Add reg test for http://crbug.com/348821

Add reg test for http://crbug.com/348821

BUG=348821
Committed: http://code.google.com/p/skia/source/detail?r=13852

[mtklein, 2014-03-07 14:47:44]

This test failed before the fix landed and now passes when synced past it.

[reed1, 2014-03-07 14:52:59]

https://codereview.chromium.org/189373008/diff/20001/tests/DashPathEffectTest...
File tests/DashPathEffectTest.cpp (right):

https://codereview.chromium.org/189373008/diff/20001/tests/DashPathEffectTest...
tests/DashPathEffectTest.cpp:13: const int count = 2;
I *thought* the bug was triggered by count < 0 ...?

https://codereview.chromium.org/189373008/diff/20001/tests/DashPathEffectTest...
tests/DashPathEffectTest.cpp:14: SkScalar phase = 1.0f/0.0f;  // Used to force
the bad fInitialDashLength = -1 path.
you may get compiler warnings about this construct.

fyi: we have SK_ScalarInfinity and SK_ScalarNaN for this purpose if you need it.

[mtklein, 2014-03-07 15:00:10]

https://codereview.chromium.org/189373008/diff/20001/tests/DashPathEffectTest...
File tests/DashPathEffectTest.cpp (right):

https://codereview.chromium.org/189373008/diff/20001/tests/DashPathEffectTest...
tests/DashPathEffectTest.cpp:13: const int count = 2;
On 2014/03/07 14:52:59, reed1 wrote:
> I *thought* the bug was triggered by count < 0 ...?

Nope.  This is the path effect:

(gdb) p *this
$1 = {
  <SkPathEffect> = {
    <SkFlattenable> = {
      <SkRefCnt> = {
        <SkRefCntBase> = {
          <SkNoncopyable> = {<No data fields>},
          members of SkRefCntBase:
          _vptr$SkRefCntBase = 0xd288648,
          fRefCnt = 2
        },
        members of SkRefCnt:
        flags_ = 1
      }, <No data fields>}, <No data fields>},
  members of SkDashPathEffect:
  fIntervals = 0x7baa61b0,
  fCount = 2,
  fInitialDashLength = -1,
  fInitialDashIndex = 0,
  fIntervalLength = 0,
  fScaleToFit = false
}
(gdb) p (SkScalar[2])this->fIntervals
$2 = {1.76934361e+36, 2.80259693e-45}

I didn't happen to record phase, so here I'm forcing that path with infinity.

https://codereview.chromium.org/189373008/diff/20001/tests/DashPathEffectTest...
tests/DashPathEffectTest.cpp:14: SkScalar phase = 1.0f/0.0f;  // Used to force
the bad fInitialDashLength = -1 path.
On 2014/03/07 14:52:59, reed1 wrote:
> you may get compiler warnings about this construct.
> 
> fyi: we have SK_ScalarInfinity and SK_ScalarNaN for this purpose if you need
it.
Duh, right.  Done.

[mtklein_C, 2014-03-18 10:55:27]

On 2014/03/07 15:00:10, mtklein wrote:
>
https://codereview.chromium.org/189373008/diff/20001/tests/DashPathEffectTest...
> File tests/DashPathEffectTest.cpp (right):
> 
>
https://codereview.chromium.org/189373008/diff/20001/tests/DashPathEffectTest...
> tests/DashPathEffectTest.cpp:13: const int count = 2;
> On 2014/03/07 14:52:59, reed1 wrote:
> > I *thought* the bug was triggered by count < 0 ...?
> 
> Nope.  This is the path effect:
> 
> (gdb) p *this
> $1 = {
>   <SkPathEffect> = {
>     <SkFlattenable> = {
>       <SkRefCnt> = {
>         <SkRefCntBase> = {
>           <SkNoncopyable> = {<No data fields>},
>           members of SkRefCntBase:
>           _vptr$SkRefCntBase = 0xd288648,
>           fRefCnt = 2
>         },
>         members of SkRefCnt:
>         flags_ = 1
>       }, <No data fields>}, <No data fields>},
>   members of SkDashPathEffect:
>   fIntervals = 0x7baa61b0,
>   fCount = 2,
>   fInitialDashLength = -1,
>   fInitialDashIndex = 0,
>   fIntervalLength = 0,
>   fScaleToFit = false
> }
> (gdb) p (SkScalar[2])this->fIntervals
> $2 = {1.76934361e+36, 2.80259693e-45}
> 
> I didn't happen to record phase, so here I'm forcing that path with infinity.
> 
>
https://codereview.chromium.org/189373008/diff/20001/tests/DashPathEffectTest...
> tests/DashPathEffectTest.cpp:14: SkScalar phase = 1.0f/0.0f;  // Used to force
> the bad fInitialDashLength = -1 path.
> On 2014/03/07 14:52:59, reed1 wrote:
> > you may get compiler warnings about this construct.
> > 
> > fyi: we have SK_ScalarInfinity and SK_ScalarNaN for this purpose if you need
> it.
> Duh, right.  Done.

Ping?

[reed1, 2014-03-18 14:33:06]

https://codereview.chromium.org/189373008/diff/40001/tests/DashPathEffectTest...
File tests/DashPathEffectTest.cpp (right):

https://codereview.chromium.org/189373008/diff/40001/tests/DashPathEffectTest...
tests/DashPathEffectTest.cpp:6: #define ASSERT(x) REPORTER_ASSERT(r, x)
Hmm, I find this confusing at the call-sites, as you're reusing a classic word
(ASSERT) that implies crash (at least in debug), when what you mean is notify
the reporter.

[mtklein, 2014-03-18 17:08:42]

On 2014/03/18 14:33:06, reed1 wrote:
>
https://codereview.chromium.org/189373008/diff/40001/tests/DashPathEffectTest...
> File tests/DashPathEffectTest.cpp (right):
> 
>
https://codereview.chromium.org/189373008/diff/40001/tests/DashPathEffectTest...
> tests/DashPathEffectTest.cpp:6: #define ASSERT(x) REPORTER_ASSERT(r, x)
> Hmm, I find this confusing at the call-sites, as you're reusing a classic word
> (ASSERT) that implies crash (at least in debug), when what you mean is notify
> the reporter.

Have another look?

[reed1, 2014-03-18 17:14:25]

lgtm

[mtklein, 2014-03-18 17:14:51]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-03-18 17:15:01]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/mtklein@chromium.org/189373008/60001

[commit-bot: I haz the power, 2014-03-18 17:38:38]

Change committed as 13852