- Added a new mode in SelectMgr_ViewerSelector for computing BVH for Select3D_SensitiveEntity in background which can be activated via method SelectMgr_ViewerSelector::SetToPrebuildBVH(). Default behavior has not been changed.
- New class SelectMgr_BVHThreadPool manages background processing of BVH building queue.
- Added Select3D_SensitiveEntity::ToBuildBVH() method that checks if BVH (if it used) is in invalidated state. Defined this method for all standard classes inherited from Select3D_SensitiveEntity.
Make RWStl_Reader::IsAscii() accepting optional argument pointing how to rewind the input stream, using unget() or seekg(). This allows writing a reader that uses a stream supporting seekg but not supporting unget.
BOPAlgo_PaveFiller::MakeBlocks() - Improve Face/Face post treatment procedure by:
* Adding more existing edge for intersection with section edges to resolve all possible intersections.
* Using tree for edges selection.
* Avoiding intersection of the existing edges among themselves.
IntPatch_ImpImpIntersection::CyCyNoGeometric - add more points to wline to make it more smooth.
In the method TObj_Object::TObj_Object::SetName(const Standard_CString theName) conversion to the HAsciiString was used instead of the conversion to the TCollection_HExtendedString. -
Classes Message_ProgressRange and Message_ProgressScope are improved to store start point of the range.
Method Message_ProgressScope::Value() is improved to compute the value in the current scope from the actual current value of the global progress.
Description of Message_ProgressScope class is improved.
Off-topic:
- method Message_ProgressScope::Relieve() is renamed to Close() for semantic consistency
- article in Upgrade Guide is revised and corrected
In order to switch off setting default names for objects of the type TObj_Partition argument "theSetName" defining addition of the default names was added
- Added the new filter SelectMgr_AndOrFilter which allows to define the context filter. By default OR selection filter is used
- Added the enumeration SelectMgr_FilterType provides filter types
- To define behavior SelectMgr_AndOrFilter use SetFilterType in AIS_InteractiveContext
- Added the test
- SelectMgr_OrFilter don't store the disabled objects, it's stored in SelectMgr_AndOrFilter
Progress indication mechanism is refactored to support incrementing progress within multithreaded algorithms.
The class Message_ProgressIndicator is only an interface to the user application.
It accumulates the progress provided by progress scopes.
The counter is protected by mutex for thread-safety.
The new class Message_ProgressScope replacing Message_ProgressSentry should be used to advance the progress.
The scopes are nested to each other to reflect the nested nature of operations.
The new class Message_ProgressRange should be used to pass the progress to sub-scopes.
All OCCT algorithms involving progress indication have been updated to new API.
Improvements in Draw_ProgressIndicator:
- Separate console mode has been added in order to make possible to put the progress into std::cout instead
or in addition to the draw interpreter, instead of trigger option "-tclOutput".
- Treatment of Ctrl-Break signal has been added.
Now any operation can be aborted by Ctrl-C or Ctrl-Break keystroke.
Added new test case 'perf fclasses progr_par' for testing of parallel work of the progress.
Added V3d_View::SetBackgroundImage() accepting Graphic3d_Texture2D on input.
Graphic3d_CView/OpenGl_View have been modified to merge Graphic3d_CView::SetBackgroundImage()
and ::SetBackgroundCubeMap() implementations into a single method.
- Update .step scanner:
- add rules to read anything within the comment
- add a function to initialize of lexical scanner
- add states for Commit & End instead of global variables modcom & modend
Added possibility to inherit existing attributes if the same persistent fields are used. All methods that allow controlling the data model changes or getting some callbacks may be overridden in successor. They may have same GUIDs as a base class or new ones.
Special macros IMPLEMENT_DERIVED_ATTRIBUTE and IMPLEMENT_DERIVED_ATTRIBUTE_WITH_TYPE must be used instead of standard Handle macro definition IMPLEMENT_STANDARD_RTTIEXT to register new derived attributes.
Using this improvement several existing attributes from TDataStd, TDataXtd and XCAFDoc packages become inherited from other base attribute-classes. XML and Bin drivers of these attributes are removed. New base attribute classes are added: TDataStd_GenericEmpty and TDataStd_GenericExtString.
This improvement does not change both present formats Bin and XML documents. The obsolete Standard scheme is not changed at all.
InitFromJson method implementation for some simple classes.
OCCT_INIT_* defines introduction to do automatic parsing of the stream into values.
Inspector is extended to visualize objects created on the dump stream if it might be created.
- Add checking for null object for edge_curve within TranslateEdgeLoop;
- Update STEP parser rule allowing empty element in the list of entity parameters;
- Update test case of bug26451 for a new scanner rule;
- Update test case of bug30273 for a new added check for null object
Font_TextFormatter inherits Standard_Transient, now it is given as a handle in functions.
Graphic3d_Text - extended with Font_TextFormatter to be able to have it filled out of text render. If it is not defined here, the default text formatter of context is used.
OpenGl_Context - has default Font_TextFormatter for rendering OpenGl_Text.
AIS_TextLabel extending with Font_TextFormatter to prepare test case for text wrapping.
Prs3d_Text returns created graphic text to be able to manage it outside.
Prs3d_ToolQuadric has been modified to return an indexed triangulation.
Added methods Prs3d_ToolQuadric::CreateTriangulation() and Prs3d_ToolQuadric::CreateTriangulation()
as more straightforward API returning generated triangulation.
Added missing const to constant methods.
Confusing method Prs3d_ToolQuadric::FillArray() filling both
Graphic3d_ArrayOfTriangles and Poly_Triangulation at once has been marked deprecated.
V3d_Trihedron, AIS_ViewCube and AIS_Manipulator
now set Closed flag to groups with shaded sphere and arrows.
Added script adm/scripts/android_build.bat and template android_custom.bat.template
intended for automating building routines targeting Android platform.
OpenCASCADEConfig.cmake has been extended to detect "$INSTALL_DIR/libs/$CMAKE_ANDROID_ARCH_ABI/cmake/opencascade" location.
The scripts cmake_gen.bat and cmake_gen.sh have been added to facilitate batch building of OCCT correspondingly on Windows and Linux platforms.
The files build.bat and install.bat are copied to the build directory by cmake configure procedure to enable batch building and installation on Windows platform.
The scripts occ_build_wasm.bat and occ_build_wasm.sh have been added
to simplify batch building of OCCT with Emscripten (Web Assembly packages) correspondingly on Windows and Linux platforms.
These scripts are configurable through calling files with custom environment wasm_custom_env.bat and wasm_custom_env.sh.
Added public method Font_FontMgr::AddFontAlias() for registering custom aliases and fallback fonts.
vfont command has been extended with arguments -aliases, -addAlias, -removeAlias, -clearAlias and -clearAliases.
Scripts adm/cmake/bison.cmake and adm/cmake/flex.cmake are refactored to enable actual search for bison and flex.
Apart of standard locations, also sub-folders of 3RDPARTY_DIR whose names contain "bison" and "flex", respectively, are added to search.
Cache variables 3RDPARTY_BISON_EXECUTABLE and 3RDPARTY_FLEX_EXECUTABLE are removed to avoid confusion (they duplicated similar variables without "3RDPARTY_" prefix).
Lex and Yacc files are corrected to match changes made manually in generated files during last years:
- StepFile/step.yacc: correction missing from #22972
- StepFile/step.lex: corrected for compilation (broken by #31060)
- MSVC-specific code is synchronized between StepFile/step.lex and ExprIntrp/ExprIntrp.lex
- Old commented code and duplicate code blocks removed
Commands for execution of Flex and Bison tools in CMake scripts are tweaked to avoid embedding line numbers (with local paths) in generated files.
Scanners and parsers are regenerated from updated source files with modified options.
Note that lex.ExprIntrp.c is regenerated with multiple differences because option -f (fast scanner) was used for generation of previous version (by WOK).
Draw_Interpretor, CommandCmd() - catched exceptions and messages put into Tcl string result
before throwing a Tcl exception (return 1) are now print in intense red (using Message::SendFail()).
Duplication of exception message in std::cout and Tcl output has been removed.
Draw Harness plugins have been updated to use either Message::SendFail() or theDI instead of std::cout/std::cerr
for printing colored error message before throwing a Tcl exception.
Handle "VTK::" prefix instead of "vtk" used by previous VTK versions for targets.
Corrected unexpected location of endif() and broken indentation.
Obsolete $VTK_USE_FILE is no more included (basing on VTK version check).
Standard_WarningsDisable.hxx/Standard_WarningsRestore.hxx pair is now used to suppress VTK warnings instead of partial disabling.
BRepFill_Pipe.cxx: protection against wrong shape type is added
GeomFill_Sweep.cxx: protection against using 0-vector to create Direction is added
bug26568: test script is corrected: "TODO ..." is removed
Added command "dputs" similar to "puts" but with extra arguments modifying text color/intensity.
Command "help" now prints the name of command with intense style.
BinMDataXtd_PresentationDriver and XmlMDataXtd_PresentationDriver have been corrected
to convert old (OCCT 7.4.0) Quantity_NameOfColor enumeration values to new ones.
Added missing NULL checks before accessing GetInteractor().
Errors are now reported using Message::SendFail().
Not found actor names are now reported as syntax error instead of silently skipping them.
ivtkinit now accepts aguments -srgb and -msaa for overriding defaults.
Default MSAA settings have been changed from 8 to 0 to match vinit behavior.
Added command ivtkclose closing a Vtk view.
Added commands ivtkaxo, ivtkfront, ivtkback, ivtkleft, ivtkright,
ivtktop and ivtkbottom assigning view projection similar to sonamed commands in ViewerTest.
Added commands ivtksetcolor and ivtksettransparency assigning
color properties to shaded presentation.
Methods ParseOnOff()/ParseColor() have been moved from package ViewerTest to Draw.
Command "vlight -color" now accepts RGB values, not only name.
Implementation of pload command has been cleaned up.
Expr_NamedUnknown.cxx - wrong comparing of named unknown is fixed
QABugs_20.cxx - new QAcommand is created
QABugs_11.cxx - wrong command is corrected
bug902 - wrong test is corrected
bug31697 - new test is added
SelectMgr_ViewerSelector3d::ClearSensitive() - fixed crash on removing presentations (via command vsensera).
Added new parameters -sync and -reset to vrenderparams command synchronizing parameters across Views.
Parameters -raster and -rayTrace now accept optional on|off values.
Improved command description.
Add support of Message_ProgressIndicator in BinTools classes.
Add support of Message_ProgressIndicator with possibility of user break in methods of opening and saving TDocStd_Application.
Add tests of ProgressIndicator in TDocStd_Applacation.
# Add test to verify import
* Note: Each layer contains a valid name, but only one contains a link to the shape, so .step file contains only one valid Layer.
gp_Trsf::SetRotationPart() - added method replacing rotation matrix without reseting other components,
similar to existing SetTraslationPart() and SetScaleFactor().
Transformation multiplication order has been fixed
for vlocation arguments -rotate, -translate, -mirror and -scale.
Added -prerotate, -pretranslate, -premirror and -prescale options following previous behavior.
vlocation -setRotation now uses new method gp_Trsf::SetRotationPart()
for consistency with methods -setLocation and -setScale.
OpenGl_Context now skips loading functions related to mapping buffer,
which are required by OpenGL ES 3.0 specs but not provided by WebGL 2.0.
Message_PrinterSystemLog does not use a broken emscripten_log() anymore, which corrupted UNICODE strings.
WasmOcctView::initWindow() - callbacks now set using EMSCRIPTEN_EVENT_TARGET_WINDOW
instead of 0 used by older Emscripten API.
Mouse callbacks now track canvas element and use
EmscriptenMouseEvent::targetX/targetY instead of ::canvasX/canvasY
as the latter was broken.
Added emscripten_set_main_loop() setup to shut up eglSwapInterval() error message.
Fixed missing \0 at the end of string converted by toUtf8Array().
Improve projection of ellipse and circle on a plane in case of the same parametrization of the original curve and the projected one is not necessary. Now the projection is a canonical curve instead of B-spline.
Set safe exit status for:
Standard_NullObject exception,
Standard_NullObject: BRep_Tool:: TopoDS_Vertex hasn't gp_Pnt,
BRep_Tool: no parameter on edge,
BRepOffset_MakeOffset::TrimEdge no projection
Fixed a couple of misprints, usage of changed classes / methods.
Added references to other supported 3d displays.
Added reference to PBR metallic-roughness material properties in addition to Common.
Cosmetics: removed numerous trailing spaces, double spaces.
Added missing {.cpp} language qualifier to code quotations.
Sentences within paragraphs have been split into multiple lines in source file.
Extrema_ExtCS.cxx: treatment of small line segments is added;
Extrema_GenExtCS.cxx: treatment of particular cases curve-quadric and conic-surface are added
Extrema_GlobOptFuncCQuadric, Extrema_GlobOptFuncConicS: new distance functions for particular cases are added
BOPAlgo_PaveFiller_5.cxx : treatment of large common parts edge-face is improved
ElSLib.cxx : method TorusParameters(...) is modified to avoid divide by zero
math_PSOParticlesPool.cxx : initialization of array is added
OpenGl_ShaderManager::GetBgCubeMapProgram() and OpenGl_BackgroundArray::createCubeMapArray()
have been corrected to draw cube in straightforward way instead of a screen-quad.
Graphic3d_Camera::SetCustomStereoProjection() now recieves decomposed projection + head-to-eye matrices.
Added method Graphic3d_Camera::StereoProjection() returning projection matrix without translation part.
OpenGl_BackgroundArray::Render() now applies stereoscopic projection matrix in case of VR output,
but keeps using mono projection matrix in case of common 3D displays.
PrsMgr_PresentationManager::displayImmediate() now uses proper ZLayer for a shadow presentation
of computed presentation (fixes dynamic highlighting on mouse move).
Graphic3d_CView::SetComputedMode() - added invalidation of bounding box on turning OFF compute mode.
AIS_ViewController::contextLazyMoveTo() now ignores MoveTo() while rotating
view with temporary disabled compute mode (leading to undefined results and crashes).
Graphic3d_Structure::Graphic3d_Structure() - do not copy Graphic3d_TOS_COMPUTED visual
from linked structure, as Graphic3d_CView::SetComputedMode() does not compute such structures.
SelectMgr_SelectingVolumeManager::SetViewClipping() now updates clipping range
using picking ray in world coordinates, as clipping planes are always defined in world space.
The text label is selected as a rectangle (adds a sensitive object - Select3D_SensitiveFace).
The bounding box has been resized to fit the sensitive rectangle.
Transform persistent has been added to AIS_TextLabel to correctly position the sensitive rectangle.
The findConnectedObject function has also been changed to correctly set transform persistence.
Changes in Technical Overview:
Corrected reference of "Mesh" as dedicated module, which is actually a Toolkit within "Modeling Algorithms" module.
Reduced references to "Configurable optimized memory manager",
which has not been updated for a long time and no more encouraged to be used as replacement of standard memory allocator of C library.
VRML/STL converters have been moved from section Mesh to Data Exchange.
Visualization chapter has been slightly reworded and reordered;
added references to PBR metallic-roughness material model.
Data Exchange chapter - reworded sentence "The exchanges run smoothly regardless of the quality of external data..."
to "This module handles various problems of interoperability between CAD systems...";
added AP242 to the list of STEP capabilities.
Changes in User Guide for Foundation Classes:
Removed references to "physical quantities" in package Quanity, deprecated since #0028799 and not actually used anywhere.
Adjusted description of string classes to avoid duplicated statements.
Collections section has been refactored to replace obsolete TCollection references by NCollection.
Remove obsolete statement "furthermore *catch()* statement does not allow passing exception object as argument".
Fixed broken nested enumeration in a couple of places.
Fixed usage of inappropriate quote symbols in code snippets.
Fixed references to non-existing classes like ZeroDivide instead of Standard_DivideByZero.
Removed suggested usage of Standard_Failure::Raise() instead of "throw Standard_Failure()" following #0026937.
Code snippets now suggest "catch(const Standard_Failure& )" instead of "catch(Standard_Failure )" (#0026937).
Fixed broken code snippets with lost symbols like <<.
Cosmetics: removed numerous trailing spaces, double spaces.
Added missing {.cpp} language qualifier to code quotations.
Sentences within paragraphs have been split into multiple lines in source file.
WNT_HIDSpaceMouse - added auxiliary class for processing 3d mouse Raw HID input.
AIS_ViewController::Update3dMouse() - added default processor for 3d mouse input event.
ViewerTest now redirects WM_INPUT to AIS_ViewController::Update3dMouse().
Aspect_VKey enumeration has been extended by 3D view buttons.
WNT_Window::RegisterRawInputDevices() has beend added as a small wrapper
over WinAPI RegisterRawInputDevices() for common HID input devices.
AIS_ViewCube now stores animation duration within AIS_AnimationCamera instead of dedicated duplicating class property.
Adaptor3d_TopolTool.cxx, IntTools_TopolTool.cxx:
Anisotropy of BSpline surface along U and V direction is taken in account for calculation of numbers of sample points;
bug31616: test case added
AIS_ViewController::handleNavigationKeys() - added an interface for processing navigation keys.
ViewerTest_EventManager now maps WASD+Arrows navigation keys.
Axonometric view hotkey A has been replaced by Backspace.
Shaded/Wireframe are now mapped with hotkeys W+Ctrl/S+Ctrl.
Hotkey D (reset view to undefined default state) has been removed.
SelectMgr_RectangularFrustum now handles degenerated triangle as a segment or a point.
Triangle orthogonal to view direction is now handled as a segment.
myViewRayDir field is now defined as normalized vector to avoid confusing math.
For the case when the segment and myViewRayDir are almost parallel,
the segmentSegmentDistance function may set the depth to zero, although this will not
be the correct value, therefore it is better to pass a segment that will not
be parallel to myViewRayDir as an argument to the function.
vpoint command has been extended by arguments -2d and -nosel
for displaying picking mouse position as on-screen point in pixels.
Fixed double-binding-map error in case of displaying point with already used name.
bugs/vis/bug31412: test case added.
Animation sample has been updated to:
- use reuse AIS_ViewController for general viewer manipulations;
- update animation using elapsed time;
- do not block camera manipilations;
- get rid of redundant controls.
Multi-touch input is now redirected to AIS_ViewController.
GLSurfaceView.RENDERMODE_WHEN_DIRTY is now used by 3D Viewer.
AIS_ViewCube is now displayed instead of trihedron.
Project structure and project files have been updated to use Android Studio 4.0, gradle and CMake.
Redundant OcctJni_Window has been replaced by Aspect_NeutralWindow.
SD Card permissions are not dynamically requested for compatibility with Android API level 26+.
In the TDocStd_Owner keep simple pointer to TDocStd_Document instead of Handle. This causes automatic destruction of the document without explicit call of Close.
In Standard_Type added a static variable theType that initializes theRegistry map earlier. Otherwise exit from Draw interpreter crashes in many test-cases because not-closed transactions are aborted on document handle release from Draw theVariables map.
Corrected method for test OCC159bug due to the fact that Owner does not add a ref count now
Close the document in the end of bugs xde bug22776 otherwise double remove of visualization objects (on library exit and on visualization attributes remove from the document) causes crash on exit from draw
Added a new test bugs caf bug31075
BRepLib_MakeWire.cxx:
in method BRepLib_BndBoxVertexSelector::Accept() accepting tolerance criterion is changed
BRepLib_MakeWire.hxx:
correctionof class field name
bug31611: test case added
CLang is now checked before MINGW within occt_defs_flags.cmake.
Added "-Wl,-s" linker flag when using CLang for stripping symbols consistently to GCC builds.
Added new condition for first indicated element at Draw_ProgressIndicator::Show(),
because it’s more logical that at start progress starts at 0
Changed usage of Next() to Next(step) for increment progress to IND_THRESHOLD in RWStl::writeASCII() and RWStl::writeBinary()
Changed condition for continuation of writing and add interrupt to this function
Added possibility to use Progress indicator in writestl
Changed paremeter in constructor Message_ProgressSentry aPS() IND_THRESHOLD to 1
Changed test
Added OpenGl_Context::GetBufferSubData() implementing getBufferSubData() based on capabilities of various APIs.
Added OpenGl_VertexBuffer::GetSubData() similar to OpenGl_VertexBuffer::SubData().
OpenGl_Texture::InitCubeMap() now avoids using sized internal format in case of GLES2.
OpenGl_Sampler::applySamplerParams() now checks OpenGL version before setting GL_TEXTURE_WRAP_R.
Graphic3d_TextureRoot::GetCompressedImage() - added new interface for fetching compressed texture image.
Default implementation detects DDS image files using Image_DDSParser parser.
Graphic3d_TextureRoot::GetImage() has been extended with new parameter
- the list of image formats supported by OpenGL driver.
Graphic3d_TextureRoot::convertToCompatible() implicitly converts
BGRA image to RGBA on OpenGL ES, which normally does not support BGR formats.
OpenGl_Caps::isTopDownTextureUV - new property defines how application defines
UV texture coordinates in primitive arrays.
OpenGl_Context::SetTextureMatrix() compares this flag with OpenGl_Texture::IsTopDown()
and automatically flips V coordinate in case of mismatch.
OpenGl_Texture now holds exact number of mipmap levels
instead of Boolean flag indicating that they are defined.
This allows loading DDS files with incomplete mipmap level set
by setting GL_TEXTURE_MAX_LEVEL to appropriate value instead of default 1000
(causing black textures in case if mipmap levels are not defined till 1x1).
Fixed order of texture coordinates transformation within GLSL program to match FFP matrix:
Rotate -> Translate -> Scale (previously Rotation was applied afterwards).
Fixed -Wrestrict warning within OSD::SetSignal().
Fixed -Wdeprecated-copy warning by adding missing counterpart
in pair copy constructor / assignment operator or by removing trivial constructor.
AdvApp2Var_Node, AdvApp2Var_Patch are now declared as Handle.
AdvApp2Var_Iso is now passed by Handle.
Disabled operator= for TDF_Transaction and TDF_IDFilter.
Standard.cxx - fixed GCC version mischeck causing building failure with experimental GCC versions.
TopOpeBRepDS_EXPORT.cxx - fixed -Wmaybe-uninitialized warnings.
Do not classify edges originated from combination of edges and vertices as the classification will likely be unstable.
Mark faces containing valid inverted edges as invalid but do not remove them from valid splits.
Extend the debug command DBRep_SetComp to allow different containers for creating the TopoDS_Compound.
Adjusting parameters of algorithm depending on axes and parameters of cylinders is added in order to reduce computation time
Test case added: tests/lowalgos/intss/bug31552
tests/perf/modalg/bug26310_1: test case corrected according to current state of algorithm
Extrema_ExtCs.cxx: number of subdivision is incresed for curve;
Modifications to fix regressions of some tests
IntTools_BeanFaceIntersector.cxx: restriction of curve, which is used in Extrema_ExtCS
ProjLib_CompProjectedCurve.cxx: tolerance is reduced for Extrema_ExtCS
BRepOffset_Tool.cxx: checking common vertex if it is Null
tests/bugs/modalg_7/bug27087 : test case added.
tests/bugs/modalg_5/bug25232_8 : test is modified according to new behavior of algorithms
- Issue is done in previous version OCCT
- Add test to verify the creating of the label structure using plain OCAF/XDE API
- Situation when the assembly label has no shape attached is correct
The argument putEndl has been removed from Message_Messenger::Send() and Message_Printer::Send() methods.
Message_Printer interface has been changed, so that sub-classes have to implement new method
Message_Printer::send() accepting TCollection_AsciiString.
Old three Message_Printer::Send() methods remain available without putEndl argument
and redirecting to new send() method by default.
Removed dummy Message_PrinterOStream::GetUseUtf8() property.
Message_PrinterOStream, Message_PrinterSystemLog and Draw_Printer
now implement single method Message_Printer::send() instead of triplet.
Operators << for Handle(Message_Messenger) are removed; their use is replaced by use of Message_Messenger::StreamBuffer
Message_Messenger has been replaced by Standard_OStream within Interface_InterfaceModel,IFSelect,IGES,STEP dump interfaces.
Class Standard_BaseDriver is inherited from Standard_Transient, its descendants are updated accordingly.
Handle is used to manipulate objects of this class and its descendants (instead of references or raw pointers) to ensure automatic destruction.
Added test bugs caf bug31546
Related:
- Standard_OVERRIDE is added in declarations of virtual methods in descendants of Storage_BaseDriver
- Methods operator << and operator >> are removed in descendants of Storage_BaseDriver (they repeat the same methods inherited from the base class)
- Typedef PCDM_BaseDriverPointer is marked as deprecated
- Unused class DDI_Ostream is removed
- Private field Standard_Transient::count is renamed to myRefCount_ to avoid compiler warnings if the same name is used within the scope of a descendant class
- Output of meaningful error messages is restored in DRAW commands fsdread and fsdwrite
V3d_View::AutoZFit() is now called only before redraw
within methods V3d_View::Redraw() and V3d_View::Update().
Graphic3d_CView now holds Aspect_ExtendedRealitySession object.
Aspect_OpenVRSession implements new interface via optional OpenVR library.
Graphic3d_CView::ProcessInput() - added new interface method
which should be called for processing positional input (head tracking).
Graphic3d_Camera now allows setting custom stereoscopic Projection matrices.
OpenGl_Context::Camera() - context now holds Camera object
in addition to active camera matrices.
genproj.tcl has been extended to handle optional CSF_OpenVR dependency.
SelectMgr_SortCriterion::Normal, SelectBasics_PickResult::SurfaceNormal() - added new property.
SelectMgr_RectangularFrustum::Overlaps() for triangle sets new normal property.
gp_GTrsf::GetMat4() - added conversion into NCollection_Mat4 similar to gp_Trsf::GetMat4().
Added short-cuts methods in Message_Messenger for sending message with specified gravity,
and stream buffer class for using stream-like interface for that.
Similar short-cuts to DefaultMessenger() are added in Message package.
ViewerTest has been updated to send messages to Message::DefaultMessenger()
instead of direct output to std::cout/std::cerr.
Off-topic: spelling error (duplicate "and") is corrected in two places
Added test bugs fclasses bug31189
Before the patch, if the values of the isReversedX/Y/Z variables were set to false, then the WorldView matrix did not change and local transformation was not applied, which caused errors.
In order to correctly set the local transformation in case the text does not have its own attach point, the local transformation is set in OpenGl_Text, and the ModelWorld matrix is changed in OpenGl_Flipper, instead of the WorldView matrix. In this case, local transformation will always be applied.
bugs/vis/bug31193: test case added
Graphic3d_CullingTool::IsCulled() has been extended with theIsInside argument for full inclusion test.
Graphic3d_Layer::UpdateCulling() now avoids frustum culling tests for BVH children for parent nodes completely included into frustum.
Graphic3d_CullingTool::SetViewVolume() has been extended by optional model-world matrix.
A special mode for the selecting by polygon is added to select only completely overlapping objects.
In order to track the sensitives that were included completely by defined polygon, the boundary points of the polygonal frustrum are stored in the variable myBoundaryPoints.
If an sensitive intersects with at least one of the frustrums from myFrustums, then checking whether this object intersects with borders using the isIntersectBoundary method; if not, then the sensitive were included completely by defined polygon.
Because the polygon can be concave, then to check the sensitive were included completely by defined polygon, it is not enough to check of all its points, it is necessary that the edges of the sensitive do not intersect polygonal frustrum. To do this, for polygonal selection, a call to the Overlaps method for a point was replaced by a call to a segment where necessary.
bugs/vis/bug31440: test case added
OpenGl_Element now provide methods ::UpdateMemStats() and ::UpdateDrawStats()
for unified request of statistics used by OpenGl_FrameStats instead of direct handling of OpenGl_PrimitiveArray.
Graphic3d_FrameStatsCounter counters list has been extended by Graphic3d_FrameStatsCounter_NbLinesNotCulled.
- Fix recursion in EntityCluster's methods :
- Value
- SetValue
- Append
- Add destructor in the EntityCluster
- Add check null grade in the STEPCAFControl_GDTProperty::GetDimClassOfTolerance
Approx_ComputeCLine.gxx : new class field is added to manage hang checking.
ProjLib/ProjLib_ComputeApprox.cxx : unset hang checking for analytical surfaces and curves
Poly_Connect::Load() now uses NCollection_IncAllocator instead of new/delete
for allocation temporary sequence elements.
StdPrs_ShadedShape, fillFaceBoundaries() now uses NCollection_IncAllocator for temporary points.
BRepFeat_Form.cxx: methods IsDeleted, Modified, Generated are corrected
BRepTest_FeatureCommands.cxx: history treatment is added in methods DEFIN and PERF
bugs/modalg_7/bug31469: test case added
Eliminate numerical instability by ensuring that the tolerance of intersection entities is slightly grater than the actual distance to the shapes creating the entity.
BRepMesh: Add new mesh parameter *AllowQualityDecrease* which affects the criteria used for checking of the consistency of the existing mesh to new meshing parameters.
So if set to true it will force the meshing of the shape if current deflection strongly vary from the new one, no matter in which side.
BRepTools::Clean: Keep triangulation on non-geometric shapes (faces with no surface or edges with no curves).
Prs3d::GetDeflection() has been moved to StdPrs_ToolTriangulatedShape::GetDeflection().
Prs3d_ShapeTool has been moved to StdPrs_ShapeTool.
Code collecting free edges on Poly_Triangulation
has been moved out from StdPrs_WFShape to Prs3d::AddFreeEdges().
StdPrs_BndBox has been moved to Prs3d_BndBox.
Geom_Transformation has been replaced by TopLoc_Datum3D within visualization classes.
Select3D_SensitiveCircle constructor now takes gp_Circ instead of Geom_Circle.
StdSelect_ViewerSelector3d has been moved to SelectMgr_ViewerSelector3d.
Methods ::GetPoint3d() and ::ArrayBounds() has been moved
from subclass Select3D_SensitiveCircle to the base class Select3D_SensitiveCurve.
StdSelect_ViewerSelector3d::computeSensitivePrs() has been moved to SelectMgr::ComputeSensitivePrs().
Removed unused declarations StdSelect_Prs, StdSelect_DisplayMode,
StdSelect_SensitivityMode, StdSelect_TypeOfResult, SelectMgr_SOPtr, TColQuantity.
Package Graphic3d has been moved from TKV3d to TKService.
The following classes and their subclasses within AIS package
have been moved to new package PrsDim and renamed:
- AIS_Relation -> PrsDim_Relation;
- AIS_Dimension -> PrsDim_Dimension,
Support of CIE Lab and Lch color spaces is introduced in Quantity_Color:
- Enumeration Quantity_TypeOfColor is extended by new values representing CIE Lab and Lch color spaces (with D65 2 deg illuminant).
- Conversion of RGB color to and from these color spaces is implemented in Quantity_Color class (within existing API).
- Color difference calculation using CIE Delta E 200 formula is implemented in method DeltaE2000().
New methods MakeUniformColors() and SetUniformColors() are added in class AIS_ColorScale, generating and setting color scale based on colors of the same lightness in CIE Lch color model.
DRAW commands vcolorconvert and vcolordiff are added to perform conversions and compute difference, respectively.
A new option -uniform is added in DRAW command vcolorscale, to set uniform color scale.
Added test grid v3d colors (color conversions and difference), test bugs vis bug31454 (uniform color scale)
Translation of the entities AXIS2_PLACEMENT_3D including in the entity CONSTRUCTIVE_GEOMETRY_REPRESENTATION_RELATIONSHIP was added.
Entity AXIS2_PLACEMENT_3D is translated to the planar face with axis equal to AXIS2_PLACEMENT_3D.
New parameter "read.step.constructivegeom.relationship" is added to enable this translation (by default it is OFF).
Added test bugs step bug29803
- Added Right() method to BRepFeat_SplitShape;
- Added an option in command buc60854, which allows to select Left() or Right() method;
- Modified method BUC60854 in QABugs_14.cxx;
- Added new option to the help of command buc60854;
- Modified test case buc60854. Now its state is "OK";
- Added check Right() in buc60854 test case.
Useless casts are removed.
Off-topic: corrected description of constructor of TCollection_HAsciiString class from TCollection_HExtendedString for the case when input string contains non-Ascii character.
For small edges fully covered by tolerances of vertices BSpline periodic curves are converted to not periodic BSpline curves by cutting segment before writing.
In the class TopoDSToStep_MakeStepEdge permute of edge vertices for closed curves was added for the case when the range obtained from the vertex projection contradicts the edge range in order to avoid computation of the invalid edge range after reading.
Second test script for bug 0031301
Changes made for #31233 are reverted.
Correction in the ShapeFix_ComposeShell:
Modification of the method SplitByLine in order to find all points of the intersection for case when initial curve shifts in the positive or negative direction
to avoid splitting edge when split point lies in the limits of the tolerance of the vertex.
to correct shift of the pcurves for case when initial shift is more than 1 in the method ShapeFix_ComposeShell::SplitByLine
In the method ShapeFix_Face::FixMissingSeam added removing small edges having length less than working precision and removing wires having area less than working precision
Modified test cases; added test for related issue: bugs step bug31301
In constructor of Geom_RectangularTrimmedSurface, skip check if an offset surface has insufficient continuity. The reason is that this surface has already passed the check once created. So, here we must not apply more strict check than that first time.
Prs3d_Drawer, AIS_Shape, AIS_InteractiveContext - removed properties HLRDeviationCoefficient() and HLRAngle().
Prs3d_Drawer::HLRAngle() is kept as alias to Prs3d_Drawer::DeviationAngle() with deprecated flag.
Prs3d_Drawer::DeviationAngle() default value is changed from 12 to 20 degrees
to match Prs3d_Drawer::HLRAngle() which has been previously used in majority of cases.
Removed unused property HLRBRep_PolyAlgo::Angle().
Multiple PrsMgr_PresentableObject::Compute() methods have been replaced by single one
PrsMgr_PresentableObject::computeHLR().
Dummy implementations of previously declared methods have been removed
from AIS classes not implementng HLR presentation.
Class Prs3d_Projector has been removed and Graphic3d_Camera is now passed directly,
so that redundant dependency from HLRAlgo_Projector in Prs3d interfaces has been eliminated.
StdPrs_HLRShape and StdPrs_HLRPolyShape now implement common interface StdPrs_HLRShapeI.
BRepTools::Triangulation() has been extended with a new parameter for checking Poly_Polygon3D presense within free Edges.
StdPrs_WFShape::Add() now performs auto-triangulation in the same way as StdPrs_ShadedShape::Add().
StdSelect_BRepSelectionTool::GetEdgeSensitive() now creates Select3D_SensitiveSegment instead of Select3D_SensitiveCurve for tessellated segment.
Select3D_SensitiveSegment default sensitivity factor has been changed to 3 pixels to match Select3D_SensitiveCurve.
Test case bug23625_1, added workaround for out-of-range crash in HLRBRep_PolyAlgo on re-triangulated shape.
Calculations in Adaptor2d_OffsetCurve are unified with similar calculations in Geom2d_OffsetCurve using methods extracted from Geom2dEvaluator_OffsetCurve to Geom2dEvaluator.cxx
BRepFill_OffsetWire.cxx, Geom2dGcc_Circ2d2TanRadGeo.cxx, Geom2dGcc_Circ2dTanOnRadGeo.cxx, MAT2d_Circuit.cxx are modified to satisfy changing offset direction.
