platform/upstream/libSkiaSharp.git
9 years agoAdd mixed samples support to XPs
cdalton [Mon, 8 Jun 2015 22:11:04 +0000 (15:11 -0700)]
Add mixed samples support to XPs

BUG=skia:

Review URL: https://codereview.chromium.org/1164973002

9 years agoRestore support for 10.6 SDK in header.
bungeman [Mon, 8 Jun 2015 21:47:06 +0000 (17:47 -0400)]
Restore support for 10.6 SDK in header.

Must use ApplicationServices, can't use CoreText directly.

9 years agoRemove attempted bug mitigation in Mac typeface.
bungeman [Mon, 8 Jun 2015 21:32:24 +0000 (14:32 -0700)]
Remove attempted bug mitigation in Mac typeface.

An attempt was made to hold onto the 'sbix' table when it existed
when drawing to mitigate an issue with vanishing font data. This
code has not affected the crash rate, so remove it.

Review URL: https://codereview.chromium.org/1168793003

9 years agoFix regression in GrXferProcessor
cdalton [Mon, 8 Jun 2015 20:43:33 +0000 (13:43 -0700)]
Fix regression in GrXferProcessor

Fix segfault introduced in https://codereview.chromium.org/1170553002/

BUG=skia:

Review URL: https://codereview.chromium.org/1167243002

9 years agoProvide tag-along for SkCreateTypefaceFromCTFont.
bungeman [Mon, 8 Jun 2015 20:39:12 +0000 (13:39 -0700)]
Provide tag-along for SkCreateTypefaceFromCTFont.

This provides a means of keeping an object alive for the entire
lifetime of an SkTypeface.

Review URL: https://codereview.chromium.org/1163573007

9 years agoSkPDF: fix name escape code
halcanary [Mon, 8 Jun 2015 19:14:56 +0000 (12:14 -0700)]
SkPDF: fix name escape code

BUG=494913

Review URL: https://codereview.chromium.org/1152283014

9 years agoApply coverage in XP base class when using dst reads
cdalton [Mon, 8 Jun 2015 19:14:44 +0000 (12:14 -0700)]
Apply coverage in XP base class when using dst reads

Moves the coverage logic into GrGLXferProcessor for XPs that perform
dst reads. XPs that don't use a dst read are still responsible to
handle coverage on their own.

BUG=skia:

Review URL: https://codereview.chromium.org/1170553002

9 years agochange SkDraw and all Blitters to use pixmap instead of bitmap
reed [Mon, 8 Jun 2015 17:47:12 +0000 (10:47 -0700)]
change SkDraw and all Blitters to use pixmap instead of bitmap

BUG=skia:

Review URL: https://codereview.chromium.org/1148793007

9 years agoAdd benjaminwagner@ to CQ_COMMITTERS.
rmistry [Mon, 8 Jun 2015 17:22:34 +0000 (10:22 -0700)]
Add benjaminwagner@ to CQ_COMMITTERS.

BUG=skia:
NOTRY=true

Review URL: https://codereview.chromium.org/1164933007

9 years agoRevert "Revert of move erase into SkPixmap (patchset #1 id:1 of https://codereview...
reed [Mon, 8 Jun 2015 17:22:22 +0000 (10:22 -0700)]
Revert "Revert of move erase into SkPixmap (patchset #1 id:1 of https://codereview.chromium.org/1161033005/)"

This reverts commit cca9b3ac908cf3e9a0c5774a5ffd118d8429e08a.

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1153473012

9 years agoRevert of move erase into SkPixmap (patchset #1 id:1 of https://codereview.chromium...
djsollen [Mon, 8 Jun 2015 17:03:49 +0000 (10:03 -0700)]
Revert of move erase into SkPixmap (patchset #1 id:1 of https://codereview.chromium.org/1161033005/)

Reason for revert:
broken builds

Original issue's description:
> move erase into SkPixmap
>
> harvested from https://codereview.chromium.org/1148793007
>
> BUG=skia:
> TBR=
>
> Committed: https://skia.googlesource.com/skia/+/0963f5dab079627c5523ce6a443af27a33e361f7

TBR=reed@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/1157303008

9 years agomove erase into SkPixmap
reed [Mon, 8 Jun 2015 16:16:53 +0000 (09:16 -0700)]
move erase into SkPixmap

harvested from https://codereview.chromium.org/1148793007

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1161033005

9 years agoCap color index values
robertphillips [Mon, 8 Jun 2015 14:10:49 +0000 (07:10 -0700)]
Cap color index values

In the provided example the color table has 10 entries but some of the pixels in the bitmap overflow. This CL goes through the pixel values and caps them to the max index.

An alternate approach would be to just have the color table always have 256 entries but zero out the unused ones.

BUG=492265

Review URL: https://codereview.chromium.org/1165493003

9 years agoAdd bigtileimagefilter GM & improvements to toString
robertphillips [Mon, 8 Jun 2015 13:21:14 +0000 (06:21 -0700)]
Add bigtileimagefilter GM & improvements to toString

This is all the ancillary code from https://codereview.chromium.org/1152553006/ (Fix dst bound reported by SkTileImageFilter).

TBR=reed@google.com, senorblanco@google.com

Review URL: https://codereview.chromium.org/1169713003

9 years agoAdded CQ config to skia repo
sergiyb [Mon, 8 Jun 2015 12:12:56 +0000 (05:12 -0700)]
Added CQ config to skia repo

This must land in sync with https://chromereviews.googleplex.com/204307013 and
revert in sync.

