- Ensure that the subshapes from the original shape will not be changed after nurbsconvert operation.
For that create the new vertexes (NewPoint(..)) as well as new curves (NewCurve(..)) and new surfaces (NewSurface(..)) before calling of Rebuild() method. Make copies of all vertexes impacted by curve or surface modifications. This eliminates necessity of creation of new vertices during recursive rebuilding of the entire shape.
- Compatibility with the old behavior of BRepTools_Modifier has been retained as an option. For that the new flag MutableInput has been added in the interface of the class.
- This patch also impacts other operations based on BRepTools_Modifier, in the sense that they also become safe regarding input shapes.
- Create new test cases. Some test cases with nurbsconvert command are changed to lock input shapes from modifications.
fix regressions
The algorithm of finding of extrema solutions of a point and a torus in Extrema_ExtPElS has been corrected for the case of torus having major radius equal to zero.
Test cases are added.
The static method CorrectSurfaceBoundaries() in the class IntTools_FaceFace,
which is used for correction of the boundaries of the faces before their intersection,
contains two inconsistent adjustments of the boundaries of the faces based on periodic surfaces.
One of these adjustments has been removed. The remaining adjustment, the one that is based on the
2D bounding boxes of the closed curves, should be sufficient.
Test case for the issue.
Intersection of curves and surfaces in HLRBRep is corrected: parametric limits of face are used instead of parametric limits of surface
Building of outlines (Contap_Contour) is corrected: method ComputeCloseLine is now used in all cases.
Correction of HLRBRep_Surface: usage of UVbounds-option
Correction of test cases according to the new behavior.
The max extension value of the faces in offset operation has been decreased
to 1.e+7 value to avoid the lack of precision in the calculations.
Update of test cases according to the new behavior
1) BRepBuilderAPI_MakeWire::Add (const TopTools_ListOfShape &L) method have been completely rewritten. The order of edges is not significant now.
2) The geometric proximity of free vertices from already existing wire and from input list of edges are also have been taken into account. If such vertices are coincident with each other then they are fused into the one. The original wire remains untouched topologically (yet the tolerances and points can be modified).
3) UBTreeFiller is used to speed up the process of picking of coincident vertices.
4) BRepLib now contains the 'new' method - BoundingVertex(..). The implemenation of this method are taken from BOPTools_AlgoTools::MakeVertex(..).
5) The '-unsorted' argument have been added to 'wire' command.
Conflicts:
src/QABugs/QABugs_20.cxx
Add missing include.
Eliminate warning.
1. IntTools_ShrunkRange::Perform
The algorithm of building shrunk range on the edge has been redesigned to make this range as big as possible.
By new definition it has to have the length not less than Precision::Confusion().
Although, the possibility of splitting of the edge has been taken into account. If it is impossible to put
vertex on edge in such a way that the tolerance spheres of the edge's vertices do not intersect the tolerance
sphere of the putting vertex the edge cannot be split. This possibility is saved in the new field
IntTools_ShrunkRange::myIsSplittable. It can be checked by the corresponding method IntTools_ShrunkRange::IsSplittable().
It returns TRUE if the shrunk range is computed successfully and it has the length more than the value of sum of two tolerance
values of the edge (for putting vertex, as its tolerance should not be less than the tolerance value of edge) and
two Precision::Confusion() values (to make two new edges valid by Shape validity criteria).
IntTools_ShrunkRange::myErrorStatus and IntTools_ShrunkRange::ErrorStatus() have been replaced by
the IntTools_ShrunkRange::myIsDone and IntTools_ShrunkRange::IsDone(). IntTools_ShrunkRange::IsDone() returns TRUE
if the shrunk range has been computed and it has length more than Precision::Confusion().
All computations of the parameters are performed using the GCPnts_AbscissaPoint, but if AbscissaPoint is unable to
compute the parameters the Resolution of the curve is used.
2. Boolean Operations algorithm now partially works with the edges that cannot be split or does not have the ShrunkData
at all (previously such edges have been considered as micro and just ignored). If by the result of some intersection such
edges should be split, i.e. the intersection vertex should be created, the algorithm just ignores them and no vertex is created.
But if such edges coincide with other shapes (other edges or faces) the algorithm uses them for creation of common blocks.
The information of the possibility for the edges to be split is saved in its PaveBlocks, in the new BOPDS_PaveBlock::myIsSplittable field.
It can be retrieved by the BOPDS_PaveBlock::IsSplittable() or BOPDS_PaveBlock::ShrunkData(), but these methods
make sense only after filling of the shrunk data for the pave block.