Use correct ranges of circles when processing the concentric case. Repeat the range comparison 3 times shifting each time for a half-period to process the extrema between boundary points of arcs.
Approx/Approx_SameParameter.cxx,hxx:
Class Approx_SameParameter refactoring. Logic is changed in many places to unify usage, simplify maintenance.
Method Curve2d() is changed to return Geom2d_Curve instead of Geom2d_BSplineCurve. Corresponding message is added to the upgrade guide.
.lxx file is merged into .hxx.
Tangent computation is extracted into special method.
Comparing number of sample points after CheckSameParameter(...) is added to define cases with projection fails.
Undesirable behavior when curves are not same parameterized is fixed.
Geom2dAdaptor/Geom2dAdaptor.cxx: treatment of offset curve is added
Adaptor3d/Adaptor3d_TopolTool.cxx: minor improvement of performance for BSpline surfaces with huge number of knots
Tests were modified according to new behavior of sameparameter algorithm
Get rid of requirement for the arguments of Boolean operations of type CUT and COMMON to be one-dimensional shape. For FUSE operation this requirement is kept.
OpenGl_FrameStats::updateStructures() now relies on OpenGl_Element inteface
instead of handling OpenGl_PrimitiveArray specifically
for collecting Graphic3d_FrameStatsCounter_EstimatedBytesGeom counter.
OpenGl_PrimitiveArray and OpenGl_Text implement new interface.
- 'Search Qt4' is renamed into 'Search Qt'. Configuration uses Qt5 from now. For Qt4 use CMake configuration.
- 'Build Inspector' check box is added. It causes 'tools' toolkits compilation. Activates automatically 'Search Qt' control.
- Many Tcl procedures are extended with additional parameter to giving a directory of processed toolkits. The default directory is 'src', inspector directory is 'tools'. This way is similar Inspector compilation implemented in Cmake build procedure.
- Dependency libraries in DEBUG are stored in __TKDEP_DEBUG__. Qt requires it as names of debug and release libraries are different.
A case of reading of a very small value was skipped, for example: 2.1219957909652723e-314.
Fixed in XmlMDataStd_RealArrayDriver.cxx and XmlMDataStd_RealListDriver.cxx
// A test file bug29452.xml is extended to check this skipped case.
Message_PrinterOStream::SetConsoleTextColor() skips color tags in case of Emscripten.
Message_PrinterSystemLog now implements log via emscripten_log().
OpenGl_Sampler::applySamplerParams() has been corrected to check for OpenGL ES 3.0
when applying GL_TEXTURE_BASE_LEVEL/GL_TEXTURE_MAX_LEVEL parameters.
-Create a new package BRepPreviewAPI, inherited from BRepPrimAPI to create both, valid and degenerative shapes
-Create a new class BRepPreviewAPI_MakeBox for working with a box
Preview can be vertex, edge, rectangle or box
-BRepPrim_GWedge: in the case of non-valid data, the exception does not happen in the constructor, but at the moment of access to the shape
-BRepPrimAPI_MakeBox: myWedge is now not private, but protected, because used in BRepPreviewAPI_MakeBox which inherits from BRepPrimAPI_MakeBox
-Add tests for checking of a creation a preview in tests/geometry/preview (vertex, edge, rectangle, box)
-Update a command "box": add new parameters
- Convert package creation. It is used to prepare some auxiliary presentations/methods to prepare variables for inspector based on occt classes.
- ViewControl package improvement:
Classes for property view, table, table model are implemented to visualize view of properties.
This view is filled by selection in tree if the selected object has implemented DumpJson functionality.
- TreeModel package improvements:
Method InitColumn is implemented in TreeModel_TreeModel to fill default columns in base model. The default columns are:
Name, Visibility, Rows. Additional columns should be added in successors.
Container of root items is moved into the base class
- View package imrovements:
Store preferences: display mode, fit all checked state and type of external context
View_DisplayPreview is added to process preview for selected in tree view objects
hide actions Clear, Multi, Single by default (as selection in tree view or Visibility state of item define which presentations should be displayed or erased)
- ShapeView plugin improvements:
Property View content based on DumpJson (columns in tree view for properties are removed),
Properties for TopoDS_Shape are displayed
Explode shape action in tree view to see content of shape by selected shape type. Type of displayed shapes is only preview, it is not used in export action.
Export shape into BREP. The previous implementation about BREP files in some directory and default view is removed.
- VInspector plugin imrovements:
Property View content based on DumpJson (columns in tree view for properties are removed),
Properties for AIS_InteractiveContext and AIS_InteractiveObject are displayed
obsolete classes to provide properties are removed.
- DFBrowser plugin improvements:
Property View content based on DumpJson is not used yet. But there is the USE_DUMPJSON macro. It's switch off by default.
If switch ON, it creates a Property View filled with DumpJson. It's a way of DFBrowser moving on this property view.
Init uses static variable CDF_Session::CurrentSession()->CurrentApplication to fill plugin if application in plugin is not set but OCAF application exists
- OCC_VERSION_HEX variable used for previous OCCT versions support (before 6.9.0 or 7.2.0) is removed
Graphic3d_Layer::UpdateCulling() now takes into account BVH_Tree::EndPrimitive().
'vstatprofiler structs' now properly prints number of structures in case of no culling.
Standard_ReadBuffer has been extended with a flag for reading an interleaved data.
RWGltf_TriangulationReader::readBuffer() uses new flag for reading vertex attributes.
- Provide DumpJson for geometrical, ocaf and visualization classes;
- Change depth parameter of DumpJson (constant is not obligate here)
- Introduce a new macro for transient objects to be called as the first row in DumpJson: OCCT_DUMP_TRANSIENT_CLASS_BEGIN. We need not put the class name in the macro, using get_type_name of Standard_Transient for it.
- change implementation of OCCT_DUMP_CLASS_BEGIN and OCCT_DUMP_TRANSIENT_CLASS_BEGIN. It is not an sentry more and it does not create a new hierarchy level. It appends a new row into the output stream: "className": <className>
- OCCT_DUMP_* does not require semicolon
- class header is included first in source files of TDataStd, TDocStd, TCAFDoc
Base normal map support has been implemented in path tracing (just geometry normal replacement).
Smooth normal adaptation has been implemeneted in order to avoid black areas artefacts during path tracing.
Tangent normal space calcuation has been moved to separate unified function.
Tangent space orthogonalization algorithm has been changed in order to handle all orientations and combinations of UV coordinates.
The normal after tangent space calculation has been reverted in order to be in right orientation.
The tangent space calculation has been modified in order to handle back faces with inverted normal.
Changes:
- in StepData_StepReaderData::ReadEntity, make sure that entity handle is not null before calling IsKind() of it.
- in StepData_StepReaderData::SetEntityNumbers, do not invert the sign of entity number of referenced parameter if the second pass is not required. Otherwise, the number -1 (obviously invalid) becomes 1 (invalid, but not obviously) without further correction of it.
- in StepData_UndefinedEntity::ReadRecord, consider the case when the entity number is positive but refers to non-existing entity. In this case literal value will be used instead of null entity for the problematic parameter.
"Range" parameter of point light sources (positional and spot) concidering in PBR has been added.
Angular attenuation parameter of spot light in PBR has been reimplemented based on existing "concentration" parameter.
Ray-Tracing now handles IBL and background cubemaps in similar way as PBR.
Ambient light source is handled by Ray-Tracing in similar way as PBR,
e.g. as coefficient for environment map (IBL) or as constant light source.
Test script generating PBR spheres has been moved to Samples
and now shared by PBR and Path-Tracing tests.
GeomConvert_1.cxx:
Creation periodic BSpline surfaces from trimmed periodic surface if trim is boundaries of periodic domain is allowed
BRepTools_NurbsConvertModification.cxx:
Checking domain of 2dCurves if surfaces are periodic is added
Test case tests/bugs/mesh/bug30008_2 is modified according to current behavior
Test case tests/bugs/modalg_7/bug31242 is added
Changed GLES2/gl2.h to GLES3/gl3.h.
OpenGl_Window_1.mm has been updated, now it is trying to create kEAGLRenderingAPIOpenGLES3 before trying to create kEAGLRenderingAPIOpenGLES2.
BRepClass/BRepClass_FaceExplorer.cxx:
Protection against intersection of test ray with almost parallel edges is added.
SWDRAW/SWDRAW_ShapeFix.cxx: add tolerance as parameter in command checkfclass2d
tests/bugs/modalg_7/bug31203:
test case is added
Draw command GetNDStrings was improved to sort the properties by the keys and
skip conversions of the strings to type TCollection_AsciiString.
Certain Draw commands were corrected to use the Draw interpreter instead of "std::cout".
A Tcl procedure was created to compare two multi-line strings.
RWMesh_NodeAttributes has been extended with NamedData property for passing Metadata.
RWMesh_CafReader::addShapeIntoDoc() has been corrected to avoid
adding calling XCAFDoc_ShapeTool::AddSubShape() for sub-shape labels.
Attributes for Products and Instances are now set independently to shape labels.
Sharing of Compounds (previously always duplicated) as sub-Components has been fixed.
TDataStd_NamedData has been extended with interface for deferred data loading.
Setters have been corrected to avoid duplicating lookups within map.
OpenGl_ShaderManager - metallic-roughness, emissive, occlusion
and normal texture maps are now supported by PBR.
Emissive, occlusion and normal texture maps are now supported by Phong shading model.
Path-Tracing now handles metallic-roughness and emissive texture maps.
Graphic3d_TextureUnit enumeration has been extended by
new values corresponding to supported texture maps.
OpenGl_TextureSet and OpenGl_ShaderProgram have been extended with
bitmask Graphic3d_TextureSetBits identifying texture slots read from GLSL Program
and slots defined within Texture Set to avoid undefined behavior by binding mock textures.
OpenGl_TextureSet now duplicates texture unit information to handle
textures shared across multiple slots (like Occlusion [R] + Metallic-Roughness [GB]).
OpenGl_Context::BindTextures() has been extended with active GLSL program paramter
to set mock textures to texture units used by program but undefined by texture set.
OpenGl_Workspace::ApplyAspects() has been extended with parameter to avoid bining texture set.
Extend the Offset operation (Join Type "Intersection", mode "Complete") to allow different offset values on adjacent co-planar faces.
The gap between adjacent faces is closed by creating artificial face perpendicular to the face.
Adding test cases.
Add Transform() and Transformed() methods by Mat4 for BVH_Box (only for 3D dimension).
Add PointBoxProjection() method to project point to box and RayBoxIntersection() method to compute hit time of ray-box intersection to BVH_Tools.
Add BVH_Ray class to define ray object.
Message_PrinterOStream now uses colorized output by default.
DRAWEXE has been corrected to disabled colorized output
when it is called with -f argument to avoid spoiling testgrid HTML logs.
Fixed bugs that occurred when using the default std::cout from Message::DefaultMessenger() instead of Draw_Printer
A dedicated option was added to Draw_ProgressIndicator, for outputting data to the tcl when performing tests
Added -tclOutput parameter to XProgress command
Added new class RWGltf_CafWriter for exporting XCAF document into glTF file
as well as Draw Harness command WriteGltf.
Added auxiliary method OSD_Path::FileNameAndExtension() splitting file name into Name and Extension.
AIS_Trihedron - removed extra fields holding highlight styles; standard styles are now used instead;
fixed unexpected modification of global aspects;
fixed unhighligting of selected plane within Shaded trihedron.
AIS_InteractiveContext now sets highlight color to highlight aspects, not only base color to drawer itself.
Standard_ReadLineBuffer now supports a processing of the special multi-line case with \ at the end of the line.
Standard_RedLineBuffer was used to load Stl files
System console is configured at DRAW start to use UTF-8 encoding, for cout and cin to deal correctly with Unicode symbols.
Use of std::wcout is avoided as it leads to corrupted output.
Command testgrid is improved to enforce UTF-8 encoding in child DRAW processes to preserve Unicode symbols in captured output.
Test bugs fclasses bug22125 is refactored:
- avoid dependency on external data file
- avoid producing snapshot
- check that Unicode name of the file created by OCCT procedure matches the name interpreted by Tcl functions
Font_SystemFont now defines Face Id in addition to file path.
Font_FontMgr::CheckFont() has been extended to return the list of Faces from font collections.
Correct methods Build and Generated:
use BRepTools_WireExplorer instead of TopExp_Explorer or TopoDS_Iterator, for compatibility with previous computations.
Exception in the algorithm Geom2dGcc_Circ2dTanOnRadGeo raised because the number of solutions exceeded the size of reserved array for solutions. A check has been added to forbid adding solutions outside of allocated place.
The similar protection has been done in other algorithms of this package.
Extended Style name is now appended to Family name for unique identification of the font.
vfont -find now accepts mask to search the font and returns family name.
BinTools_ShapeSet now defines maps with proper types instead of Standard_Transient.
BinTools_ShapeSet now avoids allocation of temporary arrays while reading
triangulation and polygonal data.
Classes Poly_Triangle, Poly_PolygonOnTriangulation, Poly_Polygon3D and Poly_Polygon2D
have been cleared from .lxx files and extended by new methods for preallocating and filling array,
as alternative to passing arrays by copy.
wavefront command - export "f" instead of obsolete "fo" keys into file.
BinTools - added missing tools for reading/writing short reals.
RWObj_CafReader::BindNamedShape(), added missing bind to material map.
XCAFDoc_VisMaterial::ConvertToCommonMaterial()/::ConvertToPbrMaterial(), added copying of base color texture.
XCAFPrs_DocumentExplorer now properly fills XCAFDoc_VisMaterial property of XCAFPrs_Style.
New optional environment variable CSF_CPULIMIT_FACTOR is introduced, allowing scaling the argument given to DRAW command cpulimit by specified factor.
Command testgrid is extended by two new options:
- -skipped: to re-run tests skipped in specified log
- -skip: to skip specified number of tests in the beginning of test sequence
Temporary Tcl scripts generated by test system are now removed from the test log directories immediately after the test execution, to save disk space.
Advises on use of CLang sanitizers are added in guide "Debugging hints and tips".
Standard_EXPORT is added in declaration of virtual methods where it was missing, to allow creation of sub-classes in other DLLs.
In some classes in OpenGl package Standard_EXPORT is added for all methods to make these classes usable in other DLLs.
Build tools are updated to support building with Visual Studion 2019 toolset ClangCL (LLVM clang-cl):
- In CMake builds, use of unsupported option "-std=c++0x" is avoided
- Tool genconf recognizes availability of ClangCL toolset and proposes it in the list
- Tool genproj recognizes new compiler specification VCVER=vclang and generates VS 2019 projects for it
Note that vclang configuration puts binaries to folder "vc14" as it is compatible with vc14 CRT.
Code is updated to build with Clang for Windows without errors and warnings:
- In BVH classes, pure virtual destructors are replaced by empty ones, to avoid bogus compiler warning
- In .lex files, pragmas are added to disable warnings in parser code
- In OSD, signature of function _osd_wnt_set_error() is corrected to avoid warning on incorrect use of va_args (undefined behavior)
- In OSD_Host, OSD_Parallel_TBB, warnings on usage of deprecated function are disabled
- In Quantity_ColorRGBA and ViewerTest_CmdParser, incorrect use of predefined macro __FUNCTION__ assuming it was a string literal is avoided
- In Standard_Macro.hxx, code is reordered to handle Clang for Windows appropriately
- In AdvApp2Var_SysBase.hxx, inclusion of obsolete (MSVC-specific) header file is replaced by Standard_TypeDef.hxx
- In Standard_ErrorHandler.hxx, empty statements added to avoid warning on unused class fields
- In IVtkDraw and ViewerTest, interface is corrected to use actual type avoiding unnecessary type casts
Off-topic: corrections of some errors (uninitialized class fields) reported by VS Code Analysis in TKernel
AIS_ViewCube has been extended by additional properties (previously hard-coded):
- radius of axes of the trihedron
- radius of cones of the axes of the trihedron
- radius of a sphere (central point) of the trihedron
ViewerTest command vviewcube has been extended to support new parameters:
-axesradius, -axesconeradius, -axessphereradius.
The interface of V3d_View and V3d_Viewer has been simplified.
For the fields myDefinedViews, myActiveViews, myDefinedLights, myActiveLights were added appropriate methods returning the internal raw data.
Make the next methods deprecated:
IfMoreLights(), InitActiveLights(), MoreActiveLights(), NextActiveLights(), ActiveLight() and
InitActiveViews(), MoreActiveViews(), NextActiveViews(), ActiveView(), InitDefinedViews(), MoreDefinedViews(), NextDefinedViews(), DefinedView(),
InitActiveLights(), MoreActiveLights(), NextActiveLights(), ActiveLight(), InitDefinedLights(), MoreDefinedLights(), NextDefinedLights(), DefinedLight().
Remove deprecated methods added in scope of tasks 0029290 and 0028987 (Target Version 7.3.0):
SelectMgr_SelectableObject: Init(), More(), Next(), CurrentSelection().
SelectMgr_Selection: Init(), More(), Next(), Sensitive().
V3d_AmbientLight: one constructor.
V3d_DirectionalLight: two constructors.
V3d_PositionalLight: one constructor.
V3d_SpotLight: two constructors.
Metallic-Roughness shading model Graphic3d_TOSM_PBR has been implemented.
New materials descriptors Graphic3d_PBRMaterial have been added to Graphic3d_MaterialAspect.
PBR shading model requires OpenGL 3.0+ or OpenGL ES 3.0+ hardware.
Environment cubemap is expected to be provided for realistic look of metallic materials.
occLight_IsHeadlight() now returns bool instead of int.
Avoid using lowp for enumerations to workaround occLight_IsHeadlight()
ignorance on Adreno 308 caused by some GLSL optimizator bugs.
OpenGl_Texture::EstimatedDataSize() - fixed estimation for Cubemap textures.
OpenGl_Sampler::applySamplerParams() - fixed uninitialized GL_TEXTURE_WRAP_R in case of GL_TEXTURE_CUBE_MAP target.
Warnings issued by gcc 8.1.0 are eliminated.
New Standard_WarningDisableFunctionCast.hxx header file is introduced to disable GCC warning "-Wcast-function-type" in those files *.cxx where it is issued. This warning is issued when the conversion from the pointer to one function type to the pointer to another function type takes places, it was added in gcc 8.1.0.
The function MyVISEDG in SWDRAW_ShapeAnalysis.cxx is removed as it does not seem to do anything useful and relevant DRAW command K_VISEDG is never used in tests.
DRAW implementation of progress indicator is corrected to print "finished" instead of 1e100 for the end of infinite range.
Added test bugs fclasses bug31092
Handled __EMSCRIPTEN__ macros to:
- Workaround atomics (__GCC_HAVE_SYNC_COMPARE_AND_SWAP_4 is undefined, but GCC atomics are provided).
- Suppress non-standard header <sys/signal.h> warning.
- Return OSD_LinuxREDHAT.
- Avoid inclusion of XLib headers.
- Skip fontconfig library.
- Enable EGL+GLES path (translated by Emscripten into WebGL).
- Skip eglCreatePbufferSurface() not implemented by Emscripten EGL.
Fixed Graphic3d_Vec4.hxx usage within Quantity_ColorRGBA.hxx.
OpenGl_ShaderManager::defaultGlslVersion() now prefers GLSL 300 es when WebGL 2.0 is available,
as there no any OpenGL ES greater than 3.0 emulation so far.
Shaders_Declarations.glsl - added workaround for GLSL compilation on WebGL 1.0
by defining Light properties accessors as macros instead of functions
('[]' : Index expression must be constant).
OpenGl_FrameBuffer::Init() - added workaround for initialization of GL_DEPTH24_STENCIL8
depth-stencil attachment on WebGL 1.0 + GL_WEBGL_depth_texture extension.
OpenGl_Context::Vec4FromQuantityColor() now considers myIsSRgbActive flag
to handle use case, when Immediate Layer is drawn directly into window buffer,
which is not sRGB-ready.
Added new sample - OCCT WebGL viewer.
Font_FTFont::FindAndCreate() now loads embedded fallback font within Font_StrictLevel_Any level.
Font_FontMgr::FindFont() - fixed misprint in message, and customized message for range fallback error.
Font_FontMgr - added one more fallback Korean font for Linux.
vfont command has been extended with options to clear Font Manager content.
When collecting the points to build plane use the points on the curve instead of poles (in case of BSpline and Bezier curves).
Side effect changes:
Changes in Geom2dHatch_Elements are to increase the chance of correct 2d classification (by the means of Geom2dHatch_Classifier) by taking more than just one point on the edge to construct the classification ray and by checking if this ray is not tangent to the edge itself.
AIS_ConnectedInteractive::connect() now throws exception on connecting already displayed object.
AIS_TextLabel::AcceptDisplayMode() - added missing Display Mode filter.
Commands vconnect, vconnectto and vaddconnected have been corrected to
replace dispalyed AIS_Shape with AIS_ConnectedInteractive in context.
Introduced new attribute XCAFDoc_VisMaterial storing visualization material definition.
XCAFPrs_Style has been exteneded Material() property.
XCAFPrs_AISObject::DispatchStyles() maps new XCAFPrs_Style::Material() property onto graphics aspects.
RWGltf_GltfJsonParser and RWObj_CafReader now put Material definition into XCAF document instead of a color label.
RWGltf_MaterialMetallicRoughness - added missing properties AlphaMode, AlphaCutOff and IsDoubleSided;
fixed default values in constructor for Metallic and Roughness.
Added commands XGetAllVisMaterials, XGetVisMaterial, XAddVisMaterial,
XRemoveVisMaterial, XSetVisMaterial, XUnsetVisMaterial for working with
new visualization materials table in the document.
OpenGL rendering is now done into sRGB framebuffer.
OpenGl_ShaderManager::prepareStdProgramFboBlit() has been extended
by programs resolving MSAA texture and applying gamma correction as fallbacks.
Quantity_Color definition has been modified to store RGB components
in linear color space within Quantity_TOC_RGB type.
Standard colors defined by Quantity_NameOfColor enumeration has been updated accordingly.
New Quantity_TOC_sRGB type has been introduced to handle RGB components in non-linear sRGB color space.
OpenGl_TextureFormat class definition has been moved to dedicated files.
New method OpenGl_TextureFormat::FindFormat() replaces OpenGl_Texture::GetDataFormat().
New method OpenGl_TextureFormat::FindSizedFormat() replaces OpenGl_FrameBuffer::getColorDataFormat()
and OpenGl_FrameBuffer::getDepthDataFormat().
Graphic3d_TextureRoot::IsColorMap() - introduced new property defining
if RGB(A)8 image formats should be loaded as sRGB(A) textures or as data RGB(A) textures.
OpenGl_Texture initialization methods have been extended with new theIsColorMap argument.
vreadpixel - added argument -sRGB printing color in sRGB color space.
Test cases have been updated to new sRGB rendered results.
The table of named colors has been compressed and moved out
from Quantity_Color.cxx into Quantity_ColorTable.pxx.
Quantity_NameOfColor - grayscale enumeration values have been re-ordered to fix discontinuity.
Duplicating colors has been merged within enumeration:
CHARTREUSE=CHARTREUSE1, GOLD=GOLD1, GREEN=GREEN1, ORANGE=ORANGE1,
ORANGERED=ORANGERED1, RED=RED1, TOMATO=TOMATO1, YELLOW=YELLOW1.
Added aliases to several other common colors:
BLUE=BLUE1, CYAN=CYAN1, LIGHTCYAN=LIGHTCYAN1, MAGENTA=MAGENTA1.
Quantity_Color class definition has been cleaned to follow OCCT coding style.
Quantity_Color now stores NCollection_Vec3<float> as class field instead of separate components.
Removed unused class Quantity_ColorDefinitionError.
New methods Quantity_Color::Convert_LinearRGB_To_sRGB() and Quantity_Color::Convert_sRGB_To_LinearRGB()
converting RGB components from linear to non-linear sRGB colorspace and vice versa.
Image_PixMap::PixelColor() and Image_PixMap::SetPixelColor() methods have been extended
with an optional argument for performing linearization/delinearization of 8-bit sRGB pixel formats.
Draw Harness command AISColor has been corrected to take color name instead of enumeration index.
1. Correct building history: the case of closed spine.
2. Rollback method GeomFill_CorrectedFrenet::InitInterval - correct processing singularities on spine.
3. Correct test cases.
Changes for using Axis Placement(not used during creation of assembly) as transformation for root Shape Representation.
Test for bug 29068 and auxilary command are added.
When splitting the shell/face with internal faces/edges use the 'internal' criteria of the face to choose the way to create loops.
Side effect changes:
- When performing Boolean operation - move the objects located far from Origin to the Origin to increase the accuracy of intersections.
1. BRepLib.cxx: calculation of 2d tolerance is changed in method BRepLib::SameParameter(Edge..)
2. Geom2dConvert.cxx: incorrect comparing
"SquareDistance < tolerance"
is replaced by
"SquareDistance < tolerance*tolerance"
because tolerance is linear value.
1. src\BRepSweep\BRepSweep_NumLinearRegularSweep.cxx
Fix bug by adding result in list of generated shapes, if initial shape is vertex, edge or face.
2. src\BRepLib\BRepLib.cxx
Add protection against treatment not geometric edge in BRepLib::UpdateInnerTolerances(...)
3. Add test case for bug and correct test for bug 30346 according to new behavior of algorithm
Methods of classes Geom_ElementarySurface, Geom_Conic, and Geom2d_Conic setting or returning values of fields are made inline and return const& to avoid copying
Updated OCC logo.
Replaced some external links to https.
Added references to VS2019 support.
Added glTF, OBJ, IFC, JT into listing within appropriate sections.
Replaced dead URLs.
Fixed misleading information.
BRepMesh_ModelHealer - recognize a small face with 1 wire and 2 small edges as a face for refinement.
BRepMesh_DefaultRangeSplinner - reduce tolerance on tiny faces to improve the meshing results.
Added new test cases: bugs mesh bug25044_*.
Existing test cases updated to reflect improvements.
Modifications added to upgrade guide:
- Font_FontAspect enum digits have different values (0030439),
- Graphic3d_PolygonOffset default value change (0030716),
- AddZLayer inserts layer before Graphic3d_ZLayerId_Top, the previous implementation inserted it above all layers(0030668).
Undefined behavior caused by left shift operations in TopLoc_Location::HashCode() function is fixed.
HashCode() function overload for Standard_Size type is made available only if Standard_Size and "unsigned int" are different types (it is usually true for 64-bit platforms). The overload for "unsigned int" is made simple non-templated function (so it behaves the same on 32-bit and 64-bit platforms).
HashCode() function overload for Standard_Utf32Char type is made available only if Standard_Utf32Char and "unsigned int" are different types (it is needed for some old compilers).
Added missing HAVE_RAPIDJSON to env.sh.
genconf.tcl has been extended with new option HAVE_E57
which enables looking for E57 and xerces libraries.
Added new libraries CSF_E57 and CSF_xerces to genproj.tcl.
ViewerTest::ParseColor() - fixed alpha component defined as 255 for RGB-only input.
Quantity_Color::ColorToHex() now rounds-up float values to closest integer.
Quantity_ColorRGBA::ColorToHex() - added method formatting color into hex with alpha component consistent to Quantity_Color::ColorToHex() for RGB.
vdrawtext command now uses ViewerTest::ParseColor() for parsing color argument, so that it accepts hex.
vreadpixel command now has been extended with -hex argument for printing color in hex format.
Commands XSetColor, XGetColor, XGetShapeColor, XGetAllColors, XAddColor, XRemoveColor, XFindColor,
XUnsetColor, XGetInstanceColor, XSetInstanceColor have been corrected to properly report syntax input errors
and to accept color names and hex (by reusing ViewerTest::ParseColor()).
The visualization code drawing iso-lines has been protected against getting iso-lines of infinite offset surface. In such case the surface is trimmed by UV bounds of the face.
Test case added
Method ShapeFix_Wire::FixNotchedEdges() is corrected to handle the case of closed notched edges: on such edges split parameter may fall to wrong end of the curve due to projection.
1. new file Standard_Dump to prepare and parse Dump in JSON format for OCCT objects
2. some presentations cover the proposed dump functionality.
3. 'bounding', 'vaspects' has '-dumpJson' field to see the DumpJson result
4. Bnd_Box constructor with min/max points is implemented to use Dump of this class in Dump BVH_Box
5. Limitation (some classes of Graphic3d, Prs3d has not full filling for DumpJson)
Implementation of auxiliary class in OpenGl_Text.cxx is corrected to avoid storing reference to temporary object.
Off-topic: test parse rules are corrected to recognize situation when test is killed by elapsed time, and report it as such.
genproj procedure has been extended with new parameters -solution, set to OCCT by default.
Added missing HAVE_RAPIDJSON to env.sh.
Added implicit file path normalization for PRODUCTS_PATH variable,
to avoid genconf.tcl script malfunction on backslashed.
Case-sensitive method strncmp used in the REStl_Reader cosidered register of letter was replaced on the method _strnicmp for Windows and strncasecmp for Linux and MacOs
Test script for case 0030113
In class ShapeFix_ComposeShell the U- and V-closedness of input face is taken immediately avoiding the error when the face is small in any of these dimensions.
Test case for bug 0030831
New overload for method OSD::SetSignal() is added accepting argument specifying how to set or unset handlers.
New enum OSD_SignalMode describes different modes of signal handlers setting used in extended version of OSD::SetSignal().
Method OSD::SignalMode() returns mode set by the last call to SetSignal().
Method OSD::IsFloatingSignalSet() is changed to return value based on actual floating point exception flags (rather than on variable set by previous call to SetSignal()).
Added new method OSD::SetThreadLocalSignal() intended to setup thread-specific handlers (e.g. _set_se_translator() on Windows) and FPE settings.
OSD_ThreadPool and Media_PlayerContext now use new method instead of OSD::SetSignal(), to avoid overriding global handlers.
dsetsignal syntax has been extended to support choice of signal handling mode.
DRAW command OCC30775 is added allowing to test signal handling in multithreaded process.
Do not limit number of available threads by number of items processed within the root scope due to possibility of spawning of an additional threads within the subscopes.
Fix Newton's minimum projection algorithm onto boundaries in case of conditional optimization.
Add possibility to detect several optimal points at initialization of the math_GlobOptMin.
Protections to avoid exception in the methods RWStepVisual_RWDraughtingModel::ReadStep and RWStepVisual_RWPresentationLayerAssignment::ReadStep for case when assigned items are absent were added
Added work-around against specific problem of ORIENTED_EDGE referring to another ORIENTED_EDGE instead of EDGE_CURVE.
Added handling of edges with null EDGE_CURVE, to allow passing the translation further instead of exception.
Off-topic: text of warning message is corrected in genproj
Added test bugs step bug29979
Corrected misprint in the method CheckPCurves of the class StepToTopoDS_TranslateEdgeLoop for case when higher bound of the range of pcurve on the edge greater then last parameter of the pcurve.
Added check on the not periodic flag for pcurve when range pcurve on the edge is adjusted to parameters pcurve.
- Fix FixSmallAreaWire for the reversed faces.
- Add nullptr check for FixOrient...
- Test bugs heal bug25923 is corrected to avoid relying on wrong orientation of face after fixing small wire
Decomposition of Ambient, Diffuse, Specular and Emissive properties has been eliminated within *Graphic3d_MaterialAspect* definition,
so that following methods of *Graphic3d_MaterialAspect* class have been removed:
SetReflectionMode(), SetReflectionModeOn(), Ambient(), Diffuse(), Emissive(), Specular(), SetAmbient(), SetDiffuse(), SetSpecular(), SetEmissive().
Graphic3d_NOM_NEON_GNC non-physical material definition has been corrected
and define WHITE emission color coefficient instead of YELLOW.
Added method Graphic3d_ClipPlane::SetCappingColor() for more straight-forward modification of Capping color in case of non-physical material.
Prs3d_ShadingAspect::Color() has been fixed to return InteriorColor instead of material color coefficient in case of non-physical material.
OpenGl_Workspace now disables at initialization time
OpenGl_View::SetImmediateModeDrawToFront() property in case of "GDI Generic" OpenGL implementation.
OpenGl_View::redrawImmediate() - fixed no buffer swap in case of
myTransientDrawToFront without offscreen FBO.
Introduced two new commands vlistcolors listing named colors and vlistmaterials listing standard materials.
These commands also can generate HTML table listing colors/materials with their properties.
vlistmaterials can generate OBJ model for comparing OCCT materials in different viewers.
New test cases use these commands for generating HTML tables.
Graphic3d_MaterialAspect::MaterialType() - added missing getter.
Corrected name of unlit shading model within vaspects and vrenderparams commands.
Method BRepOffset::Surface() is used in simple offset algorithm to create offset surface, instead of direct and unconditional construction of new Geom_OffsetSurface instance.
Added test bugs modalg_7 bug29334
Refactoring of the Line-Line intersection method strengthening the parallel and coincidence criteria to allow almost parallel lines have an intersection point.
Fixed license statements within files BRepGProp_MeshProps.cxx, BRepGProp_MeshProps.hxx,
OSD_ThreadPool.cxx and OSD_ThreadPool.hxx.
Removed non-ASCII char from BRepMesh_ModelHealer.hxx.
Graphic3d_Group::Text(...) are obsolete, AddText() should be used instead of these methods.
Graphic3d_Text is a new class for parameters necessary to fill OpenGl_Text. All parameters of Graphic3d_Group::Text() are moved into this class.
OpenGl_TextParam is removed, these fields were moved into Graphic3d_Text.
OpenGl_Text constructors/Init with OpenGl_TextParam parameter were removed. Constructor with Graphic3d_Text should be used instead of it.
Using OpenGl_Text Init() with OpenGl_TextParam should be now replaced on two cases. The first case is setting values into Graphic3d_Text and the second case is calling Reset() after. As example, look at modification in OpenGl_FrameStatsPrs.
- Fixed condition to allow proceeding of nested compounds;
- Added new key to "reshape" draw command to state a level of type until which requests are taken into account;
- Test case added.
Clipping range has been moved from SelectMgr_RectangularFrustum to SelectMgr_SelectingVolumeManager
and passed to frustum as an argument to Overlap() methods.
This fixes an issue when Clipping is customized per-object within SelectMgr_ViewerSelector::traverseObject()
in case when shallow copy of SelectMgr_SelectingVolumeManager is created
(frustums are copied from global frustum manager by Handle).
The regression caused by the patch for 0030534 has been eliminated. For that the method InsertWiresIntoFaces() has been corrected: the face to which a wire is added is first oriented as forward.
"endl" manipulator for Message_Messenger is renamed to "Message_EndLine".
The following entities from std namespace are now used
with std:: explicitly specified (from Standard_Stream.hxx):
std::istream,std::ostream,std::ofstream,std::ifstream,std::fstream,
std::filebuf,std::streambuf,std::streampos,std::ios,std::cout,std::cerr,
std::cin,std::endl,std::ends,std::flush,std::setw,std::setprecision,
std::hex,std::dec.