R=rmistry@chromium.org, tandrii@chromium.org
NOTRY=true
NOTREECHECKS=true
BUG=chromium:471767

Review URL: https://codereview.chromium.org/1153093004

9 years agoUpdate SKP version
skia.buildbots [Sun, 7 Jun 2015 07:20:33 +0000 (00:20 -0700)]
Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=
NO_MERGE_BUILDS

Review URL: https://codereview.chromium.org/1167563004

9 years agoadd extractSubset and SkAutoPixmapStorage
reed [Fri, 5 Jun 2015 21:33:17 +0000 (14:33 -0700)]
add extractSubset and SkAutoPixmapStorage

extracted from larger CL in progress: https://codereview.chromium.org/1148793007

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1162013008

9 years agoclarify that accessPixels is for write-access, and peekPixels is RO
reed [Fri, 5 Jun 2015 21:11:32 +0000 (14:11 -0700)]
clarify that accessPixels is for write-access, and peekPixels is RO

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1162203005

9 years agoAdd cmap 3,0 (symbol) as Unicode fallback.
bungeman [Fri, 5 Jun 2015 20:38:12 +0000 (13:38 -0700)]
Add cmap 3,0 (symbol) as Unicode fallback.

FreeType automatically selects the 'most Unicode' cmap available,
otherwise the default cmap is NULL. When this happens the fallback
should include the 3,0 symbol cmap, as is outlined in the Apple
TrueType documentation. This cmap should effectively be a Unicode
mapping, but exclusively in the private use area.

BUG=skia:1873,chromium:489452

Review URL: https://codereview.chromium.org/1163833003

9 years agoOpt for advanced blend at creation time in CrCustomXfermode
cdalton [Fri, 5 Jun 2015 20:26:20 +0000 (13:26 -0700)]
Opt for advanced blend at creation time in CrCustomXfermode

Modifies GrCustomXfermode to decide at creation time whether it will
use advanced blend, instead of deciding during getOptimizations.

BUG=skia:

Review URL: https://codereview.chromium.org/1161913003

9 years agodelete (unused) TransparentShader
reed [Fri, 5 Jun 2015 18:31:26 +0000 (11:31 -0700)]
delete (unused) TransparentShader

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1153193014

9 years agoAdd ColorMatrix filter tests from Android CTS.
djsollen [Fri, 5 Jun 2015 16:41:18 +0000 (09:41 -0700)]
Add ColorMatrix filter tests from Android CTS.

BUG=skia:3848

Review URL: https://codereview.chromium.org/1134753008

9 years agoUpdate how we package skia resource for android testing
djsollen [Fri, 5 Jun 2015 16:39:30 +0000 (09:39 -0700)]
Update how we package skia resource for android testing

Review URL: https://codereview.chromium.org/1119283004

9 years agouse pixmaps for dst in sprites -- NO BITMAPS
reed [Fri, 5 Jun 2015 13:58:22 +0000 (06:58 -0700)]
use pixmaps for dst in sprites -- NO BITMAPS

BUG=skia:
TBR=scroggo@google.com

Review URL: https://codereview.chromium.org/1143173011

9 years agoRemove memcmp from GrProgramDesc op== and Less
bsalomon [Thu, 4 Jun 2015 22:34:34 +0000 (15:34 -0700)]
Remove memcmp from GrProgramDesc op== and Less

Works around an ASAN complaint.

BUG=skia:3891

Review URL: https://codereview.chromium.org/1154773007

9 years agoPresubmit: add PRESUBMIT=false to non-master branch changes
borenet [Thu, 4 Jun 2015 22:15:42 +0000 (15:15 -0700)]
Presubmit: add PRESUBMIT=false to non-master branch changes

BUG=skia:

Review URL: https://codereview.chromium.org/1161973005

9 years agoStop running tiles_rt-gpu on the bots
bsalomon [Thu, 4 Jun 2015 22:10:45 +0000 (15:10 -0700)]
Stop running tiles_rt-gpu on the bots

BUG=skia:

Review URL: https://codereview.chromium.org/1152003009

9 years agoRemove unused PackBits methods and fix length checks
jschuh [Thu, 4 Jun 2015 22:10:37 +0000 (15:10 -0700)]
Remove unused PackBits methods and fix length checks

Also a bit of general cleanup.

BUG=chromium:486944

Review URL: https://codereview.chromium.org/1152163004

9 years agoAdd prelog flag to DM
bsalomon [Thu, 4 Jun 2015 21:15:33 +0000 (14:15 -0700)]
Add prelog flag to DM

Review URL: https://codereview.chromium.org/1158263004

9 years agoswitch bitmapshader internals over to pixmap
reed [Thu, 4 Jun 2015 21:12:25 +0000 (14:12 -0700)]
switch bitmapshader internals over to pixmap

BUG=skia:
NOTRY=True

Review URL: https://codereview.chromium.org/1158273007

9 years agoAdd documentation for SkiaLab
borenet [Thu, 4 Jun 2015 20:57:10 +0000 (13:57 -0700)]
Add documentation for SkiaLab

BUG=skia:3850
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=1154623006

Review URL: https://codereview.chromium.org/1154623006

9 years agoDisable unused attributes when using copy shader
bsalomon [Thu, 4 Jun 2015 20:57:00 +0000 (13:57 -0700)]
Disable unused attributes when using copy shader

BUG=chromium:496540

Review URL: https://codereview.chromium.org/1157943007

9 years agoRevert of Fix dst bound reported by SkTileImageFilter (patchset #3 id:40001 of https...
robertphillips [Thu, 4 Jun 2015 18:15:27 +0000 (11:15 -0700)]
Revert of Fix dst bound reported by SkTileImageFilter (patchset #3 id:40001 of https://codereview.chromium.org/1152553006/)

Reason for revert:
Blink

Original issue's description:
> Fix dst bound reported by SkTileImageFilter
>
> In the example from the bug we had the filter DAG:
>
> color filter (table)
>     0: xfermode filter (arith)
>         0: tile filter [0,80,34,114] -> [0,80,800,480]
>              0: color filter (table)
>                  0: bitmap src 34x34 -> [0,80,34,114]
>         1: color filter (table)
>              0: picture filter [0, 80, 800, 480]
>
> computeFastBounds was coming out of the DAG with a bound of [0,80,34,114] which didn't represent the pixels that would be drawn.
>
> This CL updates SkTileImageFilter to correctly set the bound for the pixels it will hit.
>
> BUG=493783
>
> Committed: https://skia.googlesource.com/skia/+/05be93bbdf09576f7903130e3b106b0a8c7c4b4e
>
> Committed: https://skia.googlesource.com/skia/+/0be685755f942baea26c66a87226b569fc17e960

