Most changes stem from working on an examples bracketed
by #if DEBUG_UNDER_DEVELOPMENT // tiger
These exposed many problems with coincident curves,
as well as errors throughout the code.
Fixing these errors also fixed a number of fuzzer-inspired
bug reports.
* Line/Curve Intersections
Check to see if the end of the line nearly intersects
the curve. This was a FIXME in the old code.
* Performance
Use a central chunk allocator.
Plumb the allocator into the global variable state
so that it can be shared. (Note that 'SkGlobalState'
is allocated on the stack and is visible to children
functions but not other threads.)
* Refactor
Let SkOpAngle grow up from a structure to a class.
Let SkCoincidentSpans grow up from a structure to a class.
Rename enum Alias to AliasMatch.
* Coincidence Rewrite
Add more debugging to coincidence detection.
Parallel debugging routines have read-only logic to report
the current coincidence state so that steps through the
logic can expose whether things got better or worse.
More functions can error-out and cause the pathops
engine to non-destructively exit.
* Accuracy
Remove code that adjusted point locations. Instead,
offset the curve part so that sorted curves all use
the same origin.
Reduce the size (and influence) of magic numbers.
* Testing
The debug suite with verify and the full release suite
./out/Debug/pathops_unittest -v -V
./out/Release/pathops_unittest -v -V -x
expose one error. That error is captured as cubics_d3.
This error exists in the checked in code as well.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2128633003
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2128633003
Review-Url: https://codereview.chromium.org/2128633003
Fail out in a couple of new places when the input data is very
large and exceeds the limits of the pathops machinery.
Most of the change here plumbs in a way to exclude an assert in
one of these exceptional cases. The current SkAddIntersection
implementation and the inner functions it calls has no way to
report an error to the root caller for an early exit, so rather
than add that in, exclude the assert when the test that would
trigger it runs (allowing the test to otherwise ensure that it
properly fails).
TBR=reed@google.com
BUG=617586,617635
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2046713003
Review-Url: https://codereview.chromium.org/2046713003
This fix is a tradeoff. It changes intersection to
treat a case where one coincident run is intersected at one point
and the other edge is not as continuing to be a span.
The old code tried to treat this as a single point.
The old code is probably right, but this change alone
made the data structures inconsistent. Later, extending
the coincident runs would fail by incorrectly discarding
the single point intersection.
As a result, this fixes the security test and one other, but
makes a different test fail. Isolating the failure uncovered
a reduced case that fails with and without the change, so
there are more serious problems here. Those problems are
addressed in a separate CL.
Many of the test edits below remove ill-thought out debugging
messaging that fire off global state, which isn't usable
in a multi-threaded test environment.
In the end, with this fix, all existing tests (modulo one
new failure and one new non-failure) pass in debug and
in the extended release test suites.
TBR=reed@google.com
BUG=614248
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2018513003
Review-Url: https://codereview.chromium.org/2018513003
Iterating through the 903K skps that represent the
imagable 1M top web pages triggers a number of
bugs, some of which are addressed here.
Some web pages trigger intersecting cubic
representations of arc with their conic
counterparts. This exposed a flaw in coincident
detection that caused an infinite loop. The loop
alternatively extended the coincident section and,
determining the that the bounds of the curve pairs
did not overlap, deleted the extension.
Track the number of times the coincident detection
is called, and if it exceeds an empirically found
limit, assume that the curves are coincident and
force it to be so.
The loop count limit can be determined by enabling
DEBUG_T_SECT_LOOP_COUNT and running all tests. The
largest count is reported on completion.
Another class of bugs was caused by concident
detection duplicating nearly identical points that
had been merged earlier. To track these bugs, the
'handle coincidence' code was duplicated as a
const debug variety that reported if one of a
dozen or so irregularities are present; then it is
easier to see when a block of code that fixes one
irregularity regresses another.
Creating the debug const code version exposed some
non-debug code that could be const, and some that
was experimental and could be removed. Set
DEBUG_COINCIDENCE to track coincidence health and
handling.
