[PDF] Always set the symbolic Flag for Fonts.

[PDF] Always set the symbolic Flag for Fonts.

This should fix Chrome bug 124572

Committed: https://code.google.com/p/skia/source/detail?r=8136

[bungeman, 2013-03-12 21:20:46]

https://codereview.appspot.com/7631045/diff/1/src/pdf/SkPDFFont.cpp
File src/pdf/SkPDFFont.cpp (right):

https://codereview.appspot.com/7631045/diff/1/src/pdf/SkPDFFont.cpp#newcode946
src/pdf/SkPDFFont.cpp:946: SkAdvancedTypefaceMetrics::kSymbolic_Style;
So it really depends on what you mean by symbolic here. It should be documented
and then the font hosts changed to set this bit correctly. FreeType always sets
this, GDI and DW set it when it's truetype, and Mac seems to set it when
kCTFontSymbolicClass is true, which is totally wrong. The issue (I think) is
that there has been confusion between symbol fonts (with glyph shapes which
don't correspond to the 'normal' glyph for a character code) and pdf 'symbolic'
fonts which are a different beast entirely.

If we're just always going to set this, then we should get rid of it from
advancedtypefacemetrics altogether, and remove the cruft from the font hosts.

[Steve VanDeBogart, 2013-03-12 22:36:01]

https://codereview.appspot.com/7631045/diff/1/src/pdf/SkPDFFont.cpp
File src/pdf/SkPDFFont.cpp (right):

https://codereview.appspot.com/7631045/diff/1/src/pdf/SkPDFFont.cpp#newcode946
src/pdf/SkPDFFont.cpp:946: SkAdvancedTypefaceMetrics::kSymbolic_Style;
On 2013/03/12 21:20:46, bungeman wrote:
> So it really depends on what you mean by symbolic here. It should be
documented
> and then the font hosts changed to set this bit correctly. FreeType always
sets
> this, GDI and DW set it when it's truetype, and Mac seems to set it when
> kCTFontSymbolicClass is true, which is totally wrong. The issue (I think) is
> that there has been confusion between symbol fonts (with glyph shapes which
> don't correspond to the 'normal' glyph for a character code) and pdf
'symbolic'
> fonts which are a different beast entirely.
> 
> If we're just always going to set this, then we should get rid of it from
> advancedtypefacemetrics altogether, and remove the cruft from the font hosts.

Done. (removed cruft)

[bungeman, 2013-03-13 02:37:36]

lgtm

[Steve VanDeBogart, 2013-03-13 20:01:59]

Committed patchset #3 manually as r8136 (presubmit successful).