TBR=reed@google.com,senorblanco@google.com,senorblanco@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=493783

Review URL: https://codereview.chromium.org/1156583004

9 years agomake test hittestpath random calls deterministic
caryclark [Thu, 4 Jun 2015 16:27:43 +0000 (09:27 -0700)]
make test hittestpath random calls deterministic

R=bsalomon@google.com

Review URL: https://codereview.chromium.org/1157943006

9 years agoTrivial change to bump minimum block size for Buffer pool
joshualitt [Thu, 4 Jun 2015 16:13:46 +0000 (09:13 -0700)]
Trivial change to bump minimum block size for Buffer pool

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/1146073004

9 years agoDisable dual source blending support when GLSL version is too old
bsalomon [Thu, 4 Jun 2015 15:49:34 +0000 (08:49 -0700)]
Disable dual source blending support when GLSL version is too old

BUG=skia:3891

Review URL: https://codereview.chromium.org/1160923007

9 years agoFallback bitmap devices should observe the computed pixel geometry
fmalita [Thu, 4 Jun 2015 14:52:41 +0000 (07:52 -0700)]
Fallback bitmap devices should observe the computed pixel geometry

When onCreateDevice() fails, internalSaveLayer() attempts to create a
fallback bitmap device instead - but it ignores the computed pixel geometry
in this case.

BUG=496054
R=reed@google.com

Review URL: https://codereview.chromium.org/1144153007

9 years agoRemove SK_ANDROID_FREETYPE_HAS_MM from build.
bungeman [Thu, 4 Jun 2015 14:38:52 +0000 (07:38 -0700)]
Remove SK_ANDROID_FREETYPE_HAS_MM from build.

This flag is now no longer used now that GX variations now seem
to have landed for good.

Review URL: https://codereview.chromium.org/1155303007

9 years agoAdd GM that tests clipping annotations
halcanary [Thu, 4 Jun 2015 14:26:54 +0000 (07:26 -0700)]
Add GM that tests clipping annotations

TBR=reed@google.com

BUG=skia:3880

Review URL: https://codereview.chromium.org/1159273003

9 years agoFix dst bound reported by SkTileImageFilter
robertphillips [Thu, 4 Jun 2015 13:53:37 +0000 (06:53 -0700)]
Fix dst bound reported by SkTileImageFilter

In the example from the bug we had the filter DAG:

color filter (table)
    0: xfermode filter (arith)
        0: tile filter [0,80,34,114] -> [0,80,800,480]
             0: color filter (table)
                 0: bitmap src 34x34 -> [0,80,34,114]
        1: color filter (table)
             0: picture filter [0, 80, 800, 480]

computeFastBounds was coming out of the DAG with a bound of [0,80,34,114] which didn't represent the pixels that would be drawn.

This CL updates SkTileImageFilter to correctly set the bound for the pixels it will hit.

BUG=493783

Committed: https://skia.googlesource.com/skia/+/05be93bbdf09576f7903130e3b106b0a8c7c4b4e

Review URL: https://codereview.chromium.org/1152553006

9 years agoFix edge-line artifacts issue on Mali
joel.liang [Thu, 4 Jun 2015 13:52:29 +0000 (06:52 -0700)]
Fix edge-line artifacts issue on Mali

Some GMs show edge-line artifacts on Mali at the bottom/right of the canvas.

Change the 'rtAdjustment' uniform precision from 'midiump' to 'highp' if the GPU vendor is ARM.

NOTE: Vertex Shader should have default 'highp' precision, but now we explicit declared some uniform/attribute precision as 'midiump'.

This patch should fix these issues:

Issue 3693: Some GMs produce random artifacts on Mali in MSAA4
    https://code.google.com/p/skia/issues/detail?id=3693

Issue 3685: srcmode GM has top/right line artifacts on Mali T604 in MSAA4
    https://code.google.com/p/skia/issues/detail?id=3685

BUG=skia:

Review URL: https://codereview.chromium.org/1154163012

9 years agorefactor bitmapshader to use a controller
reed [Thu, 4 Jun 2015 13:31:31 +0000 (06:31 -0700)]
refactor bitmapshader to use a controller

BUG=skia:

Review URL: https://codereview.chromium.org/1153123003

9 years agoremove stray printf
reed [Thu, 4 Jun 2015 13:27:44 +0000 (06:27 -0700)]
remove stray printf

BUG=skia:
TBR=junov@google.com

Review URL: https://codereview.chromium.org/1149083008

9 years agoRemove SK_FREETYPE_HAS_MM.
bungeman [Wed, 3 Jun 2015 19:58:40 +0000 (12:58 -0700)]
Remove SK_FREETYPE_HAS_MM.

The SK_FREETYPE_HAS_MM define was added to delay the changes to
support GX variation fonts on platforms which did not compile ftmm.c
into their FreeType build. Now that all builds should have this file
the define can be removed.

Committed: https://skia.googlesource.com/skia/+/4f61fee53af956a8beb4c56fa6479cb6b1aa5159

Review URL: https://codereview.chromium.org/1143133006

9 years agoReenable reordering
joshualitt [Tue, 2 Jun 2015 20:48:38 +0000 (13:48 -0700)]
Reenable reordering

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/1160063003

9 years agoBegin tracking driver info in GrGLContextInfo
cdalton [Tue, 2 Jun 2015 20:05:52 +0000 (13:05 -0700)]
Begin tracking driver info in GrGLContextInfo

BUG=skia:

Review URL: https://codereview.chromium.org/1165463005

9 years agoAdd adb_wait_for_device script
borenet [Tue, 2 Jun 2015 19:38:16 +0000 (12:38 -0700)]
Add adb_wait_for_device script

BUG=skia:3718

Review URL: https://codereview.chromium.org/1149163003

9 years agoFixing leaky handling of SkImage in SkDeferredCanvas.
junov [Tue, 2 Jun 2015 18:47:45 +0000 (11:47 -0700)]
Fixing leaky handling of SkImage in SkDeferredCanvas.

