====================================================================
When the ellipse to be projected and the sketch plane are parallel, the original code
by shermelin provided for a translation of the original ellipse, which would be the best solution
if it weren't because the Sketcher, internally, works under the assumption of a normal vector to the
sketcher plane being (0,0,1). If the original ellipse is parallel to the sketch plane, but the sketch
plane is not the XY plane, the copy and translation would result in a ellipse not in the XY plane of the
Sketcher. Then the sketcher internals will not properly consider its dimensions.
The solution applied here is to default to the general method for non-parallel planes.
It solves:
https://forum.freecadweb.org/viewtopic.php?f=3&t=55284#p477522
========================================================
Users chennes and hyarion made me aware of this covereity issue:
Fixes Coverity: geoit can be end() when dereferenced
https://github.com/FreeCAD/FreeCAD/pull/4429/files#
When analysing the block where the dereferrencing appears, it
appears that it is a left-over that no longer makes sense:
- The algorithm classifies block constraints into those that are
not affected by any other driving constraint and those that are
affected by other driving constraints.
- The offending block deals with internal aligned geometry, thus
per definition has a driving internal alignment constraint, for which
the previous block already set the need of post-analysis.
- No matter what, the geometries, the complex one and the internal one
will have at least the driving internal alignment constraint, so they
cannot become "not affected by any other driving constraint".
- If the geometry had a block constraint on it, it was already added for
post-analysis in the previous block. If it did not have one block constraint,
the fact that it is internal aligned geometry is an irrelevant consideration.
Probably there was a point during development when this made sense, but with
the current post-analysis, it does not appear to make sense anymore. So the
block was removed.
This commit adds a unit test for blocked geometry (new block constraint).
Fixes Coverity issue:
CID 316539 (#1 of 1): Uninitialized scalar field (UNINIT_CTOR)
uninit_member: Non-static class member lastHasPartialRedundancies is not initialized in this constructor nor in any functions that it calls.
=================================================
Knot position is not calculated by the solver, but by OCCT when updating the
b-spline to conform to given pole positions, as mandated by the solver. Before
this commit, all constraints driving and non-driving operating on the knots required
and extra solve (from advanced solver dialog, or from the Python console), or a recompute
to be recomputed.
This commit introduces transparently re-solving at Sketch.cpp level if B-Splines are present,
so that when the Sketcher mandated solve returns, the geometry is fully solved.
========================================
The geometry state stored in the geometryFacade is modified following a mutable model
(without setting the Geometry property on Constraint change), in order to avoid coupling
the addition/removal of a constraint with a change of the Geometry Property.
This design decision however interferes with the ability of the Geometry property to restore
the correct geometry state upon redo/undo.
While such a situation is rare in the case of Internal Alignment geometry, because constraint
addition/removal is performed with the corresponding geometry addition/removal (within the same
transaction. That is not the case with the Block constraint (or another future general case where
the geometry state may be applied).
This commit leverages the synchronisation mechanism already in use for non-properties (e.g. external geometry or
vertex indices) to check and synchronise geometry state upon undo/redo and restore.
Bonus:
- addGeometryState is refactored to separate the checking logic from the setting logic.
Currently the sketch fillet tool deletes any constraints associated with
the two lines to be filleted. By leaving a vertex at the intersection,
we can instead preserve most constraints in reasonable ways.
Sketch fillet horizontal and vertical point-to-point constraint support
Also better future compatibility for point constraints, and some minor tweaks.
============================================================================
This specific constraint removes the free parameter of the previous implementation. This solves:
https://tracker.freecadweb.org/view.php?id=4501fixes#4501
However, this implementation of equal size produces zero gradients when coordinates of lines are aligned,
e.g. vertical or horizontal. These zero gradients, which are mathematically right ruin the diagnosis, which
regards corresponding elements as fully constraint (because they are locked from a solver point of view), when
they are simply locked, but are movable and constrainable. For this, when the rightful gradient is small enough
(<1e-10) it is substituted by a surrogate gradient of 1e-10, which solves the problem with the diagnose, which
treats as zero only values under 1e-13 (pivot threshold used in QR decomposition).
This special behaviour fixes the wrong detection here:
https://forum.freecadweb.org/viewtopic.php?f=8&t=53466&start=40#p464168
It also fixes this one:
https://forum.freecadweb.org/viewtopic.php?p=468585#p468587
==================================================================================================
Coin documentation indicates the size of a font is:
SoSFFloat SoFont::size Size of font. Defaults to 10.0.
For 2D rendered bitmap fonts (like for SoText2), this value is the height of a character in screen pixels. For 3D text, this value is the world-space coordinates height of a character in the current units setting (see documentation for SoUnits node).
However, with hdpi monitors, the coin font labels do not respect the size passed in pixels:
https://forum.freecadweb.org/viewtopic.php?f=3&t=54347&p=467610#p467610https://forum.freecadweb.org/viewtopic.php?f=10&t=49972&start=40#p467471
Because I (abdullah) have 96 dpi logical, 82 dpi physical, and I see a 35px font setting for a "1" in a datum label as 34px,
and I see kilsore and Elyas screenshots showing 41px and 61px in higher resolution monitors for the same configuration, I think
that coin pixel size has to be corrected by the logical dpi of the monitor. The rationale is that: a) it obviously needs dpi
correction, b) with physical dpi, the ratio of representation between kilsore and me is too far away.
This means that the following correction does not have a documented basis, but appears necessary so that the Sketcher is usable in
HDPI monitors.
========================================================
With the possibility to increase icon sizes via font configuration, the
default hardcoded 2.5 step for seeking the constraint position becomes too
small, causing constraints, specially when multi-icon, to superimpose geometry points,
removing the ability to pick them.
Changing the default gives some increased room for icon size.
==============================================================
A previous commit assumed values for lineWidths and pointSizes were integers. This commit fixes this.
=========================================================================================
- Long overdue refactor to avoid repetition during save/restore and copy.
- New interface to notify an extension when it is attached. It also enables the extension to gain
a pointer to the geometry container. This is intended to extend the functionality already existing
in Part::Geometry.
