platform/upstream/libSkiaSharp.git
13 years agoMove PointTest.cpp out of tests/core into tests/ ; the other arrangement was
epoger@google.com [Wed, 15 Jun 2011 18:49:21 +0000 (18:49 +0000)]
Move PointTest.cpp out of tests/core into tests/ ; the other arrangement was
"clever" but was causing build problems.

git-svn-id: http://skia.googlecode.com/svn/trunk@1604 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agofix all_dependent_settings gyp bug (affected Mac only)
epoger@google.com [Wed, 15 Jun 2011 18:20:41 +0000 (18:20 +0000)]
fix all_dependent_settings gyp bug (affected Mac only)

git-svn-id: http://skia.googlecode.com/svn/trunk@1603 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoImplement SkPoint::Normalize() for SK_SCALAR_IS_FIXED and add performance warning...
epoger@google.com [Wed, 15 Jun 2011 18:04:58 +0000 (18:04 +0000)]
Implement SkPoint::Normalize() for SK_SCALAR_IS_FIXED and add performance warning to its declaration

http://codereview.appspot.com/4582043/

git-svn-id: http://skia.googlecode.com/svn/trunk@1602 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoImplement edge AA for concave polys in the tesselated path renderer.
senorblanco@chromium.org [Wed, 15 Jun 2011 17:52:09 +0000 (17:52 +0000)]
Implement edge AA for concave polys in the tesselated path renderer.

Review URL:  http://codereview.appspot.com/4571072/

git-svn-id: http://skia.googlecode.com/svn/trunk@1600 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agochange GC imagedecoder to use the same colorspace between image and context,
reed@google.com [Wed, 15 Jun 2011 17:33:29 +0000 (17:33 +0000)]
change GC imagedecoder to use the same colorspace between image and context,
so we get reproducible, unchanged, values.

use GC imagedecoder in gyp for mac

git-svn-id: http://skia.googlecode.com/svn/trunk@1599 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoAdd information to bench output.
bungeman@google.com [Wed, 15 Jun 2011 17:25:46 +0000 (17:25 +0000)]
Add information to bench output.
http://codereview.appspot.com/4602057/

git-svn-id: http://skia.googlecode.com/svn/trunk@1598 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoIncrease compatibility with clang, mostly around SK_RESTRICT.
tomhudson@google.com [Wed, 15 Jun 2011 16:50:27 +0000 (16:50 +0000)]
Increase compatibility with clang, mostly around SK_RESTRICT.
Attempting to fix up our use of __restrict__ so Chrome can enable it.

git-svn-id: http://skia.googlecode.com/svn/trunk@1597 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoImprovements to the SampleApp (primarily Android).
Scroggo [Wed, 15 Jun 2011 16:49:08 +0000 (16:49 +0000)]
Improvements to the SampleApp (primarily Android).

Reviewed at http://codereview.appspot.com/4587042/

Android
- Added buttons for interaction without a keyboard.
- Added the ability to zoom in to a specific point (roughly).
- Added event handling (for showing a slideshow, for example).
- Allow changing screen orientation
- Updated README file, explaining how to build

Multiplatform changes
- Added SampleApp header file
- Remove FPS when turning off measure FPS mode

git-svn-id: http://skia.googlecode.com/svn/trunk@1596 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agofix gm on Mac, and Linux buildbot build, by adding zlib source to third_party
epoger@google.com [Wed, 15 Jun 2011 16:05:37 +0000 (16:05 +0000)]
fix gm on Mac, and Linux buildbot build, by adding zlib source to third_party
http://codereview.appspot.com/4620047/

I will try to find a better way to keep all this working (remove zlib and
libpng from third_party)

git-svn-id: http://skia.googlecode.com/svn/trunk@1595 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoSet texture dirty flag only when texture has really changed
bsalomon@google.com [Wed, 15 Jun 2011 14:00:23 +0000 (14:00 +0000)]
Set texture dirty flag only when texture has really changed
Review URL: http://codereview.appspot.com/4571073/

git-svn-id: http://skia.googlecode.com/svn/trunk@1592 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoadd libpng source to repo and make Unix build use it
epoger@google.com [Wed, 15 Jun 2011 13:45:10 +0000 (13:45 +0000)]
add libpng source to repo and make Unix build use it
http://codereview.appspot.com/4572067

git-svn-id: http://skia.googlecode.com/svn/trunk@1591 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agostyle and warnings
reed@google.com [Tue, 14 Jun 2011 19:31:24 +0000 (19:31 +0000)]
style and warnings

git-svn-id: http://skia.googlecode.com/svn/trunk@1590 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoadd
reed@google.com [Tue, 14 Jun 2011 19:22:21 +0000 (19:22 +0000)]
add

git-svn-id: http://skia.googlecode.com/svn/trunk@1589 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoadd SkDataRef (for future use)
reed@google.com [Tue, 14 Jun 2011 19:18:26 +0000 (19:18 +0000)]
add SkDataRef (for future use)

git-svn-id: http://skia.googlecode.com/svn/trunk@1588 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoUpdate mac gpu baseline for SSAA on inverse paths
bsalomon@google.com [Tue, 14 Jun 2011 18:49:29 +0000 (18:49 +0000)]
Update mac gpu baseline for SSAA on inverse paths

git-svn-id: http://skia.googlecode.com/svn/trunk@1585 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoEnable SSAA on inverse filled paths
bsalomon@google.com [Tue, 14 Jun 2011 18:46:24 +0000 (18:46 +0000)]
Enable SSAA on inverse filled paths

Review URL: http://codereview.appspot.com/4613044/

git-svn-id: http://skia.googlecode.com/svn/trunk@1584 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agodelete all old (non-Gyp) buildfiles
epoger@google.com [Tue, 14 Jun 2011 18:35:08 +0000 (18:35 +0000)]
delete all old (non-Gyp) buildfiles

git-svn-id: http://skia.googlecode.com/svn/trunk@1583 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoFBO probing at Context startup is very expensive. This patch removes the
tomhudson@google.com [Tue, 14 Jun 2011 18:16:52 +0000 (18:16 +0000)]
FBO probing at Context startup is very expensive. This patch removes the
test of whether FBOs work at all, and adds parameters to the GrGLInterface
so that values can be passed into the Context creation routines instead
of executing the probes.