Long lived SkImageHeap objects currently accumulate refs indefinitely.
This leads to massive memory leaks in the gpu-accelerated 2D canvas
code path. This CL does not implement a general fix for SkGPipe, but
it resolves the leak in SkDeferredCanvas (currently the only user
of SkGPipe) by resetting the image heap when the deferral queue is
flushed. This change also fixes the accounting of bytes allocated
by referenced images in order to trigger flushing heuristics
appropriately.

BUG=crbug.com/494148

Review URL: https://codereview.chromium.org/1145893007

9 years agoRevert of Fix dst bound reported by SkTileImageFilter (patchset #3 id:40001 of https...
robertphillips [Tue, 2 Jun 2015 17:57:26 +0000 (10:57 -0700)]
Revert of Fix dst bound reported by SkTileImageFilter (patchset #3 id:40001 of https://codereview.chromium.org/1152553006/)

Reason for revert:
breaking tests

Original issue's description:
> Fix dst bound reported by SkTileImageFilter
>
> In the example from the bug we had the filter DAG:
>
> color filter (table)
>     0: xfermode filter (arith)
>         0: tile filter [0,80,34,114] -> [0,80,800,480]
>              0: color filter (table)
>                  0: bitmap src 34x34 -> [0,80,34,114]
>         1: color filter (table)
>              0: picture filter [0, 80, 800, 480]
>
> computeFastBounds was coming out of the DAG with a bound of [0,80,34,114] which didn't represent the pixels that would be drawn.
>
> This CL updates SkTileImageFilter to correctly set the bound for the pixels it will hit.
>
> BUG=493783
>
> Committed: https://skia.googlesource.com/skia/+/05be93bbdf09576f7903130e3b106b0a8c7c4b4e

TBR=reed@google.com,senorblanco@google.com,senorblanco@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=493783

Review URL: https://codereview.chromium.org/1143083006

9 years agoFix dst bound reported by SkTileImageFilter
robertphillips [Tue, 2 Jun 2015 17:52:43 +0000 (10:52 -0700)]
Fix dst bound reported by SkTileImageFilter

In the example from the bug we had the filter DAG:

color filter (table)
    0: xfermode filter (arith)
        0: tile filter [0,80,34,114] -> [0,80,800,480]
             0: color filter (table)
                 0: bitmap src 34x34 -> [0,80,34,114]
        1: color filter (table)
             0: picture filter [0, 80, 800, 480]

computeFastBounds was coming out of the DAG with a bound of [0,80,34,114] which didn't represent the pixels that would be drawn.

This CL updates SkTileImageFilter to correctly set the bound for the pixels it will hit.

BUG=493783

Review URL: https://codereview.chromium.org/1152553006

9 years agocheck (runtime) for null-pixels even when lock succeeds
reed [Tue, 2 Jun 2015 17:47:42 +0000 (10:47 -0700)]
check (runtime) for null-pixels even when lock succeeds

BUG=492818
TBR=

Review URL: https://codereview.chromium.org/1159953006

9 years agoUpdate XPF invariant info to not account for conflation
cdalton [Tue, 2 Jun 2015 17:43:39 +0000 (10:43 -0700)]
Update XPF invariant info to not account for conflation

Renames getInvariantOutput to getInvariantBlendedColor on GrXPFactory
and redefines it to not account for coverage conflation. This is the
information that all the callsites actually wanted to know.

BUG=skia:

Review URL: https://codereview.chromium.org/1161273005

9 years agowhitespace change to trigger android bots
joshualitt [Tue, 2 Jun 2015 17:35:04 +0000 (10:35 -0700)]
whitespace change to trigger android bots

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/1160853005

9 years agowhitespace change to trigger android bots
joshualitt [Tue, 2 Jun 2015 17:10:28 +0000 (10:10 -0700)]
whitespace change to trigger android bots

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/1164813002

9 years agodisable reordering until android bots clear up
joshualitt [Tue, 2 Jun 2015 16:10:31 +0000 (09:10 -0700)]
disable reordering until android bots clear up

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/1167693006

9 years agowhitespace change to get android bots to cycle
joshualitt [Tue, 2 Jun 2015 15:12:08 +0000 (08:12 -0700)]
whitespace change to get android bots to cycle

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/1155293008

9 years agofix dm crash
joshualitt [Tue, 2 Jun 2015 14:26:43 +0000 (07:26 -0700)]
fix dm crash

TBR=bsalomon@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/1149243006

9 years agore-enable reordering
joshualitt [Tue, 2 Jun 2015 14:09:31 +0000 (07:09 -0700)]
re-enable reordering

BUG=skia:

Review URL: https://codereview.chromium.org/1164763002

9 years agoRevert of Remove SK_FREETYPE_HAS_MM. (patchset #1 id:1 of https://codereview.chromium...
reed [Tue, 2 Jun 2015 01:40:52 +0000 (18:40 -0700)]
Revert of Remove SK_FREETYPE_HAS_MM. (patchset #1 id:1 of https://codereview.chromium.org/1143133006/)

