Replace use of deprecated CG methods.

Replace use of deprecated CG methods.

OSX10.9 and iOS7.0 deprecated CGContextShowGlyphsAtPoint so a new API
should be used. OSX10.7 and iOS4.2 replace CGContextShowGlyphsAtPoint with
CTFontDrawGlyphs. OSX10.5 and iOS2.0 have CGContextShowGlyphsAtPositions
which works similarly to CTFontDrawGlyphs and has not yet been deprecated.
This change allows the use of CTFontDrawGlyphs when it is available,
falling back to CGContextShowGlyphsAtPositions when it isn't.

Committed: https://skia.googlesource.com/skia/+/3490263287e9432119c501884e45840de4d986bc

[bungeman-skia, 2014-12-08 21:17:23]

bungeman@google.com changed reviewers:
+ mtklein@google.com, reed@google.com

[bungeman-skia, 2014-12-08 21:17:24]

At PS6 this 'works' in the sense that the results would be acceptable if we
didn't already have an implementation. Unfortunately, due to all the mucking
about with matrices, there are many low bit differences in the floats so there
will be a lot of imperceptible differences in the glyphs produced. I'll have to
see how bad it will be in Blink.

Take a look and make sure I'm not doing anything too silly, and if you can dream
up a way to do less matrix math (particularly inverting things) you get a gold
star.

[reed1, 2014-12-08 21:24:44]

at some point we might factor some of the helper functions out -- the file is
getting pretty big

https://codereview.chromium.org/770383002/diff/100001/src/ports/SkFontHost_ma...
File src/ports/SkFontHost_mac.cpp (right):

https://codereview.chromium.org/770383002/diff/100001/src/ports/SkFontHost_ma...
src/ports/SkFontHost_mac.cpp:1872: AutoCFRelease<CFArrayRef>
tags(CTFontCopyAvailableTables(fFontRef,kCTFontTableOptionNoOptions));
should we lazily do this and cache the answer in the scalercontext?

[bungeman-skia, 2014-12-08 21:33:23]

Yep, this file is big and needs to be split into a SkFontMgr, SkTypeface, and
SkScalerContext like DirectWrite already is. It's on the list of things that
need to be done.

https://codereview.chromium.org/770383002/diff/100001/src/ports/SkFontHost_ma...
File src/ports/SkFontHost_mac.cpp (right):

https://codereview.chromium.org/770383002/diff/100001/src/ports/SkFontHost_ma...
src/ports/SkFontHost_mac.cpp:1872: AutoCFRelease<CFArrayRef>
tags(CTFontCopyAvailableTables(fFontRef,kCTFontTableOptionNoOptions));
On 2014/12/08 21:24:44, reed1 wrote:
> should we lazily do this and cache the answer in the scalercontext?

This is onFilterRec, so it is being cached in the scalercontext (in the sense
that it's how the color mask format is set). However, you're right in that this
could (and probably should) be on the typeface (an 'I have an sbix table' bit).

[reed1, 2014-12-08 21:48:21]

On 2014/12/08 21:33:23, bungeman1 wrote:
> Yep, this file is big and needs to be split into a SkFontMgr, SkTypeface, and
> SkScalerContext like DirectWrite already is. It's on the list of things that
> need to be done.
> 
>
https://codereview.chromium.org/770383002/diff/100001/src/ports/SkFontHost_ma...
> File src/ports/SkFontHost_mac.cpp (right):
> 
>
https://codereview.chromium.org/770383002/diff/100001/src/ports/SkFontHost_ma...
> src/ports/SkFontHost_mac.cpp:1872: AutoCFRelease<CFArrayRef>
> tags(CTFontCopyAvailableTables(fFontRef,kCTFontTableOptionNoOptions));
> On 2014/12/08 21:24:44, reed1 wrote:
> > should we lazily do this and cache the answer in the scalercontext?
> 
> This is onFilterRec, so it is being cached in the scalercontext (in the sense
> that it's how the color mask format is set). However, you're right in that
this
> could (and probably should) be on the typeface (an 'I have an sbix table'
bit).

ah, that's a slow place for it to be (until we can cache it). Even a side-car
cache might work (map of typeface-ID to this bool).

[bungeman-skia, 2014-12-10 16:16:23]

Please take a look, this is down to 42 Skia GM changes (2 Windows, 40 Mac) and 4
LayoutTest changes (1 Windows, 3 Mac). It should also fix a few currently
suppressed LayoutTest expectations around zero sized text.

[reed1, 2014-12-10 16:45:17]

https://codereview.chromium.org/770383002/diff/260001/src/core/SkScalerContex...
File src/core/SkScalerContext.cpp (right):

https://codereview.chromium.org/770383002/diff/260001/src/core/SkScalerContex...
src/core/SkScalerContext.cpp:746: if (diag[0].lengthSqd() < SK_ScalarNearlyZero
|| diag[1].lengthSqd() < SK_ScalarNearlyZero) {
1. most often when we look at len^2, we compare it to nearlyzero^2
2. might change it to <=, incase nearly zero was zero.

[bungeman-skia, 2014-12-10 17:06:16]

https://codereview.chromium.org/770383002/diff/260001/src/core/SkScalerContex...
File src/core/SkScalerContext.cpp (right):

https://codereview.chromium.org/770383002/diff/260001/src/core/SkScalerContex...
src/core/SkScalerContext.cpp:746: if (diag[0].lengthSqd() < SK_ScalarNearlyZero
|| diag[1].lengthSqd() < SK_ScalarNearlyZero) {
On 2014/12/10 16:45:16, reed1 wrote:
> 1. most often when we look at len^2, we compare it to nearlyzero^2
> 2. might change it to <=, incase nearly zero was zero.

Done and done.

[reed1, 2014-12-10 18:16:09]

lgtm

[bungeman-skia, 2014-12-11 05:34:04]

The CQ bit was checked by bungeman@google.com

[commit-bot: I haz the power, 2014-12-11 05:34:24]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/770383002/300001

[commit-bot: I haz the power, 2014-12-11 05:43:35]

Committed patchset #16 (id:300001) as
https://skia.googlesource.com/skia/+/3490263287e9432119c501884e45840de4d986bc