For running on Chrome, DEBUG_VERIFY checks the
result of pathops against the same operation
using SkRegion to verify that the results are
nearly the same.
When visualizing the pathops work using
tools/pathops_visualizer.htm, set
DEBUG_DUMP_ALIGNMENT to see the curves after
they've been aligned for coincidence.
Other bugs fixed include detecting when a
section of a pair of curves have devolved into
lines and are coincident.
TBR=reed@google.com
Review URL: https://codereview.chromium.org/1394503003
Unfortunately, immintrin.h (which is also included by SkTypes)
includes xmmintrin.h which includes mm_malloc.h which includes
stdlib.h for malloc even though, from the implementation, it is
difficult to see why.
Fortunately, arm_neon.h does not seem to be involved in such
shenanigans, so building for Android will keep things sane.
TBR=reed@google.com
Doesn't change Skia API, just moves an include.
Review URL: https://codereview.chromium.org/1313203003
This should be a drop-in replacement for most for-loops to make them run in parallel:
for (int i = 0; i < N; i++) { code... }
~~~>
sk_parallel_for(N, [&](int i) { code... });
This is just syntax sugar over SkTaskGroup to make this use case really easy to write.
There's no more overhead that we weren't already forced to add using an interface like batch(),
and no extra heap allocations.
I've replaced 3 uses of SkTaskGroup with sk_parallel_for:
1) My unit tests for SkOnce.
2) Cary's path fuzzer.
3) SkMultiPictureDraw.
Performance should be the same. Please compare left and right for readability. :)
BUG=skia:
No public API changes.
TBR=reed@google.com
Review URL: https://codereview.chromium.org/1184373003
Confirm that no path ops tests are flaky, and clean up errors around
that. The test framework was incorrectly checking for >= MAX_ERRORS for
failure and <= MAX_ERRORS for success.
TBR=reed@google.com
Review URL: https://codereview.chromium.org/1140563003
Improve line/curve coincident detection and resolution. This fixed the remaining simple failures.
When an edge is unsortable, use the ray intersection to determine the angles' winding.
Deal with degenerate segments.
TBR=reed@google.com
BUG=skia:3588,skia:3762
Review URL: https://codereview.chromium.org/1140813002
This replacement shoots axis-aligned rays through all intersecting edges to find the outermost one either horizontally or vertically. The resulting code is smaller and twice as fast.
To support this, most of the horizontal / vertical intersection code was rewritten and standardized, and old code supporting the top-directed winding was deleted.
Contours were pointed to by an SkTDArray. Instead, put them in a linked list, and designate the list head with its own class to ensure that methods that take lists of contours start at the top. This change removed a large percentage of memory allocations used by path ops.
TBR=reed@google.com
BUG=skia:3588
Review URL: https://codereview.chromium.org/1111333002
All but 17 extended tests work.
A helper function is privately added to SkPath.h to permit a test to modify a given point in a path.
BUG=skia:3588
Review URL: https://codereview.chromium.org/1107353004
Extended tests (150M+) run to completion in release in about 6 minutes; the standard test suite exceeds 100K and finishes in a few seconds on desktops.
TBR=reed
BUG=skia:3588
Review URL: https://codereview.chromium.org/1037953004
Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer.
Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering.
Unify t and point values so that small intervals can be eliminated in one pass.
Break cubics up front to eliminate loops and cusps.
Make the Simplify and Op code more regular and eliminate arbitrary differences.
Add a builder that takes an array of paths and operators.
Delete unused code.
BUG=skia:3588
R=reed@google.com
Review URL: https://codereview.chromium.org/1037573004
Use the form SkDebugf("%s", arbitraryString) instead of
SkDebugf(arbitraryString).
Fixes the case where SkString::appendf-ing a string with "%%" and then
printing the string with SkDebugf would cause uninitialized read and
corrupted debug print.
ninja -C out/Debug tools && valgrind --leak-check=full
./out/Debug/render_pictures --config gpu -w q -r ...