Reason for revert:
seems to be breaking DEPS roll

FAILED: /b/build/goma/gomacc ../../third_party/android_tools/ndk/toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/arm-linux-androideabi-g++ -shared -Wl,--fatal-warnings -fPIC -Wl,-z,noexecstack -Wl,-z,now -Wl,-z,relro -Wl,-z,defs -fuse-ld=gold -Wl,--no-undefined -Wl,--exclude-libs=libgcc.a -Wl,--exclude-libs=libc++_static.a -Wl,--exclude-libs=libvpx_assembly_arm.a -Wl,--icf=safe -Wl,--warn-shared-textrel -nostdlib --sysroot=/b/build/slave/android_chromium_gn/build/src/third_party/android_tools/ndk/platforms/android-14/arch-arm  -Wl,--warn-shared-textrel -Wl,-O1 -Wl,--gc-sections -Wl,--as-needed -Wl,--version-script=/b/build/slave/android_chromium_gn/build/src/build/android/android_no_jni_exports.lst -L../../third_party/android_tools/ndk/sources/cxx-stl/llvm-libc++/libs/armeabi-v7a -o ./libskia.so -Wl,-soname=libskia.so @./libskia.so.rsp && { /b/build/goma/gomacc ../../third_party/android_tools/ndk/toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/arm-linux-androideabi-readelf -d ./libskia.so | grep SONAME ; /b/build/goma/gomacc ../../third_party/android_tools/ndk/toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/arm-linux-androideabi-nm -gD -f p ./libskia.so | cut -f1-2 -d' '; } > ./libskia.so.tmp && if ! cmp -s ./libskia.so.tmp ./libskia.so.TOC; then mv ./libskia.so.tmp ./libskia.so.TOC; fi && mkdir -p lib.stripped && ../../third_party/android_tools/ndk/toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/arm-linux-androideabi-strip --strip-unneeded -o lib.stripped/libskia.so.tmp libskia.so && if ! cmp -s lib.stripped/libskia.so.tmp lib.stripped/libskia.so; then mv lib.stripped/libskia.so.tmp lib.stripped/libskia.so; fi
../../third_party/skia/src/ports/SkFontHost_FreeType.cpp:283: error: undefined reference to 'FT_Get_MM_Var'
../../third_party/skia/src/ports/SkFontHost_FreeType.cpp:302: error: undefined reference to 'FT_Set_Var_Design_Coordinates'
../../third_party/skia/src/ports/SkFontHost_FreeType.cpp:1766: error: undefined reference to 'FT_Get_MM_Var'
collect2: error: ld returned 1 exit status

Original issue's description:
> Remove SK_FREETYPE_HAS_MM.
>
> The SK_FREETYPE_HAS_MM define was added to delay the changes to
> support GX variation fonts on platforms which did not compile ftmm.c
> into their FreeType build. Now that all builds should have this file
> the define can be removed.
>
> Committed: https://skia.googlesource.com/skia/+/4f61fee53af956a8beb4c56fa6479cb6b1aa5159

TBR=djsollen@google.com,bungeman@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1148303005

9 years agoFix for crashing bots
joshualitt [Mon, 1 Jun 2015 23:17:03 +0000 (16:17 -0700)]
Fix for crashing bots

TBR=bsalomon@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/1159183006

9 years agoSet GeometryBufferMapThreshold defaults
joshualitt [Mon, 1 Jun 2015 21:17:47 +0000 (14:17 -0700)]
Set GeometryBufferMapThreshold defaults

TBR=bsalomon@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/1161543003

9 years agosuppress chatty warning about Funkster.ttf not be supported (on mac at least)
reed [Mon, 1 Jun 2015 19:51:23 +0000 (12:51 -0700)]
suppress chatty warning about Funkster.ttf not be supported (on mac at least)

BUG=skia:

Review URL: https://codereview.chromium.org/1146273003

9 years agoRemove SK_FREETYPE_HAS_MM.
bungeman [Mon, 1 Jun 2015 19:05:23 +0000 (12:05 -0700)]
Remove SK_FREETYPE_HAS_MM.

The SK_FREETYPE_HAS_MM define was added to delay the changes to
support GX variation fonts on platforms which did not compile ftmm.c
into their FreeType build. Now that all builds should have this file
the define can be removed.

Review URL: https://codereview.chromium.org/1143133006

9 years agoRevert "Revert of Move copy-surface-as-draw fallback to GrGLGpu. (patchset #12 id...
bsalomon [Mon, 1 Jun 2015 17:41:49 +0000 (10:41 -0700)]
Revert "Revert of Move copy-surface-as-draw fallback to GrGLGpu. (patchset #12 id:220001 of https://codereview.chromium.org/1144433002/)"

This reverts commit 404b3b264b833eb7cffdc833c7e2ebbd708e123a.

Review URL: https://codereview.chromium.org/1161063003

9 years agoFix NULL access
bsalomon [Mon, 1 Jun 2015 17:22:48 +0000 (10:22 -0700)]
Fix NULL access

BUG=skia:

TBR=egdaniel@google.com

Review URL: https://codereview.chromium.org/1157363004

9 years agoanother double square
caryclark [Mon, 1 Jun 2015 17:21:31 +0000 (10:21 -0700)]
another double square

This is also a case where the square term is passed to a helper
function that squares the error term a second time.

R=reed@google.com
BUG=skia:3877

Review URL: https://codereview.chromium.org/1152623005

9 years agoExpand VisualBench to a real benching tool
joshualitt [Mon, 1 Jun 2015 17:03:54 +0000 (10:03 -0700)]
Expand VisualBench to a real benching tool

BUG=skia:

Review URL: https://codereview.chromium.org/1159213002

9 years ago/infra directory
nodir [Mon, 1 Jun 2015 16:39:11 +0000 (09:39 -0700)]
/infra directory

Added /infra directory, it will contain chrome-infra specific files, in
particular configs for chrome infra services.

/infra/project-config/cr-buildbucket.cfg defines buckets on
cr-buildbucket.appspot.com

R=borenet@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/1164633002

9 years agodon't square error term twice
caryclark [Mon, 1 Jun 2015 15:40:48 +0000 (08:40 -0700)]
don't square error term twice

The cubic code erroneously passed the error square term to a function
which in turn squares the error term.