New folder "hlr" with subfolders "exact_hlr" and "poly_hlr" has been created.
Draw command "vcomputehlr" has been corrected.
QA command "OCC27341" has been deleted as "vcomputehlr" does its job.
Presentation computation has been modified to merge triangulations per box part type.
Prs3d_Drawer::FaceBoundaryDraw() flag is now handled for drawing box sides edges.
Graphic3d_ArrayOfPrimitives has been extended by auxiliary methods:
- ::AddSegmentEdges(), ::AddTriangleEdges(), ::AddQuadEdges() raising exception on wrong primitive type.
- ::AddQuadTriangleEdges() splitting quad into indexed triangulation;
- ::AddTriangleStripEdges() splitting triangle strip into indexed triangulation;
- ::AddTriangleFanEdges() splitting triangle fan into indexed triangulation;
- ::AddPolylineEdges() splitting polyline into indexed segments.
1. Add method TransformPCurves to ShapeUpgrade_UnifySameDomain - to rotate and translate existing pcurves instead of projecting.
2. Modify method ShapeUpgrade_UnifySameDomain::IntUnifyFaces - now it does not apply ShapeFix to new wires and new faces.
3. In the method UnifyEdges: add merging degenerated edges.
Graphic3d_Camera has been revised to store camera Direction from Eye explicitly
and Distance to Center instead of auxiliary Center point.
This allows setting camera Distance to 0 (for flat objects) without side effects
like broken FitAll, wobbling on rotating.
OpenGl_ShaderManager now provides fallback stipple lines GLSL code for obsolete OpenGL ES 2.0 devices.
OpenGl_Caps has been extended by flags disallowing OpenGL extensions usage
and restricting upper OpenGL version to be used by OCCT for testing purposes.
Fixed error in math converting Y-down mouse coordinates into Y-up within AIS_ViewController::handleZoom().
Added flags ToStickToRayOnZoom()/ToStickToRayOnRotation() enabled by default,
which will project picked point onto ray at mouse cursor to preserve focus on zoom out.
OpenGl_Text now discards (redundant) translation part from camera orientation matrix.
Text anchor point alignment to integer coordinates is now not performed for 3D-oriented text.
Added new class AIS_ViewCube implementing interactive cube
displaying orientation of the main axes of the model space in the viewer.
Each side, edge, or corner of the cube corresponds to particular orientation of the camera,
and the class provides methods to move the camera to corresponding position (with animation if needed).
AIS_InteractiveContext::LastActiveView(), added new property returning the last View processed by MoveTo() event.
AIS_InteractiveContext::BoundingBoxOfSelection(), added method returning bounding box of selected objects.
SelectMgr_EntityOwner::HandleMouseClick(), added new callback for handling
mouse clicks by owner itself without automatic highlighting and clearing previous selection.
Called by AIS_InteractiveContext::Select() method.
AIS_ViewController::ViewAnimation() has been extened with camera animation propery,
which can be bound to AIS_ViewCube for smooth embedding into event loop.
Prs3d_ToolDisk has been extended with parameters specifying angle range.
Graphic3d_MaterialAspect now initializes all coefficients to 1.0
when Graphic3d_NOM_UserDefined is passed to class constructor.
AIS_AnimationCamera::update() now sets the end camera position if animation duration is 0.
Prs3d_DatumAspect, added missing setters.
New command vviewcube has been added.
The problem occurred when the algorithm IntCurvesFace_Intersector was called for a face occupying zero extension along V direction. This is obviously a bad face, but we need to have the algorithm not crashing the application. So, the algorithm has been protected against such input data. At attempt to make calculations it establishes NotDone status.
Bad test case tests/de/step_3/D9 has been updated.
False memory leak detection has been suppressed in tests/bugs/fclasses/bug7287_2.
V3d_Viewer::AddZLayer() has been replaced by
V3d_Viewer::InsertLayerBefore() and V3d_Viewer::InsertLayerAfter() methods
allowing to specify Layer settings and index of existing layer to insert new one before/after.
The behavior of V3d_Viewer::AddZLayer() has been adjusted to append
new Layer before Graphic3d_ZLayerId_Top
(between Graphic3d_ZLayerId_Default and Graphic3d_ZLayerId_Top).
Graphic3d_Layer now provides LayerId() property.
Graphic3d_ZLayerSettings has been extended with IsRaytracable() property;
Ray-Tracing now processes multiple layers instead of Graphic3d_ZLayerId_Default.
Moved some methods from OpenGl_View to the base class Graphic3d_CView.
OSD_OpenFileDescriptor(), added truncating flag to fopen() flags for std::ios_base::out,
since truncating is implied by C++ output flag.
This fixes misbehavior of OSD_OpenStream() on MinGW-w64 platform.
New classes BRepMesh_ConstrainedBaseMeshAlgo, BRepMesh_CustomBaseMeshAlgo and BRepMesh_CustomDelaunayBaseMeshAlgo are added.
These classes allow to add any custom triangulation algorithm to BRepMesh and perform post-processing and optimization of base mesh generated by those algorithms.
BRepMesh_Delaun: added possibility to process constraints when base mesh is generated by different algorithm.
BRepMesh_DelaunayNodeInsertionMeshAlgo: added PreProcessSurfaceNodes flag controlling addition of surface nodes (either before creation of base mesh or after) to gain maximum performance from triangulation algorithms.
Minor changes:
Use simple algorithm for cylinders when internal vertices mode is switched off to speed up computations.
BRepMesh_IncrementalMesh: added Perform method allowing to execute algorithm using manually created Context.
A new flag myToPropagateVisualState is introduced for PrsMgr_PresentableObject: by default it is true, it means that the visual state (display/erase/color) should be propagated to all children. If false, the visual state is not propagated.
The flag can be set via the method ToPropagateVisualState() or via Draw command: "vparent <parent> -ignoreVisu"
BRepFill_TrimShellCorner::CheckAndOrientEdges() - When orienting next edge in a sequence take into account the Orientation of the previous edge.
Test cases for the issue.
Protect IntTools_EdgeEdge::FindSolutions method from cases where splitting edge on parts does not give bounding box decreasing.
Use at least Epsilon from edges parameters as a step in BRepLib::FindValidRange.
Test case for the issue.
Added new parameter KeepSmallEdges to IMeshTools_Parameters allowing to compute min size for each edge locally, depending on the length of particular edge.
Parameter -smalledges now available for incmesh command.
Correct the behavior of the command 2dapprox for the case of points input in the command line.
Restore the work of the command 2dinterpole (implemented in the same method as 2dapprox).
Add test cases.
Correct generation of snapshots for the tests lowalgos/2dinter/*.
IMeshTools_ModelAlgo and IMeshTools_ModelBuilder have been changed to provide exception protected interfaces for performing the operations.
Protect single Edge/Face discretization methods from raising exceptions to skip broken Edges/Faces and allow mesh construction on the whole model.
Font_FTFont now uses fallback fonts for characters from unsupported Unicode subsets,
managed by Font_FTFont::ToUseUnicodeSubsetFallback()
and Font_FontMgr::ToUseUnicodeSubsetFallback() option, enabled by default.
The fallback list includes common font families for Chinese, Korean and Japanese languages.
Font_FTFont::RenderGlyph() now supports FT_PIXEL_MODE_MONO input format used by some CJK fonts.
OpenGl_Font::createTexture() now limits single texture size to circa 4096 glyphs.
test/testgrid now expects test scripts being in UTF-8 encoding in sync with "DRAWEXE -f script.tcl".
AIS::InitFaceLength() - fixed usage of uninitialized result.
Add possibility of construction of the Optimal Oriented Bounding Box from set of points (the case of shape with triangulation).
The interface of the BRepBndLib::AddOBB method is not changed, but the option <theIsOptimal> now controls also the construction of the OBB from Set of points.
The slightly modified DiTo algorithm will be used, checking all possible axes created by the extreme points.
The performance of the construction of the Optimal OBB is lower but the quality is usually much higher (can't be worse by definition).
Test cases for the issue.
Protect Expand compounds against problems with located subshapes.
Add method XCAFDoc_ShapeTool::AddSubShape() with Boolean output parameter.
Speed up XCAFDoc_ShapeTool::Expand().
Speed up XCAFDoc_ShapeTool::FindSubShape().
Avoid CDF_Stroe to find default folder since there is no case where this folder does not reset to exact value given by the user.
Also, remove the exception raising for the case this method is called from outside.
AIS_InteractiveContext::Load/ClearGlobal is modified for correct processing SetSelected for child objects.
As Display/Erase of object with children displays/erases these children, then the children status inside context should also be updated.
ViewerTest_EventManager now inherits AIS_ViewController.
Platform-dependent user input handling within ViewerTest has been revised
to process events in common way through AIS_ViewController.
The mouse navigation has been changed, so that left mouse clicked
without modifers now rotates View.
The rubber-band selection can be activated via Alt+LeftMouseButton.
Selection is now done on mouse unclick and keyboard short-cuts take effect on unclick.
Aspect_Window::SetTitle() - added new method configuring Window title.
Introduced new types Aspect_Touch, Aspect_VKey, Aspect_ScrollDelta
for processing window events in platform-independent way.
SelectMgr_RectangularFrustum::Overlaps method for computing intersection of box with frustum was reworked. Now the nearest non-clipped point is calculated.
AIS_InteractiveContext::IsHilighted() behavior has been changed.
It should use the same logic for owner processing like when it is highlighted. IsSelected of owner is used for selection state, not highlight.
Without this correction, object is never shown as selected by click (Select(bool)):
if (!HighlightStyle (theOwner, aCustomStyle) ||
(!aCustomStyle.IsNull() && aCustomStyle != anObjSelStyle))
IsSelected is already true, so the highlight style is equal to selection style, nothing to do.
RWMesh_CafReader - added new interface class for common workflow for reading mesh data files into XDE document.
OSD_Path - added auxiliary methods splitting path into folder+file pair
and checking relative/absolute path semantically:
OSD_Path::FolderAndFileFromPath(), ::IsRelativePath(), ::IsAbsolutePath().
V3d_TypeOfOrientation enumeration has been extended with aliases
(like front/left) for Z-up and Y-up conventions.
V3d_View::SetProj() now accepts argument for asking Y-up instead of Z-up.
Added command vviewproj defining standard camera direction.
Commands vaxo, vleft, vright, vtop, vbottom, vfront, vbottom now redirect to vviewproj.
TCollection_AsciiString::SubString() now uses Standard_OutOfRange_Always_Raise_if() to suppress GCC warning.
Eliminated gcc 4.4 compilation errors within Standard_OutOfRange_Raise_if,Standard_RangeError_Raise_if.
SEARCH_TEMPLATES was modified to find EGL and GLES2 libraries in different folders (angle-gles, egl-gles, egl, angle, etc).
Find 3rdparty products procedure allows to define 3RDPARTY_LIBRARY_DIR or 3RDPARTY_DLL_DIR manually to find necessary libraries and dlls.
Added an argument to OpenGl_GraphicDriver::GetSharedContext() for returning only currently bound OpenGL context.
SelectMgr_SelectingVolumeManager::GetMousePosition() - added getter returning mouse coordinates.
OpenGl_CappingPlaneResource::updateTransform() now takes into account ZLayer origin.
Fixed VT_ProcessKeyPress() passing arbitrary input to Draw::Atoi() leading
to messages in console like "unclosed braces".
Provide the easy to use interfaces for selection of the elements from BVH tree.
The selection rules should be implemented in the selector class derived from *BVH_Traverse* or in *BVH_PairTraverse* in Reject/Accept methods.
The *BVH_Traverse* is used for selection of the elements from the tree.
The *BVH_PairTraverse* is used for selection of the pairs of elements from two BVH trees.
Auxiliary changes:
- Two methods BVH_Box::IsOut(OtherBox) and BVH_Box::IsOut(Point) have been added;
- Added methods for conversion of Bnd boxes to BVH boxes
Added new class *BVH_Tools* containing useful static methods operating on BVH points and boxes.
The classes BRepExtrema_OverlapTool and BVH_DistanceField have been rebased to use the new traverse methods.
STEPCAFControl_Reader::SourceCodePage() - added property defining text encoding
for converting names within STEPCAFControl_Reader::ReadNames() method.
Added associated "read.stepcaf.codepage" parameter,
which default value is Resource_UTF8, preserving current behavior.
Resource_FormatType enumeration has been extended by UTF8 and SystemLocale values.
Resource_Unicode - added conversion methods taking Resource_FormatType as argument.
GetName command has been corrected to NOT replace non-Latin symbols.
Cache subshapes list into map instead of re-iterating it multiple times within the loop.
Avoid iteration at all if modified state has been already detected.
XCAFDoc now returns GUIDs by const reference instead of a copy.
Modification in static method BoundedArc of IntStart_SearchOnBoundaries:
add exact intersection of canonical curve-surface (when Func is IntPatch_ArcFunction).
Add protection from return of not top-level shape by FindShape method.
Add protection against located roots into FindMainShape method.
Add new Draw command for FindMainShape.
Add flag findInstance to Draw command FindShape
The following improvements have been made in Boolean operations algorithm:
1. Added possibility to update FaceInfo structure for many faces at once which helps to avoid nested loops.
2. Improve Point-Face classification procedure by caching the FaceExplorer for a face.
SelectBasics_EntityOwner has been merged into SelectMgr_EntityOwner.
Unused property SelectMgr_EntityOwner::ResetLocation() has been removed.
SelectBasics package has been moved from TKService to TKV3d.
SelectBasics_SensitiveEntity has been merged into Select3D_SensitiveEntity.
The problem caused by the changing of HashCode() method of TColStd_PackedMapOfInteger::TColStd_intMapNode class done as a part of 30550 changes is solved.
AIS_InteractiveContext::EraseGlobal() now erases all object owners
(in the same way as AIS_InteractiveContext::ClearGlobal()).
Removed redundant fields AIS_InteractiveContext::myLastinMain and myWasLastMain.
Evaluator of offset surface has been protected against evaluation at infinite parameters. Now it throws exception when evaluating such data.
The methods IsUClosed and IsVClosed of the class ShapeAnalysis_Surface have been corrected to avoid evaluation of the surface at infinite parameters (fighting with regressions "parasolid doc_3 E3" and "parasolid doc_2 A3" in products).
New methods AddChildWithCurrentTransformation(), RemoveChildWithRestoreTransformation() are implemented to keep the current global transformation of the object (child).
In DRAW a new command vchild is provided for testing low-level connections between presentations.
The matter is that when the binary OCAF file contained a triangulation with 0 nodes and 0 triangles the method Paste would cause construction error of Poly_Triangulation.
This patch protects this method against exception.
XDisplay - added new command displaying XDE document in ViewerTest map of object:
> XDisplay Doc [label1 [label2 [...]]] [-explore {on|off}] [-docPrefix {on|off}] [-names {on|off}]
> [-noupdate] [-dispMode Mode] [-highMode Mode]
vstate - added NULL checks.
vdisplay now prints error on attempt to specify unsupported display mode.
verase, vremove now raise Tcl exception on attempt to hide non-existing object.
verase, vremove, vdir now accept name masks like "verase b*".
vremove and vdir output is now consistent to verase command (space-separated list of removed objects).
Added generalized move constructor and assignment operator to initialize TopoDS_Shape by object of this or derived type.
Added move constructor and assignment operator for TopLoc_SListOfItemLocation.
Macro OCCT_NO_RVALUE_REFERENCE is used in Standard_Handle.hxx instead of direct check of compiler version.
The capabilities of the class BRepOffsetAPI_MakeEvolved have been extended with involving alternate algorithm of resolving the loops in the piped shape. The new option theIsVolume is added for that in the constructor.
New class BRepFill_AdvancedEvolved has been created in order to provide new OCCT-algorithm combining BRepFill_PipeShell and BOPAlgo_MakerVolume.
A change in BOPAlgo_PaveFiller.cxx has been made in order to solve a specific problem of Boolean operation.
The interface of DRAW-command "evolved" has been updated to add the new option. DRAW-command "evolvedsolid" has been deleted. Now it is replaced with DRAW-command "evolved" with the option "-solid".
Testgrid "evolved" has been created.
BRepPrimAPI_MakeRevol.cxx, hxx
Insert checking of self-intersecting of generated surfaces in MakeRevol
BRepTest_SweepCommands.cxx
Adjusting command revol to current behavior of MakeRevol algorithm.
Removed redundant proxy class Prs3d_Presentation - Graphic3d_Structure now should be used directly.
Removed auxiliary class PrsMgr_ModedPresentation holding Structure and Display Mode index (field has been moved to Structure itself),
so that PrsMgr_Presentations (AIS_InteractiveObject::Presentations()) now holds Structures directly.
PrsMgr_Prs proxy class has been removed, and PrsMgr_Presentation now inherits Graphic3d_Structure.
Graphic3d_Structure, maps declarations have been corrected to use class instead of void*, which allowed to avoid redundant casts.
Several methods have been also modified to avoid creating Handle from this pointer.
AIS package headers have been cleaned up from forward declaration of Prs3d_Presentation class.
Fixed dereference of destroyed presentations within connected and HLR computed objects:
- AIS_InteractiveContext::Disconnect() now erases connected presentation.
- Graphic3d_CView::Clear() now removes destructed structures from Computed list.
- fix inaccessibility of PrsMgr_Presentation methods
Fixed inconsistency of compiler flags with genproj.
warn_on+QMAKE_CXXFLAGS_WARN_ON is now used instead of warn_off+QMAKE_CXXFLAGS.
Added No_Exception to DEFINES.
AIS_Manipulator::Transform() now considers object parent transformation.
Added new command vchild for easier testing of low-level connections between presentations.
AIS_InteractiveObject/SelectMgr_SelectableObject/PrsMgr_PresentableObject
- added missing getters;
- removed redundant private sections;
- improved documentation;
- AIS_InteractiveObject::Compute() marked as pure interface;
- methods within this hierarchy have been moved to the base class, when reasonable.
Removed redundant typedefs to pointers (PrsMgr_PresentationPointer, PrsMgr_PresentableObjectPointer).
genconfdeps.tcl now puts %PRODUCTS_PATH%/$PRODUCTS_PATH into generated custom.bat/custom.sh scripts.
Added missing CSF_winmm definition to OccToolkit.pri
Method Segment() of B-spline curve and surface has been extended by parameter theTolerance (theUTolerance and theVTolerance for surface), which defines the proximity between knots of a NURBS and boundaries of cutting segment. The default value of the tolerance is Precision::PConfusion().
Test cases have been added to check segmenting of B-spline surface and curves both 2D and 3D.
draw_test_harness.md - description of new options in Draw commands
AppDef_BSplineCompute.hxx, BRepApprox_TheComputeLineOfApprox.hxx, GeomInt_TheComputeLineOfWLApprox.hxx, Approx_BSplComputeLine.gxx - implementation of method SetPeriodic(...) and implementation periodic boundary conditions for multiline in order to get periodic multicurve.
GeomAPI_PointsToBSplineSurface.hxx, GeomAPI_PointsToBSplineSurface.cxx - adding new parameter for methods Init(...) and Interpolate(...), implementation of building periodic tangents for first and last AppDef_MultiPointConstraint of multiline for U direction of surface.
GeometryTest_APICommands.cxx - implementation of new functionality in Draw command surfapp and surfint
GeomFill_NSections.cxx
Fixing problem with bugs modalg_3 bug606_2
A new command vbackground is created. Old background commands are made aliases for the newly created command (including vsetdefaultbg).
Tests are modified using newly added command vbackground.
Font_FTFont now defines shear transformation to synthesize italic style for fonts having no such style.
Font_FontMgr::FindFont() and command "vfont -find" have been extended with -strict option
to check whether the given font is actually registered or not.
Font_FTFont::Init() - added constructor from memory buffer.
Second Font_FTFont::Init() override has been renamed to Font_FTFont::FindAndInit()
to avoid ambiguity between two similar methods taking full font path and font name as string.
1. Approx_ComputeCLine.gxx, Approx_FitAndDivide.hxx, Approx_FitAndDivide2d.hxx, BRepFill_ComputeCLine.hxx
It is base modification, which allows improve performance of approximation with help of Approx_ComputeCLine. The main idea of improvement is using degree selection by inverse order - from maxdegree to mindegree. If tolerance for maxdegree is not reached, there is no sense to make approximation for current number of knots with lower degree, it is necessary to cut parametric interval.
2. ProjLib_ComputeApprox, ProjLib_ComputeApproxOnPolarSurface, ProjLib_ComputeApproxOnPolarSurface, ProjLib_ProjectOnPlane
It is additional modification of methods using Approx_ComputeCLine.
Mainly, modifications concern to more optimal choosing parameters for approximation algorithm.
3. BRepCheck_Face
Small improvement of method Intersect(...), which intersects two wires on face.
4. BRepTopAdaptor_FClass2d
Impovement of treatment infinitely narrow faces.
5. ChFi3d/ChFi3d_Builder_6.cxx
Small improvement, which forbids extension of singular boundary of surface.
It was TODO problem in tests/bugs/modalg_7/bug27711_3
6. IntTools_EdgeEdge.cxx
Improvement of performance for cases of searching common parts between line and analytical curve
7. GeomliteTest_CurveCommands.cxx
Adding Draw command fitcurve. This command is analog of approxcurve, but uses Approx_FitAndDivide algorithm.
Mainly to have direct draw command for testing Approx_ComputeCLine.
8. Extrema_ExtElC.cxx
Treatment of case "infinite solutions" for extrema line-ellipse
9. Modification of some tests according to new behavior of algorithm.
10. tests/perf/moddata/bug30435
Test for new improved algorithm.
11. Implementation QAcommand OCC30435 in QABugs_20.cxx used in test bug30435
0030551: Foundation Classes - Integer overflow in NCollection_CellFilter HashCode
Signed integers are not used in hash code functions now to prevent undefined behavior on left shift operations with signed integers.
A possibility of negative values of hash codes is eliminated.
INT_MAX → IntegerLast() in hash code functions.
All found hash code functions behaves uniformly now: they return a value in the range [1, theUpperBound]. Relevant comments are added to such functions.
Method for calculation of correct number of sampling points is added for
Adaptor2d_OffsetCurve.
Now number of sampling points correspond with similar method for
Geom2dAdaptor_Curve for curve type GeomAbs_OffsetCurve
Test case added
ViewerTest::ParseColor() function is improved to be able to parse the following set of input arguments:
- "Red Green Blue [Alpha]", where Red, Green, Blue, Alpha must be integers within the range [0, 255] or reals within the range [0.0, 1.0]. Note that "0 0 1" triple is parsed as "0.0 0.0 1.0" and will be interpreted as a blue color.
- "ColorName [Alpha]", where ColorName is one of WHITE, BLACK, RED, GREEN, BLUE, etc. (look at Quantity_NameOfColor enumeration for all possible variants). Alpha may be integer or real, as described at the previous list item.
- #HHH, [#]HHH[H], [#]HHHHHH[HH], where H is a hexadecimal digit (0 .. 9, a .. f, or A .. F). There are a short hexadecimal RGB, RGBA formats, and a usual RGB[A], respectively.
Constructor of GeomFill_SectionPlacement has been corrected: add correction of boundaries of curve to avoid mistake of LocateU in the case when the curve is BSpline.
Introduced package Media wrapping FFmpeg structures.
Media_PlayerContext class manages video decoding into Media_IFrameQueue interface.
Graphic3d_MediaTextureSet implements Media_IFrameQueue interface and can be used as texture source.
AIS_MediaPlayer implements simple AIS object displaying video.
Renamed classes (moved from TKOpenGl to TKV3d):
- OpenGl_Layer -> Graphic3d_Layer;
- OpenGl_BVHClipPrimitiveSet -> Graphic3d_BvhCStructureSet;
- OpenGl_BVHClipPrimitiveTrsfPersSet -> Graphic3d_BvhCStructureSetTrsfPers;
- OpenGl_BVHTreeSelector -> Graphic3d_CullingTool.
Method OpenGl_Layer::Render() has been moved to OpenGl_LayerList::renderLayer().
Standard Z-layers list definition has been moved
from OpenGl_GraphicDriver to base class Graphic3d_GraphicDriver.
Plane/Torus intersection: take into account the Plane-Torus distance sign to compute the correct location of the section curve.
Test case for the issue.
OSD_Parallel::ToUseOcctThreads() - new flag allowing to use OCCT threads implementation even when compiled with TBB (for testing).
Added new command dparallel for managing default Thread Pool.
OSD_Parallel::For() now avoid creation of universal iterator in simplest case.
BOPTools_Parallel - eliminated redundant typedefs/explicit instantiations of templates.
Added functor using array of per-thread context instead of a map.
Added adm/qmake/Occt.pro for building OCCT using Qt Creator.
genproj/genconf scripts have been extended with an option defining project format.
Draw_Appli now tries resolving CASROOT/DRAWDEFAULT/CSF_OCCTResourcePath/CSF_OCCTSamplesPath
locations within usual OCCT development/installation environment.
Added missing TKShHealing dependency to TKDraw/DRAWEXE (inherited from TKMesh).
OSD_Process::ExecutableFolder() - added method returning executable folder.
AIS_Manipulator crashes on SetPart method when 2 inputs (AIS_ManipulatorMode, Standard_Boolean) has been fixed up.
Test case for this bug has been added.
Added protections to avoid exceptions for null objects.
For reading entity StepShape_ToleranceValue was added reading lower and upper bounds represented by entities "StepRepr_ReprItemAndMeasureWithUnit" or "StepRepr_MeasureRepresentationItem"
Aspect_Window::InvalidateContent() - added new virtual method for invalidating window content using platform-specific API.
TKDraw, tkLoop() on Window platform has been changed so that to prevent
continuous input window events blocking terminal input
(e.g. in case if processing events is not fast enough or
if another continuously sends new events).
TKViewerTest, on X11 platform has been fixed a message processing
so that to avoid messages being not processed.
Added aggregation of Exposer and ConfigureNotify events.
Fixed aggregation MotionNotify events.
The following improvements have been implemented in BRepLib_MakeWire class to make it more reliable:
1. Quit adding edges of a wire into result if the current edges have not been added. Return NotDone status.
2. Allow merging not only bounding vertices (connected to only one edge) but also the middle ones, so the multi-connected vertices may appear in the result.
Test case for the issue.
Approx_SweepApproximation::Perform() now creates empty arrays.
GeomAdaptor_SurfaceOfRevolution::UTrim() - fixed No_Exception misuse.
StdPrs_ShadedShape - fixed defining an invalid reference to Poly_Triangulation::UVNodes().
BSplCLib::MovePoint() and BSplCLib::MovePointAndTangent() now take optional
weights parameter as pointer consistent to other methods like BSplCLib::BuildEval().
Approx_ComputeCLine.gxx - criterium to stop interval cutting is increased.
ProjLib_ProjectedCurve.cxx - fix regression for bugs modalg_5 bug25886
Other tests are modified according to current state of algorithm
Fixed incorrect array allocation within GeomEvaluator_OffsetSurface::CalculateDN()
and protection against assigning out-of-range values in case of non-square
number of U and V derivatives on surface.
Use .get() method instead of operator->() where result can be NULL.
STEPCAFControl_Reader::ReadLayers(), XCAFDoc_ColorTool,
XCAFDoc_LayerTool::SetVisibility() - fixed static functions syntax.
Method IGESToBRep_BRepEntity::TransferEdge() is fixed to record result even if it is wire rather than edge.
This allows translation of EdgeLoop entities referring to CompositeCurve items in EdgeList entities.
Curves that fail to translate are recorded in result as Null shapes (rather than skipping them) to keep indexation of curves in EdgeList valid for remaining ones.
Method IGESToBRep_BRepEntity::TransferLoop() is improved to skip failed edges thus avoiding exception and allowing translation to complete for transferable entities.
Warning messages are corrected to be more meaningful.
Added test bugs iges bug30544
Added missing NULL and empty result checks.
Updated test cases using offset_wire_092.brep shape containing all Edges
with TopAbs_INTERNAL orientation (hense no boundary to make offset).
Geom2d_BSplineCurve::LocateU(),Geom_BSplineCurve::LocateU, Law_BSpline::LocateU()
Geom_BSplineSurface::LocateU() and Geom_BSplineSurface::LocateV()
have been corrected with missing range checks.
BSplCLib::Hunt() documentation has been corrected to reflect its actual implementation.
OpenGl_AspectsTextureSet now appends Sprite texture to the end of texture set.
OpenGl_Context::init() - fixed usage of GL_MAX_COMBINED_TEXTURE_IMAGE_UNITS on too old GL context.
Default sampler names occSampler0, occActiveSampler and occSamplerBaseColor
are now excluded from header with common declarations for build-in GLSL programs.
OpenGl_ShaderObject::LoadAndCompile() and OpenGl_ShaderProgram::Link()
now includes program Id within error log.
OpenGl_ShaderManager now supports mapping texture onto point primitive within built-in programs
(texture transformation is ignored).
testdiff now looks for "IMAGE_COLOR_TOLERANCE: Tol" statements in log
to execute diffimage with -toleranceOfColor Tol parameter.
Image difference is first performed with 0 tolerance and then optionally with non-zero tolerance,
so that the log always includes information that some images are different,
but with two different messages:
- "IMAGE differs" for images failed comparison;
- "IMAGE is similar" for images passed comparison only with non-zero tolerance.
v3d/materials - disabled MSAA and removed duplicate.
demo/samples/materials - compare screenshots with 0.8% color tolerance (2/255).
1. Rewrite the method BRepTools::DetectClosedness.
2. Correct the method BRepOffset_Tool::EnLargeFace to process properly non-periodic but closed faces.
Removed excess mechanism for invalidating presentations AIS_InteractiveObject::myRecomputeEveryPrs.
PrsMgr_PresentableObject::Update() has been deprecated and replaced by PrsMgr_PresentableObject::UpdatePresentations()
recomputing only explicitly invalidated presentations.
Graphic3d_AspectFillArea3d, Graphic3d_AspectLine3d, Graphic3d_AspectMarker3d
and Graphic3d_AspectText3d have been merged into new class Graphic3d_Aspects.
The old classes are preserved as dummy sub-classes of Graphic3d_Aspects
preserving different per-aspect defaults.
Methods IsGroupPrimitivesAspectSet(), GroupPrimitivesAspect(), FillAreaAspect(),
LineAspect() and MarkerAspect() have been removed from Graphic3d_Group.
Instead, a new method Graphic3d_Group::ReplaceAspects() has been introduced
for replacing existing group aspects.
AIS_Shape now uses new method AIS_InteractiveObject::replaceAspects()
for updating computed groups with new aspects without presentation recomputation
in places where SynchronizeAspects() is not applicable.
OpenGl_AspectFace, OpenGl_AspectLine, OpenGl_AspectMarker
and OpenGl_AspectText have been merged into new class OpenGl_Aspects.
ViewerTest::parseColor() - fix uninitialized alpha component.
Graphic3d_AspectText3d/Prs3d_TextAspect - removed unused properties Space, ExpansionFactor, Angle.
Remove getters Values() deprecated since OCCT 7.1.0.
The fix is intended to obtain the correct result of offset operation in which the result contained only the small inverse part of the shape, which clearly should not have been included in the result at all.
The fix adds treatment of the inverted edges while removing the inside faces (collapsed ones) - RemoveInsideFaces method:
- Removing the invalid hanging blocks containing the inverted edges.
- Considering the face containing the inverted edges as the invalid one.
Test cases for the issue.
Added hints about ZLayer consideration and changing highlight colors to documentation
of AIS_InteractiveContext::SetHighlightStyle() method.
AIS_InteractiveContext::MoveTo() now checks if highlighting style does not use immediate layer
and performs full Redraw() instead of RedrawImmediate().
Improved parsing of ZLayer names within vdisplay, vzlayer and vselprops commands.
vreadpixel now reports syntax errors on wrong input.
Added two new Image_AlienPixMap::Load() methods, taking std::istream
and memory buffer (pointer, length) arguments.
This allows reading image from application memory or from file with non-zero offset.
Added Image_AlienPixMap::IsTopDownDefault() static property allowing to query rows order used by linked image library,
so that application might generate UV texture coordinates accordingly.
Added missing Release() to IWICImagingFactory instance.
Added new flag Graphic3d_AspectFillArea3d::ToDrawSilhouette() activating silhouette (outline) rendering.
The new feature can simulate fake HLR look-n-feel,
when combined with Aspect_IS_HIDDENLINE interior style (filling object with background color),
face boundary edges (with most continuity flag set to c2 or lower).
Silhouette GLSL program is very simple - it displaces model alongside vertex normal,
so that it is applicable only to smooth surfaces and closed volumes,
and produces visual artifacts at sharp corners, especially when face boundary is disabled.
OpenGl_SetOfShaderPrograms has been modified so that to reduce
dimensions of the grid of static size based on amount of program combinations.
OpenGl_PrimitiveArray no more allocates VBO resources if primitive array
is marked to be not drawn via interior style / line type / marker type.
The previous code has a condition to avoid processing the same faces if the face has a location.
The similar condition should be applied during the edges processing.
If not doing this, in the previous implementation, IMeshData_Edge instances are created for all edges(even located), but edges of faces located are not filled with curves.
As a result we had wrong local selection of edges.
Limit addition of edges to data model by ones with unique TShape and location using edges map already available in BRepMesh_ShapeVisitor.
A new flag Prs3d_Drawer::FaceBoundaryUpperContinuity() has been introduced
handled by StdPrs_ShadedShape::FillFaceBoundaries() method to exclude edges
of higher continuity class (e.g. to skip seam edges).
Draw Harness command vshowfaceboundary has been replaced by vaspects:
> vaspects -setFaceBoundaryDraw 1 -setFaceBoundaryColor RED -setFaceBoundaryType DASH.
Added macro Standard_NODISCARD equivalent to C++17 attribute [[nodiscard]] for compilers that support this.
Using Standard_NODISCARD macro for methods that create new object in gp, math, Geom, Bnd packages.
Marked equivalent operators with Standard_NODISCARD, if they are defined close to relevant methods.
Corrected code where warnings on unused result of calls to methods creating new objects are generated.
In most cases it looks like spelling errors (e.g. Normalised() instead of Normalise())
vdisplay no longer opens BREP file in case if Draw variable is not found.
Draw::Get() unused Complain argument has been removed from method definition.
Added Draw::GetExisting() and DBRep::GetExisting() commands never performing picking.
A new tool for debugging shaders has been introduced.
The new method OpenGl_ShaderProgram::UpdateDebugDump() allows dynamically
dumping/restoring certain shader program (vertex and fragment shaders) to/from external file.
The file name is generated from the program's name with suffix ".vs" or ".fs" for shader type.
The environment variable CSF_ShadersDirectoryDump specifies the folder for dumps.
If the file does not exist (first frame), then it is automatically saved.
When the file date/time is changed in comparison with recent cached date,
then the file will be automatically loaded from external file,
thus this file can be modified in any editor.
OpenGl_ShaderManager now generates a human-readable resource ids for standard GLSL programs.
Draw Harness command vshader has been extended with arguments -list,-dump and -reload
for debugging shaders within OpenGl_Context.
