fuzzer generated a cubic with a very tiny loop at one end.
Breaking the cubic in two generated a piece so small it was
discarded, passing the original along.
Instead, use the loop detection as a hint that passing
the original may fail further along. Keep the big piece,
which is more linear than the original;
and change its end to match the original input if the
small piece isn't worth keeping.
R=kjlubick
Bug: skia:7480
Change-Id: If4c5cd46dbf7e16526269f2854e4f2278f427461
Reviewed-on: https://skia-review.googlesource.com/103100
Reviewed-by: Kevin Lubick <kjlubick@google.com>
Commit-Queue: Cary Clark <caryclark@skia.org>
Also adds a presubmit to prevent adding trailing whitespace to source
code in the future.
Change-Id: I41a4df81487f6f00aa19b188f0cac6a3377efde6
Reviewed-on: https://skia-review.googlesource.com/57380
Reviewed-by: Ravi Mistry <rmistry@google.com>
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
Moves cubic root finding logic out of GrPathUtils and
PathOpsCubicIntersectionTest, and unifies it in SkGeometry.
"Normalizes" the homogeneous parameter values of the roots, rather
than the cubic inflection function. Does this normalization by
twiddling the exponents instead of division (which causes a loss of
precision).
Abandons the built-in derivatives in GrCubicEffect. These don't have
high enough precision on many mobile gpus. Instead we pass the KLM
matrix to the vertex shader via uniform, where we can use it to set up
new linear functionals from which the fragment shader can calculate
the gradient of the implicit function.
Bug: skia:4410
Change-Id: Ibd64e999520adc8cdef7803a492d3699995aef5a
Reviewed-on: https://skia-review.googlesource.com/19017
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
- Updates the logic to reflect the Loop-Blinn paper instead of the GPU
gems website.
- Removes the threshold for detecting local cusps. The serpentine
codepath works for these cusps anyway, so what we really want to know
is whether the discriminant is negative.
- Makes sure to not scale the inflection function by 1/0.
- Shifts the inflection function coefficients in d[] so they match the
paper.
- Stores the cubic discriminant in d[0].
Bug: skia:
Change-Id: I909a522a0fd27c9c8dfbc27d968bc43eeb7a416f
Reviewed-on: https://skia-review.googlesource.com/13304
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
The path contains a cubic with a very tight curve.
Split the cubic into pieces so that the individual
curves are better behaved.
Use both inflections and max curvature to
potentially split cubics. Since this may require
a bit of work, preflight to ignore cubics that
monotonically change in x and y.
Only one of the three tests referred to by the bug
below repro'd. Use path.dumpHex() instead of
path.dump() to capture the crashing data.
TBR=reed@google.com
BUG=skia:6041
Change-Id: I29a264f87242cacc7c421e7685b90aca81621c74
Reviewed-on: https://skia-review.googlesource.com/5702
Reviewed-by: Cary Clark <caryclark@google.com>
Commit-Queue: Cary Clark <caryclark@google.com>
Many old pathops-related fuzz failures have built up while
the codebase was under a state a flux. Now that the code
is stable, address these failures.
Most of the CL plumbs the debug global state to downstream
routines so that, if the data is not trusted (ala fuzzed)
the function can safely exit without asserting.
TBR=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2426173002
Review-Url: https://chromiumcodereview.appspot.com/2426173002
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
If a curve has the identical start and control points, the
initial or final tangent can't be trivally determined. The
perpendicular to the tangent is used to measure coincidence.
Add logic for cubics, quadratics, and conics, to use the
secondary control points or the end points if the initial
control point alone can't determine the tangent.
Add debugging (currently untriggered by exhaustive testing)
to detect zero-length tangents which are not at the curve
endpoints.
Increase the number of temporary intersecions gathered from
10 to 12 but reduce the max passed in by cubic intersection from
27 to 12. Also, add checks if the max passed exceeds the
storage allocated.
When cleaning up parallel lines, choose the intersection which
is on the end of both segments over the intersection which
is on the end of a single segment.
