skia2/samplecode/SampleImageFilterDAG.cpp

352 lines
14 KiB
C++
Raw Normal View History

/*
* Copyright 2019 Google LLC
*
* Use of this source code is governed by a BSD-style license that can be
* found in the LICENSE file.
*/
#include "samplecode/Sample.h"
#include "include/core/SkCanvas.h"
#include "include/core/SkColor.h"
#include "include/core/SkColorFilter.h"
#include "include/core/SkFont.h"
#include "include/core/SkImage.h"
#include "include/core/SkImageFilter.h"
#include "include/core/SkImageInfo.h"
#include "include/core/SkPaint.h"
#include "include/core/SkPoint.h"
#include "include/core/SkRect.h"
#include "include/core/SkSurface.h"
#include "include/effects/SkDashPathEffect.h"
#include "include/effects/SkGradientShader.h"
#include "include/effects/SkImageFilters.h"
#include "src/core/SkImageFilter_Base.h"
#include "src/core/SkSpecialImage.h"
#include "tools/ToolUtils.h"
namespace {
struct FilterNode {
// Pointer to the actual filter in the DAG, so it still contains its input filters and
// may be used as an input in an earlier node. Null when this represents the "source" input
sk_sp<SkImageFilter> fFilter;
// FilterNodes wrapping each of fFilter's inputs. Leaf node when fInputNodes is empty.
SkTArray<FilterNode> fInputNodes;
// Distance from root filter
int fDepth;
// The source content rect (this is the same for all nodes, but is stored here for convenience)
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
skif::ParameterSpace<SkRect> fContent;
// The mapping for the filter dag (same for all nodes, but stored here for convenience)
skif::Mapping fMapping;
// Cached reverse bounds using device-space clip bounds (e.g. no local bounds hint passed to
// saveLayer). This represents the layer calculated in SkCanvas for the filtering.
skif::LayerSpace<SkIRect> fUnhintedLayerBounds;
// Cached input bounds using the local draw bounds (e.g. saveLayer with a bounds rect, or
// an auto-layer for a draw with image filter). This represents the layer bounds up to this
// point of the DAG.
skif::LayerSpace<SkIRect> fHintedLayerBounds;
// Cached output bounds based on local draw bounds. This represents the output up to this
// point of the DAG.
skif::LayerSpace<SkIRect> fOutputBounds;
FilterNode(const SkImageFilter* filter,
const skif::Mapping& mapping,
const skif::ParameterSpace<SkRect>& content,
int depth)
: fFilter(sk_ref_sp(filter))
, fDepth(depth)
, fContent(content)
, fMapping(mapping) {
this->computeInputBounds();
this->computeOutputBounds();
if (fFilter) {
fInputNodes.reserve_back(fFilter->countInputs());
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
for (int i = 0; i < fFilter->countInputs(); ++i) {
fInputNodes.emplace_back(fFilter->getInput(i), mapping, content, depth + 1);
}
}
}
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
private:
void computeOutputBounds() {
if (fFilter) {
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
// For visualization purposes, we want the output bounds in layer space, before it's
// been transformed to device space. To achieve that, we mock a new mapping with the
// identity matrix transform.
Pre-construct device-to-layer matrix for skif::Mappings SkCanvas and SkDevice were using SkM44 and its definition of invert(), but it was slightly more generous than SkMatrix::invert() so the fuzzer caught a case where the layer's SkDevice had a valid transform but then converting it to a SkMatrix in skif::Mapping was no longer invertible. This modifies it so that skif::Mapping no longer tries to invert the matrices. In almost all cases, the inverse of the layer-to-device matrix can be constructed directly from a matrix multiply (that's what device->getRelativeTransform() does). When the matrices are ill-conditioned the constructed inverse may be inaccurate (hence why SkMatrix::invert reports false), but in practice this happens for ridiculously large transforms and the error isn't significant compared to the precision range of the matrices anyways. Other cases explicitly want to use the identity matrix for the layer to device matrix, so I added a helper in the few places that would have had to pass SkMatrix::I() twice instead. The last case is drawImage() that creates its own skif::Mapping, now it just calculates the inverse that skif::Mapping() would have done and if it fails it drops the draw since it means the canvas matrix is bad. Bug: chromium:1276525 Change-Id: Ib516bb2fac19d5e7397bd27d80f8e3932b25b2e2 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/509396 Reviewed-by: Robert Phillips <robertphillips@google.com> Commit-Queue: Michael Ludwig <michaelludwig@google.com>
2022-02-15 18:16:53 +00:00
skif::Mapping layerOnly{fMapping.layerMatrix()};
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
skif::DeviceSpace<SkIRect> pseudoDeviceBounds =
as_IFB(fFilter)->getOutputBounds(layerOnly, fContent);
// Since layerOnly's device matrix is I, this is effectively a cast to layer space
fOutputBounds = layerOnly.deviceToLayer(pseudoDeviceBounds);
} else {
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
fOutputBounds = fMapping.paramToLayer(fContent).roundOut();
}
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
// Fill in children
for (int i = 0; i < fInputNodes.count(); ++i) {
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
fInputNodes[i].computeOutputBounds();
}
}
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
void computeInputBounds() {
// As a proxy for what the base device had, use the content rect mapped to device space
// (e.g. clipRect() was called with the same coords prior to the draw).
