Do not invoke DrawFadedString() if it knows it does not fade.

Do not invoke DrawFadedString() if it knows it does not fade.

This is intentionally tiny and intended to be merged into M41.
views::Label already knows the size of the text (GetTextSize()
returns the cached size), therefore it knows the current text
does not fade at all.

Since this will avoid the check if the text may fade or not in
gfx::Canvas::DrawFadedString(), this can potentially make
the drawing 2x faster if the text doesn't fade at all.

BUG=451297
R=sky@chromium.org
TEST=on link_freon, with the bookmark sample in crbug.com/440551
  (contains ~15 bookmarks), repaint of bookmark bar is improved
  as: ~100msec -> ~55msec

Committed: https://crrev.com/887eb46392320c209e507880f8268111c804d99a
Cr-Commit-Position: refs/heads/master@{#312942}

[sky, 2015-01-23 17:55:59]

Why can't we do this check in canvas?

[Jun Mukai, 2015-01-23 18:16:10]

On 2015/01/23 17:55:59, sky wrote:
> Why can't we do this check in canvas?

The point is, canvas doesn't know the size of the text, so it needs to recompute
the size to check in canvas.
And we already do as
https://code.google.com/p/chromium/codesearch#chromium/src/ui/gfx/canvas_skia...
GetStringWidth() also creates RenderText and so on, so it takes the same time as
other DrawString methods.

We could avoid recomputing by reusing the same RenderText object, but I'm afraid
that that would take a bit bigger refactoring of canvas itself.

[sky, 2015-01-23 20:22:33]

Ok, makes sense, LGTM

[Jun Mukai, 2015-01-23 20:31:38]

The CQ bit was checked by mukai@chromium.org

[commit-bot: I haz the power, 2015-01-23 20:32:17]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/870903002/1

[commit-bot: I haz the power, 2015-01-23 22:02:34]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-01-23 22:04:15]

Patchset 1 (id:??) landed as
https://crrev.com/887eb46392320c209e507880f8268111c804d99a
Cr-Commit-Position: refs/heads/master@{#312942}