The GrGLInterface now defaults to run FBO tests on Android (since hardware
varies), but assumes success on desktop and software platforms.

git-svn-id: http://skia.googlecode.com/svn/trunk@1582 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agodelete manually-imported Gyp source
epoger@google.com [Tue, 14 Jun 2011 17:57:16 +0000 (17:57 +0000)]
delete manually-imported Gyp source
see http://codereview.appspot.com/4610045

git-svn-id: http://skia.googlecode.com/svn/trunk@1580 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoUpdated mac baselines from gl ctx size change and new version of shadertext with...
bsalomon@google.com [Tue, 14 Jun 2011 17:20:23 +0000 (17:20 +0000)]
Updated mac baselines from gl ctx size change and new version of shadertext with txt-on-path.

git-svn-id: http://skia.googlecode.com/svn/trunk@1578 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoUpdate baselines for linux-fixed. Make shader text sample/GM work with scalar=fixed.
bsalomon@google.com [Tue, 14 Jun 2011 17:13:22 +0000 (17:13 +0000)]
Update baselines for linux-fixed. Make shader text sample/GM work with scalar=fixed.

git-svn-id: http://skia.googlecode.com/svn/trunk@1577 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoUse automatically imported third_party/externals/gyp instead of
epoger@google.com [Tue, 14 Jun 2011 16:01:04 +0000 (16:01 +0000)]
Use automatically imported third_party/externals/gyp instead of
manually imported third_party/gyp

git-svn-id: http://skia.googlecode.com/svn/trunk@1576 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoFix text-as-path with skshader on gpu
bsalomon@google.com [Mon, 13 Jun 2011 21:55:32 +0000 (21:55 +0000)]
Fix text-as-path with skshader on gpu

Review URL: http://codereview.appspot.com/4576057/

git-svn-id: http://skia.googlecode.com/svn/trunk@1575 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agouint32_t -> int for texture extents
bsalomon@google.com [Mon, 13 Jun 2011 19:28:02 +0000 (19:28 +0000)]
uint32_t -> int for texture extents

Review URL: http://codereview.appspot.com/4584053/

git-svn-id: http://skia.googlecode.com/svn/trunk@1574 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoImplement timers for linux SampleApp. http://codereview.appspot.com/4592054/
Scroggo [Mon, 13 Jun 2011 19:17:58 +0000 (19:17 +0000)]
Implement timers for linux SampleApp. codereview.appspot.com/4592054/

git-svn-id: http://skia.googlecode.com/svn/trunk@1573 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agodrop extraneous GrPrintfs
reed@google.com [Mon, 13 Jun 2011 18:54:59 +0000 (18:54 +0000)]
drop extraneous GrPrintfs

git-svn-id: http://skia.googlecode.com/svn/trunk@1572 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoset externals/gyp dependency to same version as manually imported gyp code
epoger@google.com [Mon, 13 Jun 2011 18:45:15 +0000 (18:45 +0000)]
set externals/gyp dependency to same version as manually imported gyp code

git-svn-id: http://skia.googlecode.com/svn/trunk@1571 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoModify DEPS so that it works properly for the buildbot
epoger@google.com [Mon, 13 Jun 2011 18:08:41 +0000 (18:08 +0000)]
Modify DEPS so that it works properly for the buildbot

git-svn-id: http://skia.googlecode.com/svn/trunk@1570 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoAA updates for mac gm baseline WRT r1568
bsalomon@google.com [Mon, 13 Jun 2011 18:01:57 +0000 (18:01 +0000)]
AA updates for mac gm baseline WRT r1568

git-svn-id: http://skia.googlecode.com/svn/trunk@1569 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoConsider hw render target limit for offscreen supersample and tile
bsalomon@google.com [Mon, 13 Jun 2011 17:58:13 +0000 (17:58 +0000)]
Consider hw render target limit for offscreen supersample and tile

Review URL: http://codereview.appspot.com/4575041/

git-svn-id: http://skia.googlecode.com/svn/trunk@1568 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoUpdates to the Android SampleApp
djsollen@google.com [Mon, 13 Jun 2011 16:58:40 +0000 (16:58 +0000)]
Updates to the Android SampleApp

http://codereview.appspot.com/4602049/

git-svn-id: http://skia.googlecode.com/svn/trunk@1567 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoAdd DEPS file in an attempt to make both gclient and svn update bring in
epoger@google.com [Mon, 13 Jun 2011 16:04:25 +0000 (16:04 +0000)]
Add DEPS file in an attempt to make both gclient and svn update bring in
third_party/externals.

git-svn-id: http://skia.googlecode.com/svn/trunk@1566 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agomove SkMatrix44
reed@google.com [Mon, 13 Jun 2011 14:58:44 +0000 (14:58 +0000)]
move SkMatrix44

git-svn-id: http://skia.googlecode.com/svn/trunk@1565 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoadd row/col major getters, plus unit-tests
reed@google.com [Mon, 13 Jun 2011 14:46:52 +0000 (14:46 +0000)]
add row/col major getters, plus unit-tests

git-svn-id: http://skia.googlecode.com/svn/trunk@1564 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agomove into utils
reed@google.com [Mon, 13 Jun 2011 14:02:52 +0000 (14:02 +0000)]
move into utils

git-svn-id: http://skia.googlecode.com/svn/trunk@1563 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoadd 2nd parameter to SkFontHost::NextLogicalFont()
reed@google.com [Mon, 13 Jun 2011 13:01:10 +0000 (13:01 +0000)]
add 2nd parameter to SkFontHost::NextLogicalFont()
needed by android to map different styles of fallback fonts

git-svn-id: http://skia.googlecode.com/svn/trunk@1562 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years ago#include <math.h> if SK_SCALAR_IS_FIXED but SK_CAN_USE_FLOAT
epoger@google.com [Thu, 9 Jun 2011 22:16:28 +0000 (22:16 +0000)]
#include <math.h> if SK_SCALAR_IS_FIXED but SK_CAN_USE_FLOAT

git-svn-id: http://skia.googlecode.com/svn/trunk@1561 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoWhen user specifies an output path with -w, only output files from main
tomhudson@google.com [Thu, 9 Jun 2011 20:30:03 +0000 (20:30 +0000)]
When user specifies an output path with -w, only output files from main
render (don't output -replay and -serialize files into separate
directories).
If main render fails, don't test -replay or -serialize.

Also fix misnamed variable, line-wrap.

git-svn-id: http://skia.googlecode.com/svn/trunk@1560 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoupdate SkMatrix44
reed@google.com [Thu, 9 Jun 2011 19:13:41 +0000 (19:13 +0000)]
update SkMatrix44

git-svn-id: http://skia.googlecode.com/svn/trunk@1559 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoFix signed/unsigned mismatch in rev 1553.
tomhudson@google.com [Thu, 9 Jun 2011 18:54:01 +0000 (18:54 +0000)]
Fix signed/unsigned mismatch in rev 1553.

