, 2015-03-23 05:43:00]

Ran trys, not sure you have trybot access?

[Olle Liljenzin, 2015-03-23 08:18:21]

On 2015/03/23 05:43:00, noel gordon wrote:
> Ran trys, not sure you have trybot access?

I have trybot access now, but thanks for running them anyway!

In the bug you wrote "While we wait... can we update the current patch". So are
we still waiting for something else or should I just push the commit button now?

[Olle Liljenzin, 2015-03-23 08:43:31]

https://codereview.chromium.org/1023193002/diff/1/Source/config.h
File Source/config.h (left):

https://codereview.chromium.org/1023193002/diff/1/Source/config.h#oldcode89
Source/config.h:89: #define WTF_USE_LOW_QUALITY_IMAGE_INTERPOLATION 1
On 2015/03/21 05:49:26, noel gordon wrote:
> Seems unrelated to JPEG.  Can I assume this interpolation quality flag is not
> used anywhere?

Looks like I removed one line too much by mistake. The definition is used in
Source/platform/graphics/GraphicsTypes.h.

[Noel Gordon, 2015-03-23 09:15:22]

On 2015/03/23 08:18:21, Olle Liljenzin wrote:

> In the bug you wrote "While we wait... can we update the current patch". So
are
> we still waiting for something else or should I just push the commit button
now?

Folks pinged are in different times zones and busy, so let's give them time.

[Noel Gordon, 2015-03-23 09:19:34]

LGTM

> On 2015/03/21 05:49:26, noel gordon wrote:
> > Seems unrelated to JPEG.  Can I assume this interpolation quality flag is
not
> > used anywhere?
> 
> Looks like I removed one line too much by mistake. The definition is used in
> Source/platform/graphics/GraphicsTypes.h.

Ok good. Thanks for checking.

[Noel Gordon, 2015-03-23 13:52:47]

Alexis mentioned to me off-line, that our YUV JPEG decode path should be ok, as
long as jpeg_read_raw_data() does not depend on the IDCT method, which he was
pretty sure it doesn't.
 
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...

Maybe have a scan around in that library function olle@ and let me know.

[Olle Liljenzin, 2015-03-23 15:19:17]

On 2015/03/23 13:52:47, noel gordon wrote:
> Alexis mentioned to me off-line, that our YUV JPEG decode path should be ok,
as
> long as jpeg_read_raw_data() does not depend on the IDCT method, which he was
> pretty sure it doesn't.
>  
>
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...
> 
> Maybe have a scan around in that library function olle@ and let me know.

I tried to follow the code which isn't that easy. To my limited understanding
this function is just decompressing data, not decoding it.

[Noel Gordon, 2015-03-23 16:59:18]

Let's land it then.

[Noel Gordon, 2015-03-24 03:46:10]

The CQ bit was checked by noel@chromium.org

[Noel Gordon, 2015-03-24 03:46:10]

The patchset sent to the CQ was uploaded after l-g-t-m from kbr@chromium.org,
chrishtr@chromium.org, pkasting@chromium.org, aelias@chromium.org
Link to the patchset: https://codereview.chromium.org/1023193002/#ps20001
(title: "Adding back WTF_USE_LOW_QUALITY_IMAGE_INTERPOLATION that was removed by
mistake.")

[commit-bot: I haz the power, 2015-03-24 03:46:14]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1023193002/20001

[commit-bot: I haz the power, 2015-03-24 03:50:00]

Committed patchset #2 (id:20001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=192409