BOPTools_AlgoTools::IsMicroEdge() has an additional parameter that defines whether it is necessary to take into account
the possibility for the edge to be split or not. By default it is set to TRUE, i.e. by default the edge will be considered as micro
even if the shrunk range is computed, but it is too short for the edge to be split.
3. BOPAlgo_PaveFiller::PerformEF
To avoid creation of too close intersection vertices the intersection ranges of the edges participating in Edge/Face
intersections are reduced taking into account the common ranges computed during Edge/Edge intersections.
Thus, the Edge/Face intersection vertex is not created if it gets into a common range of the Edge/Edge intersection
between that edge and one of the face's edges. The tolerance value of Edge/Edge intersection vertex is increased
to reach the Edge/Face intersection.
4. Unification of the vertices of the section edges considered as micro edges.
If by the result of some Face/Face intersection the section edge is considered as micro edge,
the vertices of this edge will be united and the edge itself will be removed.
5. Test cases for the issues.
6. Adjusting test cases for issue CR27448.
Correction of code to avoid of usage of undefined associated parameters in order not to get exception in constructor.
Creation of test case for this issue
The problem of incorrect CUT is wrong Pcurve after projection onto cylinder of a small edge orthogonal to cylinder's axis. The result of projection is a line along V direction, but it should be along U.
ProjLib_Cylinder has no specific functionality to build projection in most common case, so it should return "not done" flag to start projection based on curve approximation.
Following changes are done:
1. ProjLib_Cylinder returns isDone=false in the following cases:
* the projected line is not parallel to cylinder's axis
* the plane of projected circle is not orthogonal to cylinder's axis
2. Test case added
Minor correction of test case
Computation of correct offset values in order to make correspondence (with adjusting to periods) between Domain of WLine and surface domain.
Creation of test cases for this issue.
Small correction in the test case.
The cause of errors is a non alphabetical order of complex entity, which do not affect export.
Move messages to debug mode.
Replace fail by warning.
Correction of test case bugs/vis/bug23153 according to the new behavior
additional check have been added to BRepLib::EnsureNormalConsistency() func.
Any edges with two adjacent faces will be skipped (while updating of their normals) if the number of discretization points (nodes) will be different for each of this faces on this edge.
test case
The following modifications were done for stability:
- Precision::PConfusion() was set as parameter of Classifier instead of 0.0
- Coinciding vertices of outlines are united to exclude neighborhood of common vertex from curve-curve intersection
- Draw command "hlrin3d" was modified to prevent exception
Minor corrections
Corrections of test cases for issue CR27557
1. Useless constructors have been deleted.
2. Value returned by gp_Parab2d::Directrix() method has been corrected to exclude reversing the directrix.
3. Documentation of gp_Parab2d class has been updated (in hxx-file).
4. Upgrade Guide has been updated according to corrections made in this issue.
Creation of test cases for this issue.
Interpolation state check is fixed to mark as valid only ordered sequences of poles.
Refactoring of Approx_SameParameter.cxx.
Test case is added.
Test cases are updated to the new behavior.
Update of testing cases
Calculating the offset vertices by the superposition of intersection vertices
between pairs of edges (BRepOffset_Inter2d::FuseVertices()).
To obtain stable result when calculating the superposition of vertices
they are sorted (BOPTools_AlgoTools::MakeVertex()).
The support of vertices has been added in nexplode command.
Small correction of tests cases for issue CR27540
- removal of excess interferences in case of simple hiding face
- correct processing of boundary edges coinciding with outlines
- change of API for testing Draw Command
- correction of Draw command "build3d"
- test cases are added.
Linux reference data is changed.
Partially fixed 0027340: Insufficient exact HLR speed
Number of samples in PolyPoly intersection algorithm tuned in each instance of generic class.
Default number of the samples in Geom2dAdaptor is changed according to the Adaptor2d_Curve2d class.
Minimal number of points in case of B-Spline is changed.
Minimal number of samples support is added in intersection algorithms.
Correct the algorithm in the function MergeEdges in ShapeUpgrade_UnifySameDomain.cxx so as to take into account the orientation of the next edge added to the chain.
Test cases for issue CR27521
Treatment for single singular point is added.
Test cases are updated to the new behavior.
New test cases are added,
Message of "bad" state is corrected.
Incorrect work of HLR algorithm caused by usage of wrong tolerances, wrong sampling of 2d curve and some other bugs are fixed.
Test cases for issue CR27341
Sometimes start point of the intersection line is in the surface boundary strictly. I.e. the parameter of this point in the surface can be equal to both 0 or 2*PI equivalently. It is important to chose correct parameter value.
The algorithm of prediction is based on monotonicity property (see CylCylMonotonicity(...) function in IntPatch_ImpImpIntersection_4.gxx). Now, this function is used wrongly. The fix improves this situation.
