skia2/dm/DM.cpp

1506 lines
52 KiB
C++
Raw Normal View History

/*
* Copyright 2013 Google Inc.
*
* Use of this source code is governed by a BSD-style license that can be
* found in the LICENSE file.
*/
#include "DMJsonWriter.h"
#include "DMSrcSink.h"
#include "ProcStats.h"
#include "Resources.h"
#include "SkBBHFactory.h"
#include "SkChecksum.h"
#include "SkChromeTracingTracer.h"
#include "SkCodec.h"
#include "SkColorPriv.h"
#include "SkColorSpace.h"
Revert "Revert "Add color spin test for SkColorSpaceXformCanvas"" This reverts commit 73e21af21390c2806eb1350253233903808edd6b. Reason for revert: I will fix the broken bot next week. Original change's description: > Revert "Add color spin test for SkColorSpaceXformCanvas" > > This reverts commit cb01aec63bcb3dee52afcf3605bcd64166b873c0. > > Reason for revert: Breaks Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-SK_USE_DISCARDABLE_SCALEDIMAGECACHE > > Original change's description: > > Add color spin test for SkColorSpaceXformCanvas > > > > Also changes behavior to treat nullptr srcs as sRGB. > > > > Testing locally, it looks like 353 gms have no diffs from 8888. > > There are 269 diffs - some are fine (gms that do color space stuff) > > and some are bugs. > > > > BUG=skia: > > > > Change-Id: I55c2825f4f4b857e0b0a0ec050c6db82ac881492 > > Reviewed-on: https://skia-review.googlesource.com/9738 > > Reviewed-by: Brian Osman <brianosman@google.com> > > Commit-Queue: Matt Sarett <msarett@google.com> > > > > TBR=mtklein@google.com,msarett@google.com,brianosman@google.com,reviews@skia.org > # Not skipping CQ checks because original CL landed > 1 day ago. > BUG=skia: > > Change-Id: I70bb69f747b863d267494e37a60888a51ab0184c > Reviewed-on: https://skia-review.googlesource.com/9823 > Reviewed-by: Eric Boren <borenet@google.com> > Commit-Queue: Eric Boren <borenet@google.com> > TBR=borenet@google.com,mtklein@google.com,msarett@google.com,reviews@skia.org,brianosman@google.com # Not skipping CQ checks because original CL landed > 1 day ago. BUG=skia: Change-Id: I766382e6655f614042cded84f547f9fd5b109fca Reviewed-on: https://skia-review.googlesource.com/9879 Reviewed-by: Mike Klein <mtklein@google.com> Commit-Queue: Mike Klein <mtklein@google.com>
2017-03-18 15:36:52 +00:00
#include "SkColorSpacePriv.h"
#include "SkCommonFlags.h"
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
#include "SkCommonFlagsConfig.h"
#include "SkCommonFlagsGpuThreads.h"
#include "SkCommonFlagsPathRenderer.h"
#include "SkData.h"
#include "SkDocument.h"
#include "SkDebugfTracer.h"
#include "SkEventTracingPriv.h"
#include "SkFontMgr.h"
#include "SkGraphics.h"
#include "SkHalf.h"
#include "SkLeanWindows.h"
#include "SkMD5.h"
#include "SkMutex.h"
#include "SkOSFile.h"
#include "SkOSPath.h"
#include "SkPM4fPriv.h"
#include "SkPngEncoder.h"
#include "SkScan.h"
#include "SkSpinlock.h"
#include "SkTHash.h"
#include "SkTaskGroup.h"
#include "SkThreadUtils.h"
#include "Test.h"
#include "Timer.h"
#include "ios_utils.h"
#include "picture_utils.h"
#include "sk_tool_utils.h"
Add support for multiple frames in SkCodec Add an interface to decode frames beyond the first in SkCodec, and add an implementation for SkGifCodec. Add getFrameData to SkCodec. This method reads ahead in the stream to return a vector containing meta data about each frame in the image. This is not required in order to decode frames beyond the first, but it allows a client to learn extra information: - how long the frame should be displayed - whether a frame should be blended with a prior frame, allowing the client to provide the prior frame to speed up decoding Add a new fields to SkCodec::Options: - fFrameIndex - fHasPriorFrame The API is designed so that SkCodec never caches frames. If a client wants a frame beyond the first, they specify the frame in Options.fFrameIndex. If the client does not have the frame's required frame (the frame that this frame must be blended on top of) cached, they pass false for Options.fHasPriorFrame. Unless the frame is independent, the codec will then recursively decode all frames necessary to decode fFrameIndex. If the client has the required frame cached, they can put it in the dst they pass to the codec, and the codec will only draw fFrameIndex onto it. Replace SkGifCodec's scanline decoding support with progressive decoding, and update the tests accordingly. Implement new APIs in SkGifCodec. Instead of using gif_lib, use GIFImageReader, imported from Chromium (along with its copyright headers) with the following changes: - SkGifCodec is now the client - Replace blink types - Combine GIFColorMap::buildTable and ::getTable into a method that creates and returns an SkColorTable - Input comes from an SkStream, instead of a SegmentReader. Add SkStreamBuffer, which buffers the (potentially partial) stream in order to decode progressively. (FIXME: This requires copying data that previously was read directly from the SegmentReader. Does this hurt performance? If so, can we fix it?) - Remove UMA code - Instead of reporting screen width and height to the client, allow the client to query for it - Fail earlier if the first frame AND screen have size of zero - Compute required previous frame when adding a new one - Move GIFParseQuery from GIFImageDecoder to GIFImageReader - Allow parsing up to a specific frame (to skip parsing the rest of the stream if a client only wants the first frame) - Compute whether the first frame has alpha and supports index 8, to create the SkImageInfo. This happens before reporting that the size has been decoded. Add GIFImageDecoder::haveDecodedRow to SkGifCodec, imported from Chromium (along with its copyright header), with the following changes: - Add support for sampling - Use the swizzler - Keep track of the rows decoded - Do *not* keep track of whether we've seen alpha Remove SkCodec::kOutOfOrder_SkScanlineOrder, which was only used by GIF scanline decoding. Call onRewind even if there is no stream (SkGifCodec needs to clear its decoded state so it will decode from the beginning). Add a method to SkSwizzler to access the offset into the dst, taking subsetting into account. Add a GM that animates a GIF. Add tests for the new APIs. *** Behavior changes: * Previously, we reported that an image with a subset frame and no transparent index was opaque and used the background index (if present) to fill the background. This is necessary in order to support index 8, but it does not match viewers/browsers I have seen. Examples: - Chromium and Gimp render the background transparent - Firefox, Safari, Linux Image Viewer, Safari Preview clip to the frame (for a single frame image) This CL matches Chromium's behavior and renders the background transparent. This allows us to have consistent behavior across products and simplifies the code (relative to what we would have to do to continue the old behavior on Android). It also means that we will no longer support index 8 for some GIFs. * Stop checking for GIFSTAMP - all GIFs should be either 89a or 87a. This matches Chromium. I suspect that bugs would have been reported if valid GIFs started with "GIFVER" instead of "GIF89a" or "GIF87a" (but did not decode in Chromium). *** Future work not included in this CL: * Move some checks out of haveDecodedRow, since they are the same for the entire frame e.g. - intersecting the frameRect with the full image size - whether there is a color table * Change when we write transparent pixels - In some cases, Chromium deemed this unnecessary, but I suspect it is slower than the fallback case. There will continue to be cases where we should *not* write them, but for e.g. the first pass where we have already cleared to transparent (which we may also be able to skip) writing the transparent pixels will not make anything incorrect. * Report color type and alpha type per frame - Depending on alpha values, disposal methods, frame rects, etc, subsequent frames may have different properties than the first. * Skip copies of the encoded data - We copy the encoded data in case the stream is one that cannot be rewound, so we can parse and then decode (possibly not immediately). For some input streams, this is unnecessary. - I was concerned this cause a performance regression, but on average the new code is faster than the old for the images I tested [1]. - It may cause a performance regression for Chromium, though, where we can always move back in the stream, so this should be addressed. Design doc: https://docs.google.com/a/google.com/document/d/12Qhf9T92MWfdWujQwCIjhCO3sw6pTJB5pJBwDM1T7Kc/ [1] https://docs.google.com/a/google.com/spreadsheets/d/19V-t9BfbFw5eiwBTKA1qOBkZbchjlTC5EIz6HFy-6RI/ GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=2045293002 Review-Url: https://codereview.chromium.org/2045293002
2016-10-24 16:03:26 +00:00
#include <vector>
#ifdef SK_PDF_IMAGE_STATS
extern void SkPDFImageDumpStats();
#endif
#include "png.h"
#include <stdlib.h>
#ifndef SK_BUILD_FOR_WIN32
#include <unistd.h>
#endif
extern bool gSkForceRasterPipelineBlitter;
DECLARE_bool(undefok);
DEFINE_string(src, "tests gm skp image", "Source types to test.");
DEFINE_bool(nameByHash, false,
"If true, write to FLAGS_writePath[0]/<hash>.png instead of "
"to FLAGS_writePath[0]/<config>/<sourceType>/<sourceOptions>/<name>.png");
DEFINE_bool2(pathOpsExtended, x, false, "Run extended pathOps tests.");
DEFINE_string(matrix, "1 0 0 1",
"2x2 scale+skew matrix to apply or upright when using "
"'matrix' or 'upright' in config.");
DEFINE_bool(gpu_threading, false, "Allow GPU work to run on multiple threads?");
DEFINE_string(blacklist, "",
"Space-separated config/src/srcOptions/name quadruples to blacklist. "
"'_' matches anything. '~' negates the match. E.g. \n"
"'--blacklist gpu skp _ _' will blacklist all SKPs drawn into the gpu config.\n"
"'--blacklist gpu skp _ _ 8888 gm _ aarects' will also blacklist the aarects GM on 8888.\n"
"'--blacklist ~8888 svg _ svgparse_' blocks non-8888 SVGs that contain \"svgparse_\" in "
"the name.");
DEFINE_string2(readPath, r, "", "If set check for equality with golden results in this directory.");
DEFINE_string(uninterestingHashesFile, "",
"File containing a list of uninteresting hashes. If a result hashes to something in "
"this list, no image is written for that result.");
DEFINE_int32(shards, 1, "We're splitting source data into this many shards.");
DEFINE_int32(shard, 0, "Which shard do I run?");
DEFINE_string(mskps, "", "Directory to read mskps from, or a single mskp file.");
DEFINE_bool(forceRasterPipeline, false, "sets gSkForceRasterPipelineBlitter");
DEFINE_bool(ddl, false, "If true, use DeferredDisplayLists for GPU SKP rendering.");
#if SK_SUPPORT_GPU
DEFINE_pathrenderer_flag;
#endif
DEFINE_bool(ignoreSigInt, false, "ignore SIGINT signals during test execution");
DEFINE_string(dont_write, "", "File extensions to skip writing to --writePath."); // See skia:6821
using namespace DM;
using sk_gpu_test::GrContextFactory;
using sk_gpu_test::GLTestContext;
using sk_gpu_test::ContextInfo;
/*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~*/
static const double kStartMs = SkTime::GetMSecs();
static FILE* gVLog;
template <typename... Args>
static void vlog(const char* fmt, Args&&... args) {
if (gVLog) {
char s[64];
HumanizeMs(s, 64, SkTime::GetMSecs() - kStartMs);
fprintf(gVLog, "%s\t", s);
fprintf(gVLog, fmt, args...);
fflush(gVLog);
}
}
template <typename... Args>
static void info(const char* fmt, Args&&... args) {
vlog(fmt, args...);
if (!FLAGS_quiet) {
printf(fmt, args...);
}
}
static void info(const char* fmt) {
if (!FLAGS_quiet) {
printf("%s", fmt); // Clang warns printf(fmt) is insecure.
}
}
SK_DECLARE_STATIC_MUTEX(gFailuresMutex);
static SkTArray<SkString> gFailures;
static void fail(const SkString& err) {
SkAutoMutexAcquire lock(gFailuresMutex);
SkDebugf("\n\nFAILURE: %s\n\n", err.c_str());
gFailures.push_back(err);
}
struct Running {
SkString id;
SkThreadID thread;
void dump() const {
info("\t%s\n", id.c_str());
}
};
// We use a spinlock to make locking this in a signal handler _somewhat_ safe.
static SkSpinlock gMutex;
static int gPending;
static SkTArray<Running> gRunning;
static void done(const char* config, const char* src, const char* srcOptions, const char* name) {
SkString id = SkStringPrintf("%s %s %s %s", config, src, srcOptions, name);
vlog("done %s\n", id.c_str());
int pending;
{
SkAutoMutexAcquire lock(gMutex);
for (int i = 0; i < gRunning.count(); i++) {
if (gRunning[i].id == id) {
gRunning.removeShuffle(i);
break;
}
}
pending = --gPending;
}
// We write out dm.json file and print out a progress update every once in a while.
// Notice this also handles the final dm.json and progress update when pending == 0.
if (pending % 500 == 0) {
JsonWriter::DumpJson();
int curr = sk_tools::getCurrResidentSetSizeMB(),
peak = sk_tools::getMaxResidentSetSizeMB();
SkString elapsed = HumanizeMs(SkTime::GetMSecs() - kStartMs);
SkAutoMutexAcquire lock(gMutex);
info("\n%dMB RAM, %dMB peak, %s elapsed, %d queued, %d active:\n",
curr, peak, elapsed.c_str(), gPending - gRunning.count(), gRunning.count());
for (auto& task : gRunning) {
task.dump();
}
}
}
static void start(const char* config, const char* src, const char* srcOptions, const char* name) {
SkString id = SkStringPrintf("%s %s %s %s", config, src, srcOptions, name);
vlog("start %s\n", id.c_str());
SkAutoMutexAcquire lock(gMutex);
gRunning.push_back({id,SkGetThreadID()});
}
static void find_culprit() {
// Assumes gMutex is locked.