Aspect_IS_HOLLOW now an alias to Aspect_IS_EMPTY and Aspect_IS_HIDDENLINE does not implicitly enables mesh edges,
so that Graphic3d_AspectFillArea3d::SetDrawEdges() should be set independently.
OpenGl_ShaderManager now provides built-in GLSL programs for drawing mesh edges
in single pass (and on OpenGL ES which does not provide glPolygonMode()).
Graphic3d_RenderingParams::ToEnableAlphaToCoverage is now enabled by default
and properly handled at TKOpenGl level - enables coverage for Graphic3d_AlphaMode_Mask primitives.
OpenGl_PrimitiveArray now uses GLSL programs instead of glPolygonMode() by default,
which can be managed by OpenGl_Caps::usePolygonMode flag (desktop OpenGL only).
glPolygonMode() is also used as fallback regardless OpenGl_Caps::usePolygonMode flag
when GLSL programs are not supported (Geometry Shaders are required)
or stipple line style is required (not implemented within Face GLSL).
vaspects command has been extended by -setInterior -setDrawEdges -setEdgeColor -setEdgeType -setEdgeWidth
arguments replacing vsetinteriorstyle/vsetedgetype/vunsetedgetype commands.
vaspects now accepts arguments without "set" prefix.
ViewerTest::ParseColor() now parses RGBA color.
Redundant command BUC60738 has been removed.
AIS_ColorScale - fixed usage of uninitialized FillArea aspects.
OpenGl_View::runPathtrace() has been extended with alternative multi-pass
Adaptive Screen Sampling mode, not relying on atomic floating point operations.
Although atomic operations on floats allows single-pass rendering,
such operations leads to instability in case of different calculation order.
Atomic float operations are also currently supported only by single GPU vendor.
Fixed GLSL compilation on Intel drivers (follow ARB_shader_image_load_store
specs rather than EXT_shader_image_load_store).
Graphic3d_RenderingParams::AdaptiveScreenSamplingAtomic option has been added
to activate 1-pass Adaptive Screen Sampling mode when supported by hardware.
vfps command has been extended with -duration argument allowing to limit command execution time.
vactivate command has been extended with -noUpdate argument.
OpenGl_TileSampler has been refactored to better describe its logic:
- Offset image now defines tile index instead of offset to tile origin.
- Added 2D array defining the number of times to sample tile for straight-forward debugging.
Graphic3d_RenderingParams has been extended with property
RayTracingTileSize for testing various tile configurations.
Default behavior is the following:
- Target number of tiles (e.g. upper limit per frame): 256
- Tile size: 32x32.
OpenGl_View::runPathtrace() has been split into two methods per rendering stage.
OpenGl_Texture::Init() now returns FALSE immediately on 0 input dimensions.
Added Image_PixMapTypedData template class allowing to work with image data of known pixel format.
Occurrences of removed method vtkDataArray::InsertNextTupleValue() have been replaced by InsertNextTypedTuple().
Fixed misprint in vtkTypeMacro usage for class IVtkTools_ShapeObject.
Patch #0030452 (SelectMgr_ViewerSelector::Deactivate() raises exception when called twice) has been propagated to IVtkOCC_ViewerSelector.
OpenGl_ShaderObject::LoadAndCompile() - added method combining Load() and Compile() with optional error logging.
OpenGl_ShaderProgram::Link() now logs failures by default.
Doxygen configuration scripts are corrected to avoid '\n' symbols in the end of the @figure alias that caused multiple warning messages.
Other Doxygen warnings in Markdown files are eliminated.
In CMake builds the format of Doxygen warnings is set to use the style recognized by Visual Studio when compiler is MSVC.
Removed unused SelectMgr_SOS_Sleeping enumeration value.
SelectMgr_ViewerSelector::Deactivate() now checks activation status before updating tolerance map.
Two classes TDataXtd_Presentation and TPrsStd_AISPresentation were improved so that they accept a list of selection modes.
A new non-regression test is added: caf presentation N1
Font_FontMgr has been redesigned to:
- Store fonts in a map instead a list.
- Allow mapping multiple fonts to a single alias.
- Log informative message about usage of non-requested font (fallback).
- Register all font files within standard folders on Linux when "fonts.dir" is not found.
- Prefer specific alias ("serif") as default fallback font instead of arbitrary one in a system.
A couple of obsolete and broken font aliases have been removed;
instead, new aliases of fonts popular on Linux platform have been added.
Font_NameOfFont.hxx has been extended with more neutral aliases
"monospace", "serif", "sans-serif", "cjk" and "korean".
Font_FontAspect enumeration values have been renamed Font_FA_ -> Font_FontAspect_
with old values preserved as alias.
Font_SystemFont has been extended with a list of paths to Font_FontAspect styles,
so that entire Font Family is now defined within a single Font_SystemFont instance.
Non-resizable fonts are now ignored by Font Manager.
Add new command "vstatprofiler" to manage rendering parameters and print them.
If there are some input parameters - print corresponding statistic counters values,
else - print all performance counters set previously.
Frustum culling is now managed by Graphic3d_RenderingParams::FrustumCullingState flag
and can be switched into Graphic3d_RenderingParams::FrustumCulling_NoUpdate state
useful for debugging the algorithm.
Draw Harness command vrustumculling has been replaced by vrenderparams -frustumCulling.
1) Added method Graphic3d_Camera::FrustumPoints() returning corner points of camera frustum.
2) Refactored methods OpenGl_BVHTreeSelector::isFullOut(...) and OpenGl_BVHTreeSelector::CacheClipPtsProjections()
3) Changed computation algorithm of frustum planes (build them by corner points)
4) Added interactive object AIS_CameraFrustum to draw camera frustum.
5) Extended Draw command "vcamera" with option displaying camera frustum.
When DRAW on Windows is launched with option -c, the command is now properly transferred to Tcl thread (separate thread that runs Tcl interpretor on Windows except when DRAW is run in batch mode) for execution, instead of being evaluated in the main thread.
Execution of DRAW in batch mode (option -b) is fixed by enabling proper initialization of the Tcl interpretor and replacement of backslashes in path to startup script by straight slashes on Windows in that mode.
Declaration of global variables used for communication of console command between threads is moved to Draw_Window.hxx to ensure consistency.
Function wscpy_s is used instead of memcpy to avoid possible buffer overrun.
Implementation of method "IsDeleted(...)" for MakeRevol and MakePrism algorithms.
Problem (exception) occurs during history building and was caused by an attempt to obtain generated shape for subshape, which was really "deleted" by algorithm - this input subshape and its possible generated shape was not used in result.
OpenGl_Context::MakeCurrent()/OpenGl_Context::IsCurrent() have been modified to NOT fail
in case if myWindow is EGL_NO_SURFACE (valid off-screen rendering case within EGL).
OpenGl_GraphicDriver::InitEglContext() now finds EGL surface config in case if it has not been passed by argument.
OpenGl_Window constructor now allows wrapping an off-screen rendering surface EGL_NO_SURFACE.
However, it still creates a dummy surface eglCreatePbufferSurface() to workaround bugs in some GLES drivers (Vivante GC2000).
V3d_View::SetWindow()/V3d_View::MustBeResized()/V3d_Viewer::SetViewOn() have been modified
to avoid implicit View redraw (leading to undefined behavior/crashes in case if rendering
should be done into default FBO defined right after V3d_View initialization).
Two test cases are added. One for Boolean operation (it is good, because BO has been fixed already), and another for fixshape operation (it causes Draw crash).
Methods IsParallel() and SetParallel() have been added to BVH_Sorter, BVH_DistanceField and BVH_BuilderTransient to control parallel execution on low-level.
Fix compilation errors for old compilers without support of c++11 (std::begin, std::end)
Explicit calls to TBB in BVH_DistanceField, BVH_LinearBuilder and BVH_RadixSorter have been replaced.
Task functors have been separated on execution and state parts for the sake of usage by OSD_Parallel.
The following changes have been made for improving the offset algorithm:
1. Multi-connexity support - intersection of the faces connected to the edge belonging to more than two faces is now performed.
2. Avoid intersection of the faces connected through internal edge.
3. Filling gaps (holes) in the splits of the created offset faces to increase possibility of creation of the closed volume from these splits.
Test cases for the issue.
Implementation of the new class *BOPAlgo_MakeConnected* for making the group of touching same-dimensional shapes connected.
Provide the material association for the first sub-elements of the input shapes.
Provide possibility to make the connected shape periodic.
Draw commands for new algorithm:
* makeconnected - make the input shapes connected or glued, performs material associations;
* cmaterialson - returns the materials located on the requested side of a shape;
* cmakeperiodic - makes the connected shape periodic in requested directions;
* crepeatshape - repeats the periodic connected shape in requested directions requested number of times;
* cperiodictwins - returns all periodic twins for the shape;
* cclearrepetitions - clears all previous repetitions of the periodic shape, keeping the shape periodic.
Documentation & test cases for the new algorithm.
Implementation of the new class *BOPAlgo_MakePeriodic* for making the shape periodic in 3D space.
Periodicity of the shape means that the shape can be repeated in any periodic direction any number of times without creation of the new geometry or splits.
The idea of this algorithm is to make the shape look similarly on the opposite sides or on the period bounds of periodic directions.
It does not mean that the opposite sides of the shape will be mirrored. It just means the the opposite sides of the shape should be split by each other and obtain the same geometry on opposite sides.
Such approach will allow repeating the shape, i.e. translating the copy of a shape on the period, without creation of new geometry because there will be no coinciding parts of different dimension.
Draw commands for the new algorithm:
* makeperiodic - makes the shape periodic in required directions;
* repeatshape - repeats the periodic shape in requested periodic direction;
* periodictwins - returns the periodic twins for the shape;
* clearrepetitions - clears all previous repetitions of the periodic shape.
Documentation & test cases for the algorithm.
1. Test case for the issue has been created. The issue is fixed by the patch #26106.
2. Since now, "tricheck" command uses edge tolerance instead of edge deflection to check cross-face-errors (set of nodes of the edge on 1st face must be equal to set of nodes of the same edge on 2nd face).
Now end points of intervals calculated in GeomAdapter_Curve are not replaced by GeomAdapter_Surface.
Left bound of periodic curve less than UMin is not corrupted by LocateParameter during intervals calculating.
Implementation of capturing of output to standard streams in DRAW (see command dlog) is revised to avoid problems with command "test" executing long test scripts:
1. Method OSD_File::Capture() is removed: on Windows it was allocating a C file descriptor for a file opened using WinAPI, and never released that descriptor (once allocated, it cannot be released separately from WinAPI file handle). Direct calls to dup/dup2 are used instead.
2. In Draw_Window.cxx the standard Tcl channels are initialized manually using corrected version of Tcl internal function. This works around a problem with Tcl channels on Windows being bound to OS device handle owned by the system which can get invalidated as result of calls to dup2() (used to capture output to standard streams).
3. Temporary file for capturing is opened once and used to store whole log, thus the need to collect log in the string stream in memory is avoided
4. Possible errors of dup() and dup2() are checked and reported
Test demo draw dlog is added
Off-topic changes:
- Test demo draw getsource is corrected for VS2017 which generates file name in lowercase
- Field myFaceBounds is initialized in constructor of the class BRepAlgo_NormalProjection to avoid undefined behavior
- Test bugs modalg_5 bug24012 is corrected to use command nproject instead of custom one, and to check propertes of the resulting shape
Two new functionalities have been added in BRepFilletAPI_MakeChamfer:
- constant throat (the section of chamfer is isosceles triangle, its height is constant in all sections - this is the "throat" of the weld);
- constant throat with penetration(the section of chamfer is right-angled triangle, the first of two surfaces (where is the top of the chamfer) is virtually moved inside the solid by offset operation, the apex of the section is on the intersection curve between moved surface and second surface, right angle is at the top of the chamfer, the length of the leg from apex to top is constant - this is the "throat" of the weld).
- New abstract classes BlendFunc_GenChamfer and BlendFunc_GenChamfInv have been added;
- Class BlendFunc_Chamfer is now descended from BlendFunc_GenChamfer, class BlendFunc_ChamfInv is now descended from BlendFunc_GenChamfInv.
- New class BlendFunc_ConstThroat is descended from BlendFunc_GenChamfer, new class BlendFund_ConstThroatInv is descended from BlendFunc_GenChamfInv.
- New class BlendFunc_ConstThroatWithPenetration is descended from BlendFunc_GenChamfer, new class BlendFund_ConstThroatWithPenetrationInv is descended from BlendFunc_GenChamfInv.
- Class ChFi3d_ChBuilder has now mode of chamfer that can be ClassicChamfer, ConstThroatChamfer and ConstThroatWithPenetrationChamfer.
- Two new DRAW Test Harness commands "chamf_throat" ant "chamf_throat_with_penetration" have been added for the second mode of ChBuilder.
- The interface of DRAW Test Harness command "chamf" changed for symmetric case.
The reason of this problem is in wrong work of classifier algorithm (see the message ~0080992 to the issue #30354). Therefore, the algorithm of IntTools_FClass2d class has been improved. Namely, now orientation of the polygon is computed from area-criterion instead of angle. As result, some simplification of the method IntTools_FClass2d::Init(...) has been made.
<!break>
1. New constructor has been added to the class CSLib_Class2d. It allows applying TColgp_SequenceOfPnt2d.
2. DRAW-commands "addpolygonnode" and "polygonprops" have been created. They are covered by the test case "tests/geometry/2dpolygon/A1".
3. New method Poly::PolygonProperties(...) has been created. See help for detailed information.
4. New testgrid "lowalgos classifier" has been created.
Split seam edge of a cone according to the specified parameters.
Seam edge functor is called before check of triangulation consistency in order to keep face status consistent.
The cause is that split seam edge can set Outdated flag when Reused is set by another functor without reset which can lead to refusement of such face by triangulation procedure.
Fix compilation error on VS2008
The misprint leading to the problems was fixed.
"Draw" tests were created to check the fix.
"Draw" test bugs/modalg_7/bug28784 was corrected
for right changes of the results.
1. "Descendants" closure mode is not taken into account by TDF closure tool
2. CopyTool always creates a new label at the target ignoring the relocation table
3. XCAFDoc_GraphNode throws an exception if Child of Father is not found during the unsetting
In fact, the DRAW-command "mesh" is duplicate of "incmesh".
The difference is that the "mesh" creates DRAW-object MeshTest_DrawableMesh. However, this object is currently not applicable (e.g. we cannot display it).
DRAW-commands "mesh", "addshape", "smooth", "edges", "vertices", "medge", "mvertex", "triangle", "dumpvertex", "dumpedge", "dumptriangle" and "onetriangulation" have been removed.
The class MeshTest_DrawableMesh has been removed as useless.
Testgrids "mesh standard_mesh" and "mesh advanced_mesh" have been removed.
Before the fix, angular deflection was scaled to some coefficient in the method BRepMesh_Deflection::ComputeDeflection(...). Now, angular deflection is constant.
1. Check whether the mesh satisfies the required angular deflection has been amended. Namely normals (to the surface) in the ends of any not "frontier" link are made collinear (with the given angular tolerance).
2. New parameters AngleInterior and DeflectionInterior have been added in IMeshTools_Parameters structure.
3. In case of thin long faces with internal edges, add points of internal edges to control parameters using grabParamsOfInternalEdges() in order to avoid aberrations on its ends. Disable addition of parameters from boundary edges in case of BSpline surface. Deviation can be controlled through the deflection parameter.
4. Grab parameters from edges in case if there is just a single interval on BSpline surface along U and V direction.
Removed tight connections between data structures, auxiliary tools and algorithms in order to create extensible solution, easy for maintenance and improvements;
Code is separated on several functional units responsible for specific operation for the sake of simplification of debugging and readability;
Introduced new data structures enabling possibility to manipulate discrete model of particular entity (edge, wire, face) in order to perform computations locally instead of processing an entire model.
The workflow of updated component can be divided on six parts:
* Creation of model data structure: source TopoDS_Shape passed to algorithm is analyzed and exploded on faces and edges. For each topological entity corresponding reflection is created in data model. Note that underlying algorithms use data model as input and access it via common interface which allows user to create custom data model with necessary dependencies between particular entities;
* Discretize edges 3D & 2D curves: 3D curve as well as associated set of 2D curves of each model edge is discretized in order to create coherent skeleton used as a base in faces meshing process. In case if some edge of source shape already contains polygonal data which suites specified parameters, it is extracted from shape and stored to the model as is. Each edge is processed separately, adjacency is not taken into account;
* Heal discrete model: source TopoDS_Shape can contain problems, such as open-wire or self-intersections, introduced during design, exchange or modification of model. In addition, some problems like self-intersections can be introduced by roughly discretized edges. This stage is responsible for analysis of discrete model in order to detect and repair faced problems or refuse model’s part for further processing in case if problem cannot be solved;
* Preprocess discrete model: defines actions specific for implemented approach to be performed before meshing of faces. By default, iterates over model faces and checks consistency of existing triangulations. Cleans topological faces and its adjacent edges from polygonal data in case of inconsistency or marks face of discrete model as not required for computation;
* Discretize faces: represents core part performing mesh generation for particular face based on 2D discrete data related to processing face. Caches polygonal data associated with face’s edges in data model for further processing and stores generated mesh to TopoDS_Face;
* Postprocess discrete model: defines actions specific for implemented approach to be performed after meshing of faces. By default, stores polygonal data obtained on previous stage to TopoDS_Edge objects of source model.
Component is now spread over IMeshData, IMeshTools, BRepMeshData and BRepMesh units.
<!break>
1. Extend "tricheck" DRAW-command in order to find degenerated triangles.
2. Class BRepMesh_FastDiscret::Parameters has been declared as deprecated.
3. NURBS range splitter: do not split intervals without necessity. Intervals are split only in case if it is impossible to compute normals directly on intervals.
4. Default value of IMeshTools_Parameters::MinSize has been changed. New value is equal to 0.1*Deflection.
5. Correction of test scripts:
1) perf mesh bug27119: requested deflection is increased from 1e-6 to 1e-5 to keep reasonable performance (but still reproducing original issue)
2) bugs mesh bug26692_1, 2: make snapshot of triangulation instead of wireframe (irrelevant)
Correction in upgrade guide.
Boolean Operations - Force the face with internal edges to be treated by the BuilderFace algorithm. It is needed for the cases when internal edges of the face go from side to side and should really split the face.
Test case for the issue.
If 2D-curves are requested in intersection result but they cannot be created (by some reason) then the full set of curves (3D and two 2D) is rejected from the intersection result.
1. New testgrid "lowalgos/intss" has been created. It will contain all test cases on geometrical intersection of two surfaces ("intersect" DRAW-command) and two faces ("bopcurves" DRAW-command).
2. New test case for the issue #28493 has been created because the problem is not reproduced on MASTER.
3. Test case (lowalgos/intss/bug24472) for the issue #29501 has been modified in order to check loops of the resulting intersection curves.
Test cases have been created. No fix is needed because the described problem has not to be cause of failing of high-level OCCT-algorithms.
See the message ~80300 (issue #26509) for detail information.
Bnd_Box now keeps calculating of finite part of bounding box after specifying it to be Open in some direction.
The finite part can be retrieved using new method Bnd_Box::FinitePart().
Prs3d::GetDeflection() now uses Bnd_Box::FinitePart() when applying relative deflection.
Draw Harness command bounding has been extended with option -finite returing a finite part of AABB.
Special treatment of bspline curve of first degree is implemented in Extrema_GExtPC.gxx
Test case is added
Some test cases are modified according to actual state of algorithm
Add option "-exclude" to testgrid command.
It's possible to exclude groups, grids and test cases from test execution.
Add possibility to exclude subgroups from executed groups and test cases from executed subgroups\groups.
Graphic3d_Buffer can be now optionally initialized as non-interleaved array of vertex attributes
and provides an interface to invalidate buffer sub-range tracked by OpenGl_PrimitiveArray.
Replacing usage of BRepAlgo_DSAccess with usage of modern Boolean operations algorithms in BRepFill_Draft.
Removing BRepAlgo_BooleanOperations and BRepAlgo_DSAccess classes.
The following classes have been removed as unused:
* BRepAlgo_DataMapOfShapeBoolean
* BRepAlgo_DataMapOfShapeInterference
* BRepAlgo_EdgeConnector
* BRepAlgo_SequenceOfSequenceOfInteger
SelectMgr_SelectableObject now assigns transformation to mySelectionPrs and myHilightPrs presentations.
Removed confusing method PrsMgr_PresentableObject::UpdateTransformation() with presentation as argument.
Added new option -glslcode for vgldebug command with the following values:
- off disables glsl source code dump;
- short to dump glsl source code in short format (except common declarations);
- full to dump glsl source code in full format.
When refining the mesh to achieve required deflection (IntPolyh_Triangle::MultipleMiddleRefinement) limit the number of new triangles to avoid infinite loop.
Test case for the issue.
Implemented workaround for dump images with width >= 5462 pix on Intel OpenGl driver.
Changes according to OpenGl_Context::myVendor field in lowercase.
Now roots-references in XCAF Document can be exported to STEP without losing structure of assembly/sharing/metadata of this root. New auxiliary root assembly with each root-reference is created to save all necessary data.
Provide possibility to perform Boolean operations on open solids.
Implementation of the new method *BOPAlgo_Builder::BuildBOP* performing the construction of the result shape for the given type of Boolean operation.
This approach does not rely on the splits of solid to be correct and looks for the faces with necessary state relatively opposite solids to build the result solid.
The call to this method is performed from BOP algorithm in case there were open solids in the arguments.
Implementation of the draw command *buildbop* performing a call to the method above.
Do not allow the precision with which the valid range is found to be less than the epsilon of the max parameter of the edge's range.
Test cases for the issue.
AIS_InteractiveObject::SynchronizeAspects() now propagates event
to OpenGl_Group::SynchronizeAspects() which properly handles
all aspects defined within the group.
Add history for subshapes of spine: edges and vertices. Each edge of spine generates a shell. Each vertex of spine generates a set of edges and, possibly, faces (in the case of Round Corner).
StdPrs_ToolRFace no skips curves with NULL curves.
Code has been cleaned up from duplicated checks, redundant casts
and dummy Adaptor2d_Curve2dPtr typedef.
StdSelect_BRepSelectionTool::GetSensitiveForFace() now catches
Standard_NullObject exception to skip invalid Edges.
Added new type of graphic3d limit Graphic3d_TypeOfLimit_HasFlatShading.
Added workaround for unexpected behaviour of mobile devices with Adreno GPU.
Added new complex flag hasFlatShading to OpenGl_Context for indicating support of flat shading.
The following improvements have been made in Boolean operations algorithm in order to fix the problem:
1. Initialization of the pave blocks which vertices have acquired the SD ones.
2. Removing from Data Structure the small edges having the same vertices on both ends (either initially or acquired).
3. Avoid adding empty SD connections when one vertex points to itself.
Test case for the issue.
Correct launching custom.sh in env.sh.in.
Correct definition of FFMPEG_DIR in custom.build.bat/sh.in and generation of custom_*.sh (variables CSF_OCCTBinPath and CSF_OCCTLibPath are not empty on the first run)
Names of Windows header files are corrected to be lower case to match their actual names on Windows (except Windows.h) and MinGW.
Files COMMANDWINDOW.h and MAINWINDOW.h in Draw package are renamed to CamelCase according to names of corresponding CXX files.
BRepLib::ExtendFace method when working with analytical and periodic faces now takes into account the possible closeness of the result face.
Test cases for the issue.
1. The condition of WLine breaking (in IntWalk_IWalking algorithm) has become more independent of the input tolerance.
2. Currently the algorithm of IntPatch_Points of WLine processing depends on the algorithm of obtaining the WLine.
3. The methods IntSurf_LineOn2S::Add(...) and IntSurf_LineOn2S::SetUV(...) have become not inline (see the message ~0077431 in the issue #29866).
Detection of specific case of error in the definition of transformation matrix describing position of the component within assembly, when it has Axis Placements swapped, is corrected to handle the case when one of these Axis Placements is contained in both Shape Representations (of the assembly and its component).
This allows the problematic STEP file to be translated correctly.
Added test bugs step bug30087
BOPTools_AlgoTools::ComputeState - increase the chance of correct classification of the face relatively solid by classifying the point located inside that face instead of the point taken near the edge of that face.
Test case for the issue.
Select3D_SensitiveSet::Matches() has been improved to check if BVH node is fully included by selection volume
and pass this information to overlapsElement()/elementIsInside() interfaces
to avoid expensive partial overlapping checks for individual elements.
Select3D_SensitivePrimitiveArray implements this new interface to improve partial overlapping performance.
Select3D_SensitivePrimitiveArray::Matches() now handles rectangle selection for sub-elements when Elements map is defined.
Added missing const to SelectMgr_BaseFrustum::Overlaps() methods.
AIS_PointCloud has been extended with new selection mode for collecting selected nodes
Draw Harness command vdrawparray has been extended with an option -shape
allowing to create a triangulation from tessellated shape.
On Apple with XCode 9.4.1 and onwards, the compiler optimization is disabled for method gp_Mat::Transpose() as optimizer generates invalid code when that method is used.
Some refactoring of the code affected by this bug.
Extended SelectBasics_PickResult structure by myObjPickedPnt field, which contained the value of the 3d point on the selected object.
Changed all Overlaps methods. Parameter theDepth replaced on object of the structure SelectBasics_PickResult. This approach will be able to add new fields to SelectBasics_PickResult structure without big changes in modules which contained Overlaps method.
Methods GeomConvert::ConcatG1, GeomConvert::ConcatC1, Geom2dConvert::ConcatG1, Geom2dConvert::ConcatC1 are corrected to prevent exceeding maximum degree of BSpline curve in case of closed contour.
A boolean parameter is added with a default value to the method TPrsStd_AISPresentation::SetSelectionMode().
By default, behavior of the method is not changed. But if the parameter's value is changed to FALSE, change of selection mode for the object will not influence the transaction. In other words, the attribute will not set modified and will not put into transaction delta. This behavior suits the usual notion of the object in an application: the object has color, material, transparency, but selection mode is not a property of the object. It is a property of an operation involving the object. Therefore, it has sense to remove the method SetSelectionMode() from TPrsStd_AISPresentation. For a while, it is marked as obsolete.
In the fragment
if (aNewPnt[0] < u1min || aNewPnt[0] > u1max ||
aNewPnt[1] < v1min || aNewPnt[1] > v1max ||
aNewPnt[2] < u2min || aNewPnt[2] > u2max ||
aNewPnt[3] < v2min || aNewPnt[3] > v2max)
{
break; // Out of borders, handle this later.
}
the parameters are compared with values obtained by IntPolyh_MaillageAffinage algorithm and do not have any relations to the domain boundaries.
This misprint has been eliminated.
Before the fix, small arc of circle returned as intersection result. It was connected with the fact that the circle always was limited in 0 and 2*PI points. Currently the bounds of the circle (retrieved as an intersection curve) are computed from the real domain of the arguments (there is not a snap to the fixed 0 and 2*PI points).
Test for handles is updated according to behavior expected from modern compilers.
Performance test perf ncollection A1 is updated (valid range extended) according to measurements made on Debian 8 with GCC 4.9 and on Ubuntu 16.04 with CLang 3.8 and 6.0 and GCC 4.7, 4.9, and 5.0.
Bnd_OBB constructor from Bnd_Box is fixed for VOID input.
DRAW command bounding command is adjusted:
- Support of old syntax 'bounding shape'is restored.
- Exception on VOID bounding box is avoided.
- Arguments are handled in case-insensitive way.
- Options -s, -perfmeter, and ability to set result as first argument are removed as redundant.
- Help is moved to command definition.
Separate command OCC29311 is added (in QADraw) to do the same as option -perfmeter did; tests are updated.
The following improvements have been introduced in Boolean Operations algorithm s:
1. UBTree is replaced with EBTree in Boolean operations to be able to add/remove elements into the tree of bounding boxes.
2. Repeated (nested) intersection of sub-shapes is performed with argument vertices whose tolerances increased during the operation.
3. The algorithms of Edge/Edge and Edge/Face intersection have been improved for the cases when the intersection point is located close to the edge boundaries .
4. New procedure has been implemented to ensure forced creation of Edge/Face common blocks in cases when the edge is really close to the face.
5. Post-processing of Face/Face intersection results has been improved.
6. Extension of the planar faces for Plane/Plane intersection is avoided.
7. Builder Face now better classifies potentially internal edges relatively to new faces with filtering by bounding boxes.
Side effect changes:
1. IntTools_ShrunkRange now keeps the length of the valid range of the edge.
2. The method BOPDS_DS::UpdateEdgeTolerance() has been removed as unused (replaced by the BOPAlgo_PaveFiller::UpdateEdgeTolerance()).
Test case for the issue 0029900.
Test case for the issue 0029711.
Adjustments of the existing test cases.
Avoid using uninitialized variables.
In implementation of DRAW command "mkplane", the case labels of the switch statement are corrected to use values of the enum matching the type of the switch argument.
NCollecton_DoubleMap has been extended with two new methods Find1() and Find2()
performing lookup and returning found key withoout throwing an exception
(similar to NCollection_DataMap::Find()).
ViewerTest_DoubleMapOfInteractiveAndName has been redeclared to define
Handle(AIS_InteractiveObject) key instead of Handle(Standard_Transient).
vtriangle and vsegment commands now share presentation created by vdrawparray command.
Added missing accessors to Graphic3d_ArrayOfPrimitives properties
VertexNumberAllocated(), EdgeNumberAllocated() and BoundNumberAllocated().
Fast rebuilding of the face is now avoided if its edges have been unified during the intersection.
Full rebuilding of the face is required in this case (using BuilderFace algorithm).
IntTools_Context::IsInfiniteFace() - Make the conclusion about infinity of the face basing on its bounding box only. No need to check NaturalRestriction flag.
Various performance improvements in STEP read/write algorithms:
- Search for the label of a shape or component shape is improved using map mechanism instead of brute force iteration.
- Invariant FindEntities() is moved out of the loop in the method getStyledItem in STEPCAFControl/STEPCAFControl_Writer.cxx.
- A pointer to the end of binders chain is added in Transfer_Binder class to speed up adding a binder to the chain.
- Small fixes are added to eliminate excess copying of handles, calls of handle DownCasts and so on.
Stack overflow is removed during destruction of STEP model with long chains of Transfer_Binder.
It is possible to use the Draw commands ReadStep and WriteStep to read/write from the session without accessing the disk file (use '.' for the file name).
Performance test cases for STEP reading/writing have been added.
Implementation of classes BSplCLib_Cache and BSplSLib_Cache is revised:
- Common functionality dealing with spans along one parametric direction is separated to new struct BSplCLib_CacheParams
- Empty constructors are removed; copying is prohibited
- Code reconsidering degree and other parameters on each call to BuildCache() is eliminated; curve parameters must be the same in constructor and all calls to BuildCache()
- Extra call to BuildCache() from constructor is eliminated
Font_BRepFont now uses a dedicated algorithm for text-to-BRep transformation instead of relying on ShapeFix.
It orders wires based on wire classification, analyzes internal zones and creates a few faces (if needed).
TKService dependency from TKShHealing has been dropped.
The algorithm has been corrected to avoid exception.
Now both Draw commands "brollingball" and "rollingball" give correct results.
Test cases "offset faces_type_a A3, A4, A9" are good now.
Name of the PROPERTY_DEFINITION entity for validation properties is changed to "geometric validation property" according to the current version of the recommended practices.
For compatibility with older definitions (with underscores) the check is made using string with underscores replaced by spaces.
Output information about failure during reading of external file has been added (it can be checked by using Draw command "tpstat").
If the reader cannot find an external file at the first found location
(using entity APPLIED_EXTERNAL_IDENTIFICATION_ASSIGNMENT) it checks other information about file path
(using entity DOCUMENT_FILE related with the above-mentioned entity) and uses a new location for import of the external file.
New class OSD_ThreadPool has been introduced to define a Thread Pool for multi-threading algorithm.
Thread Pool assigns a serial number for each thread allowing Multi-Threading algorithm to allocate thread-local storage variables as an array whose size is the same as the number of threads.
OSD_ThreadPool also redirects exceptions to a thread calling parallel execution and consistently initializes FPE exception handling.
New class Standard_Condition provides a platform-independent tool similar to Event in WinAPI.
A new auxiliary function Standard_Atomic_CompareAndSwap() has been introduced
for performing atomic compare and swap of integer number.
Standard_Atomic_Increment/Standard_Atomic_Decrement fallback implementation
using ASM code for x86 processors for GCC has been dropped;
instead, it is expected that GCC should be properly configured targeting modern x86 architectures.
OSD_Signal now declares fFltExceptions as thread_local variable accessible through OSD::ToCatchFloatingSignals() property.
Standard_THREADLOCAL macro (wrapping thread_local attribute) has been moved to public header Standard_Macro.hxx.
OSD_Parallel::ForEach() has been extended with new optional parameter theNbItems and uses OSD_ThreadPool::DefaultPool().
AIS_InteractiveContext::MoveTo() and ::Select() now do not highlight entities
when ::AutomaticHilight() option is turned off.
vselprops command has been extended with new arguments -autoHighlight and -highlightSelected
(vhighlightselected command is now deprecated).
Now vertex loops on surfaces of revolution are imported from STEP format in the same way as on spheres and b-spline surfaces: in case of a single vertex loop neutral bounds will be used.
Methods SelectMgr_SelectableObject::GetHilightPresentation(), ::GetSelectPresentation() and ::UpdateClipping()
have been corrected to copy clipping planes of the object into highlighting and selection presentations.
Since now a new WLine is not created if its start point lies in another earlier computed WLine. It allows avoiding creation of duplicate WLines in the intersection result.
<!break>
1. Methods IsOutSurf1Box(...), IsOutSurf2Box(...), IsOutBox(...) for classes IntSurf_LineOn2S and IntPatch_RLine have been created.
The "end" file from folder "de" has been corrected to exclude comparing the number of edges in the shapes from current result with test data.
"TODO messages generated for differences were removed.
Data exchange tests were regenerated.
The problem when object transformation was applied to the bounding box twice has been fixed in OpenGl_Structure.
OpenGl_ShaderManager has been extended with GLSL program drawing bounding box.
OpenGl_Structure does not create extra Groups for drawing bounding box.
Interface OpenGl_RenderFilter and its implementations have been merged into OpenGl_Workspace.
This limits flexibility of interface, but simplifies logic.
- OpenGl_RenderFilter_FillModeOnly flag replaces OpenGl_CappingAlgoFilter;
- OpenGl_RenderFilter_NonRaytraceableOnly flag replaces OpenGl_RaytraceFilter;
- OpenGl_RenderFilter_OpaqueOnly flag replaces OpenGl_OpaqueFilter;
- OpenGl_RenderFilter_TransparentOnly flag replaces OpenGl_TransparentFilter.
OpenGl_CappingAlgo now:
- avoids redundant Stencil clearing when Layer contains at least one transparent Element;
- renders semitransparent capping plane within transparent elements pass.
vclipplane command has been extended by new argument -transparency.
The code has been fixed to avoid taking a reference from a field of a temporary variable
(BRepApprox_ApproxLine::Point() returns non-reference structure).
readstl syntax has been modified, so that it creates a single-face triangulation by default.