git-svn-id: http://skia.googlecode.com/svn/trunk@1558 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoDefine SK_CAN_USE_FLOAT in gyp build, just like Makefile build
epoger@google.com [Thu, 9 Jun 2011 18:51:41 +0000 (18:51 +0000)]
Define SK_CAN_USE_FLOAT in gyp build, just like Makefile build

git-svn-id: http://skia.googlecode.com/svn/trunk@1557 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoRemove SkTableMaskFilter.cpp from gyp build, because it was breaking for skia_scalar...
epoger@google.com [Thu, 9 Jun 2011 17:37:59 +0000 (17:37 +0000)]
Remove SkTableMaskFilter.cpp from gyp build, because it was breaking for skia_scalar=fixed (and it was not being exercised by the old Makefile anyway)

git-svn-id: http://skia.googlecode.com/svn/trunk@1555 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agocheck-point for unifying results between runs and mask backends for aa
reed@google.com [Thu, 9 Jun 2011 15:54:38 +0000 (15:54 +0000)]
check-point for unifying results between runs and mask backends for aa

git-svn-id: http://skia.googlecode.com/svn/trunk@1554 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoskdiff usability improvements: summary to stdout, -diff in output filenames,
tomhudson@google.com [Thu, 9 Jun 2011 15:47:10 +0000 (15:47 +0000)]
skdiff usability improvements: summary to stdout, -diff in output filenames,
better statistics on HTML page when using -white, output directory name is
optional.

codereview.appspot.com/4590048/

git-svn-id: http://skia.googlecode.com/svn/trunk@1553 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoAdding functionality for purgin graphics resource handles
junov@google.com [Wed, 8 Jun 2011 22:55:10 +0000 (22:55 +0000)]
Adding functionality for purgin graphics resource handles
This will allow a fix to WebKit for preventing a crash when the graphics
context is destroyed before the GrContext

BUG=http://code.google.com/p/chromium/issues/detail?id=85309

git-svn-id: http://skia.googlecode.com/svn/trunk@1551 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoadd skia_scalar switch to gyp script
epoger@google.com [Wed, 8 Jun 2011 20:55:53 +0000 (20:55 +0000)]
add skia_scalar switch to gyp script

git-svn-id: http://skia.googlecode.com/svn/trunk@1550 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoFix scale metric bench.
bungeman@google.com [Wed, 8 Jun 2011 19:38:53 +0000 (19:38 +0000)]
Fix scale metric bench.

git-svn-id: http://skia.googlecode.com/svn/trunk@1549 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agofix windows compile errors in "tests"
epoger@google.com [Wed, 8 Jun 2011 18:26:08 +0000 (18:26 +0000)]
fix windows compile errors in "tests"

git-svn-id: http://skia.googlecode.com/svn/trunk@1547 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agorevert to manually imported gyp source (from SVN externals)
epoger@google.com [Wed, 8 Jun 2011 15:34:21 +0000 (15:34 +0000)]
revert to manually imported gyp source (from SVN externals)

git-svn-id: http://skia.googlecode.com/svn/trunk@1541 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoDuplicates code from GrPathUtils to verify that an approximation for the number
tomhudson@google.com [Wed, 8 Jun 2011 14:46:28 +0000 (14:46 +0000)]
Duplicates code from GrPathUtils to verify that an approximation for the number
of points in a quadratic curve is within +/- 2x the value determined by the
previous expensive method.

Running a similar approximation method on the Guimark2 HTML5 Charting demo drops
the share of time spent in SkPoint::distanceToLineSegmentBetweenSqd() from 4.57%
to under 0.6%, although SkPath::Iter::next(), SkPath::lineTo(), and
GrPathUtils::quadraticPointCount() all increase a bit.

Using a similar approximation method for SampleSlides.cpp produces visually
reasonable results. Without a relevant gm (it looks like gm/pathfill.cpp doesn't
have explicit quadratics?) I'm not sure how to get a better output quality
test.

We could avoid code duplication by:
 - have two implementations in GrPathUtils (computedQuadraticPointCount() &
estimatedQuadraticPointCount() are my working titles)
 - use a #define to select between them at compile time
 - expose both of them in the header file for this test to access

git-svn-id: http://skia.googlecode.com/svn/trunk@1540 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoMake linux sample app work on my laptop. Reviewed at http://codereview.appspot.com...
Scroggo [Wed, 8 Jun 2011 14:26:00 +0000 (14:26 +0000)]
Make linux sample app work on my laptop. Reviewed at codereview.appspot.com/4568050/

git-svn-id: http://skia.googlecode.com/svn/trunk@1539 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoGCC 4.6 fix: intptr_t <-> NULL triggers an warning.
thakis@chromium.org [Tue, 7 Jun 2011 22:18:07 +0000 (22:18 +0000)]
GCC 4.6 fix: intptr_t <-> NULL triggers an warning.

BUG=chromium:87001
TEST=none

Patch by Ryan Sleevi <rsleevi@chromium.org>
Review URL: http://codereview.appspot.com/4571048/

git-svn-id: http://skia.googlecode.com/svn/trunk@1538 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoHigher resolution timers for bench.
bungeman@google.com [Tue, 7 Jun 2011 19:16:02 +0000 (19:16 +0000)]
Higher resolution timers for bench.
http://codereview.appspot.com/4548090/

git-svn-id: http://skia.googlecode.com/svn/trunk@1534 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agorebaselin from 1528
reed@google.com [Tue, 7 Jun 2011 18:48:32 +0000 (18:48 +0000)]
rebaselin from 1528

git-svn-id: http://skia.googlecode.com/svn/trunk@1533 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoApply translate param to path bounds when computing ssaa bounds
bsalomon@google.com [Tue, 7 Jun 2011 18:46:50 +0000 (18:46 +0000)]
Apply translate param to path bounds when computing ssaa bounds

Review URL: http://codereview.appspot.com/4591044/

git-svn-id: http://skia.googlecode.com/svn/trunk@1532 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agorebaseline from 1528
reed@google.com [Tue, 7 Jun 2011 18:43:05 +0000 (18:43 +0000)]
rebaseline from 1528

git-svn-id: http://skia.googlecode.com/svn/trunk@1531 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoround Y baseline for text, as we do in the raster case
reed@google.com [Tue, 7 Jun 2011 17:56:25 +0000 (17:56 +0000)]
round Y baseline for text, as we do in the raster case

http://code.google.com/p/chromium/issues/detail?id=83934

git-svn-id: http://skia.googlecode.com/svn/trunk@1528 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoTemporarily fall back to the old makefile; the new one was killing buildbot.
epoger@google.com [Tue, 7 Jun 2011 15:46:40 +0000 (15:46 +0000)]
Temporarily fall back to the old makefile; the new one was killing buildbot.

git-svn-id: http://skia.googlecode.com/svn/trunk@1527 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoSome final cleanups leading up to The Official Switch to Gyp
epoger@google.com [Tue, 7 Jun 2011 14:48:41 +0000 (14:48 +0000)]
Some final cleanups leading up to The Official Switch to Gyp
see http://codereview.appspot.com/4580043

1. Create a single public skia.gyp file that all outside projects (Chrome,
Android, etc.) should depend on from now on.  I haven't yet created targets
suitable for those projects to use, but this is where we should add them.
2. Make gyp generate its Makefiles within out/ directory, rather than polluting
directories under source control.
3. Modify trunk/Makefile to automatically run gyp_skia and then call the
generated Makefile, to ease developer transition.

git-svn-id: http://skia.googlecode.com/svn/trunk@1526 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoNew benchmarks to determine performance of matrix-point multiplication for floating...
tomhudson@google.com [Tue, 7 Jun 2011 14:31:38 +0000 (14:31 +0000)]
New benchmarks to determine performance of matrix-point multiplication for floating point vs. double matrices.