SkThreadID thisThread = SkGetThreadID();
for (auto& task : gRunning) {
if (task.thread == thisThread) {
info("Likely culprit:\n");
task.dump();
}
}
}
#if defined(SK_BUILD_FOR_WIN32)
static LONG WINAPI crash_handler(EXCEPTION_POINTERS* e) {
static const struct {
const char* name;
DWORD code;
} kExceptions[] = {
#define _(E) {#E, E}
_(EXCEPTION_ACCESS_VIOLATION),
_(EXCEPTION_BREAKPOINT),
_(EXCEPTION_INT_DIVIDE_BY_ZERO),
_(EXCEPTION_STACK_OVERFLOW),
// TODO: more?
#undef _
};
SkAutoMutexAcquire lock(gMutex);
const DWORD code = e->ExceptionRecord->ExceptionCode;
info("\nCaught exception %u", code);
for (const auto& exception : kExceptions) {
if (exception.code == code) {
info(" %s", exception.name);
}
}
info(", was running:\n");
for (auto& task : gRunning) {
task.dump();
}
find_culprit();
fflush(stdout);
// Execute default exception handler... hopefully, exit.
return EXCEPTION_EXECUTE_HANDLER;
}
static void setup_crash_handler() {
SetUnhandledExceptionFilter(crash_handler);
}
#else
#include <signal.h>
#if !defined(SK_BUILD_FOR_ANDROID)
#include <execinfo.h>
#endif
static constexpr int max_of() { return 0; }
template <typename... Rest>
static constexpr int max_of(int x, Rest... rest) {
return x > max_of(rest...) ? x : max_of(rest...);
}
static void (*previous_handler[max_of(SIGABRT,SIGBUS,SIGFPE,SIGILL,SIGSEGV)+1])(int);
static void crash_handler(int sig) {
SkAutoMutexAcquire lock(gMutex);
info("\nCaught signal %d [%s], was running:\n", sig, strsignal(sig));
for (auto& task : gRunning) {
task.dump();
}
find_culprit();
#if !defined(SK_BUILD_FOR_ANDROID)
void* stack[64];
int count = backtrace(stack, SK_ARRAY_COUNT(stack));
char** symbols = backtrace_symbols(stack, count);
info("\nStack trace:\n");
for (int i = 0; i < count; i++) {
info(" %s\n", symbols[i]);
}
#endif
fflush(stdout);
signal(sig, previous_handler[sig]);
raise(sig);
}
static void setup_crash_handler() {
const int kSignals[] = { SIGABRT, SIGBUS, SIGFPE, SIGILL, SIGSEGV };
for (int sig : kSignals) {
previous_handler[sig] = signal(sig, crash_handler);
}
if (FLAGS_ignoreSigInt) {
signal(SIGINT, SIG_IGN);
}
}
#endif
/*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~*/
struct Gold : public SkString {
Gold() : SkString("") {}
Gold(const SkString& sink, const SkString& src,
const SkString& srcOptions, const SkString& name,
const SkString& md5)
: SkString("") {
this->append(sink);
this->append(src);
this->append(srcOptions);
this->append(name);
this->append(md5);
}
struct Hash {
uint32_t operator()(const Gold& g) const {
return SkGoodHash()((const SkString&)g);
}
};
};
static SkTHashSet<Gold, Gold::Hash> gGold;
static void add_gold(JsonWriter::BitmapResult r) {
gGold.add(Gold(r.config, r.sourceType, r.sourceOptions, r.name, r.md5));
}
static void gather_gold() {
if (!FLAGS_readPath.isEmpty()) {
SkString path(FLAGS_readPath[0]);
path.append("/dm.json");
if (!JsonWriter::ReadJson(path.c_str(), add_gold)) {
fail(SkStringPrintf("Couldn't read %s for golden results.", path.c_str()));
}
}
}
/*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~*/
#if defined(SK_BUILD_FOR_WIN32)
static const char* kNewline = "\r\n";
#else
static const char* kNewline = "\n";
#endif
static SkTHashSet<SkString> gUninterestingHashes;
static void gather_uninteresting_hashes() {
if (!FLAGS_uninterestingHashesFile.isEmpty()) {
sk_sp<SkData> data(SkData::MakeFromFileName(FLAGS_uninterestingHashesFile[0]));
if (!data) {
info("WARNING: unable to read uninteresting hashes from %s\n",
FLAGS_uninterestingHashesFile[0]);
return;
}
// Copy to a string to make sure SkStrSplit has a terminating \0 to find.
SkString contents((const char*)data->data(), data->size());
SkTArray<SkString> hashes;
SkStrSplit(contents.c_str(), kNewline, &hashes);
for (const SkString& hash : hashes) {
gUninterestingHashes.add(hash);
}
info("FYI: loaded %d distinct uninteresting hashes from %d lines\n",
gUninterestingHashes.count(), hashes.count());
}
}
/*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~*/
struct TaggedSrc : public std::unique_ptr<Src> {
SkString tag;
SkString options;
};
struct TaggedSink : public std::unique_ptr<Sink> {
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
SkString tag;
};
static const bool kMemcpyOK = true;
static SkTArray<TaggedSrc, kMemcpyOK> gSrcs;
static SkTArray<TaggedSink, kMemcpyOK> gSinks;
static bool in_shard() {
static int N = 0;
return N++ % FLAGS_shards == FLAGS_shard;
}
static void push_src(const char* tag, ImplicitString options, Src* s) {
std::unique_ptr<Src> src(s);
if (in_shard() &&
FLAGS_src.contains(tag) &&
!SkCommandLineFlags::ShouldSkip(FLAGS_match, src->name().c_str())) {
TaggedSrc& s = gSrcs.push_back();
s.reset(src.release());
s.tag = tag;
s.options = options;
}
}
static void push_codec_src(Path path, CodecSrc::Mode mode, CodecSrc::DstColorType dstColorType,
SkAlphaType dstAlphaType, float scale) {
if (FLAGS_simpleCodec) {
Add support for multiple frames in SkCodec Add an interface to decode frames beyond the first in SkCodec, and add an implementation for SkGifCodec. Add getFrameData to SkCodec. This method reads ahead in the stream to return a vector containing meta data about each frame in the image. This is not required in order to decode frames beyond the first, but it allows a client to learn extra information: - how long the frame should be displayed - whether a frame should be blended with a prior frame, allowing the client to provide the prior frame to speed up decoding Add a new fields to SkCodec::Options: - fFrameIndex - fHasPriorFrame The API is designed so that SkCodec never caches frames. If a client wants a frame beyond the first, they specify the frame in Options.fFrameIndex. If the client does not have the frame's required frame (the frame that this frame must be blended on top of) cached, they pass false for Options.fHasPriorFrame. Unless the frame is independent, the codec will then recursively decode all frames necessary to decode fFrameIndex. If the client has the required frame cached, they can put it in the dst they pass to the codec, and the codec will only draw fFrameIndex onto it. Replace SkGifCodec's scanline decoding support with progressive decoding, and update the tests accordingly. Implement new APIs in SkGifCodec. Instead of using gif_lib, use GIFImageReader, imported from Chromium (along with its copyright headers) with the following changes: - SkGifCodec is now the client - Replace blink types - Combine GIFColorMap::buildTable and ::getTable into a method that creates and returns an SkColorTable - Input comes from an SkStream, instead of a SegmentReader. Add SkStreamBuffer, which buffers the (potentially partial) stream in order to decode progressively. (FIXME: This requires copying data that previously was read directly from the SegmentReader. Does this hurt performance? If so, can we fix it?) - Remove UMA code - Instead of reporting screen width and height to the client, allow the client to query for it - Fail earlier if the first frame AND screen have size of zero - Compute required previous frame when adding a new one - Move GIFParseQuery from GIFImageDecoder to GIFImageReader - Allow parsing up to a specific frame (to skip parsing the rest of the stream if a client only wants the first frame) - Compute whether the first frame has alpha and supports index 8, to create the SkImageInfo. This happens before reporting that the size has been decoded. Add GIFImageDecoder::haveDecodedRow to SkGifCodec, imported from Chromium (along with its copyright header), with the following changes: - Add support for sampling - Use the swizzler - Keep track of the rows decoded - Do *not* keep track of whether we've seen alpha Remove SkCodec::kOutOfOrder_SkScanlineOrder, which was only used by GIF scanline decoding. Call onRewind even if there is no stream (SkGifCodec needs to clear its decoded state so it will decode from the beginning). Add a method to SkSwizzler to access the offset into the dst, taking subsetting into account. Add a GM that animates a GIF. Add tests for the new APIs. *** Behavior changes: * Previously, we reported that an image with a subset frame and no transparent index was opaque and used the background index (if present) to fill the background. This is necessary in order to support index 8, but it does not match viewers/browsers I have seen. Examples: - Chromium and Gimp render the background transparent - Firefox, Safari, Linux Image Viewer, Safari Preview clip to the frame (for a single frame image) This CL matches Chromium's behavior and renders the background transparent. This allows us to have consistent behavior across products and simplifies the code (relative to what we would have to do to continue the old behavior on Android). It also means that we will no longer support index 8 for some GIFs. * Stop checking for GIFSTAMP - all GIFs should be either 89a or 87a. This matches Chromium. I suspect that bugs would have been reported if valid GIFs started with "GIFVER" instead of "GIF89a" or "GIF87a" (but did not decode in Chromium). *** Future work not included in this CL: * Move some checks out of haveDecodedRow, since they are the same for the entire frame e.g. - intersecting the frameRect with the full image size - whether there is a color table * Change when we write transparent pixels - In some cases, Chromium deemed this unnecessary, but I suspect it is slower than the fallback case. There will continue to be cases where we should *not* write them, but for e.g. the first pass where we have already cleared to transparent (which we may also be able to skip) writing the transparent pixels will not make anything incorrect. * Report color type and alpha type per frame - Depending on alpha values, disposal methods, frame rects, etc, subsequent frames may have different properties than the first. * Skip copies of the encoded data - We copy the encoded data in case the stream is one that cannot be rewound, so we can parse and then decode (possibly not immediately). For some input streams, this is unnecessary. - I was concerned this cause a performance regression, but on average the new code is faster than the old for the images I tested [1]. - It may cause a performance regression for Chromium, though, where we can always move back in the stream, so this should be addressed. Design doc: https://docs.google.com/a/google.com/document/d/12Qhf9T92MWfdWujQwCIjhCO3sw6pTJB5pJBwDM1T7Kc/ [1] https://docs.google.com/a/google.com/spreadsheets/d/19V-t9BfbFw5eiwBTKA1qOBkZbchjlTC5EIz6HFy-6RI/ GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=2045293002 Review-Url: https://codereview.chromium.org/2045293002
2016-10-24 16:03:26 +00:00
const bool simple = CodecSrc::kCodec_Mode == mode || CodecSrc::kAnimated_Mode == mode;
if (!simple || dstColorType != CodecSrc::kGetFromCanvas_DstColorType || scale != 1.0f) {
// Only decode in the simple case.
