Use explicit constructor-enums for SkMatrix44, either kUninitialized or kIdentity

Use explicit constructor-enums for SkMatrix44, either kUninitialized or kIdentity
for clarity, and to take advantage of the fact that most of the time, we actually
want kUninitialized, which is free (doesn't call setIdentity()) and we set the entire
matrix right after it comes into scope (e.g. set3x3, setTranslate). The plan is to
hide/remove the default constructor in Skia soon.
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=177163

[danakj, 2013-01-16 16:50:41]

LGTM

[danakj, 2013-01-16 16:52:02]

Please make the first line of the CL description a single sentence summary so we
get proper git logs. :)

[reed1, 2013-01-16 18:18:23]

On 2013/01/16 16:52:02, danakj wrote:
> Please make the first line of the CL description a single sentence summary so
we
> get proper git logs. :)

What does the git log show? Is it different from "Use explicit constructor-enums
for SkMatrix44, either kUninitialized or kIdentity"?

[danakj, 2013-01-16 18:26:02]

On 2013/01/16 18:18:23, reed1 wrote:
> On 2013/01/16 16:52:02, danakj wrote:
> > Please make the first line of the CL description a single sentence summary
so
> we
> > get proper git logs. :)
> 
> What does the git log show? Is it different from "Use explicit
constructor-enums
> for SkMatrix44, either kUninitialized or kIdentity"?

Ya, it would be that. And it works out reasonably well here due to where line
wrapping happens, but it's generally nicer to do

Short summary.
<blank>
Long description.

You'll see this is the general pattern in chromium as well as most git-hosted
projects.

[enne (OOO), 2013-01-16 18:31:10]

More details: http://www.tpope.net/node/106