When calling cubicTo(a, b, c) and if the distance between fPrevPt and a
is too small, b is used instead of a to calculate the first tangent,
even if the distance between fPrevPt and b is too small.
In debug mode, this is causing an assertion to fail in
SkPathStroker::preJoinTo() and, in Release, the use of an
unitialized value.
The first patch set is adding a failing test.
The second one add the fix to SkPathStroker::cubicTo()
BUG=skia:2820
R=bsalomon@chromium.org, junov@chromium.org, reed@google.com, caryclark@google.com, bsalomon@google.com
Author: piotaixr@chromium.org
Review URL: https://codereview.chromium.org/460813002
Implement support for NV_path_rendering on OpenGL ES. Use
glProgramPathFragmentInputGenNV function call instead of glPathTexGenNV to
communicate transforms to fragment shader.
The intention is that the NVPR paths will be drawn with the same shader program
as non-NVPR geometry. For NVPR calls, the GPU will skip the vertex shader and
just run the fragment shader.
After program is linked, query the locations of the fragment shader inputs with
glGetResourceLocation. The location will be used to set the transforms with
glProgramPathFragmentInputGenNV.
The functions and their workings are documented in:
glProgramPathFragmentInputGenNV
https://www.opengl.org/registry/specs/NV/path_rendering.txt
(note: addition as of API version 1.3)
glGetResourceLocation
https://www.opengl.org/registry/specs/ARB/program_interface_query.txthttp://www.opengl.org/registry/doc/glspec44.core.pdf
(function is in core Open GL 4.4)
Note: glProgramPathFragmentInputGenNV could be used also for OpenGL. However,
using seems to trigger a bug in the driver. Disable this feature on OpenGL at
least until the driver is fixed and released. The bug manifests in shadertext
test, where the lower-left text pair is missing. Valgrind catches a bad read
for the test and causes the context to OOM reproducibly.
R=bsalomon@google.com, cdalton@nvidia.com, joshualitt@google.com, joshualitt@chromium.org
Author: kkinnunen@nvidia.com
Review URL: https://codereview.chromium.org/367643004
morphology, blur.
[Reland with fixed tests.]
Because we're computing "backwards" from a clip rect of destination
pixels to be filled to the required source pixels, we should use tail
recursion rather than head recursion in onFilterBounds().
This actually only makes a difference for drop-shadow, where
the computation is non-commutative. Blur and morphology commute, but I
moved them to tail recursion anyway for clarity (so all onFilterBounds
use tail recursion).
BUG=skia:
R=bsalomon@google.com
Author: senorblanco@chromium.org
Review URL: https://codereview.chromium.org/481273005
Reason for revert:
This CL is currently breaking the Win7 and Win8 bots on some of the new tests (in DM).
Original issue's description:
> Fix recursive computation of filter bounds for drop shadow,
> morphology, blur.
>
> Because we're computing "backwards" from a clip rect of destination
> pixels to be filled to the required source pixels, we should use tail
> recursion rather than head recursion in onFilterBounds().
>
> This actually only makes a difference for drop-shadow, where
> the computation is non-commutative. Blur and morphology commute, but I
> moved them to tail recursion anyway for clarity (so all onFilterBounds
> use tail recursion).
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/28648fe4a69b0cee8df42b5966e4e645c3aabefbR=bsalomon@google.com, senorblanco@chromium.orgTBR=bsalomon@google.com, senorblanco@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Author: djsollen@google.com
Review URL: https://codereview.chromium.org/504773003
morphology, blur.
Because we're computing "backwards" from a clip rect of destination
pixels to be filled to the required source pixels, we should use tail
recursion rather than head recursion in onFilterBounds().
This actually only makes a difference for drop-shadow, where
the computation is non-commutative. Blur and morphology commute, but I
moved them to tail recursion anyway for clarity (so all onFilterBounds
use tail recursion).
BUG=skia:
R=bsalomon@google.com
Author: senorblanco@chromium.org
Review URL: https://codereview.chromium.org/481273005
In SkBBoxHierarchyRecord:
Since the bounds we pass to saveLayer are in the pre-filtering
coordinate space, they aren't correct for determining the actual
device pixels touched by the saveLayer in this case.
The easiest fix for now is to pass the clip bounds, since the final
draw done in restore() will never draw outside the clip.
In SkRecordDraw:
We do adjust the bounds passed to saveLayer, so we just need to make
sure that when we're using a paint that may affect transparent black,
we ignore the calculated bounds of draw ops and use the clip intersected
with those adjusted bounds.