The argument "trinagulation" is no more supported.
The new argument "-brep" has been introduced to generate
a compound of per-triangle faces instead (old default behavior of the command).
The algorithm has been improved for the cases when the intersection line goes through the cone apex.
<!break>
1. All special points are put to the ALine forcefully (if they are true intersection point). Currently this step has not been implemented yet.
2. Now the tolerance of IntPatch_Point (put into ALine) is computed in order to cover the distance between it and the correspond ALine.
3. Test cases have been created.
4. Procedure of trimming IntAna_Curve has been improved.
5. Criterion when the discriminant of IntAna_Curve can be considered to be equal to 0 has been improved.
6. Methods IntAna_Curve::FindParameter(...) (and IntPatch_ALine::FindParameter(...)) currently returns list of all parameters corresponding the given point (IntAna_Curve can be self-interfered curve). Before the fix, this method always returned only one (randomly chosen) parameter.
7. Interfaces of the following methods have been changed: IntAna_Curve::FindParameter(...), IntPatch_ALine::FindParameter(...), IntPatch_ALine::ChangeVertex(...), IntPatch_SpecialPoints::AddPointOnUorVIso(...), IntPatch_SpecialPoints::AddSingularPole(...), IntPatch_WLineTool::ExtendTwoWLines().
8. Following methods have been added: IntAna_Quadric::SpecialPoints(...), IntPatch_ALineToWLine::GetSectionRadius(...), IntPatch_SpecialPoints::ProcessSphere(...), IntPatch_SpecialPoints::ProcessCone(...), IntPatch_SpecialPoints::GetTangentToIntLineForCone(...).
------------------
1) tests/boolean/volumemaker/C5
tests/boolean/volumemaker/C6
tests/boolean/volumemaker/E7
They are real IMPROVEMENTS. In the FIX (in compare with MASTER), section result between pairs of faces f2&f6 (C5), f3&f7 (C6) and f1&f5 (E7) is closed. Separated test cases have been created in order to focus on the problem with section. Bug #28503 has been fixed.
Correction in test cases.
Argument of LowerCase() and UpperCase() is cast to int via unsigned char to avoid passing negative integer in the case if the argument char is in the extended part of ASCII table (which would result in undefined behavior according to C++ standard).
Check for double solutions has been added in algorithms Extrema_ELPCOfLocateExtPC and Extrema_ELPCOfLocateExtPC2d
that find extrema between point and curve.
XCAFDoc_ShapeTool::Expand method now avoids creation of subshapes under assemblies.
XCAFDoc_ShapeTool::AddSubShape now avoids creation subshapes under all labels except for simple parts.
Use of input tolerance has been eliminated in Geom2dInt_GInter algorithm called from BRepClass_Intersector::Perform(...) method.
Now the input tolerance is used only for checking ON-status of classification.
AIS_Manipulator::ObjectTransformation() has ceased to use undefined point for myStartPick initialization within AIS_MM_Translation mode.
AIS_MM_Scaling cas has been merged into AIS_MM_Translation to reduce code duplication.
Extrema_ExtElC/IntAna_IntConicQuad are now used directly instead of more generic GeomAPI_ExtremaCurveCurve/GeomAPI_IntCS.
Deprecated functionality related to Local Context has been removed.
The following classes and types have been removed:
AIS_LocalContext, AIS_DataMapOfILC, AIS_LocalStatus, SelectMgr_SequenceOfSelector, SelectMgr_DataMapOfObjectSelectors.
Value AIS_DS_Temporary has been removed from enumeration AIS_DisplayStatus.
The following methods related to Local Context have been removed from AIS_InteractiveContext:
::HasOpenedContext(), ::HighestIndex(), ::LocalContext(),
::LocalSelector(), ::OpenLocalContext(), ::CloseLocalContext(),
::IndexOfCurrentLocal(), ::CloseAllContexts(), ::ResetOriginalState(),
::ClearLocalContext(), ::UseDisplayedObjects(), ::NotUseDisplayedObjects(),
::SetShapeDecomposition(), ::SetTemporaryAttributes(),
::ActivateStandardMode(), ::DeactivateStandardMode(), ::KeepTemporary(),
::SubIntensityOn(), ::SubIntensityOff(),
::ActivatedStandardModes(), ::IsInLocal()
and ::AddOrRemoveSelected() taking TopoDS_Shape.
The methods AIS_InteractiveContext::Display() and ::Load()
taking argument theToAllowDecomposition have been marked as deprecated;
the new methods without this unused argument should be used instead.
AIS_InteractiveContext::DisplayedObjects() - removed argument theOnlyFromNeutral.
SelectMgr_SelectionManager has been simplified so that now it support only single Viewer Selector.
The argument -local has been removed from Draw Harness commands vselmode, vremove, vdisplay, vdisplayall, verase.
New algorithms calculating global properties on mesh data have been added:
- BRepGProp_MeshCinert computes the global properties of polylines represented by a set of points;
- BRepGProp_MeshProps computes the global properties of a surface mesh.
Existing tool BRepGProp now automatically uses new algorithm for triangulation-only faces.
By default, algorithm will use exact geometry objects (surfaces), when it is available (as before the patch);
this behavior can be switched by a new flag UseTriangulation, forcing usage of triangulation instead of exact geometry when both defined.
Image_Color - removed semibroken summ/difference operators.
Image_Diff now uses signed integer for computing differnce between ubyte3 components;
properly compare squared tolerance.
Image_Diff - dropped declaration of Image_ColorXXX24.
RGB color difference is now computed using Chebyshev distance instead of Euclidean distance
Image_PixMap - added methods RawValue()/ChangeRawValue() returning a pointer
to image where specified pixel data is defined.
Methods of NCollection_Vec3 and NCollection_Vec3 that returned reference to internal buffer as vector of lower dimension (non-const xy(), xyz() etc.) are eliminated.
Use of these methods could led to generation of incorrect binary code by GCC.
Instead added new method SetValues() accepting vector of lower dimension and additional value.
DRAW test command QANColTestVec4 reproducing one situation where the bug occurs is added, along with a test case.
TopoDS_Shape::NbChildren() - new method has been introduced.
TopoDS_Shape.lxx, TopoDS_TShape.lxx - inline methods have been moved to hxx.
TopoDS_TShape_Flags has been redeclared as enumeration instead of preprocessor macros.
Cyclic dependency between headers TopoDS_Shape.hxx and TopoDS_TShape.hxx eliminated.
Places where TopoDS_Iterator is used only for calculation of number of sub-shapes are updated to use NbChildren() instead
Modified method: Approx_ComputeLine::Compute
Check of multicurve is now always unconditional, the procedure of check is modified to avoid infinite loops.
Modified classes: GeomLib_CheckBSplineCurve and GeomLib_Check2dBSplineCurve
Correction of poles at the ends of curve is modified to fit the direction of tangent defined by two first points or two last points of walking line.
Also modified:
BOPAlgo_PaveFiller: modified methods PostTreatFF, RemoveUsedVertices - now unused vertices are included in the list of vertices to be absorbed by other ones.
Improve FindSubShape method in ShapeTool.
Update Expand compounds, according to changes in ShapeTool.
Add FindSubShape and AddSubShape commands for Draw.
Implementation of the method for simplification of the result of Boolean Operation on the API level.
The method BRepAlgoAPI_BuilderAlgo::SimplifyResult has been added, so the derived classes such as BooleanOpeation and Splitter can also use this method.
The result shape simplification should be called after the operation is done. The simplification is performed by the means of ShapeUpgrade_UnifySameDomain algorithm.
Draw command "bsimplify" has been added to control the simplification options.
Documentation for new functionality and draw commands controlling the options of Boolean operations.
Test cases for the new functionality.
Side-effect change:
The algorithms in Boolean component have been changed to use the BRepTools_History as a History tool.
Now it became possible to disable the collection of shapes modifications during Boolean Operations, which may be useful for performance sake (in draw the option is controlled by *setfillhistory* command).
Draw command "unifysamedom" has been changed to accept the angular tolerance in degrees instead of radians.
The reason of this issue is in incorrectly interpreted the generic class "TheExtPC" (eliminated after the fix 0024773) as Extrema_ExtPC2d. Correct interpretation must be "Extrema_EPCOfExtPC2d" class.
Now this problem has been fixed in the class Geom2dGcc_Circ2dTanCenGeo.
New testgrid "lowalgos 2dgcc" has been created.
1. Extrema algorithm calls Curve-surface intersector. This intersector returns flag about infinite solution (in spite of extrema's returning not-parallel status correctly - axes of considered cylinder and circle are not parallel). In this case, attempt of obtaining number of intersection points leads to exception.
So, the fix adds check of infinite solution after the intersection algorithm.
2. The methods IsDone(), IsParallel(), NbExt(), SquareDistance() and Points() (of Extrema_* classes) have been corrected to make them consistent to the documentation.
3. Revision of some Extrema_* classes has been made in order to avoid places with uninitialized variables.
4. Currently Extrema does not store any points in case when the arguments are parallel. It stores the distance only.
5. Some cases on Extrema-algo have been moved from "fclasses"-group to "modalg"-group.
Make the methods Append, Prepend, InsertBefore and InsertAfter, which take another sequence as an argument, copying the sequence instead of joining if the allocators are different.
Add test cases for collection classes.
1. Sphere is not V-periodic surface. See the section "Periodicity concept" written in frame of the fix 0029115. Therefore, applying ElCLib::AdjustPeriodic(...) method is unreasonable.
2. Default U work range of sphere is 0 ... 2*PI. Therefore, adjusting U parameter in range -PI/2 ... +PI/2 is unreasonable.
Substitution of degenerated edges by their copies is implemented to make edges not shared
in BRepPrimAPI_MakeRevol::Build().
Method BRepPrimAPI_MakeRevol::Generated(...) is changed in order to use history of substitutions.
Test case is added.
The Message_ProgressIndicator handle is added as a parameter to the function LU_Decompose and the the member functions of math_Gauss, Plate_Plate and GeomPlate_BuildPlateSurface classes.
Graphic3d_ClipPlane now can define a Chain of Planes (logical AND).
OpenGl_ShaderManager - added new GLSL sub-programs for clipping plane chains.
Bnd_Range::TrimFrom(), ::TrimTo() - added auxiliary methods for trimming the range.
SelectMgr_ViewClipRange now handles non-continuous clipping range.
Graphic3d_SequenceOfHClipPlane now aggregates NCollection_Sequence instead of inheritance.
OpenGl_CappingPlaneResource - triangulation has been adjusted to make front face following CCW order.
message(STATUS"Warning. ${OCCT_HEADER_FILE_OLD} is presented in the sources but not involved in FILES and will be removed from ${ROOT_TARGET_OCCT_DIR}/inc")
message(STATUS"Warning. ${OCCT_HEADER_FILE_OLD} is present in the sources but not involved in FILES and will be removed from ${ROOT_TARGET_OCCT_DIR}/inc")
@@ -8,7 +8,7 @@ This document presents guidelines for building third-party products used by Open
You need to use the same version of MS Visual Studio for building all third-party products and OCCT itself, in order to receive a consistent set of run-time binaries.
The links for downloading the third-party products are available on the web site of OPEN CASCADE SAS at http://www.opencascade.com/content/3rd-party-components.
The links for downloading the third-party products are available on the web site of OPEN CASCADE SAS at https://www.opencascade.com/content/3rd-party-components.
There are two types of third-party products used by OCCT:
@@ -17,7 +17,6 @@ There are two types of third-party products used by OCCT:
* FreeType 2.4.10 -- 2.5.3.
* Optional products:
* TBB 3.x -- 4.x;
* gl2ps 1.3.5 -- 1.3.8;
* FreeImage 3.14.1 -- 3.16.0;
* VTK 6.1.0.
@@ -33,7 +32,7 @@ Tcl/Tk is required for DRAW test harness.
@subsubsection dev_guides__building_3rdparty_win_2_1_1 Installation from sources: Tcl
Download the necessary archive from http://www.tcl.tk/software/tcltk/download.html and unpack it.
Download the necessary archive from https://www.tcl.tk/software/tcltk/download.html and unpack it.
1. In the *win* sub-directory, edit file *buildall.vc.bat*:
@@ -86,7 +85,7 @@ Download the necessary archive from http://www.tcl.tk/software/tcltk/download.ht
@subsubsection dev_guides__building_3rdparty_win_2_1_2 Installation from sources: Tk
Download the necessary archive from http://www.tcl.tk/software/tcltk/download.html and unpack it.
Download the necessary archive from https://www.tcl.tk/software/tcltk/download.html and unpack it.
Apply the same steps as described for building Tcl above, with the same INSTALLDIR.
Note that Tk produces its own executable, called *wish*.
@@ -95,7 +94,7 @@ You might need to edit default value of *TCLDIR* variable defined in *buildall.v
Attention: If Cygwin was installed on your computer, make sure that there is no path to it in the *PATH* variable to avoid possible conflicts during the configuration.
4. Launch CMake <i>(cmake-gui.exe)</i> using the Program menu.
In CMake:
* Define where the source code is.
This path must point to *gl2ps* folder.
* Define where to build the binaries.
This path must point to the folder where generated gl2ps project binaries will be placed
(for example, *gl2ps\\bin*).
Further in this document, this folder is referred to as *gl2ps_bin*.
* Expand the CMAKE group and define *CMAKE_INSTALL_PREFIX* which is the path where you want to install the build results, for example, *c:\\occ3rdparty\\gl2ps-1.3.5*.
@@ -6,7 +6,7 @@ The list of required libraries depends on what OCCT modules will be used, and yo
The typical minimum is **Freetype** (necessary for Visualization) and **Tcl/Tk** (for DRAW Test Harness).
See "Third-party libraries" section in \ref OCCT_OVW_SECTION_5 "Overview" for a full list.
On Windows, the easiest way to install third-party libraries is to download archive with pre-built binaries from http://www.opencascade.com/content/3rd-party-components.
On Windows, the easiest way to install third-party libraries is to download archive with pre-built binaries from https://www.opencascade.com/content/3rd-party-components.
On Linux and OS X, it is recommended to use the version installed in the system natively.
You can also build third-party libraries from their sources:
@@ -82,14 +82,13 @@ The following table gives the full list of environment variables used at the con
|----------|------|---------|
| CMAKE_BUILD_TYPE | String | Specifies the build type on single-configuration generators (such as make). Possible values are Debug, Release and RelWithDebInfo |
| USE_FREEIMAGE | Boolean flag | Indicates whether FreeImage product should be used in OCCT visualization module for support of popular graphics image formats (PNG, BMP, etc.) |
| USE_GL2PS | Boolean flag | Indicates whether GL2PS product should be used in OCCT visualization module for support of vector image formats (PS, EPS, etc.) |
| USE_RAPIDJSON | Boolean flag | Indicates whether RapidJSON product should be used in OCCT Data Exchange module for support of glTF mesh file format |
| USE_TBB | Boolean flag | Indicates whether TBB 3rd party is used or not. TBB stands for Threading Building Blocks, the technology of Intel Corp, which comes with different mechanisms and patterns for injecting parallelism into your application. OCCT remains parallel even without TBB product |
| USE_VTK | Boolean flag | Indicates whether VTK 3rd party is used or not. VTK stands for Visualization ToolKit, the technology of Kitware Inc intended for general-purpose scientific visualization. OCCT comes with a bridge between CAD data representation and VTK by means of its dedicated VIS component (VTK Integration Services). You may skip this 3rd party unless you are planning to use VTK visualization for OCCT geometry. See the official documentation @ref occt_user_guides__vis for the details on VIS |
| 3RDPARTY_DIR | Path | Defines the root directory where all required 3rd party products will be searched. Once you define this path it is very convenient to click "Configure" button in order to let CMake automatically detect all necessary products|
| 3RDPARTY_FREETYPE_* | Path | Path to Freetype binaries |
| 3RDPARTY_FREEIMAGE* | Path | Path to Freeimage binaries |
| 3RDPARTY_GL2PS_* | Path | Path to GL2PS binaries |
| 3RDPARTY_TBB* | Path | Path to TBB binaries |
| 3RDPARTY_VTK_* | Path | Path to VTK binaries |
| BUILD_MODULE_<MODULE>| Boolean flag | Indicates whether the corresponding OCCT module should be built or not. It should be noted that some toolkits of a module can be built even if this module is not checked (this happens if some other modules depend on these toolkits). The main modules and their descriptions can be found in @ref user_guides |
@@ -117,7 +116,6 @@ The following table gives the full list of environment variables used at the con
| INSTALL_DIR_DOC | Path | Relative path to the documentation installation directory (absolute path is ${INSTALL_DIR}/${INSTALL_DIR_DOC}) |
| INSTALL_FREETYPE | Boolean flag | Indicates whether Freetype binaries should be installed into the installation directory |
| INSTALL_FREEIMAGE* | Boolean flag | Indicates whether Freeimage binaries should be installed into the installation directory |
| INSTALL_GL2PS | Boolean flag | Indicates whether GL2PS binaries should be installed into the installation directory |
| INSTALL_TBB | Boolean flag | Indicates whether TBB binaries should be installed into the installation directory |
| INSTALL_VTK | Boolean flag | Indicates whether VTK binaries should be installed into the installation directory |
| INSTALL_TCL | Boolean flag | Indicates whether TCL binaries should be installed into the installation directory |
@@ -18,7 +18,7 @@ If you have official distribution with project files included, you can use them
Before building OCCT, make sure to have all the required third-party libraries installed.
The easiest way to install third-party libraries is to download archive with pre-built binaries, corresponding to version of Visual Studio you are using, from http://www.opencascade.com/content/3rd-party-components.
The easiest way to install third-party libraries is to download archive with pre-built binaries, corresponding to version of Visual Studio you are using, from https://www.opencascade.com/content/3rd-party-components.
You can also build third-party libraries from their sources, see @ref occt_dev_guides__building_3rdparty_windows for instructions.
@@ -37,6 +37,8 @@ If you have Visual Studio projects already available (pre-installed or generated
- *theHandlePtr* -- a pointer to the geometric variable (<i>Handle</i> to *Geom_Geometry* or *Geom2d_Curve* or descendant) to be set.
@section occt_debug_dump_json Dump OCCT objects into Json
Many OCCT classes may dump the current state into the stream. This stream contains the information about the class field into the field value/s.
It is possible to prepare recursive dump using corresponded macro for class fields. The depth of this recursion is defined by parameter of the dump.
The object defines What parameters should be presented in the Dump. The usual way is to dump all object fields.
@subsection occt_debug_dump_json_object Implementation in object
Steps to prepare dump of the object into json:
1. Create method <b>DumpJson</b>. The method should accept the output stream and the depth for the fields dump.
Depth, equal to zero means that only fields of this class should be dumped. Default value -1 means that whole tree of dump will be built recursively calling dump of all fields.
2. Put into the first row of the method <b>OCCT_DUMP_CLASS_BEGIN</b> or <b>OCCT_DUMP_TRANSIENT_CLASS_BEGIN</b> (for Standard_Transient objects).
This macro appends class name into output stream.
3. Add several macro to store field values.
The following macro are defined to cover the object parameters into json format:
| Name | Result in json |
| :-------------------------- | :--------|
| OCCT_DUMP_FIELD_VALUE_NUMERICAL | "field": value |
Visual Studio debugger provides the Command Window (can be activated from menu <b>View / Other Windows / Command Window</b>), which can be used to evaluate variables and expressions interactively in a debug session (see http://msdn.microsoft.com/en-us/library/c785s0kz.aspx). Note that the Immediate Window can also be used but it has some limitations, e.g. does not support aliases.
Visual Studio debugger provides the Command Window (can be activated from menu <b>View / Other Windows / Command Window</b>), which can be used to evaluate variables and expressions interactively in a debug session (see https://msdn.microsoft.com/en-us/library/c785s0kz.aspx). Note that the Immediate Window can also be used but it has some limitations, e.g. does not support aliases.
When the execution is interrupted by a breakpoint, you can use this window to call the above described functions in context of the currently debugged function. Note that in most cases you will need to specify explicitly context of the function by indicating the name of the DLL where it is defined.
@@ -310,3 +369,53 @@ Each counter has its name shown when the collected statistics are printed.
In DRAW, use command *dperf* to print all performance statistics.
Note that performance counters are not thread-safe.
@section occt_debug_sanitizers Use of compiler sanitizers
GCC and Clang compilers provide options for instrumenting the code with the tools intended for detection of run-time errors, called sanitizers.
This section provides some hints for using sanitizers for detecting possible errors in OCCT code.
@subsection occt_debug_sanitizers_linux Linux
Example of configuration steps for Ubuntu:
1. In CMake configuration:
- Use up-to-date version of the GCC or CLang compiler; make sure that if CMAKE_CXX_COMPILER is set to C++ compiler (e.g. "clang++-6.0") and CMAKE_C_COMPILER is set to C compiler (e.g. "clang-6.0")
- Ensure that CMAKE_LINKER is set to the C++ linker bundled with compiler (e.g. clang++-6.0); this is important to avoid linking problems
- For building with Address sanitizer, set CMAKE_CXX_FLAGS and CMAKE_C_FLAGS to "-fsanitize=address -fno-omit-frame-pointer -fno-optimize-sibling-calls"
- For building with Undefined Behavior sanitizer, set CMAKE_CXX_FLAGS and CMAKE_C_FLAGS to "-fsanitize=undefined -fno-omit-frame-pointer -fno-optimize-sibling-calls"
- Set CMAKE_BUILD_TYPE to RelWithDebInfo to get more informative stack traces on errors
2. Build as usual (make)
Be prepared that it works much slower than normal build and consumes more disk space.
3. Before running executable, make sure that "llvm-symbolizer" is in PATH; this is necessary to get human-readable stack traces. The tool must have exactly that name.
If it is installed in common folder (/usr/bin or similar) with different name, one option is to create a symlink, for instance:
Alternatively, add directory where actual llvm-symbolizer is located (such as /usr/lib/llvm-6.0/bin) to the PATH variable.
4. Set environment variable to disable memory leaks detection (they seem to be reported for every global variable at exit, not much useful):
> export ASAN_OPTIONS=detect_leaks=0
5. Set environment variable CSF_CPULIMIT_FACTOR to reasonably large number to increase the time limits for program execution (used by OCCT tests) to compensate the performance penalty introduced by sanitizers:
> export CSF_CPULIMIT_FACTOR=20
6. When using UBSan, set environment variable UBSAN_OPTIONS to get stack traces:
> export UBSAN_OPTIONS=print_stacktrace=1
7. Run DRAW and perform tests as usual, keeping in mind that running with sanitizer is much heavier than normal build:
> ./draw.sh relwithdeb <br>
> Draw[]> testgrid -parallel 0
Note that when running tests under sanitizers, behavior may be different.
Known problems (as of CLang 6.0) are:
- Software signals (access violation etc.) are not handled
- Heap memory usage always reports zero
@subsection occt_debug_sanitizers_windows Windows
Though CLang toolset is available in Visual Studio 2015 and newer, sanitizer do not seem to be available out of the box (last tested with VS 2019 16.2.3).
@@ -12,15 +12,15 @@ This document provides practical guidelines for generation and editing of OCCT u
You need to have the following software installed to generate the documentation.
**Tcl/Tk**
Version 8.5 or 8.6: http://www.tcl.tk/software/tcltk/download.html
Version 8.5 or 8.6: https://www.tcl.tk/software/tcltk/download.html
**Doxygen**
Version 1.8.4 or above: http://www.stack.nl/~dimitri/doxygen/download.html
Version 1.8.4 or above: http://www.doxygen.nl/download.html
**Dot**
Part of Graphviz software, used by Doxygen for generation of class diagrams in Reference Manual: http://www.graphviz.org/Download..php
Part of Graphviz software, used by Doxygen for generation of class diagrams in Reference Manual: https://www.graphviz.org/download/
**MiKTeX** or other package providing **pdflatex** command (only needed for generation of PDF documents): http://miktex.org/download
**MiKTeX** or other package providing **pdflatex** command (only needed for generation of PDF documents): https://miktex.org/download
**Inkscape** (only needed for generation of PDF documents containing SVG images): http://www.inkscape.org/download
@@ -134,7 +134,7 @@ Add a relative path to *svn.md* in file <i>dox/FILES.txt</i>. For instance
dev_guides/svn/svn.md
@endverbatim
**Note** that the order of paths to documents in *FILES.txt* is reproduced in the Table of Contents in the HTML output. Please, place them logically.
**Note** that the order of paths to documents in *FILES.txt* is reproduced in the Table of Contents in the HTML output, thus they need to be placed logically.
**Note** that you should specify a file tag, not the document name. See @ref OCCT_DM_SECTION_A_1 "Header and hierarchic document structure" section for details.
@@ -180,7 +180,7 @@ For example:
@subsubsection occt_ocaf_1_1_1 The document and the data framework
@endverbatim
Please, note that section names can be used for references within the document and in other documents, so it is necessary to use the common prefix indicative of the document name for all section names in the given document.
**Note** that section names can be used for references within the document and in other documents, so it is necessary to use the common prefix indicative of the document name for all section names in the given document.
For example, *occt_ocaf* for sections in Open CASCADE Application Framework manual.
The remaining part of section names in most documents consists only of numbers, for example *1_1*. Actually, the hierarchical structure of the output table of contents is not based on these numbers and is generated automatically.
@@ -11,18 +11,17 @@ Guide to installing and using Git for OCCT development {#occt_dev_guides__git_gu
to OCCT developers who are not familiar with this tool
and to facilitate the use of the official OCCT Git repository for code contribution to OCCT.
Reading this document does not exempt from the need to learn Git concepts and tools.
Please consult a book or manual describing Git to get acquainted with this tool.
Many good books on Git can be found at http://git-scm.com/documentation
It can be useful to learn more about Git concepts and tools from a book a or manual.
Many good books on Git can be found at https://git-scm.com/documentation
For the experienced Git users it can be enough to read sections 1 and 3
of this document to start working with the repository.
Please make sure to get familiar with the Contribution Workflow document
Familiarize yourselves with the @ref occt_dev_guides__contribution_workflow "Contribution Workflow document"
that describes how Git is used for processing contributions to OCCT.
This and related documents are available at the Resources page
of the OCCT development portal at http://dev.opencascade.org/index.php?q=home/resources.
of the OCCT development portal at https://dev.opencascade.org/index.php?q=home/resources.
@subsection occt_gitguide_1_2 Git URL
@@ -83,7 +82,7 @@ The official repository contains:
@subsection occt_gitguide_1_5 Version of Git
The repository is tested to work with Git 1.7.6 and above.
Please do not use versions below 1.7.1 as they are known to cause troubles.
Avoid using versions below 1.7.1 as they are known to cause troubles.
@section occt_gitguide_2 Installing Tools for Work with Git
@@ -114,7 +113,7 @@ The official repository contains:
@subsubsection occt_gitguide_2_1_2 Installation and configuration of TortoiseGit
Download TortoiseGit distributive from http://code.google.com/p/tortoisegit/downloads/list.
Download TortoiseGit distributive from https://tortoisegit.org/download/.
Launch the installation.
* Select your SSH client. Choose option
@@ -239,7 +238,7 @@ The official repository contains:
@subsection occt_gitguide_3_3 Adding public key in your account
Log in on the portal http://dev.opencascade.org and click on **My account** link to the right. If you have a Contributor status, you will see **SSH keys** tab to the right.
Log in on the portal https://dev.opencascade.org and click on **My account** link to the right. If you have a Contributor status, you will see **SSH keys** tab to the right.
Click on that tab, then click **Add a public key**, and paste the text of the public key (see above sections on how to generate the key) into the text box.
@@ -250,7 +249,7 @@ Click **Save** to input the key to the system.
It is typical to use your e-mail address or workstation name for this field; no restrictions are set by the portal.
Please note that some time (5-10 min) is needed for the system
**Note** that some time (5-10 min) is needed for the system
to update the configuration after the new key is added.
After that time, you can try accessing Git.
@@ -497,7 +496,7 @@ To rebase your branch into a single commit, you need to do the following:
The changes made in the branch can be reviewed without direct access to Git, using GitWeb interface:
* Open GitWeb in your web browser: http://git.dev.opencascade.org/gitweb/?p=occt.git
* Open GitWeb in your web browser: https://git.dev.opencascade.org/gitweb/?p=occt.git
* Locate the branch you want to review among **heads** (click ‘…’ at the bottom of the page to see the full list).
* Click **log** (or **shortlog**) to see the history of the branch.
@@ -347,7 +347,7 @@ The test group may contain *parse.rules* file. This file defines patterns used f
Each line in the file should specify a status (single word), followed by a regular expression delimited by slashes (*/*) that will be matched against lines in the test output log to check if it corresponds to this status.
The regular expressions should follow <ahref="http://www.tcl.tk/man/tcl/TclCmd/re_syntax.htm">Tcl syntax</a>, with a special exception that "\b" is considered as word limit (Perl-style), in addition to "\y" used in Tcl.
The regular expressions should follow <ahref="https://www.tcl.tk/man/tcl/TclCmd/re_syntax.htm">Tcl syntax</a>, with a special exception that "\b" is considered as word limit (Perl-style), in addition to "\y" used in Tcl.
The rest of the line can contain a comment message, which will be added to the test report when this status is detected.
@@ -622,7 +622,7 @@ The new test created for an unsolved problem should return BAD. The new test cre
@subsection testmanual_3_6 Marking BAD cases
If the test produces an invalid result at a certain moment then the corresponding bug should be created in the OCCT issue tracker located at http://tracker.dev.opencascade.org, and the problem should be marked as TODO in the test script.
If the test produces an invalid result at a certain moment then the corresponding bug should be created in the OCCT issue tracker located at https://tracker.dev.opencascade.org, and the problem should be marked as TODO in the test script.
The following statement should be added to such a test script:
~~~~~
@@ -878,7 +878,18 @@ The test grid name is constructed depending on the type of the tested chamfers.
| dist_angle_complex | | Distance from edge and given angle |
| dist_angle_sequence | | Distance from edge and given angle |
@subsubsection testmanual_5_1_7 demo
@subsubsection testmanual_5_1_7 de
This group tests reading and writing of CAD data files (iges, step) to and from OCCT.
Test cases check transfer status, shape and attributes against expected reference values.
* All occurrences of *Select3D_Projector* in the application code (if any) should be replaced with *Handle(Select3D_Projector)*.
* The code of inheritors of *Select3D_SensitiveEntity* should be updated if they override <i>Matches()</i> (this is probable, if clipping planes are used).
* Constructor for *V3d_Plane* has been changed, so the extra argument should be removed if used in the application. It is necessary to add a new plane using method *V3d_Viewer::AddPlane()* if *V3d_Viewer* has been used to manage clipping planes list (this does not affect clipping planes representation). Please, have a look at the source code for new DRAWEXE *vclipplane* command in *ViewerTest_ObjectsCommands.cxx, VClipPlane* to see how clipping planes can be managed in the application.
* Constructor for *V3d_Plane* has been changed, so the extra argument should be removed if used in the application. It is necessary to add a new plane using method *V3d_Viewer::AddPlane()* if *V3d_Viewer* has been used to manage clipping planes list (this does not affect clipping planes representation). Have a look at the source code for new DRAWEXE *vclipplane* command in *ViewerTest_ObjectsCommands.cxx, VClipPlane* to see how clipping planes can be managed in the application.
@section upgrade_652 Upgrade to OCCT 6.5.2
@@ -112,14 +112,14 @@ Porting of user applications from an earlier OCCT version to version 6.6.0 requi
It is necessary to introduce the corresponding changes in the applications for which the order of sub-shapes resulting from a Boolean operation is important. It is strongly recommended to use identification methods not relying on the order of sub-shapes (e.g. OCAF naming).
* If you need to use OCCT on Mac OS X with X11 (without Cocoa), build OCCT with defined pre-processor macro *CSF_MAC_USE_GLX11*. XLib front-end (previously the only way for unofficial OCCT builds on Mac OS X) is now disabled by default on this platform. If your application has no support for Cocoa framework you may build OCCT with XLib front-end adding *MACOSX_USE_GLX* macro to compiler options (you may check the appropriate option in WOK configuration GUI and in CMake configuration). Notice that XQuartz (XLib implementation for Mac OS X) now is an optional component and does not provide a sufficient level of integrity with native (Cocoa-based) applications in the system. It is not possible to build OCCT with both XLib and Cocoa at the same time due to symbols conflict in OpenGL functions.
* Animation mode and degeneration presentation mode (simplified presentation for animation) and associated methods have been removed from 3D viewer functionality.
Correspondingly, the code using methods *SetAnimationModeOn(), SetAnimationModeOff(), AnimationModeIsOn(), AnimationMode(), Tumble(), SetDegenerateModeOn(), SetDegenerateModeOff()* and *DegenerateModeIsOn()* of classes *V3d_View* and *Visual3d_View* will need to be removed or redesigned. Please, notice that Hidden Line Removal presentation was not affected; however, the old code that used methods *V3d_View::SetDegenerateModeOn* or *V3d_View::SetDegenerateModeOff* to control HLR presentation should be updated to use *V3d_View::SetComputedMode* method instead.
Correspondingly, the code using methods *SetAnimationModeOn(), SetAnimationModeOff(), AnimationModeIsOn(), AnimationMode(), Tumble(), SetDegenerateModeOn(), SetDegenerateModeOff()* and *DegenerateModeIsOn()* of classes *V3d_View* and *Visual3d_View* will need to be removed or redesigned. Hidden Line Removal presentation was not affected; however, the old code that used methods *V3d_View::SetDegenerateModeOn* or *V3d_View::SetDegenerateModeOff* to control HLR presentation should be updated to use *V3d_View::SetComputedMode* method instead.
* Calls of *Graphic3d_Group::BeginPrimitives()* and *Graphic3d_Group::EndPrimitives()* should be removed from the application code.
* Application functionality for drawing 2D graphics that was formerly based on *TKV2d* API should be migrated to *TKV3d* API. The following changes are recommended for this migration:
* A 2D view can be implemented as a *V3d_View* instance belonging to *V3d_Viewer* managed by *AIS_InteractiveContext* instance. To turn *V3d_View* into a 2D view, the necessary view orientation should be set up at the view initialization stage using *V3d_View::SetProj()* method, and view rotation methods simply should not be called.
* Any 2D graphic entity (formerly represented with *AIS2D_InteractiveObject*) should become a class derived from *AIS_InteractiveObject* base. These entities should be manipulated in a view using *AIS_InteractiveContext* class API.
* All drawing code should be put into *Compute()* virtual method of a custom interactive object class and use API of *Graphic3d* package. In particular, all geometry should be drawn using class hierarchy derived from *Graphic3d_ArrayOfPrimitives*. Normally, the Z coordinate for 2D geometry should be constant, unless the application implements some advanced 2D drawing techniques like e.g. multiple "Z layers" of drawings.
* Interactive selection of 2D presentations should be set up inside *ComputeSelection()* virtual method of a custom interactive object class, using standard sensitive entities from *Select3D* package and standard or custom entity owners derived from *SelectMgr_EntityOwner* base.
