ABANDONED: linux: Avoid bitmaps for Cambria, Calibri, and Consolas.

linux: Avoid bitmaps for Cambria, Calibri, and Consolas.

When antialiasing is enabled, avoid using embedded bitmaps
for several common fonts. Blink appears to incorrectly scale
the bitmaps, but it's unlikely that users with antialiasing
enabled would want to see differently-rendered bitmap text
at any sizes.

Note that subpixel positioning appears to be disabled in
Blink at sizes where embedded bitmaps are present even when
the bitmaps aren't used. I think that this is likely to be a
separate bug.

BUG=408059

[Daniel Erat, 2014-09-19 23:16:53]

derat@chromium.org changed reviewers:
+ asvitkine@chromium.org, bungeman@chromium.org

[Daniel Erat, 2014-09-19 23:16:53]

i've discussed this in more detail in http://crbug.com/408059#c25. i think that
it's desirable to avoid bitmaps for these fonts in general, but there are still
the mysteries of why the bitmaps are getting scaled and why subpixel positioning
is disabled at some sizes even when we don't use bitmaps (maybe skia forces it
off?).

[bungeman-chromium, 2014-09-22 14:43:35]

On 2014/09/19 23:16:53, Daniel Erat wrote:
> i've discussed this in more detail in http://crbug.com/408059#c25. i think
that
> it's desirable to avoid bitmaps for these fonts in general, but there are
still
> the mysteries of why the bitmaps are getting scaled and why subpixel
positioning
> is disabled at some sizes even when we don't use bitmaps (maybe skia forces it
> off?).

This is obviously just a short-term hack which hides an actual issue, and I
would prefer we didn't do this. I'll take a look into this today and see what's
actually happening.

[Daniel Erat, 2014-09-22 14:59:41]

On 2014/09/22 14:43:35, bungeman2 wrote:
> On 2014/09/19 23:16:53, Daniel Erat wrote:
> > i've discussed this in more detail in http://crbug.com/408059#c25. i think
> that
> > it's desirable to avoid bitmaps for these fonts in general, but there are
> still
> > the mysteries of why the bitmaps are getting scaled and why subpixel
> positioning
> > is disabled at some sizes even when we don't use bitmaps (maybe skia forces
it
> > off?).
> 
> This is obviously just a short-term hack which hides an actual issue, and I
> would prefer we didn't do this. I'll take a look into this today and see
what's
> actually happening.

great, thanks!

i'm still skeptical that most users want to use bitmaps at all for these fonts,
since they won't match the rendering settings that we use at other sizes.

[bungeman-chromium, 2014-09-22 20:28:30]

On 2014/09/22 14:59:41, Daniel Erat wrote:
> On 2014/09/22 14:43:35, bungeman2 wrote:
> > On 2014/09/19 23:16:53, Daniel Erat wrote:
> > > i've discussed this in more detail in http://crbug.com/408059#c25. i think
> > that
> > > it's desirable to avoid bitmaps for these fonts in general, but there are
> > still
> > > the mysteries of why the bitmaps are getting scaled and why subpixel
> > positioning
> > > is disabled at some sizes even when we don't use bitmaps (maybe skia
forces
> it
> > > off?).
> > 
> > This is obviously just a short-term hack which hides an actual issue, and I
> > would prefer we didn't do this. I'll take a look into this today and see
> what's
> > actually happening.
> 
> great, thanks!
> 
> i'm still skeptical that most users want to use bitmaps at all for these
fonts,
> since they won't match the rendering settings that we use at other sizes.

https://codereview.chromium.org/596513002 fixes the the underlying cause of the
motivation for this CL.