See originally crrev.com/497773002
BUG=skia:
R=reed@google.com, senorblanco@chromium.org, junov@chromium.org, mtklein@google.com
Author: mtklein@chromium.org
Review URL: https://codereview.chromium.org/496963003
The only substantive change in this CL is skipping atlasing
for any layers that are involved in nesting. Prior versions
have allowed nesting layers to be atlased.
--------------------------------------------------------
All times are on Windows with a repeat count of 200.
Format is: <time in ms> (<# of glBindframebuffer calls>)
How to interpret this:
For the boogie page:
both columns should be the same (since boogie has no
nested layers)
without the new API the tiled case doesn't show any real
benefit from hoisting
For the carsvg page:
The nesting change does increase the number of FBO
switches but doesn't kill performance
Because of the location & size of the layers the
tile case does show some improvement (even
without the new API)
boogie
-------
w/o nested change w/ nested change
simple 5.62 (811) N/A
tiled 7.72 (811) N/A
simple w/ hoisting 5.23 (409) 5.77 (409)
(but no caching)
tiled w/ hoisting 7.57 (809) 7.49 (809)
(but no caching)
carsvg
------
w/o nested change w/ nested change
simple 60.37 (141990) N/A
tiled 115.13 (256929) N/A
simple w/ hoisting 41.57 (64570) 42.82 (72279)
(but no caching)
tiled w/ hoisting 84.24 (154352) 84.71 (165630)
(but no caching)
R=bsalomon@google.com
Author: robertphillips@google.com
Review URL: https://codereview.chromium.org/476833004
- Remove SkData double-unref to fix assert.
- Re-enable the test with DEBUG (was disabled due to double-unref assert).
- SkRandom instead of rand() for thread-safety.
- General clean up of the test for readability.
Motivation: I am investigating turning on SkFlate::HaveFlate on Windows.
R=mtklein@google.com
Author: halcanary@google.com
Review URL: https://codereview.chromium.org/464333002
Extreme implicit quartic equations solve to roots that are different
enough that they appear to have failed. In this case, fall back on
binary searching to find an intersection.
Relax the condition when this happens; don't give up just because the
computed implicit root points aren't remotely the same.
TBR=reed
BUG=skia:2808
Author: caryclark@google.com
Review URL: https://codereview.chromium.org/456383003
For now this only creates a degenerate bounding box hierarchy where all ops
just have maximal bounds. I will flesh out FillBounds in future CL(s).
Not quite sure why QuadTree and TileGrid aren't drawing right---haven't even
looked at the diffs yet---so I've disabled those test modes for now. RTree
seems fine, so that'll at least get us coverage for all this new plumbing.
BUG=skia:
R=robertphillips@google.com, mtklein@google.com, reed@google.com
Author: mtklein@chromium.org
Review URL: https://codereview.chromium.org/454123003
The recorder optimizer's pattern matcher was accepting command sequences
when it shouldn't have.
In the submitted case, and the pattern matcher was looking for:
saveLayer, drawBitmap, restore
and in the rendering for the submitted case, the sequence of commands
were:
saveLayer, drawBitmap, drawBitmap, restore
This sequence was improperly accepted by the matcher, and the optimizer
reduced the sequence to:
drawBitmap, drawBitmap
where the opacity from the saveLayer paint argument was applied
to the first drawBitmap only.
The user-visible effect in Chrome was a flashing effect on an image
caused by incorrect (too-high) opacity.
The patch adds a Skia test to check for pixel colour values in
a similarly structured recording. All other Skia tests pass.
Blink layout tests also pass with this change.
BUG=chromium:344987
R=robertphillips@google.com, reed@google.com, mtklein@google.com
Author: dneto@chromium.org
Review URL: https://codereview.chromium.org/430503004
Add a unique ID to SkImageFilter, and use it as part
of a persistent cache of image-filtered results. This is used for
caching frame-to-frame coherent filters.
We also keep track of which filter subtrees do not reference the
src input, and use a GenID of zero for the src input in that case.
That way, subtrees which are not dependent on the filter input can be
cached independently of it.
This gives approximately a 4X speedup on
letmespellitoutforyou.com/samples/svg/filter_terrain.svg on Z620
and Nexus10. The cache key consists of the uniqueID of the filter, the
clip bounds, the CTM and the genID of the input bitmap.
Since this does not yet handle the case where the input primitives
(and part of the resulting filter tree) are unchanged, we have
to keep around the external cache for that painting case.
When the work to cache unchanging input primitives is done, the
old cache can be removed, and the new UniqueIDCache will be renamed
to Cache.