Over-the-shoulder review by reed@.

git-svn-id: http://skia.googlecode.com/svn/trunk@1525 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoadd SkFixedFloorToFixed
reed@google.com [Tue, 7 Jun 2011 13:38:39 +0000 (13:38 +0000)]
add SkFixedFloorToFixed

git-svn-id: http://skia.googlecode.com/svn/trunk@1524 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoCorrect argument order for drawSimpleRect()
thakis@chromium.org [Tue, 7 Jun 2011 04:03:17 +0000 (04:03 +0000)]
Correct argument order for drawSimpleRect()

Patch by Ryan Sleevi <rsleevi@chromium.org>
Review URL: http://codereview.appspot.com/4568055/

git-svn-id: http://skia.googlecode.com/svn/trunk@1523 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoCommitting on behalf of gman@chromium.org:
kbr@chromium.org [Tue, 7 Jun 2011 01:27:01 +0000 (01:27 +0000)]
Committing on behalf of gman@chromium.org:

Fix for putImageData

TEST=manually ran tests that were failing
BUG=84165
TBR=reed@google.com

Code review URL: http://codereview.appspot.com/4570042

git-svn-id: http://skia.googlecode.com/svn/trunk@1518 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoIncrease use of SkMatrix::hasPerspective(), merge in has_perspective().
tomhudson@google.com [Mon, 6 Jun 2011 19:11:19 +0000 (19:11 +0000)]
Increase use of SkMatrix::hasPerspective(), merge in has_perspective().
Unifies SkMatrix::hasPerspective(), has_perspective(), and manual tests of SkMatrix::getType & kPerspective_Mask.

git-svn-id: http://skia.googlecode.com/svn/trunk@1517 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoMake scalar benchmarks compile and run under SKIA_SCALAR=fixed as well as float.
tomhudson@google.com [Mon, 6 Jun 2011 17:55:11 +0000 (17:55 +0000)]
Make scalar benchmarks compile and run under SKIA_SCALAR=fixed as well as float.

git-svn-id: http://skia.googlecode.com/svn/trunk@1516 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoAttempt to correct issue with fixed point from SkMatrix changes in r1512.
tomhudson@google.com [Mon, 6 Jun 2011 17:21:44 +0000 (17:21 +0000)]
Attempt to correct issue with fixed point from SkMatrix changes in r1512.

git-svn-id: http://skia.googlecode.com/svn/trunk@1515 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoremove legacy LCDTEXT define
reed@google.com [Mon, 6 Jun 2011 17:15:46 +0000 (17:15 +0000)]
remove legacy LCDTEXT define

git-svn-id: http://skia.googlecode.com/svn/trunk@1514 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agomove code inside #ifdef to fix warning
reed@google.com [Mon, 6 Jun 2011 17:14:52 +0000 (17:14 +0000)]
move code inside #ifdef to fix warning

git-svn-id: http://skia.googlecode.com/svn/trunk@1513 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoBenchmarking suggests that replacing some of the conversions from float to int
tomhudson@google.com [Mon, 6 Jun 2011 15:18:28 +0000 (15:18 +0000)]
Benchmarking suggests that replacing some of the conversions from float to int
with direct floating-point comparisons in computeTypeMask() is a significant win
(up to 40%?).
In practice, we see a 10-15% speedup in this function (the current most
expensive Skia function) in Chrome on the FishIETank demo.

The remainder of the function looks like it ought to be amenable to similar
speedups, but we haven't found a rewrite that behaves well in benchmarking yet.

http://codereview.appspot.com/4551097/

git-svn-id: http://skia.googlecode.com/svn/trunk@1512 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoNew scalar benchmark suite, comparing SkScalarAs2sCompliment vs native float operations
tomhudson@google.com [Mon, 6 Jun 2011 15:16:31 +0000 (15:16 +0000)]
New scalar benchmark suite, comparing SkScalarAs2sCompliment vs native float operations
http://http://codereview.appspot.com/4532116/

git-svn-id: http://skia.googlecode.com/svn/trunk@1511 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoadd internal multiplier for loopcount, to reduce variance for simple/fast tests
reed@google.com [Mon, 6 Jun 2011 13:31:30 +0000 (13:31 +0000)]
add internal multiplier for loopcount, to reduce variance for simple/fast tests

git-svn-id: http://skia.googlecode.com/svn/trunk@1510 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agomake concat_float distinguishable using -match
reed@google.com [Mon, 6 Jun 2011 13:20:29 +0000 (13:20 +0000)]
make concat_float distinguishable using -match
init our float/double arrays, so we get reliable timings (NaNs are slow)

git-svn-id: http://skia.googlecode.com/svn/trunk@1509 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoMake gyp_skia script use third_party/externals/gyp instead of the deprecated
epoger@google.com [Fri, 3 Jun 2011 21:33:03 +0000 (21:33 +0000)]
Make gyp_skia script use third_party/externals/gyp instead of the deprecated
third_party/gyp

git-svn-id: http://skia.googlecode.com/svn/trunk@1506 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoAdd third_party/externals dir that brings in gyp
epoger@google.com [Fri, 3 Jun 2011 21:19:27 +0000 (21:19 +0000)]
Add third_party/externals dir that brings in gyp

git-svn-id: http://skia.googlecode.com/svn/trunk@1505 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoPut back our old gyp dir
epoger@google.com [Fri, 3 Jun 2011 21:00:42 +0000 (21:00 +0000)]
Put back our old gyp dir

git-svn-id: http://skia.googlecode.com/svn/trunk@1504 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agooops, that external dependency conflicted with the old gyp directory
epoger@google.com [Fri, 3 Jun 2011 20:59:48 +0000 (20:59 +0000)]
oops, that external dependency conflicted with the old gyp directory

git-svn-id: http://skia.googlecode.com/svn/trunk@1503 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoAdd gyp@r197 as external dependency
epoger@google.com [Fri, 3 Jun 2011 20:55:28 +0000 (20:55 +0000)]
Add gyp@r197 as external dependency

git-svn-id: http://skia.googlecode.com/svn/trunk@1502 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoMove third_party/gyp to third_party/gyp-old in anticipation of replacement
epoger@google.com [Fri, 3 Jun 2011 20:54:08 +0000 (20:54 +0000)]
Move third_party/gyp to third_party/gyp-old in anticipation of replacement

git-svn-id: http://skia.googlecode.com/svn/trunk@1501 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoRemove warnings, improve benchmark quality, complicate benchmark results:
tomhudson@google.com [Fri, 3 Jun 2011 20:32:58 +0000 (20:32 +0000)]
Remove warnings, improve benchmark quality, complicate benchmark results:
 - data on heap instead of stack
 - more closely match operations of actual SkMatrix::setConcat()
 - avoid random perturbations, do_always;
   use perf annotate to verify that assembly isn't getting optimized away

git-svn-id: http://skia.googlecode.com/svn/trunk@1500 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoadd SkAdvancedTypeface file
reed@google.com [Fri, 3 Jun 2011 20:15:56 +0000 (20:15 +0000)]
add SkAdvancedTypeface file

git-svn-id: http://skia.googlecode.com/svn/trunk@1499 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoallow the build system to enable offscreen_aa to some max dimension
reed@google.com [Fri, 3 Jun 2011 19:51:26 +0000 (19:51 +0000)]
allow the build system to enable offscreen_aa to some max dimension
default is still 0 for the moment.