return;
Add support for multiple frames in SkCodec Add an interface to decode frames beyond the first in SkCodec, and add an implementation for SkGifCodec. Add getFrameData to SkCodec. This method reads ahead in the stream to return a vector containing meta data about each frame in the image. This is not required in order to decode frames beyond the first, but it allows a client to learn extra information: - how long the frame should be displayed - whether a frame should be blended with a prior frame, allowing the client to provide the prior frame to speed up decoding Add a new fields to SkCodec::Options: - fFrameIndex - fHasPriorFrame The API is designed so that SkCodec never caches frames. If a client wants a frame beyond the first, they specify the frame in Options.fFrameIndex. If the client does not have the frame's required frame (the frame that this frame must be blended on top of) cached, they pass false for Options.fHasPriorFrame. Unless the frame is independent, the codec will then recursively decode all frames necessary to decode fFrameIndex. If the client has the required frame cached, they can put it in the dst they pass to the codec, and the codec will only draw fFrameIndex onto it. Replace SkGifCodec's scanline decoding support with progressive decoding, and update the tests accordingly. Implement new APIs in SkGifCodec. Instead of using gif_lib, use GIFImageReader, imported from Chromium (along with its copyright headers) with the following changes: - SkGifCodec is now the client - Replace blink types - Combine GIFColorMap::buildTable and ::getTable into a method that creates and returns an SkColorTable - Input comes from an SkStream, instead of a SegmentReader. Add SkStreamBuffer, which buffers the (potentially partial) stream in order to decode progressively. (FIXME: This requires copying data that previously was read directly from the SegmentReader. Does this hurt performance? If so, can we fix it?) - Remove UMA code - Instead of reporting screen width and height to the client, allow the client to query for it - Fail earlier if the first frame AND screen have size of zero - Compute required previous frame when adding a new one - Move GIFParseQuery from GIFImageDecoder to GIFImageReader - Allow parsing up to a specific frame (to skip parsing the rest of the stream if a client only wants the first frame) - Compute whether the first frame has alpha and supports index 8, to create the SkImageInfo. This happens before reporting that the size has been decoded. Add GIFImageDecoder::haveDecodedRow to SkGifCodec, imported from Chromium (along with its copyright header), with the following changes: - Add support for sampling - Use the swizzler - Keep track of the rows decoded - Do *not* keep track of whether we've seen alpha Remove SkCodec::kOutOfOrder_SkScanlineOrder, which was only used by GIF scanline decoding. Call onRewind even if there is no stream (SkGifCodec needs to clear its decoded state so it will decode from the beginning). Add a method to SkSwizzler to access the offset into the dst, taking subsetting into account. Add a GM that animates a GIF. Add tests for the new APIs. *** Behavior changes: * Previously, we reported that an image with a subset frame and no transparent index was opaque and used the background index (if present) to fill the background. This is necessary in order to support index 8, but it does not match viewers/browsers I have seen. Examples: - Chromium and Gimp render the background transparent - Firefox, Safari, Linux Image Viewer, Safari Preview clip to the frame (for a single frame image) This CL matches Chromium's behavior and renders the background transparent. This allows us to have consistent behavior across products and simplifies the code (relative to what we would have to do to continue the old behavior on Android). It also means that we will no longer support index 8 for some GIFs. * Stop checking for GIFSTAMP - all GIFs should be either 89a or 87a. This matches Chromium. I suspect that bugs would have been reported if valid GIFs started with "GIFVER" instead of "GIF89a" or "GIF87a" (but did not decode in Chromium). *** Future work not included in this CL: * Move some checks out of haveDecodedRow, since they are the same for the entire frame e.g. - intersecting the frameRect with the full image size - whether there is a color table * Change when we write transparent pixels - In some cases, Chromium deemed this unnecessary, but I suspect it is slower than the fallback case. There will continue to be cases where we should *not* write them, but for e.g. the first pass where we have already cleared to transparent (which we may also be able to skip) writing the transparent pixels will not make anything incorrect. * Report color type and alpha type per frame - Depending on alpha values, disposal methods, frame rects, etc, subsequent frames may have different properties than the first. * Skip copies of the encoded data - We copy the encoded data in case the stream is one that cannot be rewound, so we can parse and then decode (possibly not immediately). For some input streams, this is unnecessary. - I was concerned this cause a performance regression, but on average the new code is faster than the old for the images I tested [1]. - It may cause a performance regression for Chromium, though, where we can always move back in the stream, so this should be addressed. Design doc: https://docs.google.com/a/google.com/document/d/12Qhf9T92MWfdWujQwCIjhCO3sw6pTJB5pJBwDM1T7Kc/ [1] https://docs.google.com/a/google.com/spreadsheets/d/19V-t9BfbFw5eiwBTKA1qOBkZbchjlTC5EIz6HFy-6RI/ GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=2045293002 Review-Url: https://codereview.chromium.org/2045293002
2016-10-24 16:03:26 +00:00
}
}
SkString folder;
switch (mode) {
case CodecSrc::kCodec_Mode:
folder.append("codec");
break;
case CodecSrc::kCodecZeroInit_Mode:
folder.append("codec_zero_init");
break;
case CodecSrc::kScanline_Mode:
folder.append("scanline");
break;
case CodecSrc::kStripe_Mode:
folder.append("stripe");
break;
case CodecSrc::kCroppedScanline_Mode:
folder.append("crop");
break;
case CodecSrc::kSubset_Mode:
folder.append("codec_subset");
break;
Add support for multiple frames in SkCodec Add an interface to decode frames beyond the first in SkCodec, and add an implementation for SkGifCodec. Add getFrameData to SkCodec. This method reads ahead in the stream to return a vector containing meta data about each frame in the image. This is not required in order to decode frames beyond the first, but it allows a client to learn extra information: - how long the frame should be displayed - whether a frame should be blended with a prior frame, allowing the client to provide the prior frame to speed up decoding Add a new fields to SkCodec::Options: - fFrameIndex - fHasPriorFrame The API is designed so that SkCodec never caches frames. If a client wants a frame beyond the first, they specify the frame in Options.fFrameIndex. If the client does not have the frame's required frame (the frame that this frame must be blended on top of) cached, they pass false for Options.fHasPriorFrame. Unless the frame is independent, the codec will then recursively decode all frames necessary to decode fFrameIndex. If the client has the required frame cached, they can put it in the dst they pass to the codec, and the codec will only draw fFrameIndex onto it. Replace SkGifCodec's scanline decoding support with progressive decoding, and update the tests accordingly. Implement new APIs in SkGifCodec. Instead of using gif_lib, use GIFImageReader, imported from Chromium (along with its copyright headers) with the following changes: - SkGifCodec is now the client - Replace blink types - Combine GIFColorMap::buildTable and ::getTable into a method that creates and returns an SkColorTable - Input comes from an SkStream, instead of a SegmentReader. Add SkStreamBuffer, which buffers the (potentially partial) stream in order to decode progressively. (FIXME: This requires copying data that previously was read directly from the SegmentReader. Does this hurt performance? If so, can we fix it?) - Remove UMA code - Instead of reporting screen width and height to the client, allow the client to query for it - Fail earlier if the first frame AND screen have size of zero - Compute required previous frame when adding a new one - Move GIFParseQuery from GIFImageDecoder to GIFImageReader - Allow parsing up to a specific frame (to skip parsing the rest of the stream if a client only wants the first frame) - Compute whether the first frame has alpha and supports index 8, to create the SkImageInfo. This happens before reporting that the size has been decoded. Add GIFImageDecoder::haveDecodedRow to SkGifCodec, imported from Chromium (along with its copyright header), with the following changes: - Add support for sampling - Use the swizzler - Keep track of the rows decoded - Do *not* keep track of whether we've seen alpha Remove SkCodec::kOutOfOrder_SkScanlineOrder, which was only used by GIF scanline decoding. Call onRewind even if there is no stream (SkGifCodec needs to clear its decoded state so it will decode from the beginning). Add a method to SkSwizzler to access the offset into the dst, taking subsetting into account. Add a GM that animates a GIF. Add tests for the new APIs. *** Behavior changes: * Previously, we reported that an image with a subset frame and no transparent index was opaque and used the background index (if present) to fill the background. This is necessary in order to support index 8, but it does not match viewers/browsers I have seen. Examples: - Chromium and Gimp render the background transparent - Firefox, Safari, Linux Image Viewer, Safari Preview clip to the frame (for a single frame image) This CL matches Chromium's behavior and renders the background transparent. This allows us to have consistent behavior across products and simplifies the code (relative to what we would have to do to continue the old behavior on Android). It also means that we will no longer support index 8 for some GIFs. * Stop checking for GIFSTAMP - all GIFs should be either 89a or 87a. This matches Chromium. I suspect that bugs would have been reported if valid GIFs started with "GIFVER" instead of "GIF89a" or "GIF87a" (but did not decode in Chromium). *** Future work not included in this CL: * Move some checks out of haveDecodedRow, since they are the same for the entire frame e.g. - intersecting the frameRect with the full image size - whether there is a color table * Change when we write transparent pixels - In some cases, Chromium deemed this unnecessary, but I suspect it is slower than the fallback case. There will continue to be cases where we should *not* write them, but for e.g. the first pass where we have already cleared to transparent (which we may also be able to skip) writing the transparent pixels will not make anything incorrect. * Report color type and alpha type per frame - Depending on alpha values, disposal methods, frame rects, etc, subsequent frames may have different properties than the first. * Skip copies of the encoded data - We copy the encoded data in case the stream is one that cannot be rewound, so we can parse and then decode (possibly not immediately). For some input streams, this is unnecessary. - I was concerned this cause a performance regression, but on average the new code is faster than the old for the images I tested [1]. - It may cause a performance regression for Chromium, though, where we can always move back in the stream, so this should be addressed. Design doc: https://docs.google.com/a/google.com/document/d/12Qhf9T92MWfdWujQwCIjhCO3sw6pTJB5pJBwDM1T7Kc/ [1] https://docs.google.com/a/google.com/spreadsheets/d/19V-t9BfbFw5eiwBTKA1qOBkZbchjlTC5EIz6HFy-6RI/ GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=2045293002 Review-Url: https://codereview.chromium.org/2045293002
2016-10-24 16:03:26 +00:00
case CodecSrc::kAnimated_Mode:
folder.append("codec_animated");
break;
}
switch (dstColorType) {
case CodecSrc::kGrayscale_Always_DstColorType:
folder.append("_kGray8");
break;
case CodecSrc::kNonNative8888_Always_DstColorType:
folder.append("_kNonNative");
break;
default:
break;
}
switch (dstAlphaType) {
case kPremul_SkAlphaType:
folder.append("_premul");
break;
case kUnpremul_SkAlphaType:
folder.append("_unpremul");
break;
default:
break;
}
if (1.0f != scale) {
folder.appendf("_%.3f", scale);
}
CodecSrc* src = new CodecSrc(path, mode, dstColorType, dstAlphaType, scale);
push_src("image", folder, src);
}
static void push_android_codec_src(Path path, CodecSrc::DstColorType dstColorType,
SkAlphaType dstAlphaType, int sampleSize) {
SkString folder;
folder.append("scaled_codec");
switch (dstColorType) {
case CodecSrc::kGrayscale_Always_DstColorType:
folder.append("_kGray8");
break;
case CodecSrc::kNonNative8888_Always_DstColorType:
folder.append("_kNonNative");
break;
default:
break;
}
switch (dstAlphaType) {
case kPremul_SkAlphaType:
folder.append("_premul");
break;
case kUnpremul_SkAlphaType:
folder.append("_unpremul");
break;
default:
break;
}
if (1 != sampleSize) {
folder.appendf("_%.3f", 1.0f / (float) sampleSize);
}
AndroidCodecSrc* src = new AndroidCodecSrc(path, dstColorType, dstAlphaType, sampleSize);
push_src("image", folder, src);
}
static void push_image_gen_src(Path path, ImageGenSrc::Mode mode, SkAlphaType alphaType, bool isGpu)
{
SkString folder;
switch (mode) {
case ImageGenSrc::kCodec_Mode:
folder.append("gen_codec");
break;
case ImageGenSrc::kPlatform_Mode:
folder.append("gen_platform");
break;
}
if (isGpu) {
folder.append("_gpu");
} else {
switch (alphaType) {
case kOpaque_SkAlphaType:
folder.append("_opaque");
break;
case kPremul_SkAlphaType:
folder.append("_premul");
break;
case kUnpremul_SkAlphaType:
folder.append("_unpremul");
break;
default:
break;
}
}
ImageGenSrc* src = new ImageGenSrc(path, mode, alphaType, isGpu);
push_src("image", folder, src);
}
static void push_brd_src(Path path, CodecSrc::DstColorType dstColorType, BRDSrc::Mode mode,
uint32_t sampleSize) {
SkString folder("brd_android_codec");
switch (mode) {
case BRDSrc::kFullImage_Mode:
break;
case BRDSrc::kDivisor_Mode:
folder.append("_divisor");
break;
default:
SkASSERT(false);
return;
}
switch (dstColorType) {
case CodecSrc::kGetFromCanvas_DstColorType:
break;
case CodecSrc::kGrayscale_Always_DstColorType:
folder.append("_kGray");
break;
default:
SkASSERT(false);
return;
}
if (1 != sampleSize) {
folder.appendf("_%.3f", 1.0f / (float) sampleSize);
}
BRDSrc* src = new BRDSrc(path, mode, dstColorType, sampleSize);
push_src("image", folder, src);
}
static void push_brd_srcs(Path path, bool gray) {
if (gray) {
// Only run grayscale to one sampleSize and Mode. Though interesting
// to test grayscale, it should not reveal anything across various
// sampleSizes and Modes
// Arbitrarily choose Mode and sampleSize.
push_brd_src(path, CodecSrc::kGrayscale_Always_DstColorType,
BRDSrc::kFullImage_Mode, 2);
}
// Test on a variety of sampleSizes, making sure to include:
// - 2, 4, and 8, which are natively supported by jpeg
// - multiples of 2 which are not divisible by 4 (analogous for 4)
// - larger powers of two, since BRD clients generally use powers of 2
// We will only produce output for the larger sizes on large images.