R=bsalomon@google.com, mtklein@google.com
Author: senorblanco@chromium.org
Review URL: https://codereview.chromium.org/414483003
https://codereview.chromium.org/283093002 fixed some bugs in pipe memory
allocation, but also introduced one of its own: nearly every block requested
from needOpBytes() got its own 16K allocation.
The correct logic is to take the requested size, add four more bytes for a
DrawOp, make sure that's 4-byte aligned, then check to see if there's enough
space for that in the current block. If there's not, allocate at least
MIN_BLOCK_SIZE bytes to fit the request.
The bug is that I moved that round-up-to-MIN_BLOCK_SIZE step before checking
for space in the current block. This means most (all?) blocks will be 16K but
never seem to have room to fit another allocation. You need 8 bytes? You get
16K. You need 8 more bytes? Nope, can't fit that. Here's a new 16K...
This reverts the change to the test I made then, which really should have
tipped me off. It was testing exactly this bug.
BUG=372671
R=tomhudson@chromium.org, tomhudson@google.com
Author: mtklein@chromium.org
Review URL: https://codereview.chromium.org/433463003
Refactor CanvasStateTest to pull out functions which can
either be called directly or from a shared library. Add a
command line flag to pass in the path to a library to open
to call those functions from.
Separate different CanvasTest tests into separate DEF_TEST
tests. This allows them to be run in parallel.
Move the ifdefs outside of function declarations to skip
running CanvasTest tests which do nothing.
Add the canvas_state_lib target. It is a shared library that
exports some functions to be called by another version of
Skia.
BUG=b/15693384
R=djsollen@google.com, reed@google.com
Author: scroggo@google.com
Review URL: https://codereview.chromium.org/400043003
Share command flags between dm and unit tests.
Also, allow dm's core to be included by itself and iOSShell.
Command line flags that are the same (or nearly the same) in DM
and in skia_tests have been moved to common_flags. Authors,
please check to see that the shared common flag is correct for
the tool.
For iOS, the 'tool_main' entry point has a wrapper to allow multiple
tools to be statically linked in the iOSShell.
Since SkCommandLineFlags::Parse can only be called once, these calls
are disabled in the IOS build.
Since the iOS app directory is dynamically assigned a name, use '@' to
select it. (This is the same convention chosen by the Mobile Harness
iOS file system utilities.)
Move the heart of dm.gyp into dm.gypi so that it can be included by
itself and iOSShell.gyp.
Add tools/flags/SkCommonFlags.* to define and declare common
command line flags.
Add support for dm to iOSShell.
BUG=skia:
R=scroggo@google.com, mtklein@google.com, jvanverth@google.com, bsalomon@google.com
Author: caryclark@google.com
Review URL: https://codereview.chromium.org/389653004
Fixes Android framework build.
When building for the Android framework, we depend on SkUserConfig.h
to define things like SK_SUPPORT_GPU. So we cannot have the first
include inside a guard requring SK_SUPPORT_GPU to be defined. Include
SkTypes first, which will pull in SkUserConfig, define the flag, and
allow us to include GrContextFactory.h.
R=reed@google.com, djsollen@google.com
Author: scroggo@google.com
Review URL: https://codereview.chromium.org/394533003
skia_ios.mm
Get the app's Documents directory and pass use it to set the resource path.
This is a quick hack which will be replaced by a new application that is
a tiny shim around a command line tool.
SkImageEncoder.h
SkForceLinking.cpp
SkImageDecoder_CG.cpp
Add support for FORCE_LINKING so iOS sees the PNG encoder and others.
SkFloatBits.cpp
SkPoint.cpp
Handle denormalized numbers that are floored by the iOS ARM processor.
SkImageDecoder_iOS.mm
Remove empty encoder factory.
SkTouchGesture.cpp
Return early on empty state on touch rather than aborting (crashing)
JpegTest.cpp
Hal via stackoverflow.com says partial jpegs can be gray as well.
skia_test.cpp
Remove crash handler call for now to avoid link failure.
OverwriteLine.h
Remove fancy line overwrite for iOS.
Resources.cpp
Add interface to set resource directory based on runtime query.
BUG=skia:2736 skia:2737 skia:2738
R=reed@google.com, halcanary@google.com, mtklein@google.com, tfarina@chromium.org
Author: caryclark@google.com
Review URL: https://codereview.chromium.org/373383003
When building acceleration structures for SkPicture, we must transform
the primitive's bounds not only by its own SkPaint, but by the paints of
any saveLayer()s currently active above it.
We do this by pushing the SkPaint onto a stack on
saveLayer(), and popping them on restore(). We also push
a NULL paint in save(), so that the pushes and pops are
balanced.