skif::DeviceSpace<SkIRect> targetOutput(fMapping.totalMatrix()
.mapRect(SkRect(fContent))
.roundOut());
if (fFilter) {
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
fHintedLayerBounds = as_IFB(fFilter)->getInputBounds(fMapping, targetOutput, &fContent);
fUnhintedLayerBounds = as_IFB(fFilter)->getInputBounds(fMapping, targetOutput, nullptr);
} else {
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
fHintedLayerBounds = fMapping.paramToLayer(fContent).roundOut();
fUnhintedLayerBounds = fMapping.deviceToLayer(targetOutput);
}
}
};
} // anonymous namespace
static FilterNode build_dag(const SkMatrix& ctm, const SkRect& rect,
const SkImageFilter* rootFilter) {
// Emulate SkCanvas::internalSaveLayer's decomposition of the CTM.
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
skif::ParameterSpace<SkRect> content(rect);
skif::ParameterSpace<SkPoint> center({rect.centerX(), rect.centerY()});
skif::Mapping mapping;
SkAssertResult(mapping.decomposeCTM(ctm, rootFilter, center));
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
return FilterNode(rootFilter, mapping, content, 0);
}
static void draw_node(SkCanvas* canvas, const FilterNode& node) {
canvas->clear(SK_ColorTRANSPARENT);
SkPaint filterPaint;
filterPaint.setImageFilter(node.fFilter);
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
SkRect content = SkRect(node.fContent);
SkPaint paint;
static const SkColor kColors[2] = {SK_ColorGREEN, SK_ColorWHITE};
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
SkPoint points[2] = { {content.fLeft + 15.f, content.fTop + 15.f},
{content.fRight - 15.f, content.fBottom - 15.f} };
paint.setShader(SkGradientShader::MakeLinear(points, kColors, nullptr, SK_ARRAY_COUNT(kColors),
SkTileMode::kRepeat));
SkPaint line;
line.setStrokeWidth(0.f);
line.setStyle(SkPaint::kStroke_Style);
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
canvas->save();
canvas->concat(node.fMapping.layerToDevice());
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
canvas->save();
canvas->concat(node.fMapping.layerMatrix());
canvas->saveLayer(&content, &filterPaint);
canvas->drawRect(content, paint);
canvas->restore(); // Completes the image filter
// Draw content-rect bounds
line.setColor(SK_ColorBLACK);
canvas->drawRect(content, line);
// Bounding boxes have all been mapped by the layer matrix from local to layer space, so undo
// the layer matrix, leaving just the device matrix.
canvas->restore();
// The hinted bounds of the layer saved for the filtering
line.setColor(SK_ColorRED);
canvas->drawRect(SkRect::Make(SkIRect(node.fHintedLayerBounds)).makeOutset(3.f, 3.f), line);
// The bounds of the layer if there was no local content hint
line.setColor(SK_ColorGREEN);
canvas->drawRect(SkRect::Make(SkIRect(node.fUnhintedLayerBounds)).makeOutset(2.f, 2.f), line);
// The output bounds in layer space
line.setColor(SK_ColorBLUE);
canvas->drawRect(SkRect::Make(SkIRect(node.fOutputBounds)).makeOutset(1.f, 1.f), line);
// Device-space bounding box of the output bounds (e.g. what legacy DAG manipulation via
// MatrixTransform would produce).