const uint32_t sampleSizes[] = { 1, 2, 3, 4, 5, 6, 7, 8, 12, 16, 24, 32, 64 };
const BRDSrc::Mode modes[] = { BRDSrc::kFullImage_Mode, BRDSrc::kDivisor_Mode, };
for (uint32_t sampleSize : sampleSizes) {
for (BRDSrc::Mode mode : modes) {
push_brd_src(path, CodecSrc::kGetFromCanvas_DstColorType, mode, sampleSize);
}
}
}
static void push_codec_srcs(Path path) {
sk_sp<SkData> encoded(SkData::MakeFromFileName(path.c_str()));
if (!encoded) {
info("Couldn't read %s.", path.c_str());
return;
}
std::unique_ptr<SkCodec> codec = SkCodec::MakeFromData(encoded);
if (nullptr == codec.get()) {
info("Couldn't create codec for %s.", path.c_str());
return;
}
// native scaling is only supported by WEBP and JPEG
bool supportsNativeScaling = false;
SkTArray<CodecSrc::Mode> nativeModes;
nativeModes.push_back(CodecSrc::kCodec_Mode);
nativeModes.push_back(CodecSrc::kCodecZeroInit_Mode);
switch (codec->getEncodedFormat()) {
case SkEncodedImageFormat::kJPEG:
nativeModes.push_back(CodecSrc::kScanline_Mode);
nativeModes.push_back(CodecSrc::kStripe_Mode);
nativeModes.push_back(CodecSrc::kCroppedScanline_Mode);
supportsNativeScaling = true;
break;
case SkEncodedImageFormat::kWEBP:
nativeModes.push_back(CodecSrc::kSubset_Mode);
supportsNativeScaling = true;
break;
case SkEncodedImageFormat::kDNG:
break;
default:
nativeModes.push_back(CodecSrc::kScanline_Mode);
break;
}
SkTArray<CodecSrc::DstColorType> colorTypes;
colorTypes.push_back(CodecSrc::kGetFromCanvas_DstColorType);
colorTypes.push_back(CodecSrc::kNonNative8888_Always_DstColorType);
switch (codec->getInfo().colorType()) {
case kGray_8_SkColorType:
colorTypes.push_back(CodecSrc::kGrayscale_Always_DstColorType);
break;
default:
break;
}
SkTArray<SkAlphaType> alphaModes;
alphaModes.push_back(kPremul_SkAlphaType);
if (codec->getInfo().alphaType() != kOpaque_SkAlphaType) {
alphaModes.push_back(kUnpremul_SkAlphaType);
}
for (CodecSrc::Mode mode : nativeModes) {
for (CodecSrc::DstColorType colorType : colorTypes) {
for (SkAlphaType alphaType : alphaModes) {
// Only test kCroppedScanline_Mode when the alpha type is premul. The test is
// slow and won't be interestingly different with different alpha types.
if (CodecSrc::kCroppedScanline_Mode == mode &&
kPremul_SkAlphaType != alphaType) {
continue;
}
push_codec_src(path, mode, colorType, alphaType, 1.0f);
// Skip kNonNative on different native scales. It won't be interestingly
// different.
if (supportsNativeScaling &&
CodecSrc::kNonNative8888_Always_DstColorType == colorType) {
// Native Scales
// SkJpegCodec natively supports scaling to the following:
for (auto scale : { 0.125f, 0.25f, 0.375f, 0.5f, 0.625f, 0.750f, 0.875f }) {
push_codec_src(path, mode, colorType, alphaType, scale);
}
}
}
}
}
Add support for multiple frames in SkCodec Add an interface to decode frames beyond the first in SkCodec, and add an implementation for SkGifCodec. Add getFrameData to SkCodec. This method reads ahead in the stream to return a vector containing meta data about each frame in the image. This is not required in order to decode frames beyond the first, but it allows a client to learn extra information: - how long the frame should be displayed - whether a frame should be blended with a prior frame, allowing the client to provide the prior frame to speed up decoding Add a new fields to SkCodec::Options: - fFrameIndex - fHasPriorFrame The API is designed so that SkCodec never caches frames. If a client wants a frame beyond the first, they specify the frame in Options.fFrameIndex. If the client does not have the frame's required frame (the frame that this frame must be blended on top of) cached, they pass false for Options.fHasPriorFrame. Unless the frame is independent, the codec will then recursively decode all frames necessary to decode fFrameIndex. If the client has the required frame cached, they can put it in the dst they pass to the codec, and the codec will only draw fFrameIndex onto it. Replace SkGifCodec's scanline decoding support with progressive decoding, and update the tests accordingly. Implement new APIs in SkGifCodec. Instead of using gif_lib, use GIFImageReader, imported from Chromium (along with its copyright headers) with the following changes: - SkGifCodec is now the client - Replace blink types - Combine GIFColorMap::buildTable and ::getTable into a method that creates and returns an SkColorTable - Input comes from an SkStream, instead of a SegmentReader. Add SkStreamBuffer, which buffers the (potentially partial) stream in order to decode progressively. (FIXME: This requires copying data that previously was read directly from the SegmentReader. Does this hurt performance? If so, can we fix it?) - Remove UMA code - Instead of reporting screen width and height to the client, allow the client to query for it - Fail earlier if the first frame AND screen have size of zero - Compute required previous frame when adding a new one - Move GIFParseQuery from GIFImageDecoder to GIFImageReader - Allow parsing up to a specific frame (to skip parsing the rest of the stream if a client only wants the first frame) - Compute whether the first frame has alpha and supports index 8, to create the SkImageInfo. This happens before reporting that the size has been decoded. Add GIFImageDecoder::haveDecodedRow to SkGifCodec, imported from Chromium (along with its copyright header), with the following changes: - Add support for sampling - Use the swizzler - Keep track of the rows decoded - Do *not* keep track of whether we've seen alpha Remove SkCodec::kOutOfOrder_SkScanlineOrder, which was only used by GIF scanline decoding. Call onRewind even if there is no stream (SkGifCodec needs to clear its decoded state so it will decode from the beginning). Add a method to SkSwizzler to access the offset into the dst, taking subsetting into account. Add a GM that animates a GIF. Add tests for the new APIs. *** Behavior changes: * Previously, we reported that an image with a subset frame and no transparent index was opaque and used the background index (if present) to fill the background. This is necessary in order to support index 8, but it does not match viewers/browsers I have seen. Examples: - Chromium and Gimp render the background transparent - Firefox, Safari, Linux Image Viewer, Safari Preview clip to the frame (for a single frame image) This CL matches Chromium's behavior and renders the background transparent. This allows us to have consistent behavior across products and simplifies the code (relative to what we would have to do to continue the old behavior on Android). It also means that we will no longer support index 8 for some GIFs. * Stop checking for GIFSTAMP - all GIFs should be either 89a or 87a. This matches Chromium. I suspect that bugs would have been reported if valid GIFs started with "GIFVER" instead of "GIF89a" or "GIF87a" (but did not decode in Chromium). *** Future work not included in this CL: * Move some checks out of haveDecodedRow, since they are the same for the entire frame e.g. - intersecting the frameRect with the full image size - whether there is a color table * Change when we write transparent pixels - In some cases, Chromium deemed this unnecessary, but I suspect it is slower than the fallback case. There will continue to be cases where we should *not* write them, but for e.g. the first pass where we have already cleared to transparent (which we may also be able to skip) writing the transparent pixels will not make anything incorrect. * Report color type and alpha type per frame - Depending on alpha values, disposal methods, frame rects, etc, subsequent frames may have different properties than the first. * Skip copies of the encoded data - We copy the encoded data in case the stream is one that cannot be rewound, so we can parse and then decode (possibly not immediately). For some input streams, this is unnecessary. - I was concerned this cause a performance regression, but on average the new code is faster than the old for the images I tested [1]. - It may cause a performance regression for Chromium, though, where we can always move back in the stream, so this should be addressed. Design doc: https://docs.google.com/a/google.com/document/d/12Qhf9T92MWfdWujQwCIjhCO3sw6pTJB5pJBwDM1T7Kc/ [1] https://docs.google.com/a/google.com/spreadsheets/d/19V-t9BfbFw5eiwBTKA1qOBkZbchjlTC5EIz6HFy-6RI/ GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=2045293002 Review-Url: https://codereview.chromium.org/2045293002
2016-10-24 16:03:26 +00:00
{
std::vector<SkCodec::FrameInfo> frameInfos = codec->getFrameInfo();
if (frameInfos.size() > 1) {
Add animation support to SkWebpCodec TBR=reed@google.com (No change to the public API, but changed a header file) SkWebpCodec: - Implement onGetFrameCount, onGetFrameInfo, and onGetRepetitionCount - Respect the alpha reported by libwebp. Although the spec states that it is only a hint, the libwebp encoder uses it properly. Respecting allows us to draw opaque images faster and decode them to 565. This also matches other SkCodecs (and Chromium). - onGetPixels: - Decode the frame requested, recursively decoding required frame if necessary - When blending with a prior frame, use SkRasterPipeline SkCodec: - Move check for negative index to getFrameInfo - Reset the colorXform if one is not needed SkCodecAnimation: - Add new blend enum, for WebP's (and APNG's) non-blending option SkFrameHolder: - New base classes for frames and the owner of the frames, allowing code sharing between SkWebpCodec and SkGifCodec (particularly for determining whether a frame has alpha and what frame it depends on) - When moving items from SkGIFFrameContext, use Skia conventions (i.e. int instead of unsigned) - Rename "delay time" to "duration", to match e.g. SkFrameInfo:: fDuration SkGifImageReader: - Move pieces to SkFrameHolder, and adapt to changes made in the process - Make setAlphaAndRequiredFrame (now on the base class SkFrameHolder) more general to support webp, and add support for frames that do not blend - Change SkGIFFrameContext from a struct to a class, to match how we use the distinction elsewhere (i.e. struct is a small object with public fields) - Rework hasTransparentPixel (now hasTransparency, since it returns true in some cases where there is not a transparent pixel) to better fit with the modified setAlphaAndRequiredFrame. Also be more consistent when there is no transparent pixel but no color map. - Simplify an if condition that was previously simplified in 2d61e717 but accidentally got reverted in a4db9be6 CodecAnimTest: - Test new animated webp files - Rearrange the test to more cleanly print alpha type mismatches for the first frame resources: - webp-animated.webp - animated webp from Chromium - blendBG.webp - new webp file using bits of webp-animated-semitransparent4.webp from Chromium - tests required frame and alpha when using the non-blending mode - frames have the following properties: - Frame 0: no alpha, fills screen - Frame 1: alpha, fills screen - Frame 2: no alpha, fills screen - Frame 3: alpha, fills screen, blendBG - Frame 4: no alpha, fills screen, blendBG - Frame 5: alpha, blendBG - Frame 6: covers 4, has alpha, blendBG - also used to test decoding to 565 if the new frame data has alpha but blends onto an opaque frame DM.cpp: - Test animated images to non-native 8888 and unpremul DMSrcSink.cpp: - Do not test non-native 8888 decodes to f16 dst - Test unpremul decodes to f16 - Copy a frame of an animated image prior to drawing, since in unpremul mode, the DM code will premultiply first. Bug: skia: 3315 Change-Id: I4e55ae2ee5bc095b37a743bdcfac644be603b980 Reviewed-on: https://skia-review.googlesource.com/16707 Commit-Queue: Mike Reed <reed@google.com> Reviewed-by: Mike Reed <reed@google.com> Reviewed-by: Leon Scroggins <scroggo@google.com> Reviewed-by: Matt Sarett <msarett@google.com>
2017-05-23 13:37:21 +00:00
for (auto dstCT : { CodecSrc::kNonNative8888_Always_DstColorType,
CodecSrc::kGetFromCanvas_DstColorType }) {
for (auto at : { kUnpremul_SkAlphaType, kPremul_SkAlphaType }) {
push_codec_src(path, CodecSrc::kAnimated_Mode, dstCT, at, 1.0f);
}
}
Add support for multiple frames in SkCodec Add an interface to decode frames beyond the first in SkCodec, and add an implementation for SkGifCodec. Add getFrameData to SkCodec. This method reads ahead in the stream to return a vector containing meta data about each frame in the image. This is not required in order to decode frames beyond the first, but it allows a client to learn extra information: - how long the frame should be displayed - whether a frame should be blended with a prior frame, allowing the client to provide the prior frame to speed up decoding Add a new fields to SkCodec::Options: - fFrameIndex - fHasPriorFrame The API is designed so that SkCodec never caches frames. If a client wants a frame beyond the first, they specify the frame in Options.fFrameIndex. If the client does not have the frame's required frame (the frame that this frame must be blended on top of) cached, they pass false for Options.fHasPriorFrame. Unless the frame is independent, the codec will then recursively decode all frames necessary to decode fFrameIndex. If the client has the required frame cached, they can put it in the dst they pass to the codec, and the codec will only draw fFrameIndex onto it. Replace SkGifCodec's scanline decoding support with progressive decoding, and update the tests accordingly. Implement new APIs in SkGifCodec. Instead of using gif_lib, use GIFImageReader, imported from Chromium (along with its copyright headers) with the following changes: - SkGifCodec is now the client - Replace blink types - Combine GIFColorMap::buildTable and ::getTable into a method that creates and returns an SkColorTable - Input comes from an SkStream, instead of a SegmentReader. Add SkStreamBuffer, which buffers the (potentially partial) stream in order to decode progressively. (FIXME: This requires copying data that previously was read directly from the SegmentReader. Does this hurt performance? If so, can we fix it?) - Remove UMA code - Instead of reporting screen width and height to the client, allow the client to query for it - Fail earlier if the first frame AND screen have size of zero - Compute required previous frame when adding a new one - Move GIFParseQuery from GIFImageDecoder to GIFImageReader - Allow parsing up to a specific frame (to skip parsing the rest of the stream if a client only wants the first frame) - Compute whether the first frame has alpha and supports index 8, to create the SkImageInfo. This happens before reporting that the size has been decoded. Add GIFImageDecoder::haveDecodedRow to SkGifCodec, imported from Chromium (along with its copyright header), with the following changes: - Add support for sampling - Use the swizzler - Keep track of the rows decoded - Do *not* keep track of whether we've seen alpha Remove SkCodec::kOutOfOrder_SkScanlineOrder, which was only used by GIF scanline decoding. Call onRewind even if there is no stream (SkGifCodec needs to clear its decoded state so it will decode from the beginning). Add a method to SkSwizzler to access the offset into the dst, taking subsetting into account. Add a GM that animates a GIF. Add tests for the new APIs. *** Behavior changes: * Previously, we reported that an image with a subset frame and no transparent index was opaque and used the background index (if present) to fill the background. This is necessary in order to support index 8, but it does not match viewers/browsers I have seen. Examples: - Chromium and Gimp render the background transparent - Firefox, Safari, Linux Image Viewer, Safari Preview clip to the frame (for a single frame image) This CL matches Chromium's behavior and renders the background transparent. This allows us to have consistent behavior across products and simplifies the code (relative to what we would have to do to continue the old behavior on Android). It also means that we will no longer support index 8 for some GIFs. * Stop checking for GIFSTAMP - all GIFs should be either 89a or 87a. This matches Chromium. I suspect that bugs would have been reported if valid GIFs started with "GIFVER" instead of "GIF89a" or "GIF87a" (but did not decode in Chromium). *** Future work not included in this CL: * Move some checks out of haveDecodedRow, since they are the same for the entire frame e.g. - intersecting the frameRect with the full image size - whether there is a color table * Change when we write transparent pixels - In some cases, Chromium deemed this unnecessary, but I suspect it is slower than the fallback case. There will continue to be cases where we should *not* write them, but for e.g. the first pass where we have already cleared to transparent (which we may also be able to skip) writing the transparent pixels will not make anything incorrect. * Report color type and alpha type per frame - Depending on alpha values, disposal methods, frame rects, etc, subsequent frames may have different properties than the first. * Skip copies of the encoded data - We copy the encoded data in case the stream is one that cannot be rewound, so we can parse and then decode (possibly not immediately). For some input streams, this is unnecessary. - I was concerned this cause a performance regression, but on average the new code is faster than the old for the images I tested [1]. - It may cause a performance regression for Chromium, though, where we can always move back in the stream, so this should be addressed. Design doc: https://docs.google.com/a/google.com/document/d/12Qhf9T92MWfdWujQwCIjhCO3sw6pTJB5pJBwDM1T7Kc/ [1] https://docs.google.com/a/google.com/spreadsheets/d/19V-t9BfbFw5eiwBTKA1qOBkZbchjlTC5EIz6HFy-6RI/ GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=2045293002 Review-Url: https://codereview.chromium.org/2045293002
2016-10-24 16:03:26 +00:00
}
}
if (FLAGS_simpleCodec) {
return;
}
const int sampleSizes[] = { 1, 2, 3, 4, 5, 6, 7, 8 };
for (int sampleSize : sampleSizes) {
for (CodecSrc::DstColorType colorType : colorTypes) {
for (SkAlphaType alphaType : alphaModes) {
// We can exercise all of the kNonNative support code in the swizzler with just a
// few sample sizes. Skip the rest.