R=reed@google.com
BUG=skia:3877

Review URL: https://codereview.chromium.org/1163843003

9 years agoRoll third_party/externals/gyp dd831fd:08429da
bungeman [Mon, 1 Jun 2015 14:25:17 +0000 (07:25 -0700)]
Roll third_party/externals/gyp dd831fd:08429da

Summary of changes available at:
https://chromium.googlesource.com/external/gyp/+log/dd831fd..08429da

Review URL: https://codereview.chromium.org/1162453005

9 years agoStretch small textures up to 16 pixels on PowerVR 54x
bsalomon [Mon, 1 Jun 2015 14:13:42 +0000 (07:13 -0700)]
Stretch small textures up to 16 pixels on PowerVR 54x

Review URL: https://codereview.chromium.org/1161183002

9 years agoadd res to stroke bench
caryclark [Mon, 1 Jun 2015 13:30:06 +0000 (06:30 -0700)]
add res to stroke bench

R=reed@google.com

Review URL: https://codereview.chromium.org/1158183003

9 years agoRefactor GrGpu path rendering functions to GrPathRendering
kkinnunen [Mon, 1 Jun 2015 08:37:26 +0000 (01:37 -0700)]
Refactor GrGpu path rendering functions to GrPathRendering

GrGpu, GrGLGpu, GrPathRendering, GrGLPathRendering all duplicated
each others' path rendering related member functions. Make
GrPathRending the logical extension of GrGpu by removing the methods
from GrGpu. Similarly to the GL variants.

Changes includes and forward declarations due to removing
GrDrawTarget.h include from GrGpu.h. This was necessary due to
GrDrawTarget.h including GrPathRendering.h, and now GrPathRendering.h
includes GrGpu.h. (Also GrGpu.h does not need GrDrawTarget.h anymore).

Very slight runtime improvement on x86_64:
             desk_fontwipe.skp_1  836us ->  841us 1.01x
            desk_silkfinance.skp_1 2.01ms -> 2.01ms 1x
             desk_forecastio.skp_1 7.03ms -> 7.05ms 1x
                desk_weather.skp_1 3.74ms -> 3.74ms 1x
                desk_twitter.skp_1 8.02ms -> 8.01ms 1x
                 desk_mapsvg.skp_1 5.24ms -> 5.23ms 1x
            desk_pokemonwiki.skp_1 7.06ms -> 7.03ms 1x
                 desk_sfgate.skp_1 3.19ms -> 3.17ms 1x
           tabl_cuteoverload.skp_1 2.66ms -> 2.64ms 0.99x
...
           tabl_transformice.skp_1 3.06ms -> 2.98ms 0.98x
         tabl_googlecalendar.skp_1 11.3ms ->   11ms 0.97x
            tabl_gamedeksiam.skp_1 12.8ms -> 12.4ms 0.97x
               desk_samoasvg.skp_1 10.6ms -> 10.3ms 0.97x
           tabl_worldjournal.skp_1 4.44ms ->  4.3ms 0.97x

Review URL: https://codereview.chromium.org/1157683006

9 years agoUpdate SKP version
skia.buildbots [Sun, 31 May 2015 07:49:10 +0000 (00:49 -0700)]
Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=
NO_MERGE_BUILDS

Review URL: https://codereview.chromium.org/1155133003

9 years agoRevert[6] of add asserts around results from requestLock
reed [Sat, 30 May 2015 20:41:15 +0000 (13:41 -0700)]
Revert[6] of add asserts around results from requestLock

This reverts commit 11e833d91b5ba2193ed593cb74900dddbec24b6f.

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1159183003

9 years agoRevert of Revert[4] of add asserts around results from requestLock (patchset #3 id...
reed [Sat, 30 May 2015 16:20:29 +0000 (09:20 -0700)]
Revert of Revert[4] of add asserts around results from requestLock (patchset #3 id:40001 of https://codereview.chromium.org/1159733006/)

Reason for revert:
gfx_unittests (under linux_asan)

@@@STEP_LOG_LINE@RenderTextTest.HarfBuzz_HorizontalPositions@Direct leak of 368 byte(s) in 1 object(s) allocated from:@@@
@@@STEP_LOG_LINE@RenderTextTest.HarfBuzz_HorizontalPositions@    #0 0x4cc74b in __interceptor_malloc (/b/build/slave/linux_asan/build/src/out/Release/gfx_unittests+0x4cc74b)@@@
@@@STEP_LOG_LINE@RenderTextTest.HarfBuzz_HorizontalPositions@    #1 0x7f2f7060ebdf in _cairo_image_surface_create_for_pixman_image /build/buildd/cairo-1.10.2/src/cairo-image-surface.c:158@@@
@@@STEP_LOG_LINE@RenderTextTest.HarfBuzz_HorizontalPositions@@@@
@@@STEP_LOG_LINE@RenderTextTest.HarfBuzz_HorizontalPositions@Indirect leak of 256 byte(s) in 1 object(s) allocated from:@@@
@@@STEP_LOG_LINE@RenderTextTest.HarfBuzz_HorizontalPositions@    #0 0x4cc74b in __interceptor_malloc (/b/build/slave/linux_asan/build/src/out/Release/gfx_unittests+0x4cc74b)@@@
@@@STEP_LOG_LINE@RenderTextTest.HarfBuzz_HorizontalPositions@    #1 0x7f2f6c7be26a in _pixman_image_allocate /build/buildd/pixman-0.30.2/build/pixman/../../pixman/pixman-image.c:184@@@
@@@STEP_LOG_LINE@RenderTextTest.HarfBuzz_HorizontalPositions@@@@

I think they are creating a cairo surface, but it has no pixels (zero size?). In this CL, if I see no pixels, I ignore the call-back which is used to free the surface (doh).