Small optimization in the code.
Creation of test cases .
The logic of returning value by the method BoundariesComputing() has been corrected.
1. In frame of the fix for #27282 issue, we have obtained several prolonged curves, which have common point(s). Fix for this issue joins these curves if it is possible.
2. ElCLib::InPeriod(...) method has been improved. Now it has become more faster (in general cases) and more reliable (in frame of FLT_OVERFLOW and DIVISION_BY_ZERO cases processing).
Creation of test case for issue #27302
Test case tests\bugs\modalg_6\bug27282_2 has been adjusted in accordance with its new behavior.
Method BRepLib::EncodeRegularity() is improved to set regularity GeomAbs_CN for edges lying on the same-domain surfaces (where derivatives on both surfaces are equal in all points).
DRAW command getedgeregularity is added to query regularity of the edge on specified faces.
DRAW command edgeregul is removed (this functionality is provided by command encoderegularity).
Added tests: bugs modalg_6 bug27272, bug27383_1, bug27383_2
Correction of shape name in test case for issue CR27272
ProjLib_Cone.cxx - correction wrong calculation of projection line on cone
GeomInt_IntSS_1.cxx - modification of method BuildPCurves(...) - adjusting first or last knots of 2d Curve
ProjLib_ComputeApprox.cxx - modification of method Function_SetUVBounds(...) for case projecting line on cone.
Modification of tests - removing first TODO
Test case for issue #27322
Various improvements in point-solid classifier:
1) Refactoring.
2) BndBoxTree is extracted into separate class.
3) UB-tree calculation is cashed to improve point-solid classification speed.
4) Ray / curve intersection improved by trimmed parameters and correct order.
5) Fixes in logic.
6) Calculation caching at the classifier level.
3D-claasifier now takes into the account the vertex/edges tolerances. If the given point lays inside the tolerance area of vertex or edge of the solid it's classified as TopAbs_ON.
The behavior of IntCurvesFace_Intersector::Perform was changed. Now it may use optional null-tolerance to classify 2d-point relatively to the given face.
UBTreeFiller is used to speedup intersection process between edges/vertices and the point.
The test case 'boolean gdml_private ZH2' extensively uses the SolidClassifier. After this fix it returns the correct classification statuses, which leads to incorrect result shape (reported by checkshape). Yet the result shape without this fix also seems to be incorrect (one of the isolines goes out of boundary of the face). Thats why it's marked with 'TODO'.
Corrections in test cases.
Test case is added.
1. Check, if value found by math_PSO algorithm cannot be precised by math_NewtonMinimum algorithm. In this case, we call math_PSO algorithm repeatedly, however, with other parameters.
2. Some margin of edge tolerance value has been provided in IntTools_Tools class.
3. Interface of math_NewtonMinimum class has been changed (method GetStatus() has been added).
Correction of some test cases according to their new behavior.
The update of the set of boundary edges has been added for the case when faces containing milti-connected edges are excluded from merging.
Update test cases for unifysamedom to generate snapshots.
The new option AllowInternalEdges has been added in the class ShapeUpgrade_UnifySameDomain. It determines how the algorithm treats the situation when two faces can be merged together but they have connection with another face via the common edge.
With this option turned on, merging of such faces is not stopped, but the multiconnected edges are added to the merged face as INTERNAL edges.
By default it is false. In this mode such merging is forbidden, so the result shape will not contain any new INTERNAL edges.
The behavior of the algorithm has been changed so that it does not merge faces from different shells.
Added protection for cases of the overlapped edges.
Returning value for methods in the LocOpe_SplitShape::AddOpenWire(), LocOpe_SplitShape::AddClosedWire() are modified from void to boolean in order to avoid raising exceptions.
To avoid regression for test case modalg_5 bug25243 tolerance of vertices of the degenerated edges are increased to value of tolerance of the splitting vertex for case when splitting vertex was projected on the degenerated edge.
Command checkplatform was created.
All test cases were updated.
Global variable os_type was eliminated.
New option -osx (MacOS) for procedure checkplatform was added.
1. Check, if edge is same-range, is now made with some tolerance (not strictly) in GeomLib_CheckCurveOnSurface class. Default value of this tolerance is Precision::PConfusion(). However, this value can be changed with corresponding interface.
2. DRAW-command "attachpcurve" has been added to BOPTest_UtilityCommands.cxx file. This command creates p-curve of given edge on given face. It can assign 2D-curve of one of the edge already included in the face or (if it is not possible) rebuilds new 2D-curve.
Creation of test case for this issue.
Adjusting test case boolean volumemaker A8 according to its new behavior on Windows.
