Allow use of Freetype's native emboldening

This patch allows the use of the native FreeType emboldening
algorithm when "fake bold" text is used in SkPaint.

It's enabled by #defining SK_USE_FREETYPE_EMBOLDEN.

[reed, 2010-02-05 18:13:59]

I see the intent - to use a scaler's better bold algorithm. Let me think on it
for a little bit. I think I'd like a way to query the fontscaler to know if it
supports this, rather than an ifdef.

On 2010/02/03 01:20:02, Stephen White wrote:
>

[reed, 2010-02-05 22:38:40]

On second thought, that's a change I can do later (the runtime check).

LGTM

On 2010/02/05 18:13:59, reed wrote:
> I see the intent - to use a scaler's better bold algorithm. Let me think on it
> for a little bit. I think I'd like a way to query the fontscaler to know if it
> supports this, rather than an ifdef.
> 
> On 2010/02/03 01:20:02, Stephen White wrote:
> >

[Stephen White, 2010-02-05 22:42:10]

On Fri, Feb 5, 2010 at 5:38 PM, <reed@android.com> wrote:

> On second thought, that's a change I can do later (the runtime check).
>
> LGTM


I don't mind doing it.  The only trick is that SkScalerContext::MakeRec() is
static, so I can't just put a virtual "useNativeEmbolden()" on
SkScalerContext as I'd like to.  I can put a static on the FontHost, but
it'll need a dummy implementation on FontHost implementation.

Stephen


>
> On 2010/02/05 18:13:59, reed wrote:
>
>> I see the intent - to use a scaler's better bold algorithm. Let me
>>
> think on it
>
>> for a little bit. I think I'd like a way to query the fontscaler to
>>
> know if it
>
>> supports this, rather than an ifdef.
>>
>
>  On 2010/02/03 01:20:02, Stephen White wrote:
>> >
>>
>
>
>
> http://codereview.appspot.com/198077/show
>



-- 
All truth passes through three stages. First, it is ridiculed. Second, it is
violently opposed. Third, it is accepted as being self-evident. --
Schopenhauer

[Stephen White, 2010-02-05 22:43:01]

On Fri, Feb 5, 2010 at 5:42 PM, Stephen White <senorblanco@chromium.org>wrote:

> On Fri, Feb 5, 2010 at 5:38 PM, <reed@android.com> wrote:
>
>> On second thought, that's a change I can do later (the runtime check).
>>
>> LGTM
>
>
> I don't mind doing it.  The only trick is that SkScalerContext::MakeRec()
> is static, so I can't just put a virtual "useNativeEmbolden()" on
> SkScalerContext as I'd like to.  I can put a static on the FontHost, but
> it'll need a dummy implementation on FontHost implementation.
>

should be "on each FontHost impl".


>
> Stephen
>
>
>>
>> On 2010/02/05 18:13:59, reed wrote:
>>
>>> I see the intent - to use a scaler's better bold algorithm. Let me
>>>
>> think on it
>>
>>> for a little bit. I think I'd like a way to query the fontscaler to
>>>
>> know if it
>>
>>> supports this, rather than an ifdef.
>>>
>>
>>  On 2010/02/03 01:20:02, Stephen White wrote:
>>> >
>>>
>>
>>
>>
>> http://codereview.appspot.com/198077/show
>>
>
>
>
> --
> All truth passes through three stages. First, it is ridiculed. Second, it
> is violently opposed. Third, it is accepted as being self-evident. --
> Schopenhauer
>



-- 
All truth passes through three stages. First, it is ridiculed. Second, it is
violently opposed. Third, it is accepted as being self-evident. --
Schopenhauer

[reed, 2010-02-05 22:46:38]

Yea, I was imaging some sort of static filter-like entry point, where
we pass in a bunch of flags etc. (like devKerning, emboldening,
bw-vs-aa-vs-lcd), and see how much of those can be fulfilled by the
scaler.

Honestly, lets defer that until after you checkin.

On Fri, Feb 5, 2010 at 5:42 PM, Stephen White <senorblanco@chromium.org> wrote:
>
>
> On Fri, Feb 5, 2010 at 5:42 PM, Stephen White <senorblanco@chromium.org>
> wrote:
>>
>> On Fri, Feb 5, 2010 at 5:38 PM, <reed@android.com> wrote:
>>>
>>> On second thought, that's a change I can do later (the runtime check).
>>>
>>> LGTM
>>
>> I don't mind doing it.  The only trick is that SkScalerContext::MakeRec()
>> is static, so I can't just put a virtual "useNativeEmbolden()" on
>> SkScalerContext as I'd like to.  I can put a static on the FontHost, but
>> it'll need a dummy implementation on FontHost implementation.
>
> should be "on each FontHost impl".
>
>>
>> Stephen
>>>
>>>
>>> On 2010/02/05 18:13:59, reed wrote:
>>>>
>>>> I see the intent - to use a scaler's better bold algorithm. Let me
>>>
>>> think on it
>>>>
>>>> for a little bit. I think I'd like a way to query the fontscaler to
>>>
>>> know if it
>>>>
>>>> supports this, rather than an ifdef.
>>>
>>>> On 2010/02/03 01:20:02, Stephen White wrote:
>>>> >
>>>
>>>
>>>
>>> http://codereview.appspot.com/198077/show
>>
>>
>>
>> --
>> All truth passes through three stages. First, it is ridiculed. Second, it
>> is violently opposed. Third, it is accepted as being self-evident. --
>> Schopenhauer
>
>
>
> --
> All truth passes through three stages. First, it is ridiculed. Second, it is
> violently opposed. Third, it is accepted as being self-evident. --
> Schopenhauer
>