Original issue's description:
> Revert[4] of add asserts around results from requestLock
>
> This reverts commit 19663e54c017499406036746e7689193aa6417e6.
>
> BUG=skia:
> TBR=
>
> Committed: https://skia.googlesource.com/skia/+/df91b73a34e3a306c93a5e320704736255c3d9f0

TBR=reed@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/1151063005

9 years agoRevert[4] of add asserts around results from requestLock
reed [Sat, 30 May 2015 13:32:55 +0000 (06:32 -0700)]
Revert[4] of add asserts around results from requestLock

This reverts commit 19663e54c017499406036746e7689193aa6417e6.

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1159733006

9 years agoRevert of add asserts around results from requestLock (patchset #3 id:40001 of https...
reed [Sat, 30 May 2015 01:36:07 +0000 (18:36 -0700)]
Revert of add asserts around results from requestLock (patchset #3 id:40001 of https://codereview.chromium.or… (patchset #1 id:1 of https://codereview.chromium.org/1165583005/)

Reason for revert:
failure in cc_unittests (need to diagnose)

@@@STEP_LOG_LINE@ResourceProviderTests_ResourceProviderTest.TransferInvalidSoftware_1@ResourceProviderTests/ResourceProviderTest.TransferInvalidSoftware/1 (run #1):@@@
@@@STEP_LOG_LINE@ResourceProviderTests_ResourceProviderTest.TransferInvalidSoftware_1@[ RUN      ] ResourceProviderTests/ResourceProviderTest.TransferInvalidSoftware/1@@@
@@@STEP_LOG_LINE@ResourceProviderTests_ResourceProviderTest.TransferInvalidSoftware_1@[17114:17114:0529/180822:15336467671:INFO:SkPixelRef.cpp(168)] ../../third_party/skia/src/core/SkPixelRef.cpp:168: failed assertion "pixels"@@@

Original issue's description:
> add asserts around results from requestLock (patchset #3 id:40001 of https://codereview.chromium.org/1155403003/)"
>
> This reverts commit 3953d360417655b8000df0951699121383db45c3.
>
> BUG=skia:
> TBR=
>
> Committed: https://skia.googlesource.com/skia/+/6980f5a96455c8062403c995a64b654a0e9a1319

TBR=
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/1151573009

9 years agoadd asserts around results from requestLock (patchset #3 id:40001 of https://coderevi...
reed [Fri, 29 May 2015 22:46:36 +0000 (15:46 -0700)]
add asserts around results from requestLock (patchset #3 id:40001 of https://codereview.chromium.org/1155403003/)"

This reverts commit 3953d360417655b8000df0951699121383db45c3.

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1165583005

9 years agoRevert of add asserts around results from requestLock (patchset #3 id:40001 of https...
reed [Fri, 29 May 2015 21:22:05 +0000 (14:22 -0700)]
Revert of add asserts around results from requestLock (patchset #3 id:40001 of https://codereview.chromium.org/1155403003/)

Reason for revert:
asserts in ui/gfx unittests (need to investigate why)

[ RUN      ] RenderTextTest.SelectionKeepsLigatures
[14602:14602:0529/134016:16779526944:INFO:SkPixelRef.cpp(164)] ../../third_party/skia/src/core/SkPixelRef.cpp:164: failed assertion "pixels"

Original issue's description:
> add asserts around results from requestLock and lockPixels, ensuring that true always means we have non-null pixels (and non-null colortable if that matches the colortype)
>
> BUG= 491975
> TBR=
>
> Committed: https://skia.googlesource.com/skia/+/f941a68126d8fe647eaea902c244c466568b7809

TBR=scroggo@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG= 491975

Review URL: https://codereview.chromium.org/1159013006

9 years agoadd asserts around results from requestLock and lockPixels, ensuring that true always...
reed [Fri, 29 May 2015 18:39:14 +0000 (11:39 -0700)]
add asserts around results from requestLock and lockPixels, ensuring that true always means we have non-null pixels (and non-null colortable if that matches the colortype)

BUG= 491975
TBR=

Review URL: https://codereview.chromium.org/1155403003

9 years agoSkImage::NewFromYUVTexturesCopy
bsalomon [Fri, 29 May 2015 18:37:25 +0000 (11:37 -0700)]
SkImage::NewFromYUVTexturesCopy

Review URL: https://codereview.chromium.org/1149553002

9 years agoRemove GrContext from GrDrawTarget
bsalomon [Fri, 29 May 2015 17:26:19 +0000 (10:26 -0700)]
Remove GrContext from GrDrawTarget

Still exists on clip target (for now).

Review URL: https://codereview.chromium.org/1166553002

9 years agoRemove GrContext from GrClipMaskCache
bsalomon [Fri, 29 May 2015 16:45:57 +0000 (09:45 -0700)]
Remove GrContext from GrClipMaskCache

Review URL: https://codereview.chromium.org/1144013007

9 years agoAdd Em test font.
bungeman [Fri, 29 May 2015 15:42:21 +0000 (08:42 -0700)]
Add Em test font.

This test font contains
☓ U+2613 SALTIRE
⬛ U+2B1B BLACK LARGE SQUARE
⬜ U+2B1C WHITE LARGE SQUARE
All at full EM size (with standard amount above and below the
baseline).

Review URL: https://codereview.chromium.org/1134153002

9 years agouse inline version of eval quad if tangent is not needed
caryclark [Fri, 29 May 2015 15:11:16 +0000 (08:11 -0700)]
use inline version of eval quad if tangent is not needed

R=reed@google.com

Review URL: https://codereview.chromium.org/1160163002

9 years agoRemoving GrAutoMalloc
joshualitt [Fri, 29 May 2015 15:06:48 +0000 (08:06 -0700)]
Removing GrAutoMalloc

BUG=skia:

Review URL: https://codereview.chromium.org/1161993003

9 years agoAdd direct getter for GrCaps to GrContext.
bsalomon [Fri, 29 May 2015 15:02:10 +0000 (08:02 -0700)]
Add direct getter for GrCaps to GrContext.