BUG=skia:2734
R=mtklein@google.com, reed@google.com
Author: senorblanco@chromium.org
Review URL: https://codereview.chromium.org/380373003
Now that all creation of SkImageFilters goes through
factory Create() methods, there's no real reason for the
convenience constructors. Some SkImageFilter subclasses
which actually have zero DAG-able inputs were passing NULL
to the superclass constructor. This actually means 1 input,
with a NULL value, not zero inputs. This becomes more
relevant for the upcoming cache infrastructure, where this
indicates that the filter will use its src input, where in
fact some of these filters do not (they are image generators
only).
Limiting SkImageFilter to a single constructor resolves this
ambiguity.
Along the way, I removed all of the default parameters to
the constructors, since the Create methods always call them
with the full argument list.
BUG=skia:
R=reed@google.com
Author: senorblanco@chromium.org
Review URL: https://codereview.chromium.org/376953003
This CL makes it possible for pulled-forward-layers to be atlased. It currently has a couple glaring limitations (which is why it is disabled):
1) the atlased layers cannot be purged nor aged out
2) the texture backing the atlas is not pulled from (or returned to) the resource cache
#1 is on hold until we have a recycling rectanizer
A separate major limitation (the non-atlased layers aren't cached) is blocked until we can transmute entries in the resource cache from scratch to non-scratch while potentially preserving their contents.
Committed: https://skia.googlesource.com/skia/+/55e61f0ef4e5c8c34ac107deaadc9b4ffef3111bR=bsalomon@google.com
Author: robertphillips@google.com
Review URL: https://codereview.chromium.org/354533004
Reason for revert:
Sigh
Original issue's description:
> Begin atlasing
>
> This CL makes it possible for pulled-forward-layers to be atlased. It currently has a couple glaring limitations (which is why it is disabled):
>
> 1) the atlased layers cannot be purged nor aged out
> 2) the texture backing the atlas is not pulled from (or returned to) the resource cache
>
> #1 is on hold until we have a recycling rectanizer
>
> A separate major limitation (the non-atlased layers aren't cached) is blocked until we can transmute entries in the resource cache from scratch to non-scratch while potentially preserving their contents.
>
> Committed: https://skia.googlesource.com/skia/+/55e61f0ef4e5c8c34ac107deaadc9b4ffef3111bR=bsalomon@google.comTBR=bsalomon@google.com
NOTREECHECKS=true
NOTRY=true
Author: robertphillips@google.com
Review URL: https://codereview.chromium.org/359953002
This CL makes it possible for pulled-forward-layers to be atlased. It currently has a couple glaring limitations (which is why it is disabled):
1) the atlased layers cannot be purged nor aged out
2) the texture backing the atlas is not pulled from (or returned to) the resource cache
#1 is on hold until we have a recycling rectanizer
A separate major limitation (the non-atlased layers aren't cached) is blocked until we can transmute entries in the resource cache from scratch to non-scratch while potentially preserving their contents.
R=bsalomon@google.com
Author: robertphillips@google.com
Review URL: https://codereview.chromium.org/354533004
When a RRect is an oval transforming the rectangle and the radii separately can result in a non-oval result (i.e., due to numerical issues some tiny straight edges may creep in). This CL remedies the situation by computing the new radii directly from the transformed rect.
BUG=skia:2696
R=caryclark@google.com
Author: robertphillips@google.com
Review URL: https://codereview.chromium.org/354913004
It is the only file where Test is the prefix rather than the sufix.
This patch fixes that so it matches with the rest of the files
under tests/ directory.
And separate SkISize test into its own DEF_TEST().
BUG=None
TEST=make tests && out/Debug/tests
R=mtklein@google.com
Author: tfarina@chromium.org
Review URL: https://codereview.chromium.org/337783007
- support fRecord in copy constructor
- support SkDrawPictureCallback
Moved SkDrawPictureCallback to its own header so
SkRecordDraw can include it without pulling in all of
SkPicture.
Adding an SkAutoSaveRestore to SkRecordDraw was the easiest
way to match the balance guarantees of the callback, and
probably not a bad idea in general. Updated its tests.
BUG=skia:
R=robertphillips@google.com
Review URL: https://codereview.chromium.org/349973008
Implement path tight bounds using path ops machinery. This is not
as efficient as it could be; for instance, internally, it creates
a path ops structure more suited to intersection. If this shows
up as a performance bottleneck, it could be improved.
Fix path ops gyp files, which have fallen out of sync with other
tests.
