extend FontConfigInterface to start to match new fontmgr design

extend FontConfigInterface to start to match new fontmgr design
Committed: https://code.google.com/p/skia/source/detail?r=8575

[reed1, 2013-03-29 18:55:07]

trying to design this with chrome's IPC needs in mind

[bungeman-skia, 2013-03-29 19:40:36]

https://codereview.chromium.org/13297004/diff/1004/trunk/include/ports/SkFont...
File trunk/include/ports/SkFontConfigInterface.h (right):

https://codereview.chromium.org/13297004/diff/1004/trunk/include/ports/SkFont...
trunk/include/ports/SkFontConfigInterface.h:87: struct Style {
Should the style also contain a FontIdentity? Then you would have an opaque
thing to pass back to get that specific font.

[reed1, 2013-03-29 20:11:58]

https://codereview.chromium.org/13297004/diff/1004/trunk/include/ports/SkFont...
File trunk/include/ports/SkFontConfigInterface.h (right):

https://codereview.chromium.org/13297004/diff/1004/trunk/include/ports/SkFont...
trunk/include/ports/SkFontConfigInterface.h:87: struct Style {
On 2013/03/29 19:40:36, bungeman1 wrote:
> Should the style also contain a FontIdentity? Then you would have an opaque
> thing to pass back to get that specific font.

Done via matchFamilyStyle()

[reed1, 2013-03-29 20:26:29]

now with bigger, fatter, identities

[bungeman-skia, 2013-03-29 21:13:37]

I like this (PS5) the best so far, with the following musing.

https://codereview.chromium.org/13297004/diff/9001/trunk/include/ports/SkFont...
File trunk/include/ports/SkFontConfigInterface.h (right):

https://codereview.chromium.org/13297004/diff/9001/trunk/include/ports/SkFont...
trunk/include/ports/SkFontConfigInterface.h:94: };
Maybe we also want matchFamilyStyle which does 'system fallback' for styles,
whatever that may mean. This could be seen as a helper, in that the user doesn't
need to do their own matching. On the other hand, DirectWrite doesn't have the
notion of fallback or closest fit styles, it pretty much has exactly what we
have here. So we'd have to write something for that. Then again, maybe it's
better that we write that once than having everyone else implement over and
over.

[reed1, 2013-04-09 15:54:58]

Committed patchset #6 manually as r8575 (presubmit successful).