...
==7307== Conditional jump or move depends on uninitialised value(s)
==7307== at 0x6908475: __printf_fp (printf_fp.c:1180)
==7307== by 0x6904267: vfprintf (vfprintf.c:1629)
==7307== by 0x6906E53: buffered_vfprintf (vfprintf.c:2313)
==7307== by 0x690188D: vfprintf (vfprintf.c:1316)
==7307== by 0x67E8F5: SkDebugf(char const*, ...) (SkDebug_stdio.cpp:18)
==7307== by 0x7983F1: GrContext::printCacheStats() const (GrTest.cpp:54)
==7307== by 0x408ECF: tool_main(int, char**) (render_pictures_main.cpp:480)
==7307== by 0x40913E: main (render_pictures_main.cpp:511)
==7307==
Budget: 2048 items 100663296 bytes
Entry Count: current 652 (651 budgeted, 0 wrapped, 297 locked, 638 scratch 32 0.000000ull), high 652
Entry Bytes: current 51087658 (budgeted 49826658, 49 0.000000ull, 1261000 unbudgeted) high 51087658
(observe "ull" instead of "% full")
(from mtklein)
This CL is not editing public API.
TBR=reed@google.com
Review URL: https://codereview.chromium.org/943453002
The fixes include
- detect when finding the active top loops between two possible answers
- preflight chasing winding to ensure answer is consistent
- binary search more often when quadratic intersection fails
- add more failure paths when an intersect is missed
While this fixes the chrome bug, reenabling path ops in svg should be deferred until additional fixes are landed.
TBR=
BUG=421132
Committed: https://skia.googlesource.com/skia/+/6f726addf3178b01949bb389ef83cf14a1d7b6b2
Review URL: https://codereview.chromium.org/633393002
Reason for revert:
Compile errors on bots
Original issue's description:
> These tests stress pathops by describing the union of circle-like paths that have tiny line segments embedded and double back to create near-coincident conditions.
>
> The fixes include
> - detect when finding the active top loops between two possible answers
> - preflight chasing winding to ensure answer is consistent
> - binary search more often when quadratic intersection fails
> - add more failure paths when an intersect is missed
>
> While this fixes the chrome bug, reenabling path ops in svg should be deferred until additional fixes are landed.
>
> TBR=
> BUG=421132
>
> Committed: https://skia.googlesource.com/skia/+/6f726addf3178b01949bb389ef83cf14a1d7b6b2TBR=caryclark@google.com
NOTREECHECKS=true
NOTRY=true
BUG=421132
Review URL: https://codereview.chromium.org/686843002
The fixes include
- detect when finding the active top loops between two possible answers
- preflight chasing winding to ensure answer is consistent
- binary search more often when quadratic intersection fails
- add more failure paths when an intersect is missed
While this fixes the chrome bug, reenabling path ops in svg should be deferred until additional fixes are landed.
TBR=
BUG=421132
Review URL: https://codereview.chromium.org/633393002
SkTaskGroup is like SkThreadPool except the threads stay in
one global pool. Each SkTaskGroup itself is tiny (4 bytes)
and its wait() method applies only to tasks add()ed to that
instance, not the whole thread pool.
This means we don't need to bring up new thread pools when
tests themselves want to use multithreading (e.g. pathops,
quilt). We just create a new SkTaskGroup and wait for that
to complete. This should be more efficient, and allow us
to expand where we use threads to really latency sensitive
places. E.g. we can probably now use these in nanobench
for CPU .skp rendering.
Now that all threads are sharing the same pool, I think we
can remove most of the custom mechanism pathops tests use
to control threading. They'll just ride on the global pool
with all other tests now.
This (temporarily?) removes the GPU multithreading feature
from DM, which we don't use.