if (CodecSrc::kNonNative8888_Always_DstColorType == colorType && sampleSize > 3) {
continue;
}
push_android_codec_src(path, colorType, alphaType, sampleSize);
}
}
}
const char* ext = strrchr(path.c_str(), '.');
if (ext) {
ext++;
static const char* const rawExts[] = {
"arw", "cr2", "dng", "nef", "nrw", "orf", "raf", "rw2", "pef", "srw",
"ARW", "CR2", "DNG", "NEF", "NRW", "ORF", "RAF", "RW2", "PEF", "SRW",
};
for (const char* rawExt : rawExts) {
if (0 == strcmp(rawExt, ext)) {
// RAW is not supported by image generator (skbug.com/5079) or BRD.
return;
}
}
static const char* const brdExts[] = {
"jpg", "jpeg", "png", "webp",
"JPG", "JPEG", "PNG", "WEBP",
};
for (const char* brdExt : brdExts) {
if (0 == strcmp(brdExt, ext)) {
bool gray = codec->getEncodedInfo().color() == SkEncodedInfo::kGray_Color;
push_brd_srcs(path, gray);
break;
}
}
}
// Push image generator GPU test.
push_image_gen_src(path, ImageGenSrc::kCodec_Mode, codec->getInfo().alphaType(), true);
// Push image generator CPU tests.
for (SkAlphaType alphaType : alphaModes) {
push_image_gen_src(path, ImageGenSrc::kCodec_Mode, alphaType, false);
#if defined(SK_BUILD_FOR_MAC) || defined(SK_BUILD_FOR_IOS)
if (SkEncodedImageFormat::kWEBP != codec->getEncodedFormat() &&
SkEncodedImageFormat::kWBMP != codec->getEncodedFormat() &&
kUnpremul_SkAlphaType != alphaType)
{
push_image_gen_src(path, ImageGenSrc::kPlatform_Mode, alphaType, false);
}
#elif defined(SK_BUILD_FOR_WIN)
if (SkEncodedImageFormat::kWEBP != codec->getEncodedFormat() &&
SkEncodedImageFormat::kWBMP != codec->getEncodedFormat())
{
push_image_gen_src(path, ImageGenSrc::kPlatform_Mode, alphaType, false);
}
#endif
}
}
template <typename T>
void gather_file_srcs(const SkCommandLineFlags::StringArray& flags, const char* ext) {
for (int i = 0; i < flags.count(); i++) {
const char* path = flags[i];
if (sk_isdir(path)) {
SkOSFile::Iter it(path, ext);
for (SkString file; it.next(&file); ) {
push_src(ext, "", new T(SkOSPath::Join(path, file.c_str())));
}
} else {
push_src(ext, "", new T(path));
}
}
}
static bool gather_srcs() {
for (const skiagm::GMRegistry* r = skiagm::GMRegistry::Head(); r; r = r->next()) {
push_src("gm", "", new GMSrc(r->factory()));
}
if (FLAGS_ddl) {
gather_file_srcs<DDLSKPSrc>(FLAGS_skps, "skp");
} else {
gather_file_srcs<SKPSrc>(FLAGS_skps, "skp");
}
gather_file_srcs<MSKPSrc>(FLAGS_mskps, "mskp");
#if defined(SK_XML)
gather_file_srcs<SVGSrc>(FLAGS_svgs, "svg");
#endif
SkTArray<SkString> images;
if (!CollectImages(FLAGS_images, &images)) {
return false;
}
for (auto image : images) {
push_codec_srcs(image);
}
SkTArray<SkString> colorImages;
if (!CollectImages(FLAGS_colorImages, &colorImages)) {
return false;
}
for (auto colorImage : colorImages) {
ColorCodecSrc* src = new ColorCodecSrc(colorImage, ColorCodecSrc::kBaseline_Mode,
kN32_SkColorType);
push_src("colorImage", "color_codec_baseline", src);
src = new ColorCodecSrc(colorImage, ColorCodecSrc::kDst_HPZR30w_Mode, kN32_SkColorType);
push_src("colorImage", "color_codec_HPZR30w", src);
// TODO (msarett):
// Should we test this Dst in F16 mode (even though the Dst gamma is 2.2 instead of sRGB)?
src = new ColorCodecSrc(colorImage, ColorCodecSrc::kDst_sRGB_Mode, kN32_SkColorType);
push_src("colorImage", "color_codec_sRGB_kN32", src);
src = new ColorCodecSrc(colorImage, ColorCodecSrc::kDst_sRGB_Mode, kRGBA_F16_SkColorType);
push_src("colorImage", "color_codec_sRGB_kF16", src);
}
return true;
}
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
static void push_sink(const SkCommandLineConfig& config, Sink* s) {
std::unique_ptr<Sink> sink(s);
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
// Try a simple Src as a canary. If it fails, skip this sink.
struct : public Src {
Error draw(SkCanvas* c) const override {
c->drawRect(SkRect::MakeWH(1,1), SkPaint());
return "";
}
SkISize size() const override { return SkISize::Make(16, 16); }
Name name() const override { return "justOneRect"; }
} justOneRect;
SkBitmap bitmap;
SkDynamicMemoryWStream stream;
SkString log;
Error err = sink->draw(justOneRect, &bitmap, &stream, &log);
if (err.isFatal()) {
info("Could not run %s: %s\n", config.getTag().c_str(), err.c_str());
exit(1);
}
TaggedSink& ts = gSinks.push_back();
ts.reset(sink.release());
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
ts.tag = config.getTag();
}
static bool gpu_supported() {
#if SK_SUPPORT_GPU
return FLAGS_gpu;
#else
return false;
#endif
}
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
static Sink* create_sink(const GrContextOptions& grCtxOptions, const SkCommandLineConfig* config) {
#if SK_SUPPORT_GPU
Revert of Add config options to run different GPU APIs to dm and nanobench (patchset #21 id:400001 of https://codereview.chromium.org/1490113005/ ) Reason for revert: The Test-Win8-MSVC-ShuttleB-GPU-HD4600-x86_64-Debug builder fails after this CL. Links to specific builds: http://build.chromium.org/p/client.skia/builders/Test-Win8-MSVC-ShuttleB-GPU-HD4600-x86_64-Debug/builds/1689 http://build.chromium.org/p/client.skia/builders/Test-Win8-MSVC-ShuttleB-GPU-HD4600-x86_64-Debug/builds/1690 http://build.chromium.org/p/client.skia/builders/Test-Win8-MSVC-ShuttleB-GPU-HD4600-x86_64-Debug/builds/1691 Original issue's description: > Add config options to run different GPU APIs to dm and nanobench > > Add extended config specification form that can be used to run different > gpu backend with different APIs. > > The configs can be specified with the form: > gpu(api=string,dit=bool,nvpr=bool,samples=int) > > This replaces and removes the --gpuAPI flag. > > All existing configs should still work. > > Adds following documentation: > > out/Debug/dm --help config > > Flags: > --config: type: string default: 565 8888 gpu nonrendering > Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 > nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg > xps or use extended form 'backend(option=value,...)'. > > Extended form: 'backend(option=value,...)' > > Possible backends and options: > > gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend > api type: string default: native. > Select graphics API to use with gpu backend. > Options: > native Use platform default OpenGL or OpenGL ES backend. > gl Use OpenGL. > gles Use OpenGL ES. > debug Use debug OpenGL. > null Use null OpenGL. > dit type: bool default: false. > Use device independent text. > nvpr type: bool default: false. > Use NV_path_rendering OpenGL and OpenGL ES extension. > samples type: int default: 0. > Use multisampling with N samples. > > Predefined configs: > > gpu = gpu() > msaa4 = gpu(samples=4) > msaa16 = gpu(samples=16) > nvprmsaa4 = gpu(nvpr=true,samples=4) > nvprmsaa16 = gpu(nvpr=true,samples=16) > gpudft = gpu(dit=true) > gpudebug = gpu(api=debug) > gpunull = gpu(api=null) > debug = gpu(api=debug) > nullgpu = gpu(api=null) > > BUG=skia:2992 > > Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 > GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 > > Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b > > Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd TBR=mtklein@google.com,bsalomon@google.com,joshualitt@google.com,scroggo@google.com,kkinnunen@nvidia.com NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true BUG=skia:2992 Review URL: https://codereview.chromium.org/1548683002
2015-12-22 18:22:26 +00:00
if (gpu_supported()) {
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
if (const SkCommandLineConfigGpu* gpuConfig = config->asConfigGpu()) {
GrContextFactory::ContextType contextType = gpuConfig->getContextType();
GrContextFactory::ContextOverrides contextOverrides = gpuConfig->getContextOverrides();
GrContextFactory testFactory(grCtxOptions);
if (!testFactory.get(contextType, contextOverrides)) {
info("WARNING: can not create GPU context for config '%s'. "
"GM tests will be skipped.\n", gpuConfig->getTag().c_str());
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
return nullptr;
}
if (gpuConfig->getTestThreading()) {
return new GPUThreadTestingSink(contextType, contextOverrides,
gpuConfig->getSamples(), gpuConfig->getUseDIText(),
gpuConfig->getColorType(),
gpuConfig->getAlphaType(),
sk_ref_sp(gpuConfig->getColorSpace()),
FLAGS_gpu_threading, grCtxOptions);
} else {
return new GPUSink(contextType, contextOverrides, gpuConfig->getSamples(),
gpuConfig->getUseDIText(), gpuConfig->getColorType(),
gpuConfig->getAlphaType(), sk_ref_sp(gpuConfig->getColorSpace()),
FLAGS_gpu_threading, grCtxOptions);
}
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
}
Revert of Add config options to run different GPU APIs to dm and nanobench (patchset #21 id:400001 of https://codereview.chromium.org/1490113005/ ) Reason for revert: The Test-Win8-MSVC-ShuttleB-GPU-HD4600-x86_64-Debug builder fails after this CL. Links to specific builds: http://build.chromium.org/p/client.skia/builders/Test-Win8-MSVC-ShuttleB-GPU-HD4600-x86_64-Debug/builds/1689 http://build.chromium.org/p/client.skia/builders/Test-Win8-MSVC-ShuttleB-GPU-HD4600-x86_64-Debug/builds/1690 http://build.chromium.org/p/client.skia/builders/Test-Win8-MSVC-ShuttleB-GPU-HD4600-x86_64-Debug/builds/1691 Original issue's description: > Add config options to run different GPU APIs to dm and nanobench > > Add extended config specification form that can be used to run different > gpu backend with different APIs. > > The configs can be specified with the form: > gpu(api=string,dit=bool,nvpr=bool,samples=int) > > This replaces and removes the --gpuAPI flag. > > All existing configs should still work. > > Adds following documentation: > > out/Debug/dm --help config > > Flags: > --config: type: string default: 565 8888 gpu nonrendering > Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 > nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg > xps or use extended form 'backend(option=value,...)'. > > Extended form: 'backend(option=value,...)' > > Possible backends and options: > > gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend > api type: string default: native. > Select graphics API to use with gpu backend. > Options: > native Use platform default OpenGL or OpenGL ES backend. > gl Use OpenGL. > gles Use OpenGL ES. > debug Use debug OpenGL. > null Use null OpenGL. > dit type: bool default: false. > Use device independent text. > nvpr type: bool default: false. > Use NV_path_rendering OpenGL and OpenGL ES extension. > samples type: int default: 0. > Use multisampling with N samples. > > Predefined configs: > > gpu = gpu() > msaa4 = gpu(samples=4) > msaa16 = gpu(samples=16) > nvprmsaa4 = gpu(nvpr=true,samples=4) > nvprmsaa16 = gpu(nvpr=true,samples=16) > gpudft = gpu(dit=true) > gpudebug = gpu(api=debug) > gpunull = gpu(api=null) > debug = gpu(api=debug) > nullgpu = gpu(api=null) > > BUG=skia:2992 > > Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 > GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 > > Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b > > Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd TBR=mtklein@google.com,bsalomon@google.com,joshualitt@google.com,scroggo@google.com,kkinnunen@nvidia.com NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true BUG=skia:2992 Review URL: https://codereview.chromium.org/1548683002