Please refer to the Visualization User's Guide for further details concerning OCCT 3D visualization and selection classes. See also *Viewer2D* OCCT sample application, which shows how 2D drawing can be implemented using TKV3d API.
Refer to the Visualization User's Guide for further details concerning OCCT 3D visualization and selection classes. See also *Viewer2D* OCCT sample application, which shows how 2D drawing can be implemented using TKV3d API.
* Run-time graphic driver library loading mechanism based on *CSF_GraphicShr* environment variable usage has been replaced by explicit linking against *TKOpenGl* library. The code sample below shows how the graphic driver should be created and initialized in the application code:
~~~~
// initialize a new viewer with OpenGl graphic driver
@@ -161,10 +161,10 @@ Porting of user applications from an earlier OCCT version to version 6.7.0 requi
@subsection upgrade_670_clipping Object-level clipping and capping algorithm.
* It might be necessary to revise and port code related to management of view-level clipping to use *Graphic3d_ClipPlane* instead of *V3d_Plane* instances. Please note that *V3d_Plane* class has been preserved -- as previously, it can be used as plane representation. Another approach to represent *Graphic3d_ClipPlane* in a view is to use custom presentable object.
* The list of arguments of *Select3D_SensitiveEntity::Matches()* method for picking detection has changed. Since now, for correct selection clipping, the implementations should perform a depth clipping check and return (as output argument) minimum depth value found at the detected part of sensitive. Please refer to CDL / Doxygen documentation to find descriptive hints and snippets.
* It might be necessary to revise and port code related to management of view-level clipping to use *Graphic3d_ClipPlane* instead of *V3d_Plane* instances. Note that *V3d_Plane* class has been preserved -- as previously, it can be used as plane representation. Another approach to represent *Graphic3d_ClipPlane* in a view is to use custom presentable object.
* The list of arguments of *Select3D_SensitiveEntity::Matches()* method for picking detection has changed. Since now, for correct selection clipping, the implementations should perform a depth clipping check and return (as output argument) minimum depth value found at the detected part of sensitive. Refer to CDL / Doxygen documentation to find descriptive hints and snippets.
* *Select3D_SensitiveEntity::ComputeDepth()* abstract method has been removed. Custom implementations should provide depth checks by method *Matches()* instead -- all data required for it is available within a scope of single method.
* It might be necessary to revise the code of custom sensitive entities and port *Matches()* and *ComputeDepth()* methods to ensure proper selection clipping. Please note that obsolete signature of *Matches* is not used anymore by the selector. If your class inheriting *Select3D_SensitiveEntity* redefines the method with old signature the code should not compile as the return type has been changed. This is done to prevent override of removed methods.
* It might be necessary to revise the code of custom sensitive entities and port *Matches()* and *ComputeDepth()* methods to ensure proper selection clipping. Note that obsolete signature of *Matches* is not used anymore by the selector. If your class inheriting *Select3D_SensitiveEntity* redefines the method with old signature the code should not compile as the return type has been changed. This is done to prevent override of removed methods.
@subsection upgrade_670_markers Redesign of markers presentation
@@ -210,7 +210,7 @@ If *ViewMapping* and *ViewOrientation* were used directly, this functionality ha
The current perspective model is not fully backward compatible, so the old perspective-related functionality needs to be reviewed.
Please revise application-specific custom presentations to provide proper bounding box. Otherwise object might become erroneously clipped by automatic *ZFit* or frustum culling algorithms enabled by default.
Revise application-specific custom presentations to provide a proper bounding box, otherwise the object might become erroneously clipped by automatic *ZFit* or frustum culling algorithms enabled by default.
@subsection upgrade_680_connected_objects Redesign of Connected Interactive Objects
@@ -439,7 +439,7 @@ class TColStd_Array1OfReal; -> #include <TColStd_Array1OfReal.hxx>
~~~~~
Handle_Class -> Handle(Class)
~~~~~
This change is not applied if the source or header file is recognized as containing the definition of Qt class with signals or slots, to avoid possible compilation errors of MOC files caused by inability of MOC to recognize macros (see http://doc.qt.io/qt-4.8/signalsandslots.html).
This change is not applied if the source or header file is recognized as containing the definition of Qt class with signals or slots, to avoid possible compilation errors of MOC files caused by inability of MOC to recognize macros (see https://doc.qt.io/qt-4.8/signalsandslots.html).
The file is considered as defining a Qt object if it contains strings *Q_OBJECT* and either *slots:* or *signals:*.
4. Removes forward declarations of classes with names <i>Handle(C)</i> or *Handle_C*, replacing them either by forward declaration of its argument class, or (for files defining Qt objects) <i>\#include</i> statement for a header with the name of the argument class and extension .hxx:
@@ -736,9 +736,9 @@ If you like to preserve the compatibility of your application code with OCCT ver
If your application is essentially based on CDL, and you need to upgrade it to OCCT 7.0, you will very likely need to convert your application code to non-CDL form.
This is a non-trivial effort; the required actions would depend strongly on the structure of the code and used CDL features.
The upgrade script and sources of a specialized WOK version used for OCCT code upgrade can be found in WOK Git repository in branch [CR0_700_2](http://git.dev.opencascade.org/gitweb/?p=occt-wok.git;a=log;h=refs/heads/CR0_700_2).
The upgrade script and sources of a specialized WOK version used for OCCT code upgrade can be found in WOK Git repository in branch [CR0_700_2](https://git.dev.opencascade.org/gitweb/?p=occt-wok.git;a=log;h=refs/heads/CR0_700_2).
[Contact us](http://www.opencascade.com/contact/) if you need more help.
[Contact us](https://www.opencascade.com/contact/) if you need more help.
@subsection upgrade_occt700_bspline Separation of BSpline cache
@@ -987,11 +987,11 @@ The applications that use *gp_Quaternion* to convert Yaw-Pitch-Roll angles (or o
Zoom persistent selection introduces a new structure *Graphic3d_TransformPers* to transform persistence methods and parameters and a new class *Graphic3d_WorldViewProjState* to refer to the camera transformation state. You might need to update your code to deal with the new classes if you were using the related features. Please, keep in mind the following:
Zoom persistent selection introduces a new structure *Graphic3d_TransformPers* to transform persistence methods and parameters and a new class *Graphic3d_WorldViewProjState* to refer to the camera transformation state. You might need to update your code to deal with the new classes if you were using the related features. Keep in mind the following:
* *Graphic3d_Camera::ModelViewState* has been renamed to *Graphic3d_Camera::WorldViewState*.
* Transformation matrix utilities from *OpenGl_Utils* namespace have been moved to *Graphic3d_TransformUtils* and *Graphic3d_TransformUtils.hxx* header respectively.
* Matrix stack utilities from *OpenGl_Utils* namespace have been moved to *OpenGl_MatrixStack* class and *OpenGl_MatrixStack.hxx* header respectively.
* *OpenGl_View* methods *Begin/EndTransformPersistence* have been removed. Please, use *Graphic3d_TransformPers::Apply()* instead to apply persistence to perspective and world-view projection matrices.
* *OpenGl_View* methods *Begin/EndTransformPersistence* have been removed. Use *Graphic3d_TransformPers::Apply()* instead to apply persistence to perspective and world-view projection matrices.
@subsection upgrade_occt700_correction_of_texture Texture mapping of objects
@@ -1091,7 +1091,7 @@ The following environment variables have become redundant:
* *CSF_UnitsLexicon* and *CSF_UnitsDefinition* are no more used. Units definition (*UnitsAPI/Lexi_Expr.dat* and *UnitsAPI/Units.dat*) is now embedded into source code.
* *CSF_XSMessage* and *CSF_XHMessage* are now optional.
English messages (XSMessage/*XSTEP.us* and SHMessage/*SHAPE.us*) are now embedded into source code
English messages (XSMessage/\*XSTEP.us* and SHMessage/\*SHAPE.us*) are now embedded into source code
and automatically loaded when environment variables are not set.
* *CSF_ShadersDirectory* is not required any more, though it still can be used to load custom shaders.
Mandatory GLSL resources are now embedded into source code.
@@ -1124,7 +1124,7 @@ The following classes have been changed:
* *BRepTools_Modifier* class now has two modes of work. They are defined by the boolean parameter *MutableInput*, which is turned off by default. This means that the algorithm always makes a copy of a sub-shape (e.g. vertex) if its tolerance is to be increased in the output shape. The old mode corresponds to *MutableInput* turned on. This change may impact an application if it implements a class derived from *BRepTools_Modifier*.
* The second parameter *theIsOuterWire* of method *ShapeAnalysis_Wire::CheckSmallArea* has been removed.
* In class *GeomPlate_CurveConstraint*, two constructors taking boundary curves of different types have been replaced with one constructor taking the curve of an abstract type.
* The last optional argument *RemoveInvalidFaces* has been removed from the constructor of class *BRepOffset_MakeOffset* and method *Initialize*.
* The last optional argument *RemoveInvalidFaces* has been removed from the constructor of class *BRepOffset_MakeOffset* and method *Initialize*.
* The public method *BOPDS_DS::VerticesOnIn* has been renamed into *SubShapesOnIn* and the new output parameter *theCommonPB* has been added.
@section upgrade_occt720 Upgrade to OCCT 7.2.0
@@ -1546,6 +1546,8 @@ Since the new version, the method *BRepAdaptor_CompCurve::IsPeriodic()* will alw
@subsection upgrade_730_removed Removed features
* The methods *SetDeflection*, *SetEpsilonT*, *SetDiscretize* of the class *IntTools_EdgeFace* have been removed as redundant.
* Deprecated functionality *V3d_View::Export()*, related enumerations Graphic3d_ExportFormat, Graphic3d_SortType
as well as optional dependency from gl2ps library have been removed.
@@ -1576,3 +1578,550 @@ Class *Message_PrinterOStream* can be used instead of *CDM_COutMessageDriver* to
If custom driver class is used in the application, that class shall be reimplemented inheriting from *Message_Printer* instead of *CDM_MessageDriver*.
Method *Send()* should be redefined instead of method *Write()* of *CDM_MessageDriver*.
To use the custom printer in OCAF, it can be either added to default messenger or set into the custom *Message_Messenger* object created in the method *MessageDriver()* of a class inheriting *CDF_Application*.
@section upgrade_occt740 Upgrade to OCCT 7.4.0
@subsection upgrade_740_BRepPrimAPI_MakeRevol Changes in BRepPrimAPI_MakeRevol algorithm
Previously the algorithm could create a shape with the same degenerated edge shared between some faces. Now it is prevented. The algorithm creates the different copy of this edge for each face. The method *Generated(...)* has been changed in order to apply restriction to the input shape: input shape can be only of type VERTEX, EDGE, FACE or SOLID. For input shape of another type the method always returns empty list.
@subsection upgrade_740_removed Removed features
* The following methods of the class *BRepAlgoAPI_BooleanOperation* have been removed as obsolete or replaced:
- *BuilderCanWork* can be replaced with *IsDone* or *HasErrors* method.
- *FuseEdges* removed as obsolete.
- *RefineEdges* replaced with new method *SimplifyResult*.
* The method *ImagesResult* of the class *BOPAlgo_BuilderShape* has been removed as unused. The functionality of this method can be completely replaced by the history methods *Modified* and *IsDeleted*.
* The method *TrackHistory* of the classes *BOPAlgo_RemoveFeatures* and *BRepAlgoAPI_Defeaturing* has been renamed to *SetToFillHistory*.
* The method *GetHistory* of the class *BRepAlgoAPI_Defeaturing* has been renamed to *History*.
* The classes *BRepAlgo_BooleanOperations* and *BRepAlgo_DSAccess* have been removed as obsolete. Please use the BRepAlgoAPI_* classes to perform Boolean operations.
* *BRepAlgo_DataMapOfShapeBoolean* has been removed as unused.
* *BRepAlgo_DataMapOfShapeInterference* has been removed as unused.
* *BRepAlgo_EdgeConnector* has been removed as unused.
* *BRepAlgo_SequenceOfSequenceOfInteger* has been removed as unused.
@subsection upgrade_740_localcontext Local Context removal
Previously deprecated Local Context functionality has been removed from AIS package,
so that related methods have been removed from AIS_InteractiveContext interface:
A set of deprecated methods previously related to Local Context and now redirecting to other methods has been preserved to simplify porting; they will be removed in next release.
@subsection upgrade_740_geomconvert Changes in behavior of Convert algorithms
Now methods *GeomConvert::ConcatG1*, *GeomConvert::ConcatC1*, *Geom2dConvert::ConcatG1*, *Geom2dConvert::ConcatC1* modify the input argument representing the flag of closedness.
@subsection upgrade_740_selection Changes in selection API and picked point calculation algorithm.
*SelectBasics_PickResult* structure has been extended, so that it now defines a 3D point on the detected entity in addition to Depth value along picking ray.
*SelectMgr_SelectingVolumeManager::Overlap()* methods have been corrected to fill in *SelectBasics_PickResult* structure (depth and 3D point) instead of only depth value, so that custom *Select3D_SensitiveEntity* implementation should be updated accordingly (including *Select3D_SensitiveSet* subclasses).
@subsection upgrade_740_ocafpersistence Document format version management improvement.
Previously Document format version restored by *DocumentRetrievalDriver* was propagated using static methods of the corresponding units (like *MDataStd* or *MNaming*) to static variables of these units and after that became accessible to Drivers of these units.
Now Document format version is available to drivers via *RelocationTable*. The Relocation table now keeps *HeaderData* of the document and a format version can be extracted in the following way: *theRelocTable.GetHeaderData()->StorageVersion()*.
Obsolete methods: *static void SetDocumentVersion (const Standard_Integer DocVersion)* and *static Standard_Integer DocumentVersion()* have been removed from *BinMDataStd*, *BinMNaming*, *XmlMDataStd* and *XmlMNaming*.
@subsection upgrade_740_changed_api_of_brepmesh BRepMesh - revision of the data model
The entire structure of *BRepMesh* component has been revised and separated into several logically connected classes.
In new version, deflection is controlled more accurately, this may be necessary to tune parameters of call of the BRepMesh algorithm on the application side to obtain the same quality of presentation and/or performance as before.
*BRepMesh_FastDiscret* and *BRepMesh_FastDiscretFace* classes have been removed.
The following changes have been introduced in the API of *BRepMesh_IncrementalMesh*, component entry point:
* Due to revised logic, *adaptiveMin* parameter of the constructor has been removed as meaningless;
* *BRepMesh_FastDiscret::Parameters* has been moved to a separate structure called *IMeshTools_Parameters*; the signatures of related methods have been changed correspondingly.
* Interface of *BRepMesh_Delaun* class has been changed.
Fill Area, Line and Marker aspects (classes *Graphic3d_AspectFillArea3d*, *Graphic3d_AspectLine3d*, *Graphic3d_AspectMarker3d* and *Graphic3d_AspectText3d*)
have been merged into new class *Graphic3d_Aspects* providing a single state for rendering primitives of any type.
The old per-primitive type aspect classes have been preserved as sub-classes of *Graphic3d_Aspects* with default values close to the previous behavior.
All aspects except Graphic3d_AspectFillArea3d define Graphic3d_TOSM_UNLIT shading model.
The previous approach with dedicated aspects per primitive type was handy in simplified case, but lead to confusion otherwise.
In fact, drawing points or lines with lighting applied is a valid use case, but only *Graphic3d_AspectFillArea3d* previously defined necessary material properties.
As aspects for different primitive types have been merged, Graphic3d_Group does no more provide per-type aspect properties.
Existing code relying on old behavior and putting interleaved per-type aspects into single Graphic3d_Group should be updated.
For example, the following pseudo-code will not work anymore, because all *SetGroupPrimitivesAspect* calls will setup the same property:
To solve the problem, the code should be modified to either put primitives into dedicated groups (preferred approach), or using *SetPrimitivesAspect* in proper order:
@subsection upgrade_740_materials Material definition
Decomposition of Ambient, Diffuse, Specular and Emissive properties has been eliminated within *Graphic3d_MaterialAspect* definition.
As result, the following methods of *Graphic3d_MaterialAspect* class have been removed: SetReflectionMode(), SetReflectionModeOn(), Ambient(), Diffuse(), Emissive(), Specular(), SetAmbient(), SetDiffuse(), SetSpecular(), SetEmissive().
Previously, computation of final value required the following code:
- Replace Graphic3d_MaterialAspect::Ambient(), SetAmbient(), Diffuse(), SetDiffuse(), Specular(), SetSpecular(), Emissive(), SetEmissive() with methods working with pre-multiplied color.
E.g. theMaterial.SetAmbientColor(Graphic3d_Vec3 (1.0f, 0.0f, 0.0f) * 0.2f).
- Avoid using Graphic3d_MaterialAspect::Color() and SetColor() with non-physical materials (Graphic3d_MATERIAL_ASPECT).
These materials do not include color definition, because it is taken from Graphic3d_Aspects::InteriorColor() - this has not been changed.
However, previously it was possible storing the color with SetColor() call and then fetching it with Color() by application code (the rendering ignored this value);
now SetColor() explicitly ignores call for Graphic3d_MATERIAL_ASPECT materials and Color() returns DiffuseColor() multiplication coefficients.
@subsection upgrade_740_text Changes in Graphic3d_Text and OpenGl_Text API
Parameters of *Text* in *Graphic3d_Group* are moved into a new *Graphic3d_Text* class. *AddText* of *Graphic3d_Group* should be used instead of the previous *Text*.
The previous code:
~~~~
Standard_Real x, y, z;
theAttachmentPoint.Coord(x,y,z);
theGroup->Text (theText,
Graphic3d_Vertex(x,y,z),
theAspect->Height(),
theAspect->Angle(),
theAspect->Orientation(),
theAspect->HorizontalJustification(),
theAspect->VerticalJustification());
~~~~
should be replaced by the new code:
~~~~
Handle(Graphic3d_Text) aText = new Graphic3d_Text (theAspect->Height());
Historically AIS_InteractiveObject provided two independent mechanisms invalidating presentation (asking presentation manager to recompute specific display mode or all modes):
1.*AIS_InteractiveObject::SetToUpdate()*, marking existing presentation for update.
This is main invalidation API, which is expected to be followed by *AIS_InteractiveContext::Update()* call.
This is auxiliary invalidation API, used internally by AIS_InteractiveContext::SetColor()/UnsetColor() and similar modification methods.
The latter one has been removed to avoid confusion and unexpected behavior.
In addition, two methods *AIS_InteractiveObject::Update()* have been deprecated in favor of new *AIS_InteractiveObject::UpdatePresentations()* recomputing only invalidated presentations.
Custom presentations implementing interface methods *AIS_InteractiveObject::SetColor()* and others should be revised to use *AIS_InteractiveObject::SetToUpdate()*
or updating presentation without recomputation (see *AIS_InteractiveObject::SynchronizeAspects()* and *AIS_InteractiveObject::replaceAspects()*).
* *Aspect_IS_HOLLOW* is now an alias to *Aspect_IS_EMPTY* and does not implicitly enables drawing mesh edges anymore.
Specify Graphic3d_AspectFillArea3d::SetDrawEdges(true) with Graphic3d_AspectFillArea3d::SetInteriorStyle(Aspect_IS_EMPTY) to get previous behavior of Aspect_IS_HOLLOW style.
* *Aspect_IS_HIDDENLINE* does not implicitly enables drawing mesh edges anymore.
Specify Graphic3d_AspectFillArea3d::SetDrawEdges(true) with Graphic3d_AspectFillArea3d::SetInteriorStyle(Aspect_IS_HIDDENLINE) to get previous behavior of Aspect_IS_HIDDENLINE style.
@subsection upgrade_740_modedprs PrsMgr and SelectMgr hierarchy clean up
Proxy classes *Prs3d_Presentation*, *PrsMgr_ModedPresentation* and *PrsMgr_Prs* have been removed.
Code iterating through the list of low-level structures AIS_InteractiveObject::Presentations() should be updated to access PrsMgr_Presentation directly.
Forward declarations of *Prs3d_Presentation* should be corrected, since it is now a typedef to *Graphic3d_Structure*.
Proxy classes *SelectBasics_SensitiveEntity* and *SelectBasics_EntityOwner* have been removed - *Select3D_SensitiveEntity* and *SelectMgr_EntityOwner* should be now used directly instead.
*Graphic3d_PolygonOffset* default constructor has been corrected to define Units=1 instead of Units=0.
Default polygon offset settings Mode=Aspect_POM_Fill + Factor=1 + Units=1 are intended to push triangulation
(Shaded presentation) a little bit behind of lines (Wireframe and Face Edges)
for reducing z-fighting effect of Shaded+Wireframe combination.
The change in defaults (Units changed from 0 to 1) is intended to cover scenario when camera direction is perpendicular to model plane (like 2D view).
Application observing unexpected visual difference on this change should consider customizing this property within AIS_InteractiveContext default attributes
or on per-presentation basis via *Graphic3d_Aspects::SetPolygonOffset()* methods.
@subsection upgrade_740_zlayer Adding ZLayers in given position
Interface of insertion ZLayer in the viewer has been improved with ability to insert new layer before or after existing one.
Previously undocumented behavior of *V3d_Viewer::AddZlayer()* method has been corrected to insert new layer before *Graphic3d_ZLayerId_Top*.
Applications might need revising their custom layers creation code and specify precisely their order with new methods *V3d_Viewer::InsertLayerBefore()* and *V3d_Viewer::InsertLayerAfter()*.
Applications using integer values of the following enumerations in persistence
should be corrected as these enumerations have been modified:
| Name |
| :----- |
| AIS_TypeOfAttribute |
| Aspect_InteriorStyle |
| Font_FontAspect |
@subsection upgrade_740_geproj Custom defines within env.bat
*env.bat* produced by Visual Studio project generator *genproj.bat* has been modified so that *%CSF_DEFINES%* variable is reset to initial state.
Custom building environment relying on old behavior and setting extra macros within *%CSF_DEFINES%* before env.bat should be updated
to either modify custom.bat or setup new variable *%CSF_DEFINES_EXTRA%* instead.
@subsection upgrade_740_BVH_in_BOP Switching Boolean Operations algorithm to use BVH tree instead of UB tree
Since OCCT 7.4.0 Boolean Operations algorithm uses BVH tree instead of UBTree to find the pairs of entities with interfering bounding boxes.
The following API changes have been made:
* BOPTools_BoxBndTree and BOPTools_BoxBndTreeSelector have been removed. Use the BOPTools_BoxTree and BOPTools_BoxTreeSelector instead.
* BOPTools_BoxSelector::SetBox() method now accepts the BVH_Box instead of Bnd_Box.
* Methods BOPTools_BoxSelector::Reject and BOPTools_BoxSelector::Accept have been removed as unused.
* The RunParallel flag has been removed from the list of parameters of BOPAlgo_Tools::IntersectVertices method. Earlier, it performed selection from the UB tree in parallel mode. Now all interfering pairs are found in one pass, using pair traverse of the same BVH tree.
@subsection upgrade_740_stdnamespace Standard_Stream.hxx no more has "using std::" statements
*Standard_Stream.hxx* header, commonly included by other OCCT header files, does no more add entities from *std namespace* related to streams (like *std::cout*, *std::istream* and others) into global namespace.
The application code relying on this matter should be updated to either specify std namespace explicitly (like std::cout) or add "using std::" statements locally.
@section upgrade_occt750 Upgrade to OCCT 7.5.0
@subsection upgrade_750_srgb_color RGB color definition
OCCT 3D Viewer has been improved to properly perform lighting using in linear RGB color space and then convert result into non-linear gamma-shifted sRGB color space before displaying on display.
This change affects texture mapping, material definition and color definition.
Previously *Quantity_Color* definition was provided with unspecified RGB color space.
In practice, mixed color spaces have been actually used, with non-linear sRGB prevailing in general.
Since OCCT 7.5.0, *Quantity_Color* now specifies that components are defined in linear RGB color space.
This change affects following parts:
* Standard colors defined by *Quantity_NameOfColor* enumeration have been converted into linear RGB values within Quantity_Color construction.
* Application may use new enumeration value *Quantity_TOC_sRGB* for passing/fetching colors in sRGB color space,
which can be useful for interoperation with color picking widgets (returning 8-bit integer values within [0..255] range)
or for porting colors constants within old application code without manual conversion.
* *Graphic3d_MaterialAspect* color components are now expected in linear RGB color space,
and standard OCCT materials within *Graphic3d_NameOfMaterial* enumeration have been updated accordingly.
* Texture mapping now handles new *Graphic3d_TextureRoot::IsColorMap()* for interpreting content in linear RGB or sRGB color space.
It is responsibility of user specifying this flag correctly. The flag value is TRUE by default.
* Method *Image_PixMap::PixelColor()* has been extended with a new Boolean flag for performing linearization of non-linear sRGB.
This flag is FALSE by default; application should consider passing TRUE instead for further handling *Quantity_Color* properly as linear RGB values.
Unexpected const-ness of Aspect_Window::DoResize() method has been removed, so that application classes implementing this interface should be updated accordingly.
@subsection upgrade_750_rename Renaming of types
Enumeration BRepOffset_Type is renamed to ChFiDS_TypeOfConcavity.
Properties Prs3d_Drawer::HLRAngle() and Prs3d_Drawer::HLRDeviationCoefficient() have been removed from classes *Prs3d_Drawer*, *AIS_Shape* and *AIS_InteractiveContext*.
Prs3d_Drawer::DeviationAngle() should be now used instead of Prs3d_Drawer::HLRAngle() and Prs3d_Drawer::DeviationCoefficient() instead of Prs3d_Drawer::HLRDeviationCoefficient().
The default value of Prs3d_Drawer::DeviationAngle() property has been changed from 12 to 20 degrees to match removed Prs3d_Drawer::HLRAngle(), previously used as input for triangulation algorithm.
@subsection upgrade_750_hlrprs Changes in HLR presentation API
Methods computing HLR presentation within *PrsMgr_PresentableObject::Compute()* have been renamed to *PrsMgr_PresentableObject::computeHLR()*
and now accept *Graphic3d_Camera* object instead of removed *Prs3d_Projector*.
@subsection upgrade_750_dimensions Dimension and Relation presentations moved from AIS to PrsDim
Presentation classes displaying Dimensions and Relations have been moved from *AIS* package to *PrsDim*.
Corresponding classes should be renamed in application code (like *AIS_LengthDimension* -> *PrsDim_LengthDimension*).
The method Select3D_SensitiveEntity::NbSubElements() has been changed to be constant. Select3D_SensitiveEntity subclasses at application level should be updated accordingly.
@subsection upgrade_750_Booleans Changes in Boolean operations algorithm
* TreatCompound method has been moved from *BOPAlgo_Tools* to *BOPTools_AlgoTools*. Additionally, the map parameter became optional:
@subsection upgrade_750_Adaptor2d_OffsetCurve Offset direction change
Offset direction, which used in class Adaptor2d_OffsetCurve for evaluating values and derivatives of offset curve is unified for offset direction used in class Geom2d_OffsetCurve: now offset direction points to outer ("right") side of base curve instead of the previously used inner ("left") side. Old usage of class in any application should be changed something like that:
PBR math (rasterization) {#occt_dev_guides__pbr_math}
========================
@tableofcontents
# Preface
**Empirical** illumination models like **Phong reflection model** have been used in real-time graphics for a long time due to their simplicity, convincing look and affordable performance.
Before programmable pipeline has been introduced, graphics cards implemented Gouraud shading as part of fixed-function Transformation & Lighting (T&L) hardware blocks.
Nowadays, however, numerous trade-offs of this simplicity (like lighting partially baked into object material properties and others) pushed developers to **Physically-Based Rendering** (**PBR**) illumination models.
PBR models try to fit surface shading formulas into constrains of physical laws of light propagation / absorption / reflection - hence, called "physically-based".
There are two main categories of PBR illumination:
1. Non-real-time renderer (cinematic).
2. Real-time renderer.
The main objective of cinematic renderer is uncompromised quality, so that it relies on ray-tracing (path-tracing) rendering pipeline.
Although performance of current graphics hardware does not make it possible using computationally-intensive path-tracing renderer in real-time graphics, it can be used in interactive fashion.
"Physically-based" does not necessarily mean physically-correct/precise.
The main objective of real-time PBR renderer is to be fast enough even on low-end graphics hardware.
So that in contrast, it hardly relies on rasterization rendering pipeline, various approximations and tricks making it applicable in real-time, while looking good enough and preserving some physical properties.
OCCT 3D Viewer provides both kinds of PBR renderers, and although they share some details in common, this article is devoted to real-time PBR metallic-roughness illumination model.
This article describes the math underneath PBR shading in OCCT 3D Viewer and its GLSL programs.
However, this article does not clarifies related high-level APIs nor PBR material creation pipelines, as this is another topic.
# Notation
| | | |
|-:|:-|:-|
| \f$n\f$ | normal (on surface) | \f$\|n\|=1\f$ |
| \f$v\f$ | view direction | \f$\|v\|=1\f$ |
| \f$l\f$ | light | \f$\|l\| = 1\f$ |
| \f$h=\frac{v+l}{\|v + l\|}\f$ | half vector | |
| \f$m\f$ | metallic factor | \f$[0, 1]\f$ |
| \f$r\f$ | roughness factor | \f$[0, 1]\f$ |
| \f$IOR\f$ | index of refraction | \f$[1, 3]\f$ |
| \f$c\f$ | albedo color | \f$(R, G, B)\f$ |
\f$\cos\theta_l=(n \cdot l)\f$
\f$\cos\theta_v=(n \cdot v)\f$
\f$\cos\theta_h=(n \cdot h)\f$
\f$\cos\theta_{vh}=(v \cdot h)\f$
# Illumination model
The main goal of illumination model is to calculate outgoing light radiance \f$L_o\f$ along the certain direction.
The starting point of calculation might be the view direction \f$v\f$ aimed from point on surface (or in more general case just in space) to viewer position.
Considering the point on opaque surface with normal \f$n\f$ the main equation of illumination can be defined as:
Where \f$L_i(l)\f$ is light radiance coming from \f$l\f$ direction, \f$f(v,l)\f$ is **Bidirectional Reflectance Distribution Function** (**BRDF**) and \f$H\f$ is hemisphere which is oriented regarding to the surface normal \f$n\f$.
Opaqueness of the surface mentioned earlier is important because in that case hemisphere is enough.
More general model will require to consider directions all around a whole sphere and is not observed in this paper.
\f$\cos\theta_l\f$ factor appearing is caused by affection of surface area and light direction mutual orientation to the amount of radiance coming to this area.
This is mainly due to geometric laws. The rest part of integral is the key of the whole illumination model.
BRDF defines it's complexity and optical properties of material.
It has to model all light and material interactions and also has to satisfy some following criteria in order to be physical correct:
* Positivity: \f$f(v,l) \geq 0\f$
* Helmholtz reciprocity: \f$f(v,l) = f(l, v)\f$ (follows from 2<sup>nd</sup> Law of Thermodynamics)
* Energy conservation: \f$\displaystyle \forall v \, \int\limits_H f(v,l) \cos\theta_l \, \mathrm{d}l = 1\f$ (in order not to reflect more light than came)
It is worth to be mentioned that \f$f(v,l)\f$ depends on \f$n\f$ also but it is omitted to simplify notation. BRDF is usually split into two parts:
\f[f(v,l) = f_d(v,l)+f_s(v, l)\f]
Where \f$f_s(v, l)\f$ (specular BRDF) models reflection light interaction on surface and \f$f_d(v,l)\f$ (diffuse BRDF) models other processes happening depth in material (subsurface scattering for example).
So that illumination equation might be rewritten as:
PBR theory is based on **Cook-Torrance specular BRDF**. It imagines surface as set of perfectly reflected micro faces distributed on area in different ways which is pretty good model approximation of real world materials.
If this area is small enough not to be able to recognize separate micro surfaces the results becomes a sort of averaging or mixing of every micro plane illumination contribution.
In that level it allows to work with micro faces in statistical manner manipulating only probabilities distributions of micro surfaces parameters such as normals, height, pattern, orientation etc.
In computer graphics pixels are units of images and it usually covers a relatively large areas of surfaces so that micro planes can be considered to be unrecognizable.
Going back to the BRDF the Cook-Torrance approach has the following expression:
Where \f$k=\frac{\alpha}{2}\f$, which means \f$k=\frac{r^2}{2}\f$ in terms of this paper.
But \f$G\f$ depends on many factors so that it's approximations has float nature and can be modified a little bit in some cases in order to get more pleasant visual results.
One of this modification will be described later in following chapters.
The last component \f$F\f$ shows **how much light is reflected from surface** and is called **Fresnel's factor**.
The rest amount of radiance might be absorbed or refracted by material.
The most accurate expression of it is pretty complicate for calculation so that there is a variety of approximations.
The good one with less computation efforts is **Schlick's implementation**:
\f[F=F_0+(1-F_0)(1-\cos\theta_{vh})^5\f]
Here \f$F_0\f$ is material's response coefficient at normal incidence (zero angle).
Fresnel's factor has to be calculated differently for metals and dielectric/non-metals, but PBR theory tries to come up with universal formula for all types of material.
In order to do that it is needed to be noticed that Schlick's approximation is applicable only to non-conductors and in that case \f$F_0 = F_{dielectric} = \left(\frac{1-IOR}{1+IOR}\right)^2\f$.
**Index of Refraction** \f$IOR\f$ shows the proportion between light speed in vacuum (or even in air) and in material.
The reference value of \f$IOR\f$ for plastic is **1.5**, and this value can be considered as default for all unknown dielectrics.
In practice this parameter controls reflectance ability of material.
Also it should be remembered that this approximation produces poor results with large \f$IOR\f$ values so that it is recommended to be kept in range of \f$[1, 3]\f$ in order to get plausible Fresnel's factor.
This formula might be further propagated onto metals by using \f$F_0\f$ measured specifically for certain metal.
It can be considered as some kind of a 'color' of metal and can be stored as albedo parameter \f$c\f$.
And the final step of defining Fresnel's factor formula is mixing all this \f$F_0\f$ using metallic parameter \f$m\f$ (**metalness**):
\f[F_0 = F_{dielectric}(1-m)+cm\f]
For pure dielectrics with \f$m=0\f$ exactly Schlick's approximation will be used.
For pure metals with \f$m=1\f$ it will be a little inaccurate but the same formula with measured \f$F_0\f$ values.
Everything else for \f$m \in (0, 1)\f$ is not physically correct and it is recommended to keep \f$m\f$ exactly 1 or 0.
Intermediate values may represent mixed areas for smooth transition between materials - like partially rusted metal (rust is mostly dielectric).
Also it might be useful when parameters are read from textures with filtering and smoothing.
BRDF described above has one important trait making computations easier called **isotropy**.
Isotropy in this case means independence from rotation about normal resulting from supposition of uniform micro faces distribution at any direction along a surface.
It allows to simplify random samples generation during Monte-Carlo integrals calculation and reduce dimensions of some lookup tables, which will be discussed in following chapters.
Of course, isotropic materials form only subset of all real world's materials, but this subset covers majority of cases.
There are special models considering special anisotropic traits of surfaces like a grinding of metal or other with dependency on rotation about normal;
these models require special calculation tricks and additional parameters and are out of scope of this paper.
