Note also the transform is still not good: the same parameters are applied
to convert inner arc, outer arc and middle arc of a thick arc to segments.
But these parameters depend on arc radius (or circle radius) value.
The error between the "right" shape and the polygon is slightly smaller.
The approximations of the inner arc and the outer arc are better, but not perfect.
Also add (for test and debug purpose only) the code to show the polygon when
drawing a PCB_ARC (currently disabled)
Arcs with small radius can be approximated with very few segments.
However, if the thickness is large, relative to the radius, the approximation
must be based on the external radius, not the arc radius.
The difference can be significant.
This is especially noticeable for these graphic arcs in filled zones.
When running a python script is run from a console, outside Kicad, a wxWidgets
alert was always show (even in release mode)
This is due to calling wxStandardPaths::Get() with no wxApp open( i.e. wxTheApp == O)
The fix uses a dummy wxApp in this case.
Segment approximation for arcs leaves small "indents" where the rounded
edges of two segments don't quite meet.
This gives closer approximation by calculating the inner and outer joint
points based on the radius, eliminating the indents.
This was leaking windows headers and partial wx headers to 1084 compilation units......
This also means math/util.h is leaking to 1084 compilation units which seems a bit high too.
Use the connecting straight tracks even if not exactly parallel - allow
an error margin configurable in ADVANCED_CFG (default 1 degree). Also
be less strict about end point matching and use the width of the track
as the criteria to determine suitability.
Finally, delete any short lengths of track at the end of the operation
and amend the arc end points to keep connectivity.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/7967
It appears as though there was an optimization to skip testing segments if
one of their endpoints (and only the A point) was inside. Unclear the reason
for this, but I'm going to assume that it was intended to optimize the case
where both points are inside (like the point case above it).
Before, it was possible for the canvas to steal the
focus if the cursor was only moved over it and not
clicked. Now, the canvas will only take the focus if
the frame is active.
Fixes https://gitlab.com/kicad/code/kicad/issues/7233