Using an approximation here results in different collision
results for segment<>arc than point<>arc, which
can cause odd behavior in the router.
(cherry picked from commit 5b3b0ff836)
Co-authored-by: Jon Evans <jon@craftyjon.com>
Check for vertex placement against original size. Since we are
potentially adding vertices that are out of order with the original
polygon, we only want to check against the first set of vertices when
checking for polygon collision
Fixes https://gitlab.com/kicad/code/kicad/-/issues/17664
(cherry picked from commit 8260c174e9)
Even if there are no other segments in the line chain, we still should
check for arc connections in cases of individual arcs on a line chain
(cherry picked from commit 6b915049d1)
The m_triangulationValid flag is used in several places without holding
the mutex, thus it should only ever be set when the triangulation is
guaranteed to be valid.
This can either be done by protecting both data and flag by the same
mutex, or updating the flag only after the triangulation has finished.
Also fix the case when the triangulation actually fails, the flag should
not be set in this case.
While at it, simplify the recalculation check. Only if both the
triangulation is valid, and the data hash is unchanged the recalculation
can be skipped - this is typically the case when two threads try to
update the cache concurrently, the second one will block at the mutex,
and will see the valid data after the first thread has finished.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/17180
(cherry picked from commit e07b4ce8e4)
When checking collisions, the SHAPE_POLY_SET::Collide() routine is not
const because it will regenerate the triangulation cache if it is out of
date (using a const_cast, grrr). This sidesteps the issue by assigning
a mutex to the triangulation caching
Fixes https://gitlab.com/kicad/code/kicad/-/issues/17180
(cherry picked from commit 81cb6d0c3f)
SHAPE_LINE_CHAIN_BASE can include things that are not SHAPE_LINE_CHAIN
like the triangle SHAPE_LINE primitive. To check arcs, we need to avoid
these elements as they do not have the full SHAPE_LINE_CHAIN class
available.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/17327
(cherry picked from commit 347c50b26d)
The clearance between two zones could be rather slow. This was in part
to trying to do triangle-triangle collisions between zones when we only
need outline collision and in part to the shape_line_chain collision
routine. The shape_line_chain collisions don't need to recreate
segments on each iteration and should instead create them once and using
this to check all collisions
Fixes https://gitlab.com/kicad/code/kicad/-/issues/17327
(cherry picked from commit 3cc1617f5a)
When we add vertices to the tesselation routines, we cannot reuse these
without the original vertex points.
It may be possible to copy and modify the vertices from the hint data so
that they are properly positioned but naive attempts (moving based on
first point) did not work, so for now, we disable the hint cache when
the vertex sizes do not match as this prevents OOB access
Fixes https://gitlab.com/kicad/code/kicad/-/issues/17621
(cherry picked from commit 06b199fd41)
Polygon intersections happen against the original outline, not against
the currently remaining polygon. This avoids pathalogical cases
Adds new simplification system to avoid duplicated points
Adds new edge-splitting algorithm to provide additional fall-back
Verifies that polygon cuts do not swap holes for outlines (negative
area)
Fixes https://gitlab.com/kicad/code/kicad/-/issues/17559
(cherry picked from commit c3f6a84d66)
Refactors `SHAPE_POLY_SET::fractureSingle()` to be more efficient, while
not changing the actual algorithm:
* increase cache locality by using contiguous arrays instead of what was
effectively a linked list
* reduce latency and jitter by replacing per-edge allocator calls with
ahead-of-time std::vector reserves
* increase cache efficiency by making the vertex struct smaller
* replace O(n^2) leftmost edge search with O(n log n) std::sort
* sort the polygons instead of the edges
* cut iteration count in half in the remaining O(polygons * edges) part
(cherry picked from commit e98c9f283f)
The minor triangles cost the same amount of compute time as larger ones
but do not have a material effect on the zone display. This skips these
minor triangles when earcutting
(cherry picked from commit a58e7b37ff)
Logging is useful when we find an area that cannot be triangulated.
This will be used to generated test cases.
Skipping minor untesselated areas means that the polygon will still be
considered fully tesselated (and not sent back again and again) even if
the tesselation misses an area less than the configured limit.
Currently, this is 31^2nm.
(cherry picked from commit 7e7fec69f6)
Also fixes length preview when hovering over joints.
Also prevents asserts when hovering over arcs.
Also prevents adding a tuning pattern on other tuning pattern.
Allow for overlapping points in input polygon. This is relatively
common on a grid and should be used as potential split points to shave
additional ears
Fixes https://gitlab.com/kicad/code/kicad/-/issues/16933
Failed triangulation can have multiple causes but we don't have another
approach, so skipping the cache just causes the same issue to be raised
again and again
Fixes https://gitlab.com/kicad/code/kicad/-/issues/16946
This maintains the primary goal of ensuring that bad splits do not allow
the triangulation process to continue. Instead, we boot this back up to
the top-level where we will simplify the polygon before triangulating
again.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/16624
Alex Shvartzkop
Jeff Young
Jon Evans
Seth Hillbrand
JamesJ
StefanBruens
Céleste Wouters
Roberto Fernandez Bautista