The only thing left to do is to define \f$f_d(v,l)\f$.
This part is responsible for processes happening in depth of material.
First of all the amount of input light radiance participating in these processes is needed to be calculated.
And it exactly can be realized from already known Fresnel's factor \f$F\f$ showing amount of reflected light but in negative term in this case in order to get the radiance left after reflection:
\f[1-F\f]
This part of ingoing light is assumed to be refracted in depth of surface and variety of events may happen there.
A sequence of absorptions, reflections and reemissions more or less leads to light's subsurface scattering.
Some part of this scattered light can go back outside but in modified form and in pretty unpredictable directions and positions.
For opaque materials this part is noticeable and forms it's own color.
If subsurface's paths of light are small enough and points of output are distributed locally around the input point it's possible to work in statistical way similar to the micro faces.
This assumption covers a big amount of real world opaque materials.
Other materials like skin, milk etc. with noticeable effect of subsurface scattering usually presented in form of partial translucency and some kind of self emission
have more widely distributed output points and require more accurate and complicate ways of modeling with maybe some theory and techniques from volumetric rendering.
The simple but visually enough assuming for statistically driven type of materials is just the same radiance for any direction. It results to **Lambertian's BRDF**:
\f[\frac{c}{\pi}\f]
Where \f$\pi\f$ is normalization coefficient in order to meet BRDF's criteria and \f$c\f$ is material's own color formed by adventures of light under surface.
There is one detail about light interaction bringing some physicality to the model, and that is an absence of this diffuse component in metals.
Metals reflect main part of light and the rest of it is absorbed being transformed into other form (mostly heat).
That is the main visual difference between metallic and non-metallic materials realizing of which brings model to higher level of quality in compare to older non-physical models.
So that all parts described above can be combined into united diffuse BRDF:
\f[f_d(v,l) = (1-F)(1-m)\frac{c}{\pi}\f]
\f$m\f$ is recommended to be exactly 1 or 0 but all values between can represent transition areas, as mentioned before.
In this chapter one possible implementation of illumination model reflecting main PBR principles has been defined.
The next step is using of it in practice.
# Practical application
It's time to apply deduced illumination model in practice.
And the first step of it is separation of **direction based light sources** from illumination integral.
Directional nature of such light sources means possibility to calculate it's influence to point of surface using only one direction and its intensity.
Usually sources of this type do not have physical size and are represented only by position in space (for point or spot lights) or by direction itself (direction light imagined to be too far point sources like sun).
This is just a kind of abstraction, while real world light emitters have noticeably sizes.
But sources with realistic form and size cannot be presented in discrete term and require continuous integrals calculations or special approximations in order to be accurately injected to the model.
In most cases direct based light sources in form of emitting points in space or just certain directions are good approximations and are enough for beginning.
Having finite discrete amount of it in scene and considering only single direction from every of these lights, the integral is transformed just to the sum:
Where \f$M\f$ is a number of sources, \f$l_j\f$ is a direction and \f$L_i^{direct}\f$ is an intensity related to this direction.
\f$direct\f$ label means that illumination has been computed directly from sources.
The BRDF can be used directly without any calculation problems.
The only exception might be \f$k\f$ in \f$G\f$ factor - it is recommended to be equal \f$\frac{(r+1)^2}{8}\f$ in order to get more pleasant results (that is modification mentioned in previous chapter).
And actually it is enough to finally see something.
There will be correct visualization with assumption of complete dark environment and absence of other points influence.
It is called **local illumination**. Based on this name there is also a global or **indirect illumination** and that is the rest of integral:
It includes influence of light reflected or scattered from other points and environment's contribution.
It's impossible to achieve photorealistic results without this component, but is is also very difficult to compute.
While the cross point light interaction cannot be calculated in a simple way (especially in real time rendering), the environment illumination has some options to be realized via precomputational work before visualization.
But right now lets summarize the practical application of illumination model.
At this moment the output radiance is represented as:
\f[L_o = L_{direct} + L_{indirect}\f]
Where \f$L_{direct}\f$ is direction based light sources contribution which can be directly computed just applying bare formulas.
It is enough to get some results in terms of local illumination but without \f$L_{indirect}\f$ component image will not be looked lifelike.
\f$L_{indirect}\f$ is not trivial thing for calculation and that is stumbling block for real time rendering applications.
But it can be relatively easy implemented in case of environment illumination via some precomputational work about which will be told in details in following chapters.
# Image based lighting
The next goal after \f$L_{direct}\f$ calculation is to find \f$L_{indirect}\f$.
And it would be easier if \f$L_i^{indirect}(l)\f$ was known for every \f$l\f$.
That is the main assumption of **image based lightning** (**IBL**).
In practice, it can be achieved using environment image map, which is a special image representing illumination from all possible directions.
This image might be a photo capturing a real world environment (spherical 360 degrees panoramas) or generated image baking the 3D scene itself, including in that case reflections of other objects.
Environment image might be packed in different ways - **cube maps** and equirectangular maps are the most commonly used.
Anyway, it allows \f$L_i^{indirect}(l)\f$ to be defined for every \f$l\f$ and its practical implementation in form of images gives name for this approach.
Next transformations of these expressions require understanding of numerical way to find hemisphere integral and also its performance optimization techniques.
And that the topic of the next chapter.
# Monte-Carlo numeric integration
**Monte-Carlo** is one of numeric methods to **find integral**.
It is based on idea of mathematical expectation calculation.
In one dimensional case if \f$f(x)\f$ is a function with parameter distributed according to probability density \f$p(x)\f$ the mathematical expectation of it can be found using following expression:
The main questions are choosing \f$p(l)\f$ and generating samples \f$l_i\f$.
The one of the simple ways is uniform distribution along sphere or hemisphere.
Lets realize that on sphere for example.
There are \f$4\pi\f$ possible directions in terms of sphere's areas and steradians (direction can be presented as dot on a unit sphere):
\f[\int\limits_S 1\, \mathrm{d}l = 4\pi\f]
Where \f$S\f$ is the unit sphere.
In order to be uniform \f$p(l)\f$ must be constant and satisfy normalization criteria:
\f[\int\limits_S p(l)\, \mathrm{d}l = 1\f]
So that \f$p(l) = \frac{1}{4\pi}\f$.
Usually direction \f$l\f$ is parametrized by spherical coordinates \f$\phi \in [0, 2\pi]\f$ and \f$\theta \in [0, \pi]\f$ boiling down to the 2D samples generation.
But in these terms joint \f$p(\theta, \phi)\f$ will be looked slightly different due to variables transition.
\f$l\f$ is defined in regular Cartesian coordinates \f$l=(x, y, z)\f$ with \f$\|l\| = 1\f$.
The spherical coordinates transform looks like:
\f[x = r\sin\theta\cos\phi,\, y = r\sin\theta\sin\phi,\, z = r\cos\theta\f]
Where \f$r = 1\f$.
In order to express probability density using new variables it is needed to multiply this density by Jacobian of transform:
So that joint probability density in new variables looks like:
\f[p(\theta, \phi) = \frac{\sin\theta}{4\pi}\f]
This variable transfer rule of **Probability Density Function** (**PDF**) will be useful in following chapters, when integral calculation optimization techniques will be being told.
Having \f$p(\theta, \phi)\f$ the partial single dimensional probability densities are able to be found:
The CDF maps \f$\theta\f$ values from range of \f$[0, \pi]\f$ to probability in range of \f$[0, 1]\f$.
The next step is to find inverse cumulative function which can be not so trivial sometimes but pretty obvious in current case:
\f[F^{-1}(u) = \arccos(1-2u)\f]
If substitute uniform distributed in range \f$[0, 1]\f$ values \f$u\f$ as argument of this function the values with origin probability density will appear.
Mote-Carlo integration cannot guarantee exact estimation of convergence speed with using random generated samples.
There is only probability estimation of it.
But this algorithm is pretty universal and relatively simple to be applied to almost any function using at least uniform distributed points.
Moreover special \f$p(l)\f$ can be chosen and special pseudo-random sequences can be designed in order to speed up convergence for some functions (following chapter talk about that in details).
That is why this method is widely used in computer graphics and demonstrates good results.
Also another one advantage is worth to be mentioned - possibility to iteratively accumulate computations and present intermediate results during rendering which is used in some ray tracing applications.
# Split sum
Lets go back to the image based lighting and the figure of specular component.
As was defined before that is hemisphere integral with following expression:
This integral is exact \f$L_{indirect}^s\f$ in condition when \f$L_i^{indirect}(l) = 1\f$ what just means white uniform environment.
The sum before it is kind of averaged environment illumination.
The main accomplishment after all this manipulations is possibility to calculate light and BRDF components separately.
The sum with \f$L_i^{indirect}(l_i)\f$ can be computed beforehand for every normal direction and stored to image called specular map but with some valuable details.
The problem is that \f$l_i\f$ samples must be generated according to \f$p(v, l_i)\f$ distribution depended on \f$v\f$ and \f$r\f$ as was mentioned earlier.
Variation of normal is not enough in that case and these variables are needed to be considered too.
The ways to resolve it are topic of one of the following chapters and now understanding the fact that at least this part can be precomputed before rendering is enough for now.
And it is important not to miss out that there is no more BRDF influence in this sum and only \f$p(v, l)\f$ can affect in this case.
That is why it is so important to strict to PDF during samples generation and that is why \f$p(v, l)\f$ must be correlated with BRDF somehow in this approximation approach with splitting.
For example completely mirroring materials with \f$r = 0\f$ will not be looked as expected if just uniform distribution is used
because such surfaces have only one possible direction from which light can be reflected along view direction in compare with \f$N\f$ absolutely scattered in case of uniform or many other distributions.
The rest part also can be saved to image. Lets unroll its expression:
This form may not look easier, but it has several advantages.
The first one is independence from globally defined \f$L_i^{indirect}(l)\f$, so that normal orientation does not matter and can be set in any handful way for calculations (Z axis for example).
The second one results from isotropic illumination system allowing \f$\phi\f$ component of view vector to be set arbitrarily (0 for example) and \f$\cos\theta_v\f$ will be enough to define view direction.
And the third one is scalar nature of integrals so that only two precomputed numbers are needed to find BRDF part of \f$L_{indirect}^s\f$.
Considering dependency of these integrals from \f$\cos\theta_v\f$ and \f$r\f$ both of it can be precomputed and stored to 2D look-up image variating these two parameters in range \f$[0, 1]\f$ with two channels consisting of scale and bias for \f$F_0\f$.
Current result for \f$L_{indirect}^s\f$ is computing it using 2D image for BRDF part and omnidirectional image for environment illumination.
There were a lot of words about Monte-Carlo optimizations techniques and about PDF choice which is important not only in terms of numeric integration but in terms of visual results correctness.
It's time to talk about that.
# Importance sampling
Current goal is to speed up Monte-Carlo integration of Cook-Torrance like integrals with following expression:
Where \f$g(v, l)\f$ is just arbitrary function representing Fresnel's factor itself or its components.
In order to increase convergence the samples with larger contribution (or in other words with larger function's values) have to appear more frequently than others proportionally to its contribution.
So that less significant summand with less influence to result will be considered rarely and in opposite way parts brining noticeable changes to the sum will be taken often.
That is the main idea of **importance sampling technique**.
\f$p(l)\f$ has to represent significance of sample in terms of integrated function via probability somehow.
And it will be like that if PDF is already part of original function because in that case probability density directly affects to contribution forming.
Separating this distribution component is one possible and effective way to realize importance sampling strategy.
In integral presented above PDF part already exists and that is \f$D\f$ component.
But it is distribution of micro faces normals or ideally reflection direction or \f$h\f$ in other word and not light directions distribution which is needed in fact.
Anyway that is good starting point and lets generate \f$h\f$ vectors first.
Frankly speaking \f$D(h)\f$ is called normal distribution but cannot be directly used as hemisphere distribution.
Originally it is statistical factor used to define total area of micro faces \f$\mathrm{d}A_h\f$
whose normals lie withing given infinitesimal solid angle \f$\mathrm{d}h\f$ centered on \f$h\f$ direction using the original small enough area of macro surface \f$\mathrm{d}A\f$:
\f[dA_h = D(h)\,\mathrm{d}h\, \mathrm{d}A\f]
First of all this factor must be positive:
\f[D(h) \geq 0\f]
But the total area of micro faces landscape is at least equal to origin surface but even bigger in general:
\f[1 \leq \int\limits_H D(h)\, \mathrm{d}h\f]
This trait does not allow to use \f$D\f$ as hemisphere distribution.
So that there is no more obstacles to generate \f$h\f$.
But the main goal was \f$l\f$ direction.
In order to get it the view vector \f$v\f$ has to be reflected related to already found \f$h\f$:
\f[l = 2(v \cdot h)h - v\f]
That is practical side of light direction generation.
But the theoretical one is needed to be resolved to calculate sum.
It is time to find \f$p(l)\f$ using known \f$p(h)\f$.
First of all the fact that \f$l\f$ is just transformed \f$h\f$ is needed to be understood.
In that way the light direction's PDF has following expression:
\f[p(l) = p(h)|J_T|\f]
Where \f$|J_T|\f$ is Jacobian of reflection transformation.
Lets find it.
Right now \f$n\f$ is axis from which \f$\theta\f$ spherical coordinate is encountered.
The first step is setting \f$v\f$ as starting point of \f$\theta\f$ instead of \f$n\f$.
This is linear transform so that \f$|J_T| = 1\f$.
Next step is transfer to spherical coordinate with \f$|J_T| = \sin\theta_{vh}\f$.
Due to previous step \f$\theta_{vh}\f$ is used instead of \f$\theta_h\f$.
In this coordinate system reflecting of \f$v\f$ relative to \f$h\f$ is just doubling \f$\theta_{vh}\f$ and Jacobian of it is equal to \f$\frac{1}{2}\f$.
In series of transform the Jacobians are multiplied so that currently \f$|J_T| = \frac{1}{2}\sin\theta_{vh}\f$.
And the final step is inverse transform to Cartesian coordinate system with \f$|J_T| = (\sin\theta_{vl})^{-1} = (\sin2\theta_{vh})^{-1}\f$.
Combining this all together the following expression is obtained for reflection transform Jacobian:
Here \f$G\f$ component is recommended to be calculated with original \f$k=\frac{\alpha}{2} = \frac{r^2}{2}\f$ in order to get more plausible result.
Of course, all \f$\cos\f$ must be clamped to range \f$[0, 1]\f$ because integral is calculated on a hemisphere and all expressions are defined on it.
\f$\cos\theta_v \cos\theta_{l_i}\f$ in denominator can be reduced with exactly the same part in geometric attenuation factor in order to avoid additional zero division cases.
Summarizing importance sampling strategy described above the convergence of Monte-Carlo integration can be improved using special PDF correlated with integrated function.
In case of BRDF with normal distribution functions \f$D\f$ the PDF producing procedure is defined.
Practically half vector \f$h\f$ is generated first and \f$l\f$ is obtained from it by view vector \f$v\f$ reflecting.
Due to this transformation final form of probability density used in sum is quite different but also has defined algorithm of calculation.
For isotropic Cook-Torrance BRDF the \f$\cos\theta_v\f$ and roughness \f$r\f$ are enough to start generation so that all integrals of that kind can be precalculated in 2D look-up tables variating these two parameters.
The same samples generation procedure must be used in specular map baking described in next chapter.
# Specular map
The situation with BRDF part of \f$L_{indirect}^s\f$ is clear now and \f$L_i^{indirect}(l)\f$ sum is left to be discussed.
That was called **specular map** and has following form:
\f[\frac{1}{N}\sum_{i=1}^N L_i^{indirect}(l_i)\f]
As was mentioned this sum must be calculated for every normal direction using the same samples generation principles as in numeric integration computation.
This principles require two scalar parameters \f$\cos\theta_v\f$ and \f$r\f$ but now \f$\phi\f$ really matters.
So that in fact the specular map has to be saved in 3D table consisting omnidirectional textures.
That is a big expense of computational and memory resources.
A couple of tricks helps to reduce dimensions.
First of all the \f$\cos\theta_v\f$ and \f$\phi\f$ can be just excluded.
In that way \f$v\f$ is considered to be equal to \f$n\f$.
Of course this approach produces an error and affects the final result.
It can be fixed more or less by \f$\cos\theta_{l_i}\f$ weighting:
It is not a complete solution but practice shows that it is enough to get plausible illumination with sacrificing of lengthy reflections at grazing angles which exist in fact if everything is honestly computed.
The problem is that for \f$v \neq n\f$ considering this sum to be defined related to \f$n\f$ became incorrect.
For example, for complete mirroring materials with \f$r = 0\f$ this sum must boil down to \f$L_i^{indirect}(v_r)\f$
but not to \f$L_i^{indirect}(n)\f$ where \f$v_r\f$ is just reflected \f$v\f$ or in other words \f$v_r = 2(v \cdot n)n - v\f$.
That it just mirroring reflection principle.
Assumption of \f$n = v\f$ also means that \f$n = v = v_r\f$.
In that way radiance map better to be considered as averaging of illumination coming from \f$v_r\f$.
So that it has become to be defined related to reflection direction which has to be calculated before map's fetching.
Anyway, there are just two dimensions in radiance look-up table remain.
The rest one with \f$r\f$ parameter cannot be reduced.
There is no other ways except just roughness variation but in order to simplify that computations can be done for several values and the rest ones lying between can be obtained from linear interpolation.
This is another source of visual artifacts but it also works good in practice and that is pretty common approach.
But it still requires noticeably amount of samples and that is for every pixel related to each \f$r\f$ value.
It can be appropriate for precomputations but still limits using dynamic environments in real time rendering or just even static environments but on weak devices such as mobile ones.
And there are several possible ways to improve radiance map baking performance.
The first one is using textures with smaller resolutions for larger roughnesses.
The point is that smaller \f$r\f$ values produce map saving more details from origin environment in opposite to larger ones representing lower frequency components and working as low pass filters in fact.
So less pixels in combination with linear interpolation is enough to store less detailed convolutions.
Moreover, this approach naturally works with textures levels of details in graphics API
so that every certain radiance map related to certain \f$r\f$ can be stored on its own mip level and be directly fetched with linear interpolation not only over one texture but over levels too.
As practice shows 6 levels are enough.
After reducing pixels count it is turn for samples number.
And again correlation with roughness can be noticed.
For example map for completely mirroring materials with \f$r = 0\f$ the same sample \f$l_i = v_r\f$ will be produced.
So that only one sample is enough in this case.
In opposite way directions for \f$r = 1\f$ will be scattered over almost whole hemisphere what requires as much samples as available.
The 'locality' of distribution is decreased during increasing roughness and it is possible to assume that samples number might to be proportional to this 'locality' keeping accuracy at the same level.
But how can 'locality' be interpreted in terms of probability distribution? One possible way is CDF meaning.
\f$F(\theta_h)\f$ has been already defined and by definition it shows the probability of random value \f$\theta_h\f$ to be less than argument of CDF.
In other words \f$F(\theta'_h) = u\f$ means that probability of \f$\theta_h\f$ to be in range of \f$[0, \theta'_h]\f$ is \f$u\f$.
The inverse task of range searching using given \f$u\f$ can be solved with help of \f$F^{-1}(u) = \theta'_h\f$.
If \f$u\f$ is close to 1 (exact 1 has no sense because in that case \f$\theta'_h = \max\theta_h = \frac{\pi}{2}\f$)
then \f$\theta'_h\f$ represents the range of the most probable or most frequently generated values and that can be interpreted as 'locality' of distribution.
After that if samples number of the worst case with \f$r = 1\f$ is set (\f$N(1) = \max N\f$) the other ones can be estimated using following formula:
It is approximate expression representing only estimated general proportionality so that cases of \f$r = 0\f$ and \f$r = 1\f$ must be processed separately with \f$N(0) = 1\f$ and \f$N(1) = \max N\f$.
\f$u\f$ can be parameter of this optimization strategy controlling speed of samples reducing in order to balance performance and quality (\f$u = 1\f$ disables this optimization at all).
This pretty tricky technique allows reducing calculations for every pixels without sacrificing the quality.
In addition to optimizations mentioned before another one can be applied in order to help to reduce samples numbers as previous one.
Using less samples produces image noise due to discrete nature of Monte-Carlo approximation.
But it can be slightly smoothed using some prefiltration.
The idea is that for the directions with small PDF or in other words for rare directions the samples near of it is unlikely to be generated.
So that this direction represents the averaged illumination from relatively big area on hemisphere but approximate it by just a constant.
It wold be better to get from such direction already averaged over bigger area environment.
It can be achieved using mip levels of origin \f$L_i^{indirect}\f$ whose pixels of one level is exact 4 averaged pixels from previous one.
Also mip levels generation is build in most common graphic API so there are no problems with it.
But first of all the area covered by one sample is needed to be found.
The mathematics connected with mip levels sizes lie behind it but this is out of scope of this paper.
In combination with previous optimization technique this approach allows \f$N(1)\f$ to be smaller keeping visual results good.
That is not all possible optimization tricks but at least these three significantly reduces compute efforts and brings radiance map calculation to weak devices or even to dynamic environments in real time but in reasonable limits.
In that way \f$L_{indirect}^s\f$ can be completely computed without any integral approximations.
Only 2D look-up table of BRDF part and mip mapped omnidirectional texture of irradiance map are needed.
The first one can be got even without any environment.
It was achieved using some rough approximations and assumptions but despite of that the visual result are still plausible and can be compared even with ray traced images.
In order to complete whole image based lighting the \f$L_{indirect}^d\f$ component is left to be discussed.
# Spherical harmonics
Lets go back to diffuse indirect illumination component represented by following formula:
It differs from origin one and loses accuracy a little bit but now there is no light direction inside
so that it can be considered as kind of screen space defined Fresnel's factor (\f$ss\f$ means exactly 'screen space') and can be removed from integral:
The resulted expression without \f$(1-m)\f$ and \f$(1-F_{ss})\f$ parts is pretty known entity called **irradiance**.
It can be precalculated using \f$\cos\theta_l\f$ as PDF for importance sampling (actually it is only option in this case excluding uniform distribution).
But even with that samples will be scattered almost over whole hemisphere.
As was discussed in previous chapter this case requires significant amount of samples in order to average illumination with appropriate quality.
Poor accuracy resulted from lack of summand can be noticed especially on high frequency environments having a lot of contrasting details.
It worth to be mentioned that irradiance is used not only in BRDF.
Omnidirectional diffuse illumination captured for certain point or even for several points uniformly or hierarchically distributed is base of some baking global illumination techniques.
There it is called a **light probe**. So that other way to compute and store irradiance maps was found resolving many mentioned problems.
The Fourier's decomposition analogue for spherical function allows to achieve this.
That would be easy to explain concept directly on example.
So lets start from \f$L_i^{indirect}(l)\f$.
It is spherical function because directions are just points on sphere.
@@ -32,7 +32,7 @@ Open CASCADE Technology is free software; you can redistribute it and / or modif
@ref license_lgpl_21 "GNU Lesser General Public License (LGPL) version 2.1", with additional @ref occt_lgpl_exception "exception".
Note that LGPL imposes some obligations on the application linked with Open CASCADE Technology.
If you wish to use OCCT in a proprietary application, please, pay special attention to address the requirements of LGPL section 6.
If you wish to use OCCT in a proprietary application, please pay a special attention to address the requirements of LGPL section 6.
At minimum the following should be considered:
1. Add the notice visible to the users of your application clearly stating that Open CASCADE Technology is used in this application, and that they have rights in this regard according to LGPL.
Such notice can be added in About dialog box (this is mandatory if this box contains copyright statements) or a similar place and/or in the documentation.
@@ -55,7 +55,7 @@ disclaimer.
You are hereby informed that all software is a property of its respective authors and is protected by
international and domestic laws on intellectual property and trademarks.
Should you need further information, please directly contact the authors.
Should you need further information, directly contact the authors.
**CAS.CADE** and **Open CASCADE** are registered trademarks of
OPEN CASCADE S.A.S.
@@ -80,7 +80,7 @@ and, in case you need any further information, directly contact their authors.
**Qt** is a cross-platform application framework that is widely used for developing application software
with graphical user interface (GUI). Qt is free and open source software distributed under
the terms of the GNU Lesser General Public License. In OCCT Qt is used for programming samples.
If you need further information on Qt, please, refer to Qt Homepage (https://www.qt.io/)
If you need further information on Qt, refer to Qt Homepage (https://www.qt.io/)
**Tcl** is a high-level programming language. Tk is a graphical user interface (GUI) toolkit,
with buttons, menus, listboxes, scrollbars, and so on. Taken together Tcl and Tk provide a solution
@@ -88,11 +88,7 @@ to develop cross-platform graphical user interfaces with a native look and feel.
Scriptics Corp., Sun Microsystems, and other companies. However, Tcl/Tk is an open source, and
the copyright allows you to use, modify, and redistribute Tcl/Tk for any purpose, without an
explicit license agreement and without paying any license fees or royalties.
To use Tcl/Tk, please refer to the Licensing Terms (https://www.tcl.tk/software/tcltk/license.html).
**GL2PS** is developed by Christophe Geuzaine and others. It is optionally used by OCCT to
export content of OpenGL scene to vector graphics formats (PS, PDF, EMF, SVG).
The library is licensed under GL2PS license (http://www.geuz.org/gl2ps/COPYING.GL2PS).
To use Tcl/Tk, refer to the Licensing Terms (https://www.tcl.tk/software/tcltk/license.html).
**FreeType 2** is developed by Antoine Leca, David Turner, Werner Lemberg and others.
It is a software font engine that is designed to be small, efficient, highly customizable and
@@ -113,18 +109,18 @@ implementation of 3D viewer. OpenGL specification is developed by the
Khronos group, https://www.khronos.org/opengl/. OCCT code includes header
file *glext.h* obtained from Khronos web site.
**VTK** -- The **Visualization Toolkit (VTK)** is an open-source, freely available software system for 3D computer graphics, image processing and visualization. OCCT VIS component provides adaptation functionality for visualization of OCCT topological shapes by means of VTK library. If you need further information on VTK, please, refer to VTK Homepage http://www.vtk.org/.
**VTK** -- The **Visualization Toolkit (VTK)** is an open-source, freely available software system for 3D computer graphics, image processing and visualization. OCCT VIS component provides adaptation functionality for visualization of OCCT topological shapes by means of VTK library. If you need further information on VTK, refer to VTK Homepage https://www.vtk.org/.
**Doxygen** developed by Dimitri van Heesch is open source documentation system for
C++, C, Java, Objective-C, Python, IDL, PHP and C#. This product is used in Open CASCADE Technology
for automatic creation of Technical Documentation from C++ header files.
If you need further information on Doxygen, please refer to https://www.stack.nl/~dimitri/doxygen/index.html.
If you need further information on Doxygen, refer to https://www.stack.nl/~dimitri/doxygen/index.html.
**Graphviz** is open source graph visualization software developed by John Ellson, Emden Gansner, Yifan Hu and Arif Bilgin.
Graph visualization is representiation of structured information as diagrams of abstract graphs and networks.
This product is used together with Doxygen in Open CASCADE Technology for automatic creation of Technical Documentation
(generation of dependency graphs). Current versions of Graphviz are licensed on an open source
basis under The Eclipse Public License (EPL) (http://www.graphviz.org/License.php).
basis under The Eclipse Public License (EPL) (https://www.graphviz.org/license/).
**Inno Setup** is a free script-driven installation system created in CodeGear Delphi by Jordan Russell.
In OCCT Inno Setup is used to create Installation Wizard on Windows.
@@ -134,10 +130,10 @@ It is licensed under Inno Setup License (http://www.jrsoftware.org/files/is/lice
and others used by multimedia applications. This library is developed by Hervé Drolon and Floris van den Berg.
FreeImage is easy to use, fast, multithreading safe, compatible with all 32-bit or 64-bit versions of Windows,
and cross-platform (works both with Linux and Mac OS X). FreeImage is optionally used by OCCT to work
with images, on conditions of the FreeImage Public License (FIPL) (http://freeimage.sourceforge.net/freeimage-license.txt).
with images, on conditions of the FreeImage Public License (FIPL) (https://freeimage.sourceforge.net/freeimage-license.txt).
**David M. Gay's floating point routines** (dtoa.c) are used for fast reading of floating point values from text strings.
These routines are available under MIT-like license (see http://www.netlib.org/fp/).
These routines are available under MIT-like license (see https://www.netlib.org/fp/).
**CMake** is an open-source, cross-platform family of tools designed to build, test and package software. CMake is used to control the software compilation process using simple platform and compiler independent configuration files, and generate native makefiles and workspaces that can be used in the compiler environment of your choice.
OCCT uses CMake as a build system. CMake is available under BSD 3-Clause license. See more at https://cmake.org/
@@ -154,6 +150,12 @@ FFmpeg is optionally used by OCCT for video recording, on LGPL conditions (https
for generation of User and Developer Guides in PDF format. See https://miktex.org for information
on this tool.
**RapidJSON** is an Open Source JSON parser and generator for C++.
RapidJSON is optionally used by OCCT for reading glTF files (https://rapidjson.org/).
**DejaVu** fonts are a font family based on the Vera Fonts under a permissive license (MIT-like, https://dejavu-fonts.github.io/License.html).
DejaVu Sans (basic Latin sub-set) is used by OCCT as fallback font when no system font is available.
Adobe Systems, Inc. provides **Adobe Reader**, which can be used to view files in Portable Document Format (PDF).
@section OCCT_OVW_SECTION_3 Documentation
@@ -208,13 +210,13 @@ for which OCCT is certified to work.
| Doxygen (optional for building documentation) | Doxygen 1.8.5+ https://www.stack.nl/~dimitri/doxygen/download.html |
@subsection overview_req_hw Hardware
@@ -286,13 +288,13 @@ When the installation is complete, you will find the directories for 3rd party p
@figure{/overview/images/overview_3rdparty.png}
The contents of the OCCT-7.3.0 directory (called further "OCCT root", or $CASROOT) are as follows:
The contents of the OCCT-7.4.0 directory (called further "OCCT root", or $CASROOT) are as follows:
@figure{/overview/images/overview_installation.png, "The directory tree"}
* **adm** This folder contains administration files, which allow rebuilding OCCT;
* **adm/cmake** This folder contains files of CMake building procedure;
* **adm/msvc** This folder contains Visual Studio projects for Visual C++ 2010, 2012, 2013, 2015 and 2017 which allow rebuilding OCCT under Windows platform in 32 and 64-bit mode;
* **adm/msvc** This folder contains Visual Studio projects for Visual C++ 2010, 2012, 2013, 2015, 2017 and 2019 which allow rebuilding OCCT under Windows platform in 32 and 64-bit mode;
* **data** This folder contains CAD files in different formats, which can be used to test the OCCT functionality;
* **doc** This folder contains OCCT documentation in HTML and PDF format;
* **dox** This folder contains sources of OCCT documentation in plain text (MarkDown) format;
@@ -311,7 +313,7 @@ To run any Open CASCADE Technology application you need to set the environment v
You can define the environment variables with env.bat script located in the
$CASROOT folder. This script accepts two arguments to be used:
the version of Visual Studio (vc10 -- vc141) and the architecture (win32 or win64).
the version of Visual Studio (vc10 -- vc142) and the architecture (win32 or win64).
The additional environment settings necessary for compiling OCCT libraries and samples
by Microsoft Visual Studio can be set using script custom.bat located in the same folder.
@@ -334,7 +336,7 @@ The scripts are located in the OCCT root folder.
* **CASROOT** is used to define the root directory of Open CASCADE Technology;
* **PATH** is required to define the path to OCCT binaries and 3rdparty folder;
* **LD_LIBRARY_PATH** is required to define the path to OCCT libraries (on UNIX platforms only);
* **LD_LIBRARY_PATH** is required to define the path to OCCT libraries (on UNIX platforms only; **DYLD_LIBRARY_PATH** variable in case of macOS);
* **MMGT_OPT** (optional) if set to 1, the memory manager performs optimizations as described below; if set to 2,
Intel (R) TBB optimized memory manager is used; if 0 (default), every memory block is allocated
in C memory heap directly (via malloc() and free() functions).
@@ -401,7 +403,7 @@ the specified resource file and activates the commands implemented in the plug-i
The whole process of using the plug-in mechanism as well as the instructions for extending Test Harness is described in the @ref occt_user_guides__test_harness.
Draw Test Harness provides an environment for OCCT automated testing system.
Please, consult its @ref occt_dev_guides__tests "Automated Testing System" for details.
Check its @ref occt_dev_guides__tests "Automated Testing System" for details.
Remarks:
@@ -415,8 +417,7 @@ Remarks:
**On Linux:**
1. If OCCT was built by Code::Blocks use <i>$CASROOT/draw_cbp.sh</i> file to launch *DRAWEXE* executable;
2. If OCCT was built by Automake use <i>$CASROOT/draw_amk.sh</i> file to launch *DRAWEXE* executable;
* If OCCT was built by Code::Blocks use <i>$CASROOT/draw.sh</i> file to launch *DRAWEXE* executable.
Draw[1]> prompt appears in the command window
@@ -572,3 +573,11 @@ There is a sample demonstrating usage of OCCT on iOS with Apple UIKit framework.
@figure{/overview/images/sample_ios_uikit.png}
See \subpage occt_samples_ios_uikit "iOS sample Readme" for details.
@subsubsection OCCT_OVW_SECTION_7_3_6 Web
WebGL Viewer sample demonstrating usage of OCCT 3D Viewer in Web browser with Emscripten SDK can be found in `samples/webgl`.
@figure{/overview/images/sample_webgl.png}
See \subpage occt_samples_webgl "WebGL sample Readme" for details.
@@ -26,7 +26,7 @@ This modular structure is illustrated in the diagram below.
*@ref OCCT_TOVW_SECTION_2 "Foundation Classes" module underlies all other OCCT classes;
*@ref OCCT_TOVW_SECTION_3 "Modeling Data" module supplies data structures to represent 2D and 3D geometric primitives and their compositions into CAD models;
*@ref OCCT_TOVW_SECTION_4 "Modeling Algorithms" module contains a vast range of geometrical and topological algorithms;
*@ref OCCT_TOVW_SECTION_4a "Mesh" module implements tessellated representations of objects;
*@ref OCCT_TOVW_SECTION_4a "Mesh" toolkit from "Modeling Algorithms" module implements tessellated representations of objects;
*@ref OCCT_TOVW_SECTION_5 "Visualization" module provides complex mechanisms for graphical data representation;
*@ref OCCT_TOVW_SECTION_6 "Data Exchange" module inter-operates with popular data formats and relies on @ref OCCT_TOVW_SECTION_6a "Shape Healing" to improve compatibility between CAD software of different vendors;
*@ref OCCT_TOVW_SECTION_7 "Application Framework" module offers ready-to-use solutions for handling application-specific data (user attributes) and commonly used functionality (save/restore, undo/redo, copy/paste, tracking CAD modifications, etc).
@@ -38,196 +38,207 @@ In addition, @ref OCCT_TOVW_SECTION_8 "Open CASCADE Test Harness", also called D
**Foundation Classes** module contains data structures and services used by higher-level Open CASCADE Technology classes:
* Primitive types, such as Boolean, Character, Integer or Real;
* String classes that handle ASCII and Unicode strings;
* String classes that handle Unicode strings;
* Collection classes that handle statically or dynamically sized aggregates of data, such as arrays, lists, queues, sets and hash tables (data maps).