2015-12-22 18:22:26 +00:00
}
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
#endif
#define SINK(t, sink, ...) if (config->getBackend().equals(t)) { return new sink(__VA_ARGS__); }
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Review URL: https://codereview.chromium.org/1490113005
2015-12-22 07:48:13 +00:00
if (FLAGS_cpu) {
auto srgbColorSpace = SkColorSpace::MakeSRGB();
auto srgbLinearColorSpace = SkColorSpace::MakeSRGBLinear();
SINK("565", RasterSink, kRGB_565_SkColorType);
SINK("8888", RasterSink, kN32_SkColorType);
SINK("srgb", RasterSink, kN32_SkColorType, srgbColorSpace);
SINK("f16", RasterSink, kRGBA_F16_SkColorType, srgbLinearColorSpace);
SINK("pdf", PDFSink, false, SK_ScalarDefaultRasterDPI);
SINK("skp", SKPSink);
SINK("pipe", PipeSink);
SINK("svg", SVGSink);
SINK("null", NullSink);
SINK("xps", XPSSink);
SINK("pdfa", PDFSink, true, SK_ScalarDefaultRasterDPI);
SINK("pdf300", PDFSink, false, 300);
SINK("jsdebug", DebugSink);
}
#undef SINK
return nullptr;
}
static sk_sp<SkColorSpace> adobe_rgb() {
return SkColorSpace::MakeRGB(SkColorSpace::kSRGB_RenderTargetGamma,
SkColorSpace::kAdobeRGB_Gamut);
}
Revert "Revert "Add color spin test for SkColorSpaceXformCanvas"" This reverts commit 73e21af21390c2806eb1350253233903808edd6b. Reason for revert: I will fix the broken bot next week. Original change's description: > Revert "Add color spin test for SkColorSpaceXformCanvas" > > This reverts commit cb01aec63bcb3dee52afcf3605bcd64166b873c0. > > Reason for revert: Breaks Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-SK_USE_DISCARDABLE_SCALEDIMAGECACHE > > Original change's description: > > Add color spin test for SkColorSpaceXformCanvas > > > > Also changes behavior to treat nullptr srcs as sRGB. > > > > Testing locally, it looks like 353 gms have no diffs from 8888. > > There are 269 diffs - some are fine (gms that do color space stuff) > > and some are bugs. > > > > BUG=skia: > > > > Change-Id: I55c2825f4f4b857e0b0a0ec050c6db82ac881492 > > Reviewed-on: https://skia-review.googlesource.com/9738 > > Reviewed-by: Brian Osman <brianosman@google.com> > > Commit-Queue: Matt Sarett <msarett@google.com> > > > > TBR=mtklein@google.com,msarett@google.com,brianosman@google.com,reviews@skia.org > # Not skipping CQ checks because original CL landed > 1 day ago. > BUG=skia: > > Change-Id: I70bb69f747b863d267494e37a60888a51ab0184c > Reviewed-on: https://skia-review.googlesource.com/9823 > Reviewed-by: Eric Boren <borenet@google.com> > Commit-Queue: Eric Boren <borenet@google.com> > TBR=borenet@google.com,mtklein@google.com,msarett@google.com,reviews@skia.org,brianosman@google.com # Not skipping CQ checks because original CL landed > 1 day ago. BUG=skia: Change-Id: I766382e6655f614042cded84f547f9fd5b109fca Reviewed-on: https://skia-review.googlesource.com/9879 Reviewed-by: Mike Klein <mtklein@google.com> Commit-Queue: Mike Klein <mtklein@google.com>
2017-03-18 15:36:52 +00:00
static sk_sp<SkColorSpace> rgb_to_gbr() {
float gbr[9];
gbr[0] = gSRGB_toXYZD50[1];
gbr[1] = gSRGB_toXYZD50[2];
gbr[2] = gSRGB_toXYZD50[0];
gbr[3] = gSRGB_toXYZD50[4];
gbr[4] = gSRGB_toXYZD50[5];
gbr[5] = gSRGB_toXYZD50[3];
gbr[6] = gSRGB_toXYZD50[7];
gbr[7] = gSRGB_toXYZD50[8];
gbr[8] = gSRGB_toXYZD50[6];
SkMatrix44 toXYZD50(SkMatrix44::kUninitialized_Constructor);
toXYZD50.set3x3RowMajorf(gbr);
return SkColorSpace::MakeRGB(SkColorSpace::kSRGB_RenderTargetGamma, toXYZD50);
}
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
static Sink* create_via(const SkString& tag, Sink* wrapped) {
#define VIA(t, via, ...) if (tag.equals(t)) { return new via(__VA_ARGS__); }
Revert "Revert "Add color spin test for SkColorSpaceXformCanvas"" This reverts commit 73e21af21390c2806eb1350253233903808edd6b. Reason for revert: I will fix the broken bot next week. Original change's description: > Revert "Add color spin test for SkColorSpaceXformCanvas" > > This reverts commit cb01aec63bcb3dee52afcf3605bcd64166b873c0. > > Reason for revert: Breaks Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-SK_USE_DISCARDABLE_SCALEDIMAGECACHE > > Original change's description: > > Add color spin test for SkColorSpaceXformCanvas > > > > Also changes behavior to treat nullptr srcs as sRGB. > > > > Testing locally, it looks like 353 gms have no diffs from 8888. > > There are 269 diffs - some are fine (gms that do color space stuff) > > and some are bugs. > > > > BUG=skia: > > > > Change-Id: I55c2825f4f4b857e0b0a0ec050c6db82ac881492 > > Reviewed-on: https://skia-review.googlesource.com/9738 > > Reviewed-by: Brian Osman <brianosman@google.com> > > Commit-Queue: Matt Sarett <msarett@google.com> > > > > TBR=mtklein@google.com,msarett@google.com,brianosman@google.com,reviews@skia.org > # Not skipping CQ checks because original CL landed > 1 day ago. > BUG=skia: > > Change-Id: I70bb69f747b863d267494e37a60888a51ab0184c > Reviewed-on: https://skia-review.googlesource.com/9823 > Reviewed-by: Eric Boren <borenet@google.com> > Commit-Queue: Eric Boren <borenet@google.com> > TBR=borenet@google.com,mtklein@google.com,msarett@google.com,reviews@skia.org,brianosman@google.com # Not skipping CQ checks because original CL landed > 1 day ago. BUG=skia: Change-Id: I766382e6655f614042cded84f547f9fd5b109fca Reviewed-on: https://skia-review.googlesource.com/9879 Reviewed-by: Mike Klein <mtklein@google.com> Commit-Queue: Mike Klein <mtklein@google.com>
2017-03-18 15:36:52 +00:00
VIA("adobe", ViaCSXform, wrapped, adobe_rgb(), false);
VIA("gbr", ViaCSXform, wrapped, rgb_to_gbr(), true);
SkLite* SkLiteRecorder, a new SkCanvas, fills out SkLiteDL, a new SkDrawable. This SkDrawable is a display list similar to SkRecord and SkBigPicture / SkRecordedDrawable, but with a few new design points inspired by Android and slimming paint: 1) SkLiteDL is structured as one big contiguous array rather than the two layer structure of SkRecord. This trades away flexibility and large-op-count performance for better data locality for small to medium size pictures. 2) We keep a global freelist of SkLiteDLs, both reusing the SkLiteDL struct itself and its contiguous byte array. This keeps the expected number of mallocs per display list allocation <1 (really, ~0) for cyclical use cases. These two together mean recording is faster. Measuring against the code we use at head, SkLiteRecorder trends about ~3x faster across various size pictures, matching speed at 0 draws and beating the special-case 1-draw pictures we have today. (I.e. we won't need those special case implementations anymore, because they're slower than this new generic code.) This new strategy records 10 drawRects() in about the same time the old strategy took for 2. This strategy stays the winner until at least 500 drawRect()s on my laptop, where I stopped checking. A simpler alternative to freelisting is also possible (but not implemented here), where we allow the client to manually reset() an SkLiteDL for reuse when its refcnt is 1. That's essentially what we're doing with the freelist, except tracking what's available for reuse globally instead of making the client do it. This code is not fully capable yet, but most of the key design points are there. The internal structure of SkLiteDL is the area I expect to be most volatile (anything involving Op), but its interface and the whole of SkLiteRecorder ought to be just about done. You can run nanobench --match picture_overhead as a demo. Everything it exercises is fully fleshed out, so what it tests is an apples-to-apples comparison as far as recording costs go. I have not yet compared playback performance. It should be simple to wrap this into an SkPicture subclass if we want. I won't start proposing we replace anything old with anything new quite yet until I have more ducks in a row, but this does look pretty promising (similar to the SkRecord over old SkPicture change a couple years ago) and I'd like to land, experiment, iterate, especially with an eye toward Android. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2213333002 Review-Url: https://codereview.chromium.org/2213333002
2016-08-06 19:51:51 +00:00
VIA("lite", ViaLite, wrapped);
VIA("pipe", ViaPipe, wrapped);
VIA("twice", ViaTwice, wrapped);
#ifdef TEST_VIA_SVG
VIA("svg", ViaSVG, wrapped);
#endif
VIA("serialize", ViaSerialization, wrapped);
VIA("pic", ViaPicture, wrapped);
VIA("2ndpic", ViaSecondPicture, wrapped);
VIA("sp", ViaSingletonPictures, wrapped);
VIA("defer", ViaDefer, wrapped);
VIA("tiles", ViaTiles, 256, 256, nullptr, wrapped);
VIA("tiles_rt", ViaTiles, 256, 256, new SkRTreeFactory, wrapped);
if (FLAGS_matrix.count() == 4) {
SkMatrix m;
m.reset();
m.setScaleX((SkScalar)atof(FLAGS_matrix[0]));
m.setSkewX ((SkScalar)atof(FLAGS_matrix[1]));
m.setSkewY ((SkScalar)atof(FLAGS_matrix[2]));
m.setScaleY((SkScalar)atof(FLAGS_matrix[3]));
VIA("matrix", ViaMatrix, m, wrapped);
VIA("upright", ViaUpright, m, wrapped);
}
#undef VIA
return nullptr;
}
static bool gather_sinks(const GrContextOptions& grCtxOptions, bool defaultConfigs) {
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
SkCommandLineConfigArray configs;
ParseConfigs(FLAGS_config, &configs);
for (int i = 0; i < configs.count(); i++) {
const SkCommandLineConfig& config = *configs[i];
Sink* sink = create_sink(grCtxOptions, &config);
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
if (sink == nullptr) {
info("Skipping config %s: Don't understand '%s'.\n", config.getTag().c_str(),
config.getTag().c_str());
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
continue;
}
const SkTArray<SkString>& parts = config.getViaParts();
for (int j = parts.count(); j-- > 0;) {
const SkString& part = parts[j];
Sink* next = create_via(part, sink);
if (next == nullptr) {
info("Skipping config %s: Don't understand '%s'.\n", config.getTag().c_str(),
part.c_str());
delete sink;
sink = nullptr;
break;
}
sink = next;
}
if (sink) {
push_sink(config, sink);
}
}
// If no configs were requested (just running tests, perhaps?), then we're okay.
if (configs.count() == 0 ||
// If we're using the default configs, we're okay.
defaultConfigs ||
// If we've been told to ignore undefined flags, we're okay.
FLAGS_undefok ||
// Otherwise, make sure that all specified configs have become sinks.
configs.count() == gSinks.count()) {
return true;
}
info("Invalid --config. Use --undefok to bypass this warning.\n");
return false;
}
static bool dump_png(SkBitmap bitmap, const char* path, const char* md5) {
SkPixmap pm;
if (!bitmap.peekPixels(&pm)) {
return false; // Ought to never happen... we're already read-back at this point.
