Using no compression can save up to a minute of overhead on the RPIs, for
a ~10% increase in file size to download, a great tradeoff.
This commit also regenerates svg and skimage to use no compression.
The next time RecreateSKPs is run, it will pick up the no-compression.
BUG=skia:
Change-Id: I7887e0f8152548185fe095c1f05b08696ab055ec
Reviewed-on: https://skia-review.googlesource.com/9630
Reviewed-by: Eric Boren <borenet@google.com>
Commit-Queue: Kevin Lubick <kjlubick@google.com>
In the common case, there are no intersections in the inner or outer
meshes generated for edge-AA. In that case, we don't have to build
connector edges, and we don't need to run the full simplify and
tessellate path on the full combined mesh. In order to maintain the
correspondence between inner and outer meshes, we can keep partner
pointers between inner and outer vertices instead.
So the new flow is:
- stroke the original boundaries to generate inner & outer meshes
- assign partner pointers to join inner & outer vertices
- build Edges only for Inner and Outer contours
- sort the two meshes independently
- do a complexity check on both meshes (simplified Bentley-Ottmann that
just aborts on the first found intersection)
- if neither mesh is complex, use the fast path:
- tessellate only the inner mesh
- return the outer mesh, and use the partner pointers to generate
the outer geometry triangles
- otherwise, use the complex path (as before):
- connect the inner & outer partners with Connector Edges
- merge the inner & outer meshes via sorted_merge()
- simplify and tessellate the resulting complete mesh
On a 2012 Retina MBP (Intel), this yields:
Canvas Arcs +6%
Stroke Shapes +6%
Fill Shapes +15%
On a Z620 Ubuntu w/NVidia GTX 650:
Canvas Arcs: +5.0%
Stroke Shapes: +1.8%
Fill Shapes: +17.6%
Other changes:
- implemented VertexList::append(VertexList), for use by sorted_merge()
- renamed boundary_to_aa_mesh() to stroke_boundary(), and made it append
inner & outer contours to inner & outer meshes
- the connect() loop at the bottom of stroke_boundary() now uses open
VertexLists, since it can then append them easily to the inner & outer meshes
- sort_and_simplify() changed to sort_mesh(), with merging and simplification
done explicitly by the callers
- sorted_merge() factored out of merge_sort(), for use when zipping together
the inner and outer meshes
Change-Id: Ib00f9f12a375412eff35dd2bb78ccd787d9c37ce
Reviewed-on: https://skia-review.googlesource.com/9600
Commit-Queue: Stephen White <senorblanco@chromium.org>
Reviewed-by: Brian Salomon <bsalomon@google.com>
- handle color arrays in drawPatch(), drawVertices(), drawAtlas()
- color filters
Color filter support is a one-off for SkModeColorFilter. I don't know
any other color filters that are parameterized by a color. If there are
any/many, we may want to wire up something more comprehensive here.
Change-Id: Ibc89574e3a32d38af3bc2443a7d4bac0bb52d493
Reviewed-on: https://skia-review.googlesource.com/9601
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
This builds fine on my laptop, but fails to link on the bots,
so I'm leaving out a new Build bot for now.
BUG=skia:6329
Change-Id: I4b33770f13ab9dec914d090b45d9921b19ee2c9d
Reviewed-on: https://skia-review.googlesource.com/9519
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>
This will be needed to have GrDrawOps that haven't yet built pipelines.
Change-Id: If5292aaa5dc9f98dccbe27be98960b630332158d
Reviewed-on: https://skia-review.googlesource.com/9480
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
TODO:
images
shaders
color filters
image filters
a couple stray color arrays
Change-Id: Ib91639bb0a6a00af737dd5186180011fe5120860
Reviewed-on: https://skia-review.googlesource.com/9529
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
This compacts the object so that it is more efficient for ops to store it.
It also adds a new constructor and query that will allow ops to use the analysis to also store the GrPaint's color.
This has the side effect of limiting the number of color processors on a GrProcessorSet to 64K which is just under 64K more than should ever be needed.
Change-Id: I4e6bc8e3f81bb2ff6a73af685beb6fb928a3de67
Reviewed-on: https://skia-review.googlesource.com/8972
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Always create them budgeted, and register them with the cache (not as
wrapped resources).