git-svn-id: http://skia.googlecode.com/svn/trunk@1498 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years ago<body onunload="M_unloadPage();">
tomhudson@google.com [Fri, 3 Jun 2011 19:16:56 +0000 (19:16 +0000)]
<body onunload="M_unloadPage();">

<!-- Begin help window -->
<script type="text/javascript"><!--
var xsrfToken = 'dd5d1a57ca1d5440593050fdc596d377';
var helpDisplayed = false;
document.onclick = M_clickCommon;
var media_url = "/static/";
var base_url = "/";

var publish_link = "/4571045/publish";

// -->
</script>
<div id="help" style="display: none;">

<div style="font-size: medium; text-align: center;">Keyboard Shortcuts</div>
<hr>
<table width="100%">

  <tbody><tr valign="top">
      <td>
            <table width="100%">
            <tbody><tr>
              <td></td><th>File</th>
              </tr>
              <tr>
                <td class="shortcut"><span class="letter">u</span> <b>:</b></td><td>up to issue</td>
        </tr>
        <tr>
          <td class="shortcut"><span class="letter">m</span> <b>:</b></td><td>publish + mail comments</td>
          </tr>
   <tr>
  <td class="shortcut"><span class="letter">M</span> <b>:</b></td><td>edit review message</td>
   </tr>
        <tr>
          <td class="shortcut"><span class="letter">j</span> / <span class="letter">k</span> <b>:</b></td><td>jump to file after / before current file</td>
          </tr>
          <tr>
            <td class="shortcut"><span class="letter">J</span> / <span class="letter">K</span> <b>:</b></td><td>jump to next file with a comment after / before current file</td>
            </tr>
            <tr>
              <td></td><th>Side-by-side diff</th>
              </tr>
              <tr>
                <td class="shortcut"><span class="letter">i</span> <b>:</b></td><td>toggle intra-line diffs</td>
        </tr>
        <tr>
          <td class="shortcut"><span class="letter">e</span> <b>:</b></td><td>expand all comments</td>
          </tr>
          <tr>
            <td class="shortcut"><span class="letter">c</span> <b>:</b></td><td>collapse all comments</td>
            </tr>
            <tr>
              <td class="shortcut"><span class="letter">s</span> <b>:</b></td><td>toggle showing all comments</td>
              </tr>
              <tr>
                <td class="shortcut"><span class="letter">n</span> / <span class="letter">p</span> <b>:</b></td><td>next / previous diff chunk or comment</td>
        </tr>
        <tr>
          <td class="shortcut"><span class="letter">N</span> / <span class="letter">P</span> <b>:</b></td><td>next / previous comment</td>
          </tr>
          <tr>
            <td class="shortcut"><span class="letter">&lt;Enter&gt;</span> <b>:</b></td><td>respond to / edit current comment</td>
            </tr>
          </tbody></table>
      </td>
          <td>
        <table width="100%">
        <tbody><tr>
         <td></td><th>Issue</th>
         </tr>
         <tr>
           <td class="shortcut"><span class="letter">u</span> <b>:</b></td><td>up to list of issues</td>
           </tr>
           <tr>
             <td class="shortcut"><span class="letter">m</span> <b>:</b></td><td>publish + mail comments</td>
             </tr>
      <tr>
  <td class="shortcut"><span class="letter">j</span> / <span class="letter">k</span> <b>:</b></td><td>jump to patch after / before current patch</td>
   </tr>
<tr>
  <td class="shortcut"><span class="letter">o</span> / <span class="letter">&lt;Enter&gt;</span> <b>:</b></td><td>open current patch in side-by-side view</td>
   </tr>
<tr>
  <td class="shortcut"><span class="letter">i</span> <b>:</b></td><td>open current patch in unified diff view</td>
   </tr>
        <tr><td>&nbsp;</td></tr>
        <tr><td></td><th>Issue List</th></tr>
        <tr>
          <td class="shortcut"><span class="letter">j</span> / <span class="letter">k</span> <b>:</b></td><td>jump to issue after / before current issue</td>
          </tr>
          <tr>
            <td class="shortcut"><span class="letter">o</span> / <span class="letter">&lt;Enter&gt;</span> <b>:</b></td><td>open current issue</td>
            </tr>
            <tr>
       <td class="shortcut"><span class="letter">#</span> <b>:</b></td>
     <td>close issue</td>
   </tr>
        <tr><td>&nbsp;</td></tr>
        <tr>
          <td></td><th>Comment/message editing</th>
          </tr>
          <tr>
            <td class="shortcut"><span class="letter">&lt;Ctrl&gt;</span> + <span class="letter">s</span> <b>:</b></td><td>save comment</td>
            </tr>
            <tr>
              <td class="shortcut"><span class="letter">&lt;Esc&gt;</span> <b>:</b></td><td>cancel edit</td>
              </tr>
            </tbody></table>
        </td>
  </tr>

  </tbody></table>
  </div>
  <!-- End help window -->

  <div align="right">

  <div style="float:left; font-weight:bold"><i>Rietveld</i> Code Review Tool</div>

  <b>tomhudson@google.com (TomH)</b>
  |

  <a class="novisit" href="/settings">Settings</a>
  |

  <a class="novisit" target="_blank" href="http://code.google.com/p/rietveld/wiki/CodeReviewHelp">Help</a>
  |
  <a class="novisit" target="_blank" href="http://code.google.com/p/rietveld/issues/list">Bug tracker</a>
  |
  <a class="novisit" target="_blank" href="http://groups.google.com/group/codereview-discuss">Discussion group</a>
  |
  <a class="novisit" target="_blank" href="http://code.google.com/p/rietveld">Source code</a>
  |

  <a class="novisit" href="http://codereview.appspot.com/_ah/logout?continue=https://www.google.com/accounts/Logout%3Fcontinue%3Dhttps://appengine.google.com/_ah/logout%253Fcontinue%253Dhttp://codereview.appspot.com/4571045/%26service%3Dah">Sign out</a>

  </div>

  <div class="counter">(691)</div>
  <div class="mainmenu">

      <a href="/" class="active">Issues</a>
        <a href="/repos">Repositories</a>
  <a href="/search">Search</a>