}
SkFILEWStream dst{path};
SkString description;
description.append("Key: ");
for (int i = 0; i < FLAGS_key.count(); i++) {
description.appendf("%s ", FLAGS_key[i]);
}
description.append("Properties: ");
for (int i = 0; i < FLAGS_properties.count(); i++) {
description.appendf("%s ", FLAGS_properties[i]);
}
description.appendf("MD5: %s", md5);
const char* comments[] = {
"Author", "DM dump_png()",
"Description", description.c_str(),
};
size_t lengths[] = {
strlen(comments[0])+1, strlen(comments[1])+1,
strlen(comments[2])+1, strlen(comments[3])+1,
};
SkPngEncoder::Options options;
options.fComments = SkDataTable::MakeCopyArrays((const void**)comments, lengths, 4);
options.fFilterFlags = SkPngEncoder::FilterFlag::kNone;
options.fZLibLevel = 1;
options.fUnpremulBehavior = pm.colorSpace() ? SkTransferFunctionBehavior::kRespect
: SkTransferFunctionBehavior::kIgnore;
return SkPngEncoder::Encode(&dst, pm, options);
}
static bool match(const char* needle, const char* haystack) {
if ('~' == needle[0]) {
return !match(needle + 1, haystack);
}
if (0 == strcmp("_", needle)) {
return true;
}
return nullptr != strstr(haystack, needle);
}
static bool is_blacklisted(const char* sink, const char* src,
const char* srcOptions, const char* name) {
for (int i = 0; i < FLAGS_blacklist.count() - 3; i += 4) {
if (match(FLAGS_blacklist[i+0], sink) &&
match(FLAGS_blacklist[i+1], src) &&
match(FLAGS_blacklist[i+2], srcOptions) &&
match(FLAGS_blacklist[i+3], name)) {
return true;
}
}
return false;
}
// Even when a Task Sink reports to be non-threadsafe (e.g. GPU), we know things like
// .png encoding are definitely thread safe. This lets us offload that work to CPU threads.
static SkTaskGroup gDefinitelyThreadSafeWork;
// The finest-grained unit of work we can run: draw a single Src into a single Sink,
// report any errors, and perhaps write out the output: a .png of the bitmap, or a raw stream.
struct Task {
Task(const TaggedSrc& src, const TaggedSink& sink) : src(src), sink(sink) {}
const TaggedSrc& src;
const TaggedSink& sink;
static void Run(const Task& task) {
SkString name = task.src->name();
SkString log;
if (!FLAGS_dryRun) {
SkBitmap bitmap;
SkDynamicMemoryWStream stream;
start(task.sink.tag.c_str(), task.src.tag.c_str(),
task.src.options.c_str(), name.c_str());
Error err = task.sink->draw(*task.src, &bitmap, &stream, &log);
if (!log.isEmpty()) {
info("%s %s %s %s:\n%s\n", task.sink.tag.c_str()
, task.src.tag.c_str()
, task.src.options.c_str()
, name.c_str()
, log.c_str());
}
if (!err.isEmpty()) {
if (err.isFatal()) {
fail(SkStringPrintf("%s %s %s %s: %s",
task.sink.tag.c_str(),
task.src.tag.c_str(),
task.src.options.c_str(),
name.c_str(),
err.c_str()));
} else {
done(task.sink.tag.c_str(), task.src.tag.c_str(),
task.src.options.c_str(), name.c_str());
return;
}
}
// We're likely switching threads here, so we must capture by value, [=] or [foo,bar].
SkStreamAsset* data = stream.detachAsStream().release();
gDefinitelyThreadSafeWork.add([task,name,bitmap,data]{
std::unique_ptr<SkStreamAsset> ownedData(data);
SkString md5;
if (!FLAGS_writePath.isEmpty() || !FLAGS_readPath.isEmpty()) {
SkMD5 hash;
if (data->getLength()) {
hash.writeStream(data, data->getLength());
data->rewind();
} else {
// If we're BGRA (Linux, Windows), swizzle over to RGBA (Mac, Android).
// This helps eliminate multiple 0-pixel-diff hashes on gold.skia.org.
// (Android's general slow speed breaks the tie arbitrarily in RGBA's favor.)
// We might consider promoting 565 to RGBA too.
if (bitmap.colorType() == kBGRA_8888_SkColorType) {
SkBitmap swizzle;
SkAssertResult(sk_tool_utils::copy_to(&swizzle, kRGBA_8888_SkColorType,
bitmap));
hash.write(swizzle.getPixels(), swizzle.getSize());
} else {
hash.write(bitmap.getPixels(), bitmap.getSize());
}
}
SkMD5::Digest digest;
hash.finish(digest);
for (int i = 0; i < 16; i++) {
md5.appendf("%02x", digest.data[i]);
}
}
if (!FLAGS_readPath.isEmpty() &&
!gGold.contains(Gold(task.sink.tag, task.src.tag,
task.src.options, name, md5))) {
fail(SkStringPrintf("%s not found for %s %s %s %s in %s",
md5.c_str(),
task.sink.tag.c_str(),
task.src.tag.c_str(),
task.src.options.c_str(),
name.c_str(),
FLAGS_readPath[0]));
}
if (!FLAGS_writePath.isEmpty()) {
const char* ext = task.sink->fileExtension();
if (ext && !FLAGS_dont_write.contains(ext)) {
if (data->getLength()) {
WriteToDisk(task, md5, ext, data, data->getLength(), nullptr);
SkASSERT(bitmap.drawsNothing());
} else if (!bitmap.drawsNothing()) {
WriteToDisk(task, md5, ext, nullptr, 0, &bitmap);
}
}
}
});
}
done(task.sink.tag.c_str(), task.src.tag.c_str(), task.src.options.c_str(), name.c_str());
}
static void WriteToDisk(const Task& task,
SkString md5,
const char* ext,
SkStream* data, size_t len,
const SkBitmap* bitmap) {
bool gammaCorrect = false;
if (bitmap) {
gammaCorrect = SkToBool(bitmap->info().colorSpace());
}
JsonWriter::BitmapResult result;
result.name = task.src->name();
result.config = task.sink.tag;
result.sourceType = task.src.tag;
result.sourceOptions = task.src.options;
result.ext = ext;
result.gammaCorrect = gammaCorrect;
result.md5 = md5;
JsonWriter::AddBitmapResult(result);
// If an MD5 is uninteresting, we want it noted in the JSON file,
// but don't want to dump it out as a .png (or whatever ext is).
if (gUninterestingHashes.contains(md5)) {
return;
}
const char* dir = FLAGS_writePath[0];
if (0 == strcmp(dir, "@")) { // Needed for iOS.
dir = FLAGS_resourcePath[0];
}
sk_mkdir(dir);
SkString path;
if (FLAGS_nameByHash) {
path = SkOSPath::Join(dir, result.md5.c_str());
path.append(".");
path.append(ext);
if (sk_exists(path.c_str())) {
return; // Content-addressed. If it exists already, we're done.
}
} else {
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
path = SkOSPath::Join(dir, task.sink.tag.c_str());
sk_mkdir(path.c_str());
path = SkOSPath::Join(path.c_str(), task.src.tag.c_str());
sk_mkdir(path.c_str());
if (strcmp(task.src.options.c_str(), "") != 0) {
path = SkOSPath::Join(path.c_str(), task.src.options.c_str());
sk_mkdir(path.c_str());
}
path = SkOSPath::Join(path.c_str(), task.src->name().c_str());
path.append(".");
path.append(ext);
}
if (bitmap) {
if (!dump_png(*bitmap, path.c_str(), result.md5.c_str())) {
fail(SkStringPrintf("Can't encode PNG to %s.\n", path.c_str()));
return;
}
} else {
SkFILEWStream file(path.c_str());
if (!file.isValid()) {
fail(SkStringPrintf("Can't open %s for writing.\n", path.c_str()));
return;
}
if (!file.writeStream(data, len)) {
fail(SkStringPrintf("Can't write to %s.\n", path.c_str()));
return;
}
}
}
};
/*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~*/
// Unit tests don't fit so well into the Src/Sink model, so we give them special treatment.
static SkTDArray<skiatest::Test> gParallelTests, gSerialTests;
static void gather_tests() {
if (!FLAGS_src.contains("tests")) {
return;
}
for (const skiatest::TestRegistry* r = skiatest::TestRegistry::Head(); r; r = r->next()) {
if (!in_shard()) {
continue;
}
// Despite its name, factory() is returning a reference to
// link-time static const POD data.
const skiatest::Test& test = r->factory();
if (SkCommandLineFlags::ShouldSkip(FLAGS_match, test.name)) {
continue;
}
if (test.needsGpu && gpu_supported()) {
(FLAGS_gpu_threading ? gParallelTests : gSerialTests).push(test);
} else if (!test.needsGpu && FLAGS_cpu) {
gParallelTests.push(test);
}
}
}
static void run_test(skiatest::Test test, const GrContextOptions& grCtxOptions) {
struct : public skiatest::Reporter {
void reportFailed(const skiatest::Failure& failure) override {
fail(failure.toString());
JsonWriter::AddTestFailure(failure);
}
bool allowExtendedTest() const override {
return FLAGS_pathOpsExtended;
}
bool verbose() const override { return FLAGS_veryVerbose; }
} reporter;
if (!FLAGS_dryRun && !is_blacklisted("_", "tests", "_", test.name)) {
start("unit", "test", "", test.name);
GrContextFactory factory(grCtxOptions);
test.run(&reporter, &factory);
}
done("unit", "test", "", test.name);
}
/*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~*/
#define PORTABLE_FONT_PREFIX "Toy Liberation "
static sk_sp<SkTypeface> create_from_name(const char familyName[], SkFontStyle style) {
if (familyName && strlen(familyName) > sizeof(PORTABLE_FONT_PREFIX)
&& !strncmp(familyName, PORTABLE_FONT_PREFIX, sizeof(PORTABLE_FONT_PREFIX) - 1)) {
return sk_tool_utils::create_portable_typeface(familyName, style);
}
return nullptr;
}
#undef PORTABLE_FONT_PREFIX
extern sk_sp<SkTypeface> (*gCreateTypefaceDelegate)(const char [], SkFontStyle );
int main(int argc, char** argv) {
SkCommandLineFlags::Parse(argc, argv);
initializeEventTracingForTools();
#if !defined(GOOGLE3) && defined(SK_BUILD_FOR_IOS)
cd_Documents();
#endif
setbuf(stdout, nullptr);
setup_crash_handler();
gSkUseAnalyticAA = FLAGS_analyticAA;
New analytic AA scan converter using delta (I call it DAA for now) DAA is: 1. Much simpler than AAA. SkScan_AAAPath.cpp is about 1700 lines. SkScan_DAAPath.cpp is about 300 lines. The whole DAA CL is only about 800 lines. 2. Much faster than AAA for complicated paths. The speedup applies to GL backend (including ccpr)! Here's the frame time of 'SampleApp --slide Chart' on macbook pro: AAA-raster: 33ms DAA-raster: 21ms AAA-gl: 30ms DAA-gl: 20ms AAA-ccpr: 18ms DAA-ccpr: 12ms My linux desktop doesn't have SSE3 so the speedup is smaller (~25% for Chart). I believe that DAA is so fast that I can enable it for any paths (AAA is not enabled by default for complicated paths because it is slow; hence our older supersampling scan converter is used for stroking on Chart for AAA-xxx config.) 3. The SkCoverageDelta is suitable for threaded backend with out-of-order concurrent scan conversion as commented in the source code. Maybe we can also just send deltas to GPU. 4. Similar to most analytic path renderers, the quality is on the best ground-truth level, unless there are intersections within a pixel. The intersections look good to my eyes although theoretically that could be arbitrary far from the ground truth (see my AAA slides). 5. For simple paths, such as circle, triangle, rrect, etc., DAA is slower than AAA. But DAA is faster than our older supersampling scan converter in most cases. As those simple paths usually don't constitute the bottleneck of a picture (skp or svg), I strongly recommend use DAA. 6. DAA also heavily favors blitMask so it may work quite well with SkRasterPipeline and SkRasterPipelineBlitter. Finally, please check https://skia-review.googlesource.com/c/22420/ which accelerate DAA by specializing blitCoverageDeltas for SkARGB32_Blitter and SkARGB32_Black_Blitter. It brings a little(<5%) speedup. But I couldn't figure out how to reduce the duplicate code so I don't intend to land it. Bug: skia: Change-Id: I3b7ed6a727447922e645b1acb737a506e7c09a4c Reviewed-on: https://skia-review.googlesource.com/19666 Reviewed-by: Mike Reed <reed@google.com> Reviewed-by: Cary Clark <caryclark@google.com> Commit-Queue: Yuqian Li <liyuqian@google.com>
2017-07-25 15:26:31 +00:00
gSkUseDeltaAA = FLAGS_deltaAA;
if (FLAGS_forceAnalyticAA) {
gSkForceAnalyticAA = true;
}
if (FLAGS_forceDeltaAA) {
gSkForceDeltaAA = true;
}
if (FLAGS_forceRasterPipeline) {
gSkForceRasterPipelineBlitter = true;
}
// The bots like having a verbose.log to upload, so always touch the file even if --verbose.
if (!FLAGS_writePath.isEmpty()) {
sk_mkdir(FLAGS_writePath[0]);
gVLog = fopen(SkOSPath::Join(FLAGS_writePath[0], "verbose.log").c_str(), "w");
}
if (FLAGS_verbose) {
gVLog = stderr;
}
GrContextOptions grCtxOptions;
#if SK_SUPPORT_GPU
grCtxOptions.fGpuPathRenderers = CollectGpuPathRenderersFromFlags();
grCtxOptions.fExecutor = GpuExecutorForTools();
#endif
JsonWriter::DumpJson(); // It's handy for the bots to assume this is ~never missing.
