Make SkImageFilter methods const.

Make SkImageFilter methods const.

SkImageFilter had some non-const methods that could all be made const.
This is a first step towards making SkImageFilter immutable.

BUG=skia:2097
Committed: http://code.google.com/p/skia/source/detail?r=13330

[Dominik Grewe, 2014-01-29 22:35:41]

Thought I'd upload that as a first baby step towards making SkImageFilter
immutable. Not that it really changes anything but if the methods don't change
the object we might as well mark them as const.

PTAL, thanks!

[mtklein, 2014-01-30 02:48:15]

On 2014/01/29 22:35:41, Dominik Grewe wrote:
> Thought I'd upload that as a first baby step towards making SkImageFilter
> immutable. Not that it really changes anything but if the methods don't change
> the object we might as well mark them as const.
> 
> PTAL, thanks!

lgtm

I _love_ const.  Don't downplay this sort of change.  It does an important
thing, which is make the compiler yell at anyone who messes with out plan.  :)

[mtklein, 2014-01-30 02:48:55]

On 2014/01/30 02:48:15, mtklein wrote:
> On 2014/01/29 22:35:41, Dominik Grewe wrote:
> > Thought I'd upload that as a first baby step towards making SkImageFilter
> > immutable. Not that it really changes anything but if the methods don't
change
> > the object we might as well mark them as const.
> > 
> > PTAL, thanks!
> 
> lgtm
> 
> I _love_ const.  Don't downplay this sort of change.  It does an important
> thing, which is make the compiler yell at anyone who messes with out plan.  :)

Can't spell lately.  "our plan".

[Dominik Grewe, 2014-01-30 13:46:32]

Thanks!

[Dominik Grewe, 2014-01-30 13:46:33]

Thanks!

[commit-bot: I haz the power, 2014-01-30 13:46:57]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/dominikg@chromium.org/148883011/1

[commit-bot: I haz the power, 2014-01-30 13:47:15]

Presubmit check for 148883011-1 failed and returned exit status 1.

Running presubmit commit checks ...

** Presubmit ERRORS **
Since the CL is editing public API, you must have an LGTM from one of:
('reed@chromium.org', 'reed@google.com', 'bsalomon@chromium.org',
'bsalomon@google.com', 'rmistry@google.com')

Presubmit checks took 1.2s to calculate.

Was the presubmit check useful? If not, run "git cl presubmit -v"
to figure out which PRESUBMIT.py was run, then run git blame
on the file to figure out who to ask for help.

[Dominik Grewe, 2014-01-30 13:52:22]

+reed for reviewing changes to public API.

[Dominik Grewe, 2014-02-05 17:40:35]

On 2014/01/30 13:52:22, Dominik Grewe wrote:
> +reed for reviewing changes to public API.

Any more thoughts on this?

[reed1, 2014-02-05 18:01:09]

Had I known what you wanted to try, I would have asked to stage this:

1. just change all methods to const
2. try changing parameters to const SkImageFilter*

However, these don't seem at all easy to tease apart, so I won't ask.

I don't think we need const parameters per-se to make the class immutable (e.g.
SkTypeface is completely immutable but we don't usually pass a const param).

All that said, the idea is a good one to try, so lgtm

[reed1, 2014-02-05 18:02:46]

BTW -- there are subclasses of canvas/device in Chrome, that must be modified as
well. How can we perform that w/o jeopardizing the daily DEPS roll?

I wonder if in fact just changing the methods to const would make this
synchronizing with Chrome more manageable?

[Dominik Grewe, 2014-02-05 18:56:23]

Unfortunately there are some dependencies between changing the methods and the
parameters. For example, to make SkImageFilter::filterImage() const, the
SkImageFilter parameter for SkProxy::filterImage() also needs to be const so we
can pass 'this'. I think I only changed the parameters when necessary.


On 2014/02/05 18:02:46, reed1 wrote:
> BTW -- there are subclasses of canvas/device in Chrome, that must be modified
as
> well. How can we perform that w/o jeopardizing the daily DEPS roll?

I'm not making any changes to SkCanvas. And the only subclasses of SkDevice
don't overwrite any of the methods I've modified. So I think we're lucky and
should be okay to land this :) In fact, I've applied this patch to my local
Chrome checkout and it compiles fine.

[reed1, 2014-02-05 18:57:45]

On 2014/02/05 18:56:23, Dominik Grewe wrote:
> Unfortunately there are some dependencies between changing the methods and the
> parameters. For example, to make SkImageFilter::filterImage() const, the
> SkImageFilter parameter for SkProxy::filterImage() also needs to be const so
we
> can pass 'this'. I think I only changed the parameters when necessary.

Understood. We have a handful of const_casts in the case of SkTypeface, so I
sort of anticipated this.

> 
> 
> On 2014/02/05 18:02:46, reed1 wrote:
> > BTW -- there are subclasses of canvas/device in Chrome, that must be
modified
> as
> > well. How can we perform that w/o jeopardizing the daily DEPS roll?
> 
> I'm not making any changes to SkCanvas. And the only subclasses of SkDevice
> don't overwrite any of the methods I've modified. So I think we're lucky and
> should be okay to land this :) In fact, I've applied this patch to my local
> Chrome checkout and it compiles fine.

Ah, good news.

[Dominik Grewe, 2014-02-05 19:21:40]

The CQ bit was checked by dominikg@chromium.org

[commit-bot: I haz the power, 2014-02-05 19:21:51]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/dominikg@chromium.org/148883011/70001

[commit-bot: I haz the power, 2014-02-05 19:35:19]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-02-05 19:35:19]

Retried try job too often on Build-Mac10.7-Clang-x86-Release-Trybot for step(s)
BuildGm
http://108.170.219.164:10117/buildstatus?builder=Build-Mac10.7-Clang-x86-Rele...

[Dominik Grewe, 2014-02-05 20:34:44]

The CQ bit was checked by dominikg@chromium.org

[commit-bot: I haz the power, 2014-02-05 20:34:49]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/dominikg@chromium.org/148883011/40002

[commit-bot: I haz the power, 2014-02-05 21:36:28]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-02-05 21:36:29]

Retried try job too often on Build-Mac10.7-Clang-x86-Release-Trybot for step(s)
BuildEverything
http://108.170.219.164:10117/buildstatus?builder=Build-Mac10.7-Clang-x86-Rele...

[Dominik Grewe, 2014-02-05 22:15:25]

The CQ bit was checked by dominikg@chromium.org

[commit-bot: I haz the power, 2014-02-05 22:15:28]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/dominikg@chromium.org/148883011/260001

[commit-bot: I haz the power, 2014-02-05 22:32:17]

Change committed as 13330