  </div>
  <div class="mainmenu2">

        <a href="/use_uploadpy">Create Issue</a>

        |
    <a href="/mine">My Issues</a>
        |
    <a href="/starred">Starred</a>
        |
    <a href="/all">Open</a>
        |
    <a href="/all?closed=1">All</a>

      </div>

      <div>

        <h2>
    <span id="issue-star-4571045">

        <a href="javascript:M_addIssueStar(4571045)">
<img src="/static/star-dark.gif" width="15" height="15" border="0"></a>

      </span>

      <span class="issue-close" id="issue-close-4571045">
          <a href="javascript:M_closeIssue(4571045)">
        <img src="/static/close.gif" title="Close This Issue" width="15" height="15" border="0"></a>
</span>

    Issue <a href="/4571045/" onmouseover="M_showPopUp(this, 'popup-issue');">
                4571045</a>:
    New matrix benchmarks to evaluate ::setConcat implementation options
      </h2>

        <table class="issue-details" border="0" width="100%">
    <tbody><tr valign="top">

          <td class="meta" width="20%">

            <div>

              <a class="novisit" href="/4571045/edit">
                      Edit Issue
                    </a>

            <br>

              <a class="novisit" href="/4571045/publish">
                      Publish+Mail Comments
                    </a> ('m')

                  <br>
                <a class="novisit" href="/4571045/diff/1/bench/MatrixBench.cpp">
                <b>Start Review</b>
              </a>

            </div>

            <div class="issue_details_sidebar">
              <div><b>Created:</b><br>
                  16 minutes ago by me
            </div>
              <div><b>Modified:</b><br>
                  0 minutes ago
            </div>
              <div><b>Reviewers:</b><br>
                  <a href="/user/reed1" onmouseover="M_showUserInfoPopup(this)">reed1</a>
            </div>
              <div><b>CC:</b><br>

            </div>
              <div><b>Base URL:</b><br>
                  http://skia.googlecode.com/svn/trunk/
            </div>
              <div><b>Visibility:</b><br>

                  Public.

                </div>
        </div>
      </td>

            <td style="padding-left: .8em; padding-right: .8em;" width="80%">

      <h3><a id="issue-description-pointer" href="javascript:M_toggleSection('issue-description')" class="toggled-section opentriangle">
          Description</a></h3>
    <div id="issue-description" style="margin-left:15px;">
        <pre>On platforms that use Float (instead of Fixed), SkMatrix stores its internal
values as floats, but setConcat() needs extra precision and so uses doubles to
contain intermediate values.
These three benchmarks compare the speed of float-only, double-only, and
float-cast-to-double implementations of a chunk of code extracted from the
non-perspective case of setConcat().
</pre>
  </div>

    <h3>
        <a id="ps-1-pointer" href="javascript:M_toggleSectionForPS('4571045', '1')" class="toggled-section opentriangle">
      Patch Set 1

        </a>
  </h3>

      <div id="ps-1" style="">

  <div class="issue-list">

    <div class="pagination">
        <div style="float: left;">
      <i>Created:</i> 16 minutes ago
          </div>
      <div style="float: right;">

            <a href="/download/issue4571045_1.diff">
              Download raw patch set</a>

      </div>
          <div style="clear:both;"></div>
    </div>

      <table id="queues" style="clear:both;">
          <tbody><tr align="left">
        <th colspan="2"></th>
      <th>Unified diffs</th>
            <th>Side-by-side diffs</th>
          <th>Delta from patch set</th>
        <th colspan="3">Stats</th>
      <th>Patch</th>
          </tr>

            <tr name="patch">
            <td class="first" width="14"><img src="/static/closedtriangle.gif" style="" width="12" height="9"></td>
            <td style="white-space: nowrap">M    </td>
            <td>
              <a class="noul" href="/4571045/patch/1/2">
                  bench/MatrixBench.cpp
            </a>
            </td>
            <td>
              <a class="noul" href="/4571045/diff/1/bench/MatrixBench.cpp">
                  View
            </a>
            </td>
            <td style="white-space: nowrap">

            </td>
            <td style="white-space: nowrap">4 chunks</td>
            <td style="white-space: nowrap">+114 lines, -3 lines</td>
            <td style="white-space: nowrap">

                  0 comments

        </td>
        <td>
          <a href="/download/issue4571045_1_2.diff" title="Download patch for bench/MatrixBench.cpp">
              Download
                </a>
        </td>
      </tr>

    </tbody></table>
    </div>

        </div>

    <h3>
        <a id="messages-pointer" href="javascript:M_toggleSection('messages')" class="toggled-section opentriangle">
      Messages
          </a>
    </h3>

      <div><i>Total messages: 3</i></div>

        <div id="messages">
    <div style="margin-bottom: .5em;">
          <a href="javascript:M_showAllComments('cl', 3)">
          Expand All Messages</a>
        |
      <a href="javascript:M_hideAllComments('cl', 3)">
              Collapse All Messages</a>
          </div>

            <div class="message " id="msg-agpjb2RlcmV2aWV3chwLEgVJc3N1ZRil_5YCDAsSB01lc3NhZ2UY6QcM" name="0">
            <a name="msg-agpjb2RlcmV2aWV3chwLEgVJc3N1ZRil_5YCDAsSB01lc3NhZ2UY6QcM"></a>
            <div class="header">
              <table border="0" width="100%" cellspacing="0" cellpadding="0">
                  <tbody><tr class="comment_title" onclick="M_switchChangelistComment(0)">
                <td style="padding-left: 5px; white-space: nowrap;">
                <b>me</b>
              </td>
                    <td width="100%">
                    <table style="table-layout:fixed; white-space: nowrap;" width="100%">
                      <tbody><tr>
                          <td>
                        <div style="white-space: nowrap; overflow: hidden;" class="extra" id="cl-preview-0">

                         </div>
                     </td>
                       </tr>
                       </tbody></table>
                     </td>
                   <td align="right" style="white-space: nowrap; padding-right: 5px;">
                   13 minutes ago
                 </td>
             </tr>
               </tbody></table>
               </div>

               <div id="cl-comment-0" style="display: none;">
                 <div class="message-body">
             <pre name="cl-message-0"></pre>
               </div>
                 <div class="message-actions">

                   <a href="javascript:M_replyToMessage('0', '2011/06/03 18:57:35', 'TomH')" id="message-reply-href-0">Reply</a>
                 <textarea rows="7" cols="70" name="message" style="display:none"></textarea>
               <div id="message-reply-0" style="display:none;"></div>

