Add better bounds checks for getTime to fix perf debug assert below

Add better bounds checks for getTime to fix perf debug assert below

Due to rounding, we request a length slightly larger than the total
length in MeasureBench. This will be fixed in a following CL and there
will be another CL adding unit tests for bounds checking and other
problems.

Revert "Revert  386ba54 and  4ab47e0 : perf debug assert."

This reverts commit 69aaa5a49a10454d573cbd8c5d980029d78ae459.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2233983003

Committed: https://skia.googlesource.com/skia/+/5a4b18c6f27fb010954e6435ddf22b85a3b84b94

[Harry Stern, 2016-08-10 19:06:12]

Description was changed from

==========
Add better bounds checks for getTime to fix perf debug assert below

Due to rounding, we request a length slightly larger than the total
length in MeasureBench. This will be fixed in a following CL and there
will be another CL adding unit tests for bounds checking and other
problems.

Revert "Revert  386ba54 and  4ab47e0 : perf debug assert."

This reverts commit 69aaa5a49a10454d573cbd8c5d980029d78ae459.

BUG=skia:
==========

to

==========
Add better bounds checks for getTime to fix perf debug assert below

Due to rounding, we request a length slightly larger than the total
length in MeasureBench. This will be fixed in a following CL and there
will be another CL adding unit tests for bounds checking and other
problems.

Revert "Revert  386ba54 and  4ab47e0 : perf debug assert."

This reverts commit 69aaa5a49a10454d573cbd8c5d980029d78ae459.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2233983003
==========

[Harry Stern, 2016-08-10 19:09:24]

hstern@google.com changed reviewers:
+ caryclark@google.com, reed@google.com

[Harry Stern, 2016-08-10 19:09:25]

I thought I had tested on Debug but I guess not. I was doing something like
`numPieces = 10; pieceLength = totalLength/10; getTime(pieceLength*i) for i in
range(0, numPieces)" and the divide and then multiply has some rounding errors
which gives me an input greater than the total length at the end.

I'm just clamping out of bounds lengths to the endpoints now, as we do in
SkPathMeasure.

[reed1, 2016-08-10 19:17:08]

lgtm

[Harry Stern, 2016-08-10 19:54:15]

The CQ bit was checked by hstern@google.com

[commit-bot: I haz the power, 2016-08-10 19:54:22]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-10 23:31:16]

Description was changed from

==========
Add better bounds checks for getTime to fix perf debug assert below

Due to rounding, we request a length slightly larger than the total
length in MeasureBench. This will be fixed in a following CL and there
will be another CL adding unit tests for bounds checking and other
problems.

Revert "Revert  386ba54 and  4ab47e0 : perf debug assert."

This reverts commit 69aaa5a49a10454d573cbd8c5d980029d78ae459.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2233983003
==========

to

==========
Add better bounds checks for getTime to fix perf debug assert below

Due to rounding, we request a length slightly larger than the total
length in MeasureBench. This will be fixed in a following CL and there
will be another CL adding unit tests for bounds checking and other
problems.

Revert "Revert  386ba54 and  4ab47e0 : perf debug assert."

This reverts commit 69aaa5a49a10454d573cbd8c5d980029d78ae459.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2233983003

Committed:
https://skia.googlesource.com/skia/+/5a4b18c6f27fb010954e6435ddf22b85a3b84b94
==========

[commit-bot: I haz the power, 2016-08-10 23:31:17]

Committed patchset #1 (id:1) as
https://skia.googlesource.com/skia/+/5a4b18c6f27fb010954e6435ddf22b85a3b84b94