static const SkScalar kDashParams[] = {6.f, 12.f};
line.setPathEffect(SkDashPathEffect::Make(kDashParams, 2, 0.f));
SkRect devOutputBounds = SkRect::Make(SkIRect(node.fMapping.layerToDevice(node.fOutputBounds)));
canvas->restore(); // undoes device matrix
canvas->drawRect(devOutputBounds, line);
}
static constexpr float kLineHeight = 16.f;
static constexpr float kLineInset = 8.f;
static float print_matrix(SkCanvas* canvas, const char* prefix, const SkMatrix& matrix,
float x, float y, const SkFont& font, const SkPaint& paint) {
canvas->drawString(prefix, x, y, font, paint);
y += kLineHeight;
for (int i = 0; i < 3; ++i) {
SkString row;
row.appendf("[%.2f %.2f %.2f]",
matrix.get(i * 3), matrix.get(i * 3 + 1), matrix.get(i * 3 + 2));
canvas->drawString(row, x, y, font, paint);
y += kLineHeight;
}
return y;
}
static float print_size(SkCanvas* canvas, const char* prefix, const SkIRect& rect,
float x, float y, const SkFont& font, const SkPaint& paint) {
canvas->drawString(prefix, x, y, font, paint);
y += kLineHeight;
SkString sz;
sz.appendf("%d x %d", rect.width(), rect.height());
canvas->drawString(sz, x, y, font, paint);
return y + kLineHeight;
}
static float print_info(SkCanvas* canvas, const FilterNode& node) {
SkFont font(nullptr, 12);
SkPaint text;
text.setAntiAlias(true);
float y = kLineHeight;
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
if (node.fFilter) {
canvas->drawString(node.fFilter->getTypeName(), kLineInset, y, font, text);
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
y += kLineHeight;
if (node.fDepth == 0) {
// The mapping is the same for all nodes, so only print at the root
y = print_matrix(canvas, "Param->Layer", node.fMapping.layerMatrix(),
kLineInset, y, font, text);
y = print_matrix(canvas,
"Layer->Device",
node.fMapping.layerToDevice(),
kLineInset,
y,
font,
text);
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
}
y = print_size(canvas, "Layer Size", SkIRect(node.fUnhintedLayerBounds),
kLineInset, y, font, text);
y = print_size(canvas, "Layer Size (hinted)", SkIRect(node.fHintedLayerBounds),
kLineInset, y, font, text);
} else {
Update image filter debug samples to use new APIs For clarity, this renames skif::Mapping::Make to skif::Mapping::DecomposeCTM to really emphasize when it should be used over just a constructor. Renames SampleBackdropBounds to SampleFilterBounds since the operations and coordinate spaces that it visualizes are the same for regular or backdrop filtering (we just swap the notion of src and dst devices, really). Technically, this is not quite true yet since regular filtering modifies the DAG with a matrix transform right now, but that's going away soon (tm). The "new" SampleFilterBounds example is updated to use the new bounds APIs that use the coord-space safe types in the skif namespace. It also visualizes a filter, and simplifies some of the bounds being drawn and reported to only those most interesting. The SampleImageFilterDAG has been updated to match the soon-to-be state of how SkCanvas processes regular image filters for a saveLayer. All implicit matrix-transform node code is removed. The bounds calculation code is similarly updated to use the new 'skif' types and functions. To simplify the visualization of each node, the 'isolated' versions of the bounds were dropped. Bug: skia:9282, skia:9283 Change-Id: If2ad2c302e5165ae009bba93cea52bf0566a543a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/326718 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Auto-Submit: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2020-10-14 19:59:02 +00:00
canvas->drawString("Source Input", kLineInset, y, font, text);
y += kLineHeight;
}
return y;
}
// Returns bottom edge in pixels that the subtree reached in canvas
static float draw_dag(SkCanvas* canvas, SkSurface* nodeSurface, const FilterNode& node) {
// First capture the results of the node, into nodeSurface
draw_node(nodeSurface->getCanvas(), node);
sk_sp<SkImage> nodeResults = nodeSurface->makeImageSnapshot();
// Fill in background of the filter node with a checkerboard
canvas->save();
canvas->clipRect(SkRect::MakeWH(nodeResults->width(), nodeResults->height()));
ToolUtils::draw_checkerboard(canvas, SK_ColorGRAY, SK_ColorLTGRAY, 10);
canvas->restore();
// Display filtered results in current canvas' location (assumed CTM is set for this node)
canvas->drawImage(nodeResults, 0, 0);
SkPaint line;
line.setAntiAlias(true);
line.setStyle(SkPaint::kStroke_Style);
line.setStrokeWidth(3.f);
// Text info
canvas->save();
canvas->translate(0, nodeResults->height());
float textHeight = print_info(canvas, node);
canvas->restore();
// Border around filtered results + text info
canvas->drawRect(SkRect::MakeWH(nodeResults->width(), nodeResults->height() + textHeight),
line);
static const float kPad = 20.f;
float x = nodeResults->width() + kPad;
float y = 0;
for (int i = 0; i < node.fInputNodes.count(); ++i) {
// Line connecting this node to its child
canvas->drawLine(nodeResults->width(), 0.5f * nodeResults->height(), // right of node
x, y + 0.5f * nodeResults->height(), line); // left of child
canvas->save();
canvas->translate(x, y);
y = draw_dag(canvas, nodeSurface, node.fInputNodes[i]);
canvas->restore();
}
return std::max(y, nodeResults->height() + textHeight + kPad);
}
static void draw_dag(SkCanvas* canvas, sk_sp<SkImageFilter> filter,
const SkRect& rect, const SkISize& surfaceSize) {
// Get the current CTM, which includes all the viewer's UI modifications, which we want to
// pass into our mock canvases for each DAG node.