             </div>
             </div>

           </div>

             <div class="message " id="msg-agpjb2RlcmV2aWV3chwLEgVJc3N1ZRil_5YCDAsSB01lc3NhZ2UY0Q8M" name="1">
             <a name="msg-agpjb2RlcmV2aWV3chwLEgVJc3N1ZRil_5YCDAsSB01lc3NhZ2UY0Q8M"></a>
             <div class="header">
               <table border="0" width="100%" cellspacing="0" cellpadding="0">
                   <tbody><tr class="comment_title" onclick="M_switchChangelistComment(1)">
                 <td style="padding-left: 5px; white-space: nowrap;">
                 <b>reed1</b>
               </td>
                     <td width="100%">
                     <table style="table-layout:fixed; white-space: nowrap;" width="100%">
                       <tbody><tr>
                           <td>
                         <div style="white-space: nowrap; overflow: hidden;" class="extra" id="cl-preview-1">
                           1. lets remove the "fix" from the function name. just muladdmul I think 2. we ...
                          </div>
                      </td>
                        </tr>
                </tbody></table>
              </td>
                    <td align="right" style="white-space: nowrap; padding-right: 5px;">
                    10 minutes ago
                  </td>
              </tr>
                </tbody></table>
        </div>

        <div id="cl-comment-1" style="display: none;">
          <div class="message-body">
              <pre name="cl-message-1">1. lets remove the "fix" from the function name. just muladdmul I think
      2. we can remove the return, since it will be compiled away anyway
      3. nice checkin comment!
      4. what are the results like?</pre>
                </div>
          <div class="message-actions">

                    <a href="javascript:M_replyToMessage('1', '2011/06/03 19:00:39', 'reed1')" id="message-reply-href-1">Reply</a>
                  <textarea rows="7" cols="70" name="message" style="display:none"></textarea>
                <div id="message-reply-1" style="display:none;"></div>

              </div>
              </div>

            </div>

      <div class="message " id="msg-agpjb2RlcmV2aWV3chwLEgVJc3N1ZRil_5YCDAsSB01lc3NhZ2UYuRcM" name="2">
              <a name="msg-agpjb2RlcmV2aWV3chwLEgVJc3N1ZRil_5YCDAsSB01lc3NhZ2UYuRcM"></a>
              <div class="header">
                <table border="0" width="100%" cellspacing="0" cellpadding="0">
            <tbody><tr class="comment_title" onclick="M_switchChangelistComment(2)">
                  <td style="padding-left: 5px; white-space: nowrap;">
                  <b>me</b>
                </td>
              <td width="100%">
                      <table style="table-layout:fixed; white-space: nowrap;" width="100%">
                        <tbody><tr>
                    <td>
                          <div style="white-space: nowrap; overflow: hidden; visibility: hidden;" class="extra" id="cl-preview-2">
                            &gt; 4. what are the results like? For posterity: tomhudson@tomhudson-zx600-linux:/usr/local/google/src/skia3$ out/bench/bench -match matrix_concat -repeat 1000 ...
                           </div>
                       </td>
                         </tr>
                 </tbody></table>
               </td>
                     <td align="right" style="white-space: nowrap; padding-right: 5px;">
                     0 minutes ago
                   </td>
               </tr>
                 </tbody></table>
         </div>

         <div id="cl-comment-2">
           <div class="message-body">
               <pre name="cl-message-2">&gt; 4. what are the results like?

       For posterity:

       tomhudson@tomhudson-zx600-linux:/usr/local/google/src/skia3$ out/bench/bench
       -match matrix_concat -repeat 1000
       skia bench: alpha=0xFF antialias=1 filter=0
       running bench [640 480]         matrix_concat_double  8888: msecs =  0.79   565:
       msecs =  0.79   GPU: msecs =  0.79
       running bench [640 480]    matrix_concat_floatdouble  8888: msecs =  0.97   565:
       msecs =  0.97   GPU: msecs =  0.97
       running bench [640 480]          matrix_concat_float  8888: msecs =  0.74   565:
       msecs =  0.73   GPU: msecs =  0.74</pre>
                 </div>
           <div class="message-actions">

                     <a href="javascript:M_replyToMessage('2', '2011/06/03 19:11:06', 'TomH')" id="message-reply-href-2">Reply</a>
                   <textarea rows="7" cols="70" name="message" style="display:none"></textarea>
                 <div id="message-reply-2" style="display:none;"></div>

               </div>
               </div>

             </div>

     <div>
           <a href="javascript:M_showAllComments('cl', 3)">
         Expand All Messages</a>
       |
             <a href="javascript:M_hideAllComments('cl', 3)">
             Collapse All Messages</a>
         </div>

   </div>

   <script language="JavaScript" type="text/javascript">
     <!--
         document.onkeydown = M_changelistKeyDown;
     var dashboardState = new M_DashboardState(window, 'patch', 'M_CLPatchMarker');
         M_switchChangelistCommentByAnchor();
   // -->
   </script>

     <div style="display:none;">
         <form method="POST" action="/4571045/publish" id="message-reply-form">
       <input type="hidden" name="xsrf_token" value="dd5d1a57ca1d5440593050fdc596d377">
             <div></div>
           <input type="hidden" name="subject" value="New matrix benchmarks to evaluate ::setConcat implementation options">
         <input type="hidden" name="message_only" value="1">
       <input type="submit" value="Send Message">
             <input type="button" value="Discard" name="discard">
           <input type="checkbox" name="send_mail" value="1" id="message-reply-send-mail" checked="checked">
         <label>Send mail to reviewers</label>
     </form>
       </div>
         <a id="resizer" class="resizer" style="display:none;cursor:pointer">
     <img src="/static/zippyplus.gif">
       </a>

             </td>
         </tr>
   </tbody></table>

   </div>

         <div class="popup" id="popup-issue" style="left: 104px; top: 116px; visibility: hidden; ">
       <b>Issue 4571045: New matrix benchmarks to evaluate ::setConcat implementation options
        </b><br>
      Created 16 minutes ago by me<br>
            Modified 0 minutes ago<br>
          Reviewers: reed1<br>
        Base URL: http://skia.googlecode.com/svn/trunk/<br>
      Comments: 0

        </div>

  <p></p>
  <div style="float: left;">
    <a target="_blank" href="http://code.google.com/appengine/"><img border="0" src="/static/appengine-noborder-120x30.gif" alt="Powered by Google App Engine"></a>
    </div>

    <div class="extra" style="font-size: 9pt; float: right; text-align: right;">
      <div style="height:14px;">
        <img src="/static/rss.gif" alt="RSS Feeds" width="14" height="14" align="top">
  <a href="/rss/all">Recent Issues</a>

      |
        <a href="/rss/mine/TomH">My Issues</a>
  |
    <a href="/rss/reviews/TomH">My Reviews</a>
      |
        <a href="/rss/closed/TomH">My Closed</a>

      |
        <a href="/rss/issue/4571045">This issue</a>

     </div>
       <div style="margin-top: .3em;">This is Rietveld <a href="http://code.google.com/p/rietveld/updates/list">r756
       </a></div>
       </div>