On my desktop, DM runs a little faster (57s -> 55s) in
Debug, and a lot faster in Release (36s -> 24s). The bots
show speedups of similar proportions, cutting more than a
minute off the N4/Release and Win7/Debug runtimes.
BUG=skia:
Committed: https://skia.googlesource.com/skia/+/9c7207b5dc71dc5a96a2eb107d401133333d5b6fR=caryclark@google.com, bsalomon@google.com, bungeman@google.com, mtklein@google.com, reed@google.com
Author: mtklein@chromium.org
Review URL: https://codereview.chromium.org/531653002
Reason for revert:
Leaks, leaks, leaks.
Original issue's description:
> SkThreadPool ~~> SkTaskGroup
>
> SkTaskGroup is like SkThreadPool except the threads stay in
> one global pool. Each SkTaskGroup itself is tiny (4 bytes)
> and its wait() method applies only to tasks add()ed to that
> instance, not the whole thread pool.
>
> This means we don't need to bring up new thread pools when
> tests themselves want to use multithreading (e.g. pathops,
> quilt). We just create a new SkTaskGroup and wait for that
> to complete. This should be more efficient, and allow us
> to expand where we use threads to really latency sensitive
> places. E.g. we can probably now use these in nanobench
> for CPU .skp rendering.
>
> Now that all threads are sharing the same pool, I think we
> can remove most of the custom mechanism pathops tests use
> to control threading. They'll just ride on the global pool
> with all other tests now.
>
> This (temporarily?) removes the GPU multithreading feature
> from DM, which we don't use.
>
> On my desktop, DM runs a little faster (57s -> 55s) in
> Debug, and a lot faster in Release (36s -> 24s). The bots
> show speedups of similar proportions, cutting more than a
> minute off the N4/Release and Win7/Debug runtimes.
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/9c7207b5dc71dc5a96a2eb107d401133333d5b6fR=caryclark@google.com, bsalomon@google.com, bungeman@google.com, reed@google.com, mtklein@chromium.orgTBR=bsalomon@google.com, bungeman@google.com, caryclark@google.com, mtklein@chromium.org, reed@google.com
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Author: mtklein@google.com
Review URL: https://codereview.chromium.org/533393002
SkTaskGroup is like SkThreadPool except the threads stay in
one global pool. Each SkTaskGroup itself is tiny (4 bytes)
and its wait() method applies only to tasks add()ed to that
instance, not the whole thread pool.
This means we don't need to bring up new thread pools when
tests themselves want to use multithreading (e.g. pathops,
quilt). We just create a new SkTaskGroup and wait for that
to complete. This should be more efficient, and allow us
to expand where we use threads to really latency sensitive
places. E.g. we can probably now use these in nanobench
for CPU .skp rendering.
Now that all threads are sharing the same pool, I think we
can remove most of the custom mechanism pathops tests use
to control threading. They'll just ride on the global pool
with all other tests now.
This (temporarily?) removes the GPU multithreading feature
from DM, which we don't use.
On my desktop, DM runs a little faster (57s -> 55s) in
Debug, and a lot faster in Release (36s -> 24s). The bots
show speedups of similar proportions, cutting more than a
minute off the N4/Release and Win7/Debug runtimes.
BUG=skia:
R=caryclark@google.com, bsalomon@google.com, bungeman@google.com, mtklein@google.com, reed@google.com
Author: mtklein@chromium.org
Review URL: https://codereview.chromium.org/531653002
This fixes all but one of those failures.
Major changes include:
- Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles.
- Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed.
- Line segments with ends nearly the same are treated as coincident first.
- Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident.
TBR=reed
Author: caryclark@google.com
Review URL: https://codereview.chromium.org/272153002
Mike K: please sanity check Test.cpp and skia_test.cpp
Feel free to look at the rest, but I don't expect any in depth review of path ops innards.
Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule.
This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data.
The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can.
Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct.
The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release.
R=mtklein@google.com, reed@google.com
Author: caryclark@google.com
Review URL: https://codereview.chromium.org/131103009
git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