SkAutoGraphics ag;
SkTaskGroup::Enabler enabled(FLAGS_threads);
gCreateTypefaceDelegate = &create_from_name;
{
SkString testResourcePath = GetResourcePath("color_wheel.png");
SkFILEStream testResource(testResourcePath.c_str());
if (!testResource.isValid()) {
info("Some resources are missing. Do you need to set --resourcePath?\n");
}
}
gather_gold();
gather_uninteresting_hashes();
if (!gather_srcs()) {
return 1;
}
// TODO(dogben): This is a bit ugly. Find a cleaner way to do this.
bool defaultConfigs = true;
for (int i = 0; i < argc; i++) {
static const char* kConfigArg = "--config";
if (strcmp(argv[i], kConfigArg) == 0) {
defaultConfigs = false;
break;
}
}
if (!gather_sinks(grCtxOptions, defaultConfigs)) {
return 1;
}
gather_tests();
gPending = gSrcs.count() * gSinks.count() + gParallelTests.count() + gSerialTests.count();
info("%d srcs * %d sinks + %d tests == %d tasks\n",
gSrcs.count(), gSinks.count(), gParallelTests.count() + gSerialTests.count(), gPending);
// Kick off as much parallel work as we can, making note of any serial work we'll need to do.
SkTaskGroup parallel;
SkTArray<Task> serial;
for (auto& sink : gSinks)
for (auto& src : gSrcs) {
if (src->veto(sink->flags()) ||
is_blacklisted(sink.tag.c_str(), src.tag.c_str(),
src.options.c_str(), src->name().c_str())) {
SkAutoMutexAcquire lock(gMutex);
gPending--;
continue;
}
Task task(src, sink);
if (src->serial() || sink->serial()) {
serial.push_back(task);
} else {
parallel.add([task] { Task::Run(task); });
}
}
for (auto test : gParallelTests) {
parallel.add([test, grCtxOptions] { run_test(test, grCtxOptions); });
}
// With the parallel work running, run serial tasks and tests here on main thread.
for (auto task : serial) { Task::Run(task); }
for (auto test : gSerialTests) { run_test(test, grCtxOptions); }
// Wait for any remaining parallel work to complete (including any spun off of serial tasks).
parallel.wait();
gDefinitelyThreadSafeWork.wait();
// We'd better have run everything.
SkASSERT(gPending == 0);
// Make sure we've flushed all our results to disk.
JsonWriter::DumpJson();
// At this point we're back in single-threaded land.
sk_tool_utils::release_portable_typefaces();
if (gFailures.count() > 0) {
info("Failures:\n");
for (int i = 0; i < gFailures.count(); i++) {
info("\t%s\n", gFailures[i].c_str());
}
info("%d failures\n", gFailures.count());
return 1;
}
#ifdef SK_PDF_IMAGE_STATS
SkPDFImageDumpStats();
#endif // SK_PDF_IMAGE_STATS
SkGraphics::PurgeAllCaches();
info("Finished!\n");
return 0;
}
Generate list of GPU contexts outside SurfaceTest tests Add support for feeding the tests with contexts directly to the unit test framework. This fixes the problem where tests are more complex than needed just in order to run the test code with multiple backends. Also makes it possible to change the logic how contexts are created. Instead of direct numbering, the different testable contexts may be generated from filtered cross-product of context options. For example: currently NVPR is a type of context. However, it could be also an on/off feature of any context. In order to test this kind of context, the enumeration can not be just of context type. It's simpler to move the enumeration out of the tests. A test targeting both normal and GPU backends would look like: static void test_obj_behavior(skiatest::Reporter* reporter, SkObj* obj, [other params] ) { ... test with obj and param .. } DEF_TEST(ObjBehavior, reporter) { for (auto& object : generate_object) { for (auto& other_param : generate_other_variant) { test_obj_behavior(reporter, object, other_param); } } } #if SK_SUPPORT_GPU DEF_GPUTEST_FOR_ALL_CONTEXTS(ObjBehavior_Gpu, reporter, context) { for (auto& object : generate_gpu_object) { for (auto& other_param : generate_other_variant) { test_obj_behavior(reporter, object, other_param); } } } #endif Uses the feature in SurfaceTests as an example. Moves SkSurface -related tests from ImageTest to SurfaceTest. BUG=skia:2992 Review URL: https://codereview.chromium.org/1446453003
2015-11-20 21:32:24 +00:00
// TODO: currently many GPU tests are declared outside SK_SUPPORT_GPU guards.
// Thus we export the empty RunWithGPUTestContexts when SK_SUPPORT_GPU=0.
namespace skiatest {
Generate list of GPU contexts outside SurfaceTest tests Add support for feeding the tests with contexts directly to the unit test framework. This fixes the problem where tests are more complex than needed just in order to run the test code with multiple backends. Also makes it possible to change the logic how contexts are created. Instead of direct numbering, the different testable contexts may be generated from filtered cross-product of context options. For example: currently NVPR is a type of context. However, it could be also an on/off feature of any context. In order to test this kind of context, the enumeration can not be just of context type. It's simpler to move the enumeration out of the tests. A test targeting both normal and GPU backends would look like: static void test_obj_behavior(skiatest::Reporter* reporter, SkObj* obj, [other params] ) { ... test with obj and param .. } DEF_TEST(ObjBehavior, reporter) { for (auto& object : generate_object) { for (auto& other_param : generate_other_variant) { test_obj_behavior(reporter, object, other_param); } } } #if SK_SUPPORT_GPU DEF_GPUTEST_FOR_ALL_CONTEXTS(ObjBehavior_Gpu, reporter, context) { for (auto& object : generate_gpu_object) { for (auto& other_param : generate_other_variant) { test_obj_behavior(reporter, object, other_param); } } } #endif Uses the feature in SurfaceTests as an example. Moves SkSurface -related tests from ImageTest to SurfaceTest. BUG=skia:2992 Review URL: https://codereview.chromium.org/1446453003
2015-11-20 21:32:24 +00:00
#if SK_SUPPORT_GPU
bool IsGLContextType(sk_gpu_test::GrContextFactory::ContextType type) {
return kOpenGL_GrBackend == GrContextFactory::ContextTypeBackend(type);
}
bool IsVulkanContextType(sk_gpu_test::GrContextFactory::ContextType type) {
return kVulkan_GrBackend == GrContextFactory::ContextTypeBackend(type);
}
bool IsRenderingGLContextType(sk_gpu_test::GrContextFactory::ContextType type) {
return IsGLContextType(type) && GrContextFactory::IsRenderingContext(type);
}
bool IsNullGLContextType(sk_gpu_test::GrContextFactory::ContextType type) {
return type == GrContextFactory::kNullGL_ContextType;
}
#else
bool IsGLContextType(int) { return false; }
bool IsVulkanContextType(int) { return false; }
bool IsRenderingGLContextType(int) { return false; }
bool IsNullGLContextType(int) { return false; }
Add config options to run different GPU APIs to dm and nanobench Add extended config specification form that can be used to run different gpu backend with different APIs. The configs can be specified with the form: gpu(api=string,dit=bool,nvpr=bool,samples=int) This replaces and removes the --gpuAPI flag. All existing configs should still work. Adds following documentation: out/Debug/dm --help config Flags: --config: type: string default: 565 8888 gpu nonrendering Options: 565 8888 debug gpu gpudebug gpudft gpunull msaa16 msaa4 nonrendering null nullgpu nvprmsaa16 nvprmsaa4 pdf pdf_poppler skp svg xps or use extended form 'backend(option=value,...)'. Extended form: 'backend(option=value,...)' Possible backends and options: gpu(api=string,dit=bool,nvpr=bool,samples=int) GPU backend api type: string default: native. Select graphics API to use with gpu backend. Options: native Use platform default OpenGL or OpenGL ES backend. gl Use OpenGL. gles Use OpenGL ES. debug Use debug OpenGL. null Use null OpenGL. dit type: bool default: false. Use device independent text. nvpr type: bool default: false. Use NV_path_rendering OpenGL and OpenGL ES extension. samples type: int default: 0. Use multisampling with N samples. Predefined configs: gpu = gpu() msaa4 = gpu(samples=4) msaa16 = gpu(samples=16) nvprmsaa4 = gpu(nvpr=true,samples=4) nvprmsaa16 = gpu(nvpr=true,samples=16) gpudft = gpu(dit=true) gpudebug = gpu(api=debug) gpunull = gpu(api=null) debug = gpu(api=debug) nullgpu = gpu(api=null) BUG=skia:2992 Committed: https://skia.googlesource.com/skia/+/e13ca329fca4c28cf4e078561f591ab27b743d23 GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1490113005 Committed: https://skia.googlesource.com/skia/+/c8b4336444e7b90382e04e33665fb3b8490b825b Committed: https://skia.googlesource.com/skia/+/9ebc3f0ee6db215dde461dc4777d85988cf272dd Review URL: https://codereview.chromium.org/1490113005
2015-12-23 09:33:00 +00:00
#endif
void RunWithGPUTestContexts(GrContextTestFn* test, GrContextTypeFilterFn* contextTypeFilter,
Reporter* reporter, GrContextFactory* factory) {
#if SK_SUPPORT_GPU
#if defined(SK_BUILD_FOR_UNIX) || defined(SK_BUILD_FOR_WIN) || defined(SK_BUILD_FOR_MAC)
static constexpr auto kNativeGLType = GrContextFactory::kGL_ContextType;
#else
static constexpr auto kNativeGLType = GrContextFactory::kGLES_ContextType;
#endif
for (int typeInt = 0; typeInt < GrContextFactory::kContextTypeCnt; ++typeInt) {
GrContextFactory::ContextType contextType = (GrContextFactory::ContextType) typeInt;
// Use "native" instead of explicitly trying OpenGL and OpenGL ES. Do not use GLES on
// desktop since tests do not account for not fixing http://skbug.com/2809
if (contextType == GrContextFactory::kGL_ContextType ||
contextType == GrContextFactory::kGLES_ContextType) {
if (contextType != kNativeGLType) {
continue;
}
}
ContextInfo ctxInfo = factory->getContextInfo(contextType,
GrContextFactory::ContextOverrides::kDisableNVPR);
if (contextTypeFilter && !(*contextTypeFilter)(contextType)) {
continue;
}
Revert "Revert "Add API for flushing surfaces with gpu semaphores"" This reverts commit 7292231905c34ed290ba479338f26b56ae2a7792. This change relands the original plus the follow on change: https://skia-review.googlesource.com/20059. Additionally it adds a blacklist for the mac intel bots which don't see to respect the added fences on the GPU. Original change's description: > Revert "Add API for flushing surfaces with gpu semaphores" > > This reverts commit 66366c697853e906d961ae691e2bc5209cdcfa62. > > Reason for revert: Failing test on mac bots > > Original change's description: > > Add API for flushing surfaces with gpu semaphores > > > > BUG=skia: > > > > Change-Id: Ia4bfef784cd5f2516ceccafce958be18a86f91d1 > > Reviewed-on: https://skia-review.googlesource.com/11488 > > Commit-Queue: Greg Daniel <egdaniel@google.com> > > Reviewed-by: Brian Salomon <bsalomon@google.com> > > Reviewed-by: Forrest Reiling <freiling@google.com> > > TBR=egdaniel@google.com,jvanverth@google.com,bsalomon@google.com,brianosman@google.com,freiling@google.com > > Change-Id: I75633a2732d2d48b1926f9ad818a9f1a9196d211 > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Bug: skia: > Reviewed-on: https://skia-review.googlesource.com/20063 > Commit-Queue: Greg Daniel <egdaniel@google.com> > Reviewed-by: Greg Daniel <egdaniel@google.com> TBR=egdaniel@google.com,jvanverth@google.com,bsalomon@google.com,brianosman@google.com,freiling@google.com Change-Id: I4dc6c0e1deb0398eeb165a34f0a26af7a58259f1 Reviewed-on: https://skia-review.googlesource.com/20141 Commit-Queue: Greg Daniel <egdaniel@google.com> Reviewed-by: Greg Daniel <egdaniel@google.com>
2017-06-16 13:45:32 +00:00
ReporterContext ctx(reporter, SkString(GrContextFactory::ContextTypeName(contextType)));
if (ctxInfo.grContext()) {
(*test)(reporter, ctxInfo);
ctxInfo.grContext()->flush();
}
ctxInfo = factory->getContextInfo(contextType,
GrContextFactory::ContextOverrides::kRequireNVPRSupport);
if (ctxInfo.grContext()) {
(*test)(reporter, ctxInfo);
ctxInfo.grContext()->flush();
Generate list of GPU contexts outside SurfaceTest tests Add support for feeding the tests with contexts directly to the unit test framework. This fixes the problem where tests are more complex than needed just in order to run the test code with multiple backends. Also makes it possible to change the logic how contexts are created. Instead of direct numbering, the different testable contexts may be generated from filtered cross-product of context options. For example: currently NVPR is a type of context. However, it could be also an on/off feature of any context. In order to test this kind of context, the enumeration can not be just of context type. It's simpler to move the enumeration out of the tests. A test targeting both normal and GPU backends would look like: static void test_obj_behavior(skiatest::Reporter* reporter, SkObj* obj, [other params] ) { ... test with obj and param .. } DEF_TEST(ObjBehavior, reporter) { for (auto& object : generate_object) { for (auto& other_param : generate_other_variant) { test_obj_behavior(reporter, object, other_param); } } } #if SK_SUPPORT_GPU DEF_GPUTEST_FOR_ALL_CONTEXTS(ObjBehavior_Gpu, reporter, context) { for (auto& object : generate_gpu_object) { for (auto& other_param : generate_other_variant) { test_obj_behavior(reporter, object, other_param); } } } #endif Uses the feature in SurfaceTests as an example. Moves SkSurface -related tests from ImageTest to SurfaceTest. BUG=skia:2992 Review URL: https://codereview.chromium.org/1446453003
2015-11-20 21:32:24 +00:00
}
}
#endif
}
} // namespace skiatest