This patch does not add finger joints to the vertical walls to attach the
retainer walls as the walls are likely to be to small to be able to make use
of them. In hte long run they should be added, though.
Now that the text() function has a fontsize parameter, setting the font size
on ctx before calling text() has no impact.
Resolves: #173
Thanks to Marko Luther for reporting and suggesting this fix!
TrayLayout was still trying to produce a drawing and failed. Now TrayLayout is
not using most of the infrastructure and only TrayLayout2 has the capacity to
create a drawing.
Resolves: #172
Previously, the extra space was added above the first pin and below the
(optional) second pin. This would cause the lid to misalign with the
holes in the back wall, making it harder to close the lid.
With this commit, the extra space is divided evenly above and below the
nominal position of the pin.
For this to work, the startwidth (or endwidth, depending on the side) is
slightly adjusted (since the start of the first pin slot is actually
generated by the adjacent edge). This moves the starting point, so the
location of the spring hole is adjusted to compensate for that.
This allows modifying the height of the wall that extends above the lid.
The box back wall (with holes) would already support edge_width, so this
makes it supported on all three walls where needed.
When rendering the FingerHoleEdge, the current position would be
corrected for burn (e.g. it would be on the center of the laser path,
rather than the actual edge of the final part). However, the FingerHoles
class (or actually the rectangularHole method called by it) that draws
the actual holes already corrects for burn, which caused the hole
position to be offset by the burn value.
This offset was particularly problematic when making a sliding lid,
since the holes would not align perfectly with the lid, making it hard
to close the lid. With this change, the sliding slots in the side walls
and the holes in the back wall align perfectly (though only when the lid
settings have zero play, which is a separate problem).
This commit fixes FingerHoleEdge (and its subclass LidHoleEdge), but
there might be other users of the FingerHole class (usually indirectly
through boxes.fingerHolesAt) that still have the same problem.
Marius Bergmann
Florian Festi
gmorell
Guillaume Collic
Stefan Schwarz
Matthijs Kooijman