R=mtklein@google.com, bsalomon@google.com
TBR=mtklein
BUG=skia:1712
Author: caryclark@google.com
Review URL: https://codereview.chromium.org/348343002
That way when declaring a test with DEF_TEST() macro, you don't have to
uniquify the test name because it might colide with the class it is
being testing.
For example, if you are testing SkBase64 and do:
DEF_TEST(SkBase64, reporter) {
}
That will generate an error because the macro will declare a function
named SkBase64 which colides with the type SkBase64.
By adding Test to the function name we avoid this problem.
Fixed the entries found with the following command line:
$ git grep "Test, r" | grep DEF
BUG=None
TEST=make tests && out/Debug/tests
R=mtklein@google.com
Author: tfarina@chromium.org
Review URL: https://codereview.chromium.org/345753007
This fixes the last bug discovered by iterating through the 800K
skp corpus representing the top 1M websites. For every clip on the
stack, the paths are replaced with the pathop intersection. The
resulting draw is compared with the original draw for pixel errors.
At least two prominent bugs remain. In one, the winding value is
confused by a cubic with an inflection. In the other, a quad/cubic
pair, nearly coincident, fails to find an intersection.
These minor changes include ignoring very tiny self-intersections
of cubics, and processing degenerate edges that don't connect to
anything else.
R=reed@android.com
TBR=reed
Author: caryclark@google.com
Review URL: https://codereview.chromium.org/340103002
The interesting stuff is in SkPictureRecorder.{h,cpp}. The rest is mostly moving SkRecord from its own directories into core to avoid circular dependencies in GYP.
After plumbing SkRecord all the way through in Picture, I'll delete its old entry point include/record/SkRecording.h. For now it and record.gypi need to stay where they are to keep Chrome building.
BUG=skia:
R=reed@google.com, mtklein@google.com
Author: mtklein@chromium.org
Review URL: https://codereview.chromium.org/331573004
This fixes all but one of those failures.
Major changes include:
- Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles.
- Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed.
- Line segments with ends nearly the same are treated as coincident first.
- Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident.
TBR=reed
Author: caryclark@google.com
Review URL: https://codereview.chromium.org/272153002
This CL improves the separation of the SkPicture and SkPictureRecord classes. It delays creation of the SkPicture (in SkPictureRecorder) until recording is actually completed. To accomplish this the SkRecord-derived classes now get SkPathHeap and SkPictureContentInfo members that are absorbed by the SkPicture when it is constructed.
As an ancillary change, this CL also moves the SkPictureContentInfo object from SkPicture to SkPicturePlayback. This is intended to centralize all the data in the SkPicturePlayback object.
R=mtklein@google.com, reed@google.com
Author: robertphillips@google.com
Review URL: https://codereview.chromium.org/324293004
The issue is/was that the original Picture/PictureRecorder that is being partially replayed is not guaranteed to issue any more commands before attempting to modify the existing data. Such modification is prohibited if there is a extant copy-on-write snapshot. Rather then further complicate the SkWriter32::snapshot capability for a dis-preferred use case, this CL simply copies the operation data.
R=scroggo@google.com, reed@google.com
Author: robertphillips@google.com
Review URL: https://codereview.chromium.org/316063005
Reason for revert:
Some benchmarks are written in a way that makes this change unsafe (e.g. const char* resPath = GetResourcePath().c_str(); in SkipZeroesBench) and Valgrind and ASAN caught that. We can try again after a more careful cleanup of GetResourcePath().
Original issue's description:
> Clean up resourcePath code.
>
> 1) Make the implementation of SetResourcePath/GetResourcePath of GM and SkBenchmark match with the one in Test.
> 2) Make gResourcePath a static pointer to const char and move it inside the classes.
>
> BUG=None
> TEST=make tests && out/Debug/tests
> make gm && out/Debug/gm
> make bench && out/Debug/bench
> R=mtklein@google.com
>
> Committed: https://skia.googlesource.com/skia/+/52e4f413ffe2d281f9e90ff2147db08083ffcba7R=tfarina@chromium.orgTBR=tfarina@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=None
Author: mtklein@google.com
Review URL: https://codereview.chromium.org/320733002
1) Make the implementation of SetResourcePath/GetResourcePath of GM and SkBenchmark match with the one in Test.
2) Make gResourcePath a static pointer to const char and move it inside the classes.
BUG=None
TEST=make tests && out/Debug/tests
make gm && out/Debug/gm
make bench && out/Debug/bench
R=mtklein@google.com
Author: tfarina@chromium.org
Review URL: https://codereview.chromium.org/319473003