Re-land (with fixes) of: https://skia-review.googlesource.com/9497
BUG=skia:
Change-Id: I2df7198adc99efa3eea99fc86b0b2930136f22c7
Reviewed-on: https://skia-review.googlesource.com/9544
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
We can flag the last run record instead. Run iteration is always
sequential, so no penalty.
As a side effect, we can no longer allow instantiation of zero-run text
blobs - but that seems like a good idea anyway.
Change-Id: I7ca80c4780623d5a188f92dfe6d6fe152f20f666
Reviewed-on: https://skia-review.googlesource.com/9149
Commit-Queue: Florin Malita <fmalita@chromium.org>
Reviewed-by: Mike Reed <reed@google.com>
BUG=skia:
Change-Id: Id8c90c6fb2e8f94713937b2e85666c76e96df2ed
Reviewed-on: https://skia-review.googlesource.com/9550
Reviewed-by: Eric Boren <borenet@google.com>
Commit-Queue: Eric Boren <borenet@google.com>
Adds examples of circular reveal and XY tilt.
BUG=skia:6119
Change-Id: I9e7e7729e1d74249e985bc185cb4936f70a75544
Reviewed-on: https://skia-review.googlesource.com/9540
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
This reverts commit cccda60aca.
Reason for revert: Android and Windows bot failures.
Original change's description:
> Treat cross context images as Ganesh-created resources
>
> Always create them budgeted, and register them with the cache (not as
> wrapped resources).
>
> BUG=skia:
>
> Change-Id: Id18ecf6e9e512db4be21b4f2bfd8e8c060bbe805
> Reviewed-on: https://skia-review.googlesource.com/9497
> Commit-Queue: Brian Osman <brianosman@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
>
TBR=bsalomon@google.com,robertphillips@google.com,brianosman@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Change-Id: Ib7bebcad33037dd206c9b06b5cb6c503b00accba
Reviewed-on: https://skia-review.googlesource.com/9541
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Always create them budgeted, and register them with the cache (not as
wrapped resources).
BUG=skia:
Change-Id: Id18ecf6e9e512db4be21b4f2bfd8e8c060bbe805
Reviewed-on: https://skia-review.googlesource.com/9497
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
We'll be adding GalaxyS7_G930FD soon.
NOTRY=true
BUG=skia:6359
Change-Id: I3235576957ea0c395c8d42ee09d5ee89946176d9
Reviewed-on: https://skia-review.googlesource.com/9091
Reviewed-by: Eric Boren <borenet@google.com>
Commit-Queue: Ben Wagner <benjaminwagner@google.com>
The element rearrange logic in SkTHashTable::remove() marks empty slots
as such, but does not reset their value.
When breaking out of the rearrange loop, we must also reset the last empty
slot value to avoid retaining unwanted copies.
Change-Id: I8ba2a25088c0aa5210277124e0917224cb295691
Reviewed-on: https://skia-review.googlesource.com/9533
Reviewed-by: Mike Klein <mtklein@chromium.org>
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>
We used to (incidentally) guard for this when we used bitmapdevice as our backnig.
Now that we have a (faster) nodrawdevice, we need to explicitly guard for it.
BUG=skia:
Change-Id: I9cbbf064cbfced78f0004a2e5aff60aa3ded6215
Reviewed-on: https://skia-review.googlesource.com/9530
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Mike Reed <reed@google.com>
We never adopt render targets (just borrow them).
BUG=skia:
Change-Id: Ie899b814a7a81339a8735bbd7ad9facc66e580d7
Reviewed-on: https://skia-review.googlesource.com/9525
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
One of SkImageCacherator, GrBitmapTextureMaker, GrImageTextureMaker, GrTextureAdjuster, GrTextureProducer or SkImage has to take the first step. This is probably the least odd of the options.
Change-Id: Ie167034553451f4b3633a5a1548dbd4d75839b3d
Reviewed-on: https://skia-review.googlesource.com/9488
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This allows devices (gpu, pdf) which are themselves always dynamically allocated
(since they are reference counted) to provide storage to clipstack, allowing it
to avoid calls to malloc.
Previously this was attempted by embedding the storage directly in clipstack,
but that increased the size of clipstack in all instances, even those where
it might be on the stack. This can be problematic for small-stack environments
like servers.
