Some items only used within algorithms were never cached
Because they are hard to manually clean up
But caching them does help the algorithms a lot
So this solution is the best I can think of
(cherry picked from commit d29c07a663)
While ERROR_INSIDE was good for plotting, 3D generation, etc., it's
not good for generating router hulls.
Also reverts part of the change to always use polygons for PNS::SOLIDs. A single shape in a SHAPE_COMPOUND will be faster (and
more accurate).
Fixes https://gitlab.com/kicad/code/kicad/-/issues/14898
Also give them more info that might be queried by custom DRC rules.
This currently just includes positioning info (it already included
nets and principal layer). We *could* also try to figure out all
layers for dummy vias, and/or widths for dummy tracks / arcs....
Graphic shapes (excluding text) can now have nets when on
copper layers. Shapes behave like tracks in that they will
pick up nets from connected pads, and follow track opacity
settings.
We were calculating the same thing in three locations and we missed
adding the clearance from the footprints in, resulting in bad fills and
missed drc errors (see QA addition)
This was causing intermittent shove hiccups/freezes as RULE_RESOLVER::GetClearance() could return a bogus clearance value, taking
a cached value for an item that does not exist anymore. If the allocator/stack accidentally reclaimed such item's address - we were getting
very nasty and difficult to reproduce misbehaviours of the shove algorithm. Hopefully this patch fixes this. More info in the comments.
Note to self - I'm still not fully convinced I want PNS::ITEM pointers as the cache key, probably a unique counter would be better here.
Don't open-code knockout text shape generation in several different
places.
Make sure triangulated knockout text gets clearance added when
specified.
Collapse duplicated footprint text item plot routine (they're no
longer any different from plotting pcb text items).
Change teardrop generation to rely more heavily on BOARD_CONNECTIVITY
for improved performance.
Add updating of teardrops on BOARD_COMMIT::Push().
Also converts m_CopperItemRTreeCache to std::shared_ptr.
We don't copy it around anyway, and having to create a new set
of std::unique_ptr's for each operation is likely to be more
expensive than std::shared_ptr's overhead.
About 1/3 of callers to Parent() don't care if they get the immediate
parent or not, about 1/3 want only the immediate parent, and about 1/3
want the hole parent's Parent().
I had earlier changed PNS::ITEM::HOLE to override Parent() and return
the hole parent's Parent(), but since the callers are pretty evenly
split I've reverted that and added BoardItem() for callers who want the
eventual BOARD_ITEM (whether a direct parent or a grandparent).
(Also removes a dead routine so I didn't have to figure out which of
the two it wanted....)
(While we don't currently have holes without a parent pad or via, you
could imagine using HOLEs in the future to represent a hole in a first-class
board outline, so better not to build this assumption in.)
Jeff Young
jean-pierre charras
Alex Shvartzkop
Seneral
Jon Evans
Marek Roszko
Tomasz Wlostowski
Seth Hillbrand
Mike Williams
Lucas Dumont