TBR=joshualitt@google.com

Committed: https://skia.googlesource.com/skia/+/9138c46e572085870638b6f7ad7fcdfcdf3cac99

Review URL: https://codereview.chromium.org/1149773005

9 years agoadd bench for building strokes
reed [Fri, 29 May 2015 14:17:16 +0000 (07:17 -0700)]
add bench for building strokes

BUG=skia:

Review URL: https://codereview.chromium.org/1160093004

9 years agodelete experimental path renderers
joshualitt [Fri, 29 May 2015 14:09:09 +0000 (07:09 -0700)]
delete experimental path renderers

BUG=skia:

Review URL: https://codereview.chromium.org/1160863006

9 years agoadd context override of GeometryBufferMapThreshold
joshualitt [Fri, 29 May 2015 13:46:47 +0000 (06:46 -0700)]
add context override of GeometryBufferMapThreshold

BUG=skia:

Review URL: https://codereview.chromium.org/1159713006

9 years agoadd assert that the pixelref really did success in requestlock
reed [Thu, 28 May 2015 21:06:06 +0000 (14:06 -0700)]
add assert that the pixelref really did success in requestlock

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1142343010

9 years agoRevert of small CL to experiment with 'disabling' memory mapping (patchset #1 id...
joshualitt [Thu, 28 May 2015 20:40:20 +0000 (13:40 -0700)]
Revert of small CL to experiment with 'disabling' memory mapping (patchset #1 id:1 of https://codereview.chromium.org/1160003004/)

Reason for revert:
experiment complete

Original issue's description:
> small CL to experiment with 'disabling' memory mapping
>
> TBR=bsalomon@google.com
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/2e48e1514396b363ecd19372eab637b11bfb8a82

TBR=robertphillips@google.com,joshualitt@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/1159133004

9 years agoWhitespace change to test bugdroid
rmistry [Thu, 28 May 2015 18:59:07 +0000 (11:59 -0700)]
Whitespace change to test bugdroid

BUG=493278
BUG=skia:2139

NOTRY=true
TBR=

Review URL: https://codereview.chromium.org/1166443002

9 years agosmall CL to experiment with 'disabling' memory mapping
joshualitt [Thu, 28 May 2015 18:53:49 +0000 (11:53 -0700)]
small CL to experiment with 'disabling' memory mapping

TBR=bsalomon@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/1160003004

9 years agoRevert of Add direct getter for GrCaps to GrContext. (patchset #4 id:60001 of https...
bsalomon [Thu, 28 May 2015 15:20:58 +0000 (08:20 -0700)]
Revert of Add direct getter for GrCaps to GrContext. (patchset #4 id:60001 of https://codereview.chromium.org/1149773005/)

Reason for revert:
Breaking

Original issue's description:
> Add direct getter for GrCaps to GrContext.
>
> TBR=joshualitt@google.com
>
> Committed: https://skia.googlesource.com/skia/+/9138c46e572085870638b6f7ad7fcdfcdf3cac99

TBR=joshualitt@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1164443002

9 years agoAdd direct getter for GrCaps to GrContext.
bsalomon [Thu, 28 May 2015 15:13:16 +0000 (08:13 -0700)]
Add direct getter for GrCaps to GrContext.

TBR=joshualitt@google.com

Review URL: https://codereview.chromium.org/1149773005

9 years agoBreak GrTextContext's reliance on SkGpuDevice
robertphillips [Thu, 28 May 2015 14:45:59 +0000 (07:45 -0700)]
Break GrTextContext's reliance on SkGpuDevice

This CL seems to have 2 main downsides:

1) It duplicates some code in SkBaseDevice::filterTextFlags
2) It makes it tougher to derive from SkGpuDevice

It seems reasonable (at least to me) that the TextContexts get the power to reset the LCD flags.

Review URL: https://codereview.chromium.org/1159973002

9 years agosimplify RawIter - don't return a pt in kClose
reed [Thu, 28 May 2015 02:53:25 +0000 (19:53 -0700)]
simplify RawIter - don't return a pt in kClose

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1156893003

9 years agoImplement Porter Duff XP with a blend table
cdalton [Wed, 27 May 2015 22:08:33 +0000 (15:08 -0700)]
Implement Porter Duff XP with a blend table

Removes the runtime logic used by PorterDuffXferProcessor to decide
blend coeffs and shader outputs, and instead uses a compile-time
constant table of pre-selected blend formulas. Separates out the dst
read fallback into its own XP.

Introduces a new blend strategy for srcCoeff=0 that can apply coverage
with a reverse subtract blend equation instead of dual source
blending.

Adds new macros in GrBlend.h to analyze blend formulas both runtime.

Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no
longer used.

Adds a GM that verifies all xfermodes, including arithmetic, with the
color/coverage invariants used by Porter Duff.

Adds a unit test that verifies each Porter Duff formula with every
color/coverage invariant.

Major changes:

 * Uses a reverse subtract blend equation for coverage when srcCoeff=0
   (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't
   support dual source blending no longer require a dst copy for
   dst-in and modulate.

 * Sets BlendInfo::fWriteColor to false when the blend does not modify
   the dst. GrGLGpu will now use glColorMask instead of blending for
   these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]).

 * Converts all SA blend coeffs to One for opaque inputs, and ISA to
   Zero if there is also no coverage. (We keep ISA around when there
   is coverage because we use it to tweak alpha for coverage.)

 * Abandons solid white optimizations for the sake of simplicity
   (screen was the only mode that previous had solid white opts).

Minor differences:

 * Inconsequential differences in opt flags (e.g. we now return
   kCanTweakAlphaForCoverage_OptFlag even when there is no coverage).

 * Src coeffs when the shader outputs 0.

 * IS2C vs IS2A when the secondary output is scalar.

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414

Review URL: https://codereview.chromium.org/1124373002