TBR=reed@google.com
BUG=425140,516266
Review URL: https://codereview.chromium.org/1288863004
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
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
PathOps tests internal routines direcctly. Check to make sure that
test points, lines, quads, curves, triangles, and bounds read from
arrays are valid (i.e., don't contain NaN) before calling the
test function.
Repurpose the test flags.
- make 'v' verbose test region output against path output
- make 'z' single threaded (before it made it multithreaded)
The latter change speeds up tests run by the buildbot by 2x to 3x.
BUG=
Review URL: https://codereview.chromium.org/19374003
git-svn-id: http://skia.googlecode.com/svn/trunk@10107 2bbb7eff-a529-9590-31e7-b0007b416f81
Modify line intersections to first
- match exact ends
- compute intersections
- match near ends
where the exact ends are preferred, then near matches, then
computed matches. This pulls matches towards existing end points
when possible, and keeps intersection distances consistent with
different line/line line/quad and line/cubic computations.
BUG=
Review URL: https://codereview.chromium.org/19183003
git-svn-id: http://skia.googlecode.com/svn/trunk@10073 2bbb7eff-a529-9590-31e7-b0007b416f81
Replace SkTDArray with SkTArray and use SkSTArray when
the probable array size is known.
In a couple of places (spans, chases) the arrays are
constructed using insert() so SkTArrays can't be used for
now.
Also, add an optimization to cubic subdivide if either end
is zero or one.
BUG=
Review URL: https://codereview.chromium.org/16951017
git-svn-id: http://skia.googlecode.com/svn/trunk@9635 2bbb7eff-a529-9590-31e7-b0007b416f81
This is a major change resulting from a minor
tweak. In the old code, the intersection point
of two curves was shared between them, but the
intersection points and end points of sorted edges was
computed directly from the intersection T value.
In this CL, both intersection points and sorted points
are the same, and intermediate control points are computed
to preserve their slope.
The sort itself has been completely rewritten to be more
robust and remove 'magic' checks, conditions that empirically
worked but couldn't be rationalized.
This CL was triggered by errors generated computing the clips
of SKP files. At this point, all 73M standard tests work and
at least the first troublesome SKPs work.
Review URL: https://codereview.chromium.org/15338003
git-svn-id: http://skia.googlecode.com/svn/trunk@9432 2bbb7eff-a529-9590-31e7-b0007b416f81
Try to fix the 32 bit build by making some math
decisions more robust.
Rewrite the cubic intersection special case that
detects if only end points are shared.
Rewrite the angle sort setup that computes whether
a cubic bends to the left or right.
git-svn-id: http://skia.googlecode.com/svn/trunk@8726 2bbb7eff-a529-9590-31e7-b0007b416f81
standardize tests
use SK_ARRAY_COUNT everywhere
debug why x87 differs from SIMD 64
various platform specific fixes
git-svn-id: http://skia.googlecode.com/svn/trunk@8689 2bbb7eff-a529-9590-31e7-b0007b416f81
- fix rand for Android
- build unit test on linux
- use atomic inc in test count
- add casting for Android
git-svn-id: http://skia.googlecode.com/svn/trunk@8610 2bbb7eff-a529-9590-31e7-b0007b416f81
This CL depends on
https://codereview.chromium.org/12827020/
"Add base types for path ops"
The intersection of a line, quadratic, or cubic
with another curve (or with itself) is found by
solving the implicit equation for the curve pair.
The curves are first reduced to find the simplest
form that will describe the original, and to detect
degenerate or special-case data like horizontal and
vertical lines.
For cubic self-intersection, and for a pair of cubics,
the intersection is found by recursively
approximating the cubic with a series of quadratics.
The implicit solutions depend on the root finding
contained in the DCubic and DQuad structs, and
the quartic root finder included here.
Review URL: https://codereview.chromium.org/12880016
git-svn-id: http://skia.googlecode.com/svn/trunk@8552 2bbb7eff-a529-9590-31e7-b0007b416f81