SkMatrix ctm = canvas->getTotalMatrix();
canvas->save();
// Reset the matrix so that the DAG layout and instructional text is fixed to the window.
canvas->resetMatrix();
// Process the image filter DAG to display intermediate results later on, which will apply the
// provided CTM during draw_node calls.
FilterNode dag = build_dag(ctm, rect, filter.get());
Reland "Reland "SkSurface asynchronous read APIs allow client to extend pixel lifetime"" This is a reland of 6fc04f88a89ed2c9a1b4aa48bcd28602a69a457b Original change's description: > Reland "SkSurface asynchronous read APIs allow client to extend pixel lifetime" > > This is a reland of ce240cc6fd8ec95bd051c7df2173dad2ae8f6ad6 > > Original change's description: > > SkSurface asynchronous read APIs allow client to extend pixel lifetime > > > > Previously the pixel data passed to the client was only valid during > > the client's callback. This meant if the client wanted to defer > > processing of the data a copy was necessary. > > > > Now we pass an object to the callback and the pixel lifetime is tied > > to the lifetime of that object. > > > > The object may be holding a GPU transfer buffer mapped. We don't assume > > that the object will be released on the direct GrContext thread. So > > when the object is destroyed it posts a message to a new type, > > GrClientMappedBufferManager, hanging off the direct context. The direct > > context will periodically check for messages and unmap and then unref > > buffers so that they can be reused. Currently this is done in > > GrContext::performDeferredCleanup() and GrDrawingManager::flush(). > > > > The old API is kept around for backwards compatibility but it is > > reimplemented as a bridge on top of the new mechanism. > > > > Also a utility function to SkImageInfo is added to directly make a new > > info with a specified dimensions rather than passing the width and > > height separately to makeWH(). > > > > Bug: chromium:973403 > > Bug: skia:8962 > > > > Change-Id: Id5cf04235376170142a48e90d3ecd13fd021a2a6 > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/245457 > > Reviewed-by: Brian Osman <brianosman@google.com> > > Commit-Queue: Brian Salomon <bsalomon@google.com> > > Bug: chromium:973403, skia:8962 > Change-Id: I5cecd36276c8b6dc942cf549c7095db2df88530c > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/245678 > Reviewed-by: Brian Salomon <bsalomon@google.com> > Commit-Queue: Brian Salomon <bsalomon@google.com> Bug: chromium:973403, skia:8962 Change-Id: Ie584c1c3ef8021c976f71b708e53871c693cc450 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/246057 Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: Brian Salomon <bsalomon@google.com>
2019-10-03 17:26:54 +00:00
sk_sp<SkSurface> nodeSurface =
canvas->makeSurface(canvas->imageInfo().makeDimensions(surfaceSize));
draw_dag(canvas, nodeSurface.get(), dag);
canvas->restore();
}
class ImageFilterDAGSample : public Sample {
public:
ImageFilterDAGSample() {}
void onDrawContent(SkCanvas* canvas) override {
static const SkRect kFilterRect = SkRect::MakeXYWH(20.f, 20.f, 60.f, 60.f);
static const SkISize kFilterSurfaceSize = SkISize::Make(
2 * (kFilterRect.fRight + kFilterRect.fLeft),
2 * (kFilterRect.fBottom + kFilterRect.fTop));
// Somewhat clunky, but we want to use the viewer calculated CTM in the mini surfaces used
// per DAG node. The rotation matrix viewer calculates is based on the sample size so trick
// it into calculating the right matrix for us w/ 1 frame latency.
this->setSize(kFilterSurfaceSize.width(), kFilterSurfaceSize.height());
// Make a large DAG
// /--- Color Filter <---- Blur <--- Offset
// Merge <
// \--- Blur <--- Drop Shadow
sk_sp<SkImageFilter> drop2 = SkImageFilters::DropShadow(
10.f, 5.f, 3.f, 3.f, SK_ColorBLACK, nullptr);
sk_sp<SkImageFilter> blur1 = SkImageFilters::Blur(2.f, 2.f, std::move(drop2));
sk_sp<SkImageFilter> offset3 = SkImageFilters::Offset(-5.f, -5.f, nullptr);
sk_sp<SkImageFilter> blur2 = SkImageFilters::Blur(4.f, 4.f, std::move(offset3));
sk_sp<SkImageFilter> cf1 = SkImageFilters::ColorFilter(
SkColorFilters::Blend(SK_ColorGRAY, SkBlendMode::kModulate), std::move(blur2));
sk_sp<SkImageFilter> merge0 = SkImageFilters::Merge(std::move(blur1), std::move(cf1));
draw_dag(canvas, std::move(merge0), kFilterRect, kFilterSurfaceSize);
}
SkString name() override { return SkString("ImageFilterDAG"); }
private:
using INHERITED = Sample;
};
DEF_SAMPLE(return new ImageFilterDAGSample();)