       <script type="text/javascript">
       var gaJsHost = (("https:" == document.location.protocol) ? "https://ssl." : "http://www.");
       document.write(unescape("%3Cscript src='" + gaJsHost + "google-analytics.com/ga.js' type='text/javascript'%3E%3C/script%3E"));
       </script><script src="http://www.google-analytics.com/ga.js" type="text/javascript"></script>
       <script type="text/javascript">
       var pageTracker = _gat._getTracker("UA-4803694-4");
       pageTracker._initData();
       pageTracker._trackPageview();
       </script>
On platforms that use Float (instead of Fixed), SkMatrix stores its internal
values as floats, but setConcat() needs extra precision and so uses doubles
to contain intermediate values.
These three benchmarks compare the speed of float-only, double-only, and
float-cast-to-double implementations of a chunk of code extracted from the
non-perspective case of setConcat().

git-svn-id: http://skia.googlecode.com/svn/trunk@1497 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agofix gm build on Mac
epoger@google.com [Fri, 3 Jun 2011 17:52:12 +0000 (17:52 +0000)]
fix gm build on Mac
see http://codereview.appspot.com/4571044/

git-svn-id: http://skia.googlecode.com/svn/trunk@1494 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agofix bench and gm targets on Linux
epoger@google.com [Fri, 3 Jun 2011 17:21:21 +0000 (17:21 +0000)]
fix bench and gm targets on Linux
see http://codereview.appspot.com/4530104

git-svn-id: http://skia.googlecode.com/svn/trunk@1490 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoTurn off optimization on VC debug builds.
bungeman@google.com [Fri, 3 Jun 2011 15:48:17 +0000 (15:48 +0000)]
Turn off optimization on VC debug builds.
http://codereview.appspot.com/4568052/

git-svn-id: http://skia.googlecode.com/svn/trunk@1488 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoCorrection of texture domain calculation. Review=http://codereview.appspot.com/4517126
twiz@google.com [Thu, 2 Jun 2011 20:30:02 +0000 (20:30 +0000)]
Correction of texture domain calculation.  Review=codereview.appspot.com/4517126

git-svn-id: http://skia.googlecode.com/svn/trunk@1483 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agofix gyp-generated xcode project files for SampleApp and other targets
epoger@google.com [Thu, 2 Jun 2011 19:52:14 +0000 (19:52 +0000)]
fix gyp-generated xcode project files for SampleApp and other targets
see http://codereview.appspot.com/4515181/

git-svn-id: http://skia.googlecode.com/svn/trunk@1482 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoAdd EGL context for Windows.
bungeman@google.com [Thu, 2 Jun 2011 15:38:06 +0000 (15:38 +0000)]
Add EGL context for Windows.
http://codereview.appspot.com/4517133/

git-svn-id: http://skia.googlecode.com/svn/trunk@1481 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoReduce GrGLProgram::ProgramDesc from 88B to 24B.
tomhudson@google.com [Thu, 2 Jun 2011 15:37:14 +0000 (15:37 +0000)]
Reduce GrGLProgram::ProgramDesc from 88B to 24B.
http://codereview.appspot.com/4561058.

git-svn-id: http://skia.googlecode.com/svn/trunk@1480 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agocreate GYP files for these targets: tests, gm, skhello, skimage, skdiff
epoger@google.com [Thu, 2 Jun 2011 14:38:23 +0000 (14:38 +0000)]
create GYP files for these targets: tests, gm, skhello, skimage, skdiff
see http://codereview.appspot.com/4536096/

git-svn-id: http://skia.googlecode.com/svn/trunk@1479 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoadd copyright notices
reed@google.com [Thu, 2 Jun 2011 14:08:49 +0000 (14:08 +0000)]
add copyright notices

git-svn-id: http://skia.googlecode.com/svn/trunk@1478 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoremove unneeded check for thumb. see http://codereview.appspot.com/4539048/
reed@google.com [Thu, 2 Jun 2011 13:45:33 +0000 (13:45 +0000)]
remove unneeded check for thumb. see codereview.appspot.com/4539048/

git-svn-id: http://skia.googlecode.com/svn/trunk@1477 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoAdd a missing include path for the linux build of utils.gyp. Remove the Linux
senorblanco@chromium.org [Wed, 1 Jun 2011 20:55:45 +0000 (20:55 +0000)]
Add a missing include path for the linux build of utils.gyp.  Remove the Linux
utils files from the build of views.a, since they're already being compiled into
utils.a, since it was causing unnecessary rebuilds (the same could probably be
done for linux and mac).  Remove the commented-out SkImageDecoder_libpvjpeg.cpp,
which has been removed from SVN.

http://codereview.appspot.com/4532105/

git-svn-id: http://skia.googlecode.com/svn/trunk@1476 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoFix to bug, partial fix to unit tests following on from
tomhudson@google.com [Wed, 1 Jun 2011 20:43:05 +0000 (20:43 +0000)]
Fix to bug, partial fix to unit tests following on from
http://codereview.appspot.com/4536100/

git-svn-id: http://skia.googlecode.com/svn/trunk@1475 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoinclude SkTypes.h
reed@google.com [Wed, 1 Jun 2011 20:27:53 +0000 (20:27 +0000)]
include SkTypes.h

git-svn-id: http://skia.googlecode.com/svn/trunk@1474 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agocommitting http://codereview.appspot.com/4515175/
caryclark@google.com [Wed, 1 Jun 2011 20:15:20 +0000 (20:15 +0000)]
committing http://codereview.appspot.com/4515175/

git-svn-id: http://skia.googlecode.com/svn/trunk@1473 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoSpeed up GrBinHashKey computation by replacing Adler32 with One-at-a-Time hash
tomhudson@google.com [Wed, 1 Jun 2011 19:27:31 +0000 (19:27 +0000)]
Speed up GrBinHashKey computation by replacing Adler32 with One-at-a-Time hash
function, do 32b at a time instead of 8b at a time.

git-svn-id: http://skia.googlecode.com/svn/trunk@1472 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoremove dead/abandoned file
reed@google.com [Wed, 1 Jun 2011 18:38:51 +0000 (18:38 +0000)]
remove dead/abandoned file

git-svn-id: http://skia.googlecode.com/svn/trunk@1471 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoClean up the gyp files to reflect some files that were removed and/or renamed.
senorblanco@chromium.org [Wed, 1 Jun 2011 18:28:22 +0000 (18:28 +0000)]
Clean up the gyp files to reflect some files that were removed and/or renamed.

Review URL:  http://codereview.appspot.com/4536098/

git-svn-id: http://skia.googlecode.com/svn/trunk@1470 2bbb7eff-a529-9590-31e7-b0007b416f81

13 years agoAdded code to enable the pipe module to record to a file
yangsu@google.com [Wed, 1 Jun 2011 18:03:34 +0000 (18:03 +0000)]
Added code to enable the pipe module to record to a file
http://codereview.appspot.com/4564052/

git-svn-id: http://skia.googlecode.com/svn/trunk@1469 2bbb7eff-a529-9590-31e7-b0007b416f81