See previous (reverted) CL: https://skia-review.googlesource.com/c/9522/
BUG=skia:
Change-Id: Ifc7f5ef411303f33513195b1502ea9f281e995c5
Reviewed-on: https://skia-review.googlesource.com/9508
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Mike Reed <reed@google.com>
This rolls the engine past the bad revision which was the cause of the
mentioned bug (22e413ad35481ecd49d232620e7794ce6f544958).
No expectation changes.
BUG=chromium:699379
R=borenet@google.com
Change-Id: I3b44ae54ddec3b2053af59117074b5c1332d0cdf
Reviewed-on: https://skia-review.googlesource.com/9503
Reviewed-by: Eric Boren <borenet@google.com>
Commit-Queue: Eric Boren <borenet@google.com>
This reverts commit 94cbbba96f.
Reason for revert: exceeded stack-size on g3 (in SkPDFDevice.cpp:1552
Original change's description:
> prealloc room for some number of Elements to avoid malloc
>
> I chose 16, as in my test case from android, the depth was
> at least 9. Possibly we could make it even smaller if our
> underlying impl (SkDeque) would never prune its allocations,
> so that we don't malloc repeatedly if we save/restore/save/restore
> across the boundary of the first/nth chunk...
>
> BUG=skia:
>
> Change-Id: Id3f0b900b1931f713f80a664f2b4b142f264be8d
> Reviewed-on: https://skia-review.googlesource.com/9522
> Reviewed-by: Brian Salomon <bsalomon@google.com>
> Commit-Queue: Mike Reed <reed@google.com>
>
TBR=bsalomon@google.com,robertphillips@google.com,reed@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Change-Id: I22c45970b1e3f585087ed22f75c300df00c8124d
Reviewed-on: https://skia-review.googlesource.com/9505
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
I chose 16, as in my test case from android, the depth was
at least 9. Possibly we could make it even smaller if our
underlying impl (SkDeque) would never prune its allocations,
so that we don't malloc repeatedly if we save/restore/save/restore
across the boundary of the first/nth chunk...
BUG=skia:
Change-Id: Id3f0b900b1931f713f80a664f2b4b142f264be8d
Reviewed-on: https://skia-review.googlesource.com/9522
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Mike Reed <reed@google.com>
This avoids taking the (more expensive) copy case when we don't need to.
The old behavior only took this fast case if we were "actively" using
a dynamically allocated array.
BUG=skia:
Change-Id: I0f606ba83ff4aff3a8fc282db7a3ce1b0191fb1a
Reviewed-on: https://skia-review.googlesource.com/9521
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Reed <reed@google.com>
VC++19.10.2517 does not support the C++11 change to allow the address of
internal linkage objects to be used as non-type template arguments.
BUG=skia:6351
Change-Id: I7e1f628db794f950dfba7d043cf6d2fbf0a8c453
Reviewed-on: https://skia-review.googlesource.com/9496
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
Technically, we only ever pass kAdopt along one specific code path, but
the remaining functions that have it are at least all similar. This was
another outlier that seems unlikely to ever benefit.
BUG=skia:
Change-Id: If9a1275a2a0b83417225660f5e18c2133681830a
Reviewed-on: https://skia-review.googlesource.com/9494
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
The first phase of deferring GrPipeline creation until flush will apply only to GrDrawOp subclasses that do not derive from GrMeshDrawOp. This change prepares for that by creating separate draw functions on GrRenderTargetContext for GrMeshDrawOp-derived ops. This is temporary and will incrementally be undone as pipeline-creation deferral rolls out to the GrMeshDrawOps in a later phase of this work.
Change-Id: I0f5b71fe913f3273cfe9e965f7d8bbe7f01ad0ef
Reviewed-on: https://skia-review.googlesource.com/9481
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
This is the last public reference to GrWrapOwnership, so removal paves
the way for moving it to somewhere internal.
BUG=skia:
Change-Id: I876298642ff812452a644c1d2b9519691eac44b9
Reviewed-on: https://skia-review.googlesource.com/9492
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
We can't use memcpy for copy construction, even when MEM_COPY == true.
Change-Id: I50eb369f0fbf77e8f0ad5a148c67d46df0d3ab0e
Reviewed-on: https://skia-review.googlesource.com/9487
Commit-Queue: Florin Malita <fmalita@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Reviewed-by: Ben Wagner <bungeman@google.com>