* Classes providing commonly used numerical algorithms and basic linear algebra calculations (addition, multiplication, transposition of vectors and matrices, solving linear systems etc).
* Fundamental types representing physical quantities and supporting date and time information;
* Fundamental types like color, date and time information;
* Primitive geometry types providing implementation of basic geometric and algebraic entities that define and manipulate elementary data structures.
* Exception classes that describe situations, when the normal execution of program is abandoned;
This module also provides a variety of general-purpose services, such as:
* Safe handling of dynamically created objects, ensuring automatic deletion of unreferenced objects (smart pointers);
*Configurable optimized memory manager increasing the performance of applications that intensively use dynamically created objects;
*Standard and specialized memory allocators;
* Extended run-time type information (RTTI) mechanism maintaining a full type hierarchy and providing means to iterate over it;
* Encapsulation of C++ streams;
* Automated management of heap memory by means of specific allocators;
* Basic interpreter of expressions facilitating the creation of customized scripting tools, generic definition of expressions, etc.;
* Tools for dealing with configuration resource files and customizable message files facilitating multi-language support in applications;
* Tools for dealing with configuration resource files and customizable message files facilitating multi-language support in applications;
* Progress indication and user break interfaces, giving a possibility even for low-level algorithms to communicate with the user in a universal and convenient way;
* and many others...
Please, see the details in @ref occt_user_guides__foundation_classes "Foundation Classes User's Guide"
See the details in @ref occt_user_guides__foundation_classes "Foundation Classes User's Guide"
See also: our <ahref="http://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
See also: our <ahref="https://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
@section OCCT_TOVW_SECTION_3 Modeling Data
**Modeling Data** supplies data structures to implement boundary representation (BRep) of objects in 3D. In BRep the shape is represented as an aggregation of geometry within topology. The geometry is understood as a mathematical description of a shape, e.g. as curves and surfaces (simple or canonical, Bezier, NURBS, etc). The topology is a data structure binding geometrical objects together.
**Modeling Data** supplies data structures to implement boundary representation (BRep) of objects in 3D.
In BRep the shape is represented as an aggregation of geometry within topology.
The geometry is understood as a mathematical description of a shape, e.g. as curves and surfaces (simple or canonical, Bezier, NURBS, etc).
The topology is a data structure binding geometrical objects together.
Geometry types and utilities provide geometric data structures and services for:
* Description of points, vectors, curves and surfaces:
* their positioning in 3D space using axis or coordinate systems, and
* their geometric transformation, by applying translations, rotations, symmetries, scaling transformations and combinations thereof.
* their positioning in 3D space using axis or coordinate systems, and
* their geometric transformation, by applying translations, rotations, symmetries, scaling transformations and combinations thereof.
* Creation of parametric curves and surfaces by interpolation and approximation;
* Algorithms of direct construction;
* Conversion of curves and surfaces to NURBS form;
* Computation of point coordinates on 2D and 3D curves;
* Calculation of extrema between geometric objects.
Topology defines relationships between simple geometric entities. A shape, which is a basic topological entity, can be divided into components (sub-shapes):
Topology defines relationships between simple geometric entities.
A shape, which is a basic topological entity, can be divided into components (sub-shapes):
* Vertex -- a zero-dimensional shape corresponding to a point;
* Edge -- a shape corresponding to a curve and bounded by a vertex at each extremity;
* Wire -- a sequence of edges connected by their vertices;
* Face -- a part of a plane (in 2D) or a surface (in 3D) bounded by wires;
* Shell -- a collection of faces connected by edges of their wire boundaries;
* Solid -- a finite closed part of 3D space bounded by shells;
* Compound solid -- a collection of solids connected by faces of their shell boundaries.
* Composite solid -- a collection of solids connected by faces of their shell boundaries;
* Compound -- a collection of shapes of arbitrary type.
Complex shapes can be defined as assemblies of simpler entities.
Complex shapes can be defined as assemblies (compounds) of simpler entities.
Please, see the details in @ref occt_user_guides__modeling_data "Modeling Data User's Guide"
See the details in @ref occt_user_guides__modeling_data "Modeling Data User's Guide"
3D geometric models can be stored in OCCT native BREP format.
See @ref occt_user_guides__brep_wp "BREP Format Description White Paper" for details on the format.
See also: our <ahref="http://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
See also: our <ahref="https://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
@section OCCT_TOVW_SECTION_4 Modeling Algorithms
**Modeling Algorithms** module groups a wide range of topological and geometric algorithms used in geometric modeling. Basically, there are two groups of algorithms in Open CASCADE Technology:
* High-level modeling routines used in the real design;
*Low-level mathematical support functions used as a groundwork for the modeling API;
**Modeling Algorithms** module groups a wide range of topological and geometric algorithms used in geometric modeling.
Basically, there are two groups of algorithms in Open CASCADE Technology:
*High-level modeling routines used in the real design;
* Low-level mathematical support functions used as a groundwork for the modeling API.
*Low-level geometric tools provide the algorithms, which:
* Calculate the intersection of two curves, surfaces, or a curve and a surface;
* Project points onto 2D and 3D curves, points onto surfaces and 3D curves onto surfaces;
* Construct lines and circles from constraints;
* Construct free-form curves and surfaces from constraints (interpolation, approximation, skinning, gap filling, etc);
Low-level *geometric tools* provide the algorithms, which:
* Calculate the intersection of two curves, surfaces, or a curve and a surface;
* Project points onto 2D and 3D curves, points onto surfaces and 3D curves onto surfaces;
* Construct lines and circles from constraints;
* Construct free-form curves and surfaces from constraints (interpolation, approximation, skinning, gap filling, etc).
*Low-level topological tools provide the algorithms, which:
* Tessellate shapes;
* Check correct definition of shapes;
* Determine the local and global properties of shapes (derivatives, mass-inertia properties, etc);
* Perform affine transformations;
* Find planes in which edges are located;
* Convert shapes to NURBS geometry;
* Sew connected topologies (shells and wires) from separate topological elements (faces and edges).
Low-level *topological tools* provide the algorithms, which:
* Tessellate shapes;
* Check correct definition of shapes;
* Determine the local and global properties of shapes (derivatives, mass-inertia properties, etc);
* Perform affine transformations;
* Find planes in which edges are located;
* Convert shapes to NURBS geometry;
* Sew connected topologies (shells and wires) from separate topological elements (faces and edges).
Top-level API provides the following functionality:
*Construction of Primitives:
*Boxes;
*Prisms;
* Cylinders;
*Cones;
*Spheres;
*Toruses.
*Kinematic Modeling:
*Prisms -- linear sweeps;
*Revolutions -- rotational sweeps;
*Pipes -- general-form sweeps;
* Lofting.
* Construction of Primitives:
*Boxes;
*Prisms;
*Cylinders;
* Cones;
*Spheres;
*Toruses.
*Kinematic Modeling:
*Prisms -- linear sweeps;
*Revolutions -- rotational sweeps;
*Pipes -- general-form sweeps;
*Lofting.
@figure{/technical_overview/images/0001.png "Shapes containing pipes with variable radius produced by sweeping"}
* Boolean Operations, which allow creating new shapes from the combinations of source shapes. For two shapes *S1* and *S2*:
* *Common* contains all points that are in *S1* and *S2*;
* *Fuse* contains all points that are in *S1* or *S2*;
* *Cut* contains all points in that are in *S1* and not in *S2*
* Boolean Operations, which allow creating new shapes from the combinations of source shapes. For two shapes *S1* and *S2*:
* *Common* contains all points that are in *S1* and *S2*;
* *Fuse* contains all points that are in *S1* or *S2*;
* *Cut* contains all points in that are in *S1* and not in *S2*.
See @ref occt_user_guides__boolean_operations "Boolean Operations" User's Guide for detailed documentation.
* Algorithms for local modifications such as:
* Hollowing;
* Shelling;
* Creation of tapered shapes using draft angles;
* Algorithms to make fillets and chamfers on shape edges, including those with variable radius (chord).
* Algorithms for local modifications such as:
* Hollowing;
* Shelling;
* Creation of tapered shapes using draft angles;
* Algorithms to make fillets and chamfers on shape edges, including those with variable radius (chord).
* Algorithms for creation of mechanical features, i.e. depressions, protrusions, ribs and grooves or slots along planar or revolution surfaces.
* Algorithms for creation of mechanical features, i.e. depressions, protrusions, ribs and grooves or slots along planar or revolution surfaces.
@figure{/technical_overview/images/0004.png}
Please, see the details in @ref occt_user_guides__modeling_algos "Modeling Algorithms User's Guide".
See the details in @ref occt_user_guides__modeling_algos "Modeling Algorithms User's Guide".
See also: our <ahref="http://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
See also: our <ahref="https://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
@section OCCT_TOVW_SECTION_4a Mesh
@subsection OCCT_TOVW_SECTION_4a Mesh
**Mesh**module provides the functionality to work with tessellated representations of objects in form of triangular facets. This module contains:
**Mesh**toolkit provides the functionality to work with tessellated representations of objects in form of triangular facets. This toolkit contains:
- data structures to store surface mesh data associated to shapes and basic algorithms to handle them;
- data structures and algorithms to a build triangular surface mesh from *BRep* objects (shapes);
- data structures and algorithms to build triangular surface mesh from *BRep* objects (shapes);
- tools for displaying meshes with associated pre- and post-processor data (scalars or vectors).
Open CASCADE Technology includes two mesh converters:
- VRML converter translates Open CASCADE shapes to VRML 1.0 files (Virtual Reality Modeling Language). Two representation modes are possible: shaded, which presents shapes as sets of triangles computed by the mesh algorithm, or wireframe, which presents shapes as sets of curves.
- STL converter translates Open CASCADE shapes to STL files. STL (STtereoLithography) format is widely used for rapid prototyping (3D printing).
Open CASCADE SAS also offers Advanced Mesh Products:
**Visualization** module provides ready-to-use algorithms to create graphic presentations from various objects: shapes, meshes, etc.
In Open CASCADE Technology visualization is based on the separation of CAD data and its graphical presentation. The presentations can be customized to take the specificity of your application into account.
**Visualization** module provides ready-to-use algorithms to create graphic presentations from various objects: shapes, meshes, etc.
In Open CASCADE Technology visualization is based on the separation of CAD data and its graphical presentation.
The module also supports a fast and powerful interactive selection mechanism.
The view facilities provided by OCCT range from low-level tools working with basic geometry and topology (such as NURBS visualization with control points and nodes, rendering of isolines to estimate speed and quality of parameterization, or rendering of a parametric profile of edges) to high-level tools for real time quality rendering of models using ray tracing: shades, reflections, transparency, anti-aliasing, etc.
Visualization module relies on the following key toolkits:
- *TKV3d* toolkit defines a high-level API called (Application Interactive Services* (AIS) for working with interactive objects.
- *TKService* toolkit defines a low-level API for managing and creating presentations from primitive arrays.
This toolkit defines an abstraction layer for defining an arbitrary graphic driver responsible for actual rendering.
- *TKOpenGl* toolkit implements the graphic driver using OpenGL and OpenGL ES libraries.
Here are just a few examples:
* Camera-driven view projection and orientation. It is possible to choose between perspective, orthographic and stereographic projection.
* Real-time ray tracing technique using recursive Whitted's algorithm and Bounded Volume Hierarchy effective optimization structure.
@figure{/technical_overview/images/0002.png, "Real time visualization by ray tracing method"}
* Support of GLSL shaders. The shader management is fully automatic, like with any other OpenGL resource.
* Support of standard and custom materials, defined by transparency, diffuse, ambient and specular reflection and refraction index. The latter allows implementing transparent materials, such as glass, diamond and water.
@figure{/technical_overview/images/0012.png, "Simulation of a glass cover"}
* Optimization of rendering performance through the algorithms of:
* View frustum culling, which skips the presentation outside camera at the rendering stage and
* Back face culling, which reduces the rendered number of triangles and eliminates artifacts at shape boundaries.
* Definition of clipping planes through the plane equation coefficients. Ability to define visual attributes for cross-section at the level or individual clipping planes. In the image below different parts of the rocket are clipped with different planes and hatched.
* Possibility to flexibly adjust appearance of dimensions in a 3D view. The 3D text object represents a given text string as a true 3D object in the model space.
While low-level API operates with primitive arrays (triangles, lines, points), the higher level includes services for building presentations for B-Rep shapes (shaded and wireframe).
A comprehensive list of standard interactive objects includes topological shape, mesh presentation, various dimensions, manipulators and others.
It provides a solid basis for rapid application development, while flexible and extensible API allows development of highly customized application-specific presentations.
Here are a few examples of OCCT Visualization features:
* Camera-driven view projection and orientation.
Perspective, orthographic and stereographic projections are supported.
* Support of Common (diffuse/ambient/specular) and PBR metallic-roughness material models.
* Possibility to flexibly adjust appearance of dimensions in a 3D view.
The 3D text object represents a given text string as a true 3D object in the model space.
* Definition of clipping planes through the plane equation coefficients.
Ability to define visual attributes for cross-section at the level or individual clipping planes.
In the image below different parts of the rocket are clipped with different planes and hatched.
@figure{/technical_overview/images/0008.png, "Display of shape cross-section and dimensions"}
For more details see @ref occt_user_guides__visualization "Visualization User's Guide".
* Support of built-in and application-specific GLSL shaders.
* Optimization of rendering performance through the algorithms of:
* View frustum culling, which skips the presentation outside camera at the rendering stage;
* Back face culling, which reduces the rendered number of triangles and eliminates artifacts at shape boundaries.
* Real-time ray tracing technique using recursive Whitted's algorithm and Bounded Volume Hierarchy effective optimization structure.
@figure{/technical_overview/images/0002.png, "Real time visualization by ray tracing method"}
@figure{/technical_overview/images/0012.png, "Simulation of a glass cover"}
For more details, see @ref occt_user_guides__visualization "Visualization User's Guide".
The visualization of OCCT topological shapes by means of VTK library provided by VIS component is described in a separate @ref occt_user_guides__vis "VTK Integration Services" User's Guide.
See also: our <ahref="http://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
See also: our <ahref="https://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
@section OCCT_TOVW_SECTION_6 Data Exchange
**Data Exchange** allows developing OCCT-based applications that can interact with other CAD systems by writing and reading CAD models to and from external data. The exchanges run smoothly regardless of the quality of external data or requirements to its internal representation, for example, to the data types, accepted geometric inaccuracies, etc.
**Data Exchange** allows developing OCCT-based applications that can interact with other CAD systems by writing and reading CAD models to and from external data.
@figure{/technical_overview/images/0014.png,"Shape imported from STEP"}
**Data Exchange** is organized in a modular way as a set of interfaces that comply with various CAD formats: IGES, STEP, STL, VRML, etc. The interfaces allow software based on OCCT to exchange data with various CAD/PDM software packages, maintaining a good level of interoperability.
**Data Exchange** is organized in a modular way as a set of interfaces that comply with various CAD formats: IGES, STEP, STL, VRML, etc.
The interfaces allow software based on OCCT to exchange data with various CAD/PDM software packages, maintaining a good level of interoperability.
This module handles various problems of interoperability between CAD systems, caused by differences in model validity criteria and requirements to internal representation.
* **Standardized Data Exchange** interfaces allow querying and examining the input file, converting its contents to a CAD model and running validity checks on a fully translated shape. The following formats are currently supported.
*@ref occt_user_guides__step "STEP" (AP203 : Mechanical Design, this covers General 3D CAD; AP214: Automotive Design)
*@ref occt_user_guides__iges "IGES" (up to 5.3)
*VRML and STL meshes.
* **Standardized Data Exchange** interfaces allow querying and examining the input file, converting its contents to a CAD model and running validity checks on a fully translated shape.
The following formats are currently supported:
*@ref occt_user_guides__step "STEP" (AP203: Mechanical Design, this covers General 3D CAD; AP214: Automotive Design; AP242).
*@ref occt_user_guides__iges "IGES" (up to 5.3).
* **glTF** 2.0 reader and writer.
* **OBJ** mesh file reader.
* **VRML** converter translates Open CASCADE shapes to VRML 1.0 files (Virtual Reality Modeling Language).
* **STL** converter translates Open CASCADE shapes to STL files.
STL (STtereoLithography) format is widely used for rapid prototyping (3D printing).
*@ref occt_user_guides__xde "Extended data exchange" (XDE) allows translating additional attributes attached to geometric data (colors, layers, names, materials etc).
*<ahref="http://www.opencascade.com/content/advanced-data-exchange-components">Advanced Data Exchange Components</a> are available in addition to standard Data Exchange interfaces to support interoperability and data adaptation (also using @ref OCCT_TOVW_SECTION_6a "Shape Healing") with CAD software using the following proprietary formats:
*<ahref="https://www.opencascade.com/content/advanced-data-exchange-components">Advanced Data Exchange Components</a>
are available in addition to standard Data Exchange interfaces to support interoperability and data adaptation (also using @ref OCCT_TOVW_SECTION_6a "Shape Healing") with CAD software using the following proprietary formats:
These components are based on the same architecture as interfaces with STEP and IGES.
@@ -236,22 +247,22 @@ These components are based on the same architecture as interfaces with STEP and
**Shape Healing** library provides algorithms to correct and adapt the geometry and topology of shapes imported to OCCT from other CAD systems.
Shape Healing algorithms include, but are not limited to, the following operations:
*analyze shape characteristics and, in particular, identify the shapes that do not comply with OCCT geometry and topology validity rules by analyzing geometrical objects and topology:
- check edge and wire consistency;
- check edge order in a wire;
- check the orientation of face boundaries;
- analyze shape tolerances;
- identify closed and open wires in a boundary.
*fix incorrect or incomplete shapes:
- provide consistency between a 3D curve and its corresponding parametric curve;
- repair defective wires;
- fit the shapes to a user-defined tolerance value;
- fill gaps between patches and edges.
*upgrade and change shape characteristics:
- reduce curve and surface degree;
- split shapes to obtain C1 continuity;
- convert any types of curves or surfaces to Bezier or B-Spline curves or surfaces and back;
- split closed surfaces and revolution surfaces.
*Analyze shape characteristics and, in particular, identify the shapes that do not comply with OCCT geometry and topology validity rules by analyzing geometrical objects and topology:
- check edge and wire consistency;
- check edge order in a wire;
- check the orientation of face boundaries;
- analyze shape tolerances;
- identify closed and open wires in a boundary.
*Fix incorrect or incomplete shapes:
- provide consistency between a 3D curve and its corresponding parametric curve;
- repair defective wires;
- fit the shapes to a user-defined tolerance value;
- fill gaps between patches and edges.
*Upgrade and change shape characteristics:
- reduce curve and surface degree;
- split shapes to obtain C1 continuity;
- convert any types of curves or surfaces to Bezier or B-Spline curves or surfaces and back;
- split closed surfaces and revolution surfaces.
Each sub-domain of Shape Healing has its own scope of functionality:
@@ -263,36 +274,40 @@ Each sub-domain of Shape Healing has its own scope of functionality:
| Customization | Modifies the shape representation to fit specific needs. | The shape is not modified, only the mathematical form of its internal representation is changed. |
| Processing | Mechanism of shape modification via a user-editable resource file. | |
For more details refer to @ref occt_user_guides__shape_healing "Shape Healing User's guide".
See also: our <ahref="http://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
For more details, refer to @ref occt_user_guides__shape_healing "Shape Healing User's guide".
See also: our <ahref="https://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
**Open CASCADE Application Framework** (OCAF) handles Application Data basing on the Application/Document paradigm. It uses an associativity engine to simplify the development of a CAD application thanks to the following ready-to-use features and services:
**Open CASCADE Application Framework** (OCAF) handles Application Data basing on the Application/Document paradigm.
It uses an associativity engine to simplify the development of a CAD application thanks to the following ready-to-use features and services:
* Data attributes managing the application data, which can be organized according to the development needs;
* Data storage and persistence (open/save);
* Possibility to modify and recompute attributes in documents. With OCAF it is easy to represent the history of modification and parametric dependencies within your model;
* Possibility to modify and recompute attributes in documents.
With OCAF it is easy to represent the history of modification and parametric dependencies within your model;
* Possibility to manage multiple documents;
* Predefined attributes common to CAD/CAM/CAE applications (e.g. to store dimensions);
* Undo-Redo and Copy-Paste functions.
Since OCAF handles the application structure, the only development task is the creation of application-specific data and GUIs.
OCAF differs from any other CAD framework in the organization of application data, as there the data structures are based on reference keys rather than on shapes. In a model, such attributes as shape data, color and material are attached to an invariant structure, which is deeper than the shapes. A shape object becomes the value of *Shape* attribute, in the same way as an integer number is the value of *Integer* attribute and a string is the value of *Name* attribute.
OCAF differs from any other CAD framework in the organization of application data, as there the data structures are based on reference keys rather than on shapes.
In a model, such attributes as shape data, color and material are attached to an invariant structure, which is deeper than the shapes.
A shape object becomes the value of *Shape* attribute, in the same way as an integer number is the value of *Integer* attribute and a string is the value of *Name* attribute.
OCAF organizes and embeds these attributes in a document. OCAF documents, in their turn, are managed by an OCAF application.
For more details see @ref occt_user_guides__ocaf "OCAF User's Guide".
See also: our <ahref="http://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
For more details, see @ref occt_user_guides__ocaf "OCAF User's Guide".
See also: our <ahref="https://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
@section OCCT_TOVW_SECTION_8 Draw Test Harness
**Test Harness** or **Draw** is a convenient testing tool for OCCT libraries. It can be used to test and prototype various algorithms before building an entire application. It includes:
**Test Harness** or **Draw** is a convenient testing tool for OCCT libraries.
It can be used to test and prototype various algorithms before building an entire application.
It includes:
- A command interpreter based on the TCL language;
- A number of 2D and 3D viewers;
- A set of predefined commands.
@@ -310,4 +325,4 @@ In addition, **Test Harness** provides commands to create and manipulate curves
You can add custom commands to test or demonstrate any new functionalities, which you develop.
For more details see @ref occt_user_guides__test_harness "Draw Test Harness Manual".
For more details, see @ref occt_user_guides__test_harness "Draw Test Harness Manual".
We hope that this tutorial has provided you with a feel for the industrial strength power of Open CASCADE Technology.
If you want to know more and develop major projects using Open CASCADE Technology, we invite you to study our training, support, and consulting services on our site at http://www.opencascade.com/content/technology-support. Our professional services can maximize the power of your Open CASCADE Technology applications.
If you want to know more and develop major projects using Open CASCADE Technology, we invite you to study our training, support, and consulting services on our site at https://www.opencascade.com/content/technology-support. Our professional services can maximize the power of your Open CASCADE Technology applications.
| 2.4 | Add the solids <i>(SDi)</i> to the result | |
@subsection occt_algorithms_bop_on_opensolids Boolean operations on open solids
The Boolean operations on open solids are tricky enough that the standard approach of Boolean operations for building the result, based on the splits of solids does not work.
It happens because the algorithm for splitting solids (*BOPAlgo_BuilderSolid*) always tries to create the closed loops (shells) and make solids from them. But if the input solid is not closed, what can be expected from its splits?
For performing Boolean Operations on open solids another approach is used, which does not rely on the splits of the solids to be correct, but tries to select the splits of faces, which are necessary for the given type of operation.
The point here is that the type of Boolean operation clearly defines the states for the faces to be taken into result:
- For **COMMON** operation all the faces from the arguments located inside any solid of the opposite group must be taken;
- For **FUSE** operation all the faces from the arguments located outside of all solids of the opposite group must be taken;
- For **CUT** operation all the faces from the Objects located outside of all solids of the Tools and all faces from the Tools located inside any solid of the Objects must be taken;
- For **CUT21** operation all the faces from the Objects located inside any solid of the Tools and all faces from the Tools located outside of all solids of the Objects must be taken.
From the selected faces the result solids are built. Please note, that the result may contain as normal (closed) solids as the open ones.
Even with this approach, the correct result of Boolean operation on open solids cannot be always guaranteed.
This is explained by non-manifold nature of open solids: in some cases classification of a face depends on the point of the face chosen for classification.
@section occt_algorithms_10a Section Algorithm
@subsection occt_algorithms_10a_1 Arguments
@@ -2428,7 +2443,7 @@ A lot of failures of GFA algorithm can be caused by bugs in low-level algorithms
* The Projection Algorithm is used at the Intersection step. The purpose of Projection Algorithm is to compute 2D curves on surfaces. Wrong results here lead to incorrect or missing faces in the final GFA result.
* The Classification Algorithm is used at the Building step. The bugs in the Classification Algorithm lead to errors in selecting shape parts (edges, faces, solids) and ultimately to a wrong final GFA result.
The description below illustrates some known GFA limitations. It does not enumerate exhaustively all problems that can arise in practice. Please, address cases of Algorithm failure to the OCCT Maintenance Service.
The description below illustrates some known GFA limitations. It does not enumerate exhaustively all problems that can arise in practice. Please address cases of Algorithm failure to the OCCT Maintenance Service.
@subsection occt_algorithms_10_1 Arguments
@@ -2524,7 +2539,7 @@ Let us also consider a cylinder-based *Face 2* with radii *R=3000* and *H=6000*
@figure{/user_guides/boolean_operations/images/operations_image047.png,"P-Curves for Face 2",230}
Please, pay attention to the Zoom value of the Figures.
Pay attention to the Zoom value of the Figures.
It is obvious that starting with some value of *ScF*, e.g. *ScF>1000000*, all sloped p-Curves on *Face 2* will be almost vertical. At least, there will be no difference between the values of angles computed by standard C Run-Time Library functions, such as *double acos(double x)*. The loss of accuracy in computation of angles can cause failure of some BP sub-algorithms, such as building faces from a set of edges or building solids from a set of faces.
@@ -3070,6 +3085,50 @@ generated gf2 com_hist f2
~~~~
@section occt_algorithms_simplification BOP result simplification
The API algorithms implementing Boolean Operations provide possibility to simplify the result shape by unification of the connected tangential edges and faces.
This simplification is performed by the method *SimplifyResult* which is implemented in the class *BRepAlgoAPI_BuilderAlgo* (General Fuse operation).
It makes it available for users of the classes *BRepAlgoAPI_BooleanOperation* (all Boolean Operations) and *BRepAlgoAPI_Splitter* (split operation).
The simplification is performed by the means of *ShapeUpgrade_UnifySameDom* algorithm. The result of operation is overwritten with the simplified result.
The simplification is performed without creation of the Internal shapes, i.e. shapes connections will never be broken. It is performed on the whole result shape.
Thus, if the input shapes contained connected tangent edges or faces unmodified during the operation they will also be unified.
History of the simplification is merged into the main history of operation, thus it will be accounted when asking for Modified, Generated and Deleted shapes.
Some options of the main operation are passed into the Unifier:
- Fuzzy tolerance of the operation is given to the Unifier as the linear tolerance.
- Non destructive mode here controls the safe input mode in Unifier.
For controlling this possibility in DRAW the command **bsimplify** has been implemented. See the @ref occt_draw_bop_options "Boolean Operations options" chapter in draw user guide.
@@ -9,13 +9,13 @@ The IGES interface reads IGES files and translates them to Open CASCADE Technolo
Other kinds of data such as colors and names can be read or written with the help of XDE tools <i> IGESCAFControl_Reader</i> and <i> IGESCAFControl_Writer</i>.
Please, note:
**Note**:
* an IGES model is an IGES file that has been loaded into memory.
* an IGES entity is an entity in the IGES normal sense.
* a root entity is the highest level entity of any given type, e.g. type 144 for surfaces and type 186 for solids. Roots are not referenced by other entities.
This manual mainly explains how to convert an IGES file to an Open CASCADE Technology (**OCCT**) shape and vice versa. It provides basic documentation on conversion. For advanced information on conversion, see our <ahref="http://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
This manual mainly explains how to convert an IGES file to an Open CASCADE Technology (**OCCT**) shape and vice versa. It provides basic documentation on conversion. For advanced information on conversion, see our <ahref="https://www.opencascade.com/content/tutorial-learning">E-learning & Training</a> offerings.
IGES files produced in accordance with IGES standard versions up to and including version 5.3 can be read. IGES files that are produced by this interface conform to IGES version 5.3 (Initial Graphics Exchange Specification, IGES 5.3. ANS US PRO/IPO-100-1996).
@@ -178,10 +178,10 @@ if (!Interface_Static::SetRVal ("read.maxprecision.val",0.1))
Default value is 1.
<h4>read.stdsameparameter.mode</h4>
defines the using of *BRepLib::SameParameter*. Its possible values are:
* 0 (Off) -- *BRepLib::SameParameter* is not called,
* 1 (On) -- *BRepLib::SameParameter* is called.
*BRepLib::SameParameter* is used through *ShapeFix_Edge::SameParameter*. It ensures that the resulting edge will have the lowest tolerance taking pcurves either unmodified from the IGES file or modified by *BRepLib::SameParameter*.
defines the using of *BRepLib\::SameParameter*. Its possible values are:
* 0 (Off) -- *BRepLib\::SameParameter* is not called,
* 1 (On) -- *BRepLib\::SameParameter* is called.
*BRepLib\::SameParameter* is used through *ShapeFix_Edge\::SameParameter*. It ensures that the resulting edge will have the lowest tolerance taking pcurves either unmodified from the IGES file or modified by *BRepLib\::SameParameter*.
@@ -202,7 +202,7 @@ The processor also decides to re-compute either the 3D or the 2D curve even if
* the number of sub-curves in the 2D curve is different from the number of sub-curves in the 3D curve. This can be either due to different numbers of sub-curves given in the IGES file or because of splitting of curves during translation.
* 3D or 2D curve is a Circular Arc (entity type 100) starting and ending in the same point (note that this case is incorrect according to the IGES standard).
The parameter *read.surfacecurve.mode* defines which curve (3D or 2D) is used for re-computing the other one:
The parameter *read.surfacecurve.mode* defines which curve (3D or 2D) is used for re-computing the other one:
* *Default(0)* use the preference flag value in the entity's Parameter Data section. The flag values are:
* 0: no preference given,
* 1: use 2D for 142 entities and 3D for 141 entities,
@@ -251,7 +251,7 @@ Default value is 0.01.
<h4>read.iges.bspline.approxd1.mode</h4>
This parameter is obsolete (it is rarely used in real practice). If set to True, it affects the translation of bspline curves of degree 1 from IGES: these curves (which geometrically are polylines) are split by duplicated points, and the translator attempts to convert each of the obtained parts to a bspline of a higher continuity.
<h4>read.iges.resource.name and read.iges.sequence</h4>
These two parameters define the name of the resource file and the name of the sequence of operators (defined in that file) for Shape Processing, which is automatically performed by the IGES translator. The Shape Processing is a user-configurable step, which is performed after the translation and consists in application of a set of operators to a resulting shape. This is a very powerful tool allowing to customize the shape and to adapt it to the needs of a receiving application. By default, the sequence consists of a single operator *ShapeFix* that calls Shape Healing from the IGES translator.
Please find an example of the resource file for IGES (which defines parameters corresponding to the sequence applied by default, i.e. if the resource file is not found) in the Open CASCADE Technology installation, by the path <i>%CASROOT%/src/XSTEPResource/IGES</i>.
Find an example of the resource file for IGES (which defines parameters corresponding to the sequence applied by default, i.e. if the resource file is not found) in the Open CASCADE Technology sources by the path <i>%CASROOT%/src/XSTEPResource/IGES</i>.
In order for the IGES translator to use that file, you have to define the environment variable *CSF_IGESDefaults*, which should point to the directory where the resource file resides. Note that if you change parameter *read.iges.resource.name*, you should change the name of the resource file and the name of the environment variable correspondingly. The variable should contain a path to the resource file.
IGES translator will use that file if you define the environment variable *CSF_IGESDefaults*, which should point to the directory where the resource file resides. Note that if you change parameter *read.iges.resource.name*, you should change the name of the resource file and the name of the environment variable correspondingly. The variable should contain a path to the resource file.
Default values:
* read.iges.resource.name -- IGES,
@@ -286,7 +286,7 @@ A list of entities can be formed by invoking the method *IGESControl_Reader::Gi
Handle(TColStd_HSequenceOfTransient) list = reader.GiveList();
~~~~~
Several predefined operators can be used to select a list of entities of a specific type.
To make a selection, you use the method *IGESControl_Reader::GiveList* with the selection type in quotation marks as an argument. You can also make cumulative selections. For example, you would use the following syntax:
To make a selection, use the method *IGESControl_Reader::GiveList* with the selection type in quotation marks as an argument. You can also make cumulative selections. For example, you would use the following syntax:
displays the messages that appeared during the last invocation of *Transfer* or *TransferRoots*.
If *failsonly* is *IFSelect_FailOnly*, only fail messages will be output, if it is *IFSelect_FailAndWarn*, all messages will be output. Parameter “mode” can have *IFSelect_xxx* values where *xxx* can be:
* *GeneralCount* -- gives general statistics on the transfer (number of translated IGES entities, number of fails and warnings, etc)
* *CountByItem* -- gives the number of IGES entities with their types per message.
* *ListByItem* -- gives the number of IGES entities with their type and DE numbers per message.
* *ResultCount* -- gives the number of resulting OCCT shapes per type.
* *Mapping* -- gives mapping between roots of the IGES file and the resulting OCCT shape per IGES and OCCT type.
If *failsonly* is *IFSelect_FailOnly*, only fail messages will be output, if it is *IFSelect_FailAndWarn*, all messages will be output. Parameter “mode” can have *IFSelect_xxx* values where *xxx* can be:
* *GeneralCount* -- gives general statistics on the transfer (number of translated IGES entities, number of fails and warnings, etc)
* *CountByItem* -- gives the number of IGES entities with their types per message.
* *ListByItem* -- gives the number of IGES entities with their type and DE numbers per message.
* *ResultCount* -- gives the number of resulting OCCT shapes per type.
* *Mapping* -- gives mapping between roots of the IGES file and the resulting OCCT shape per IGES and OCCT type.
@subsection occt_iges_2_4 Mapping of IGES entities to Open CASCADE Technology shapes
@@ -577,7 +577,7 @@ All methods are in charge of transferring curves from IGES curve entities <i>(
* *IGESToBRep_TopoCurve::TransferPoint* -- vertex is constructed from a Point entity with tolerance *EpsGeom*UnitFactor*.
* *IGESToBRep_TopoCurve::Transfer2dPoint* -- vertex is constructed from a Point entity with tolerance *EpsCoeff*.
* *IGESToBRep_TopoCurve::TransferCompositeCurveGeneral* -- obtains shapes (edges or wires) from other methods and adds them into the resulting wire. Two adjacent edges of the wire can be connected with tolerance up to *MaxTol*.
* *IGESToBRep_TopoCurve::TransferCurveOnFace* and *IGESToBRep_TopoCurve::TransferBoundaryOnFace* build a wire from 3D and 2D representations of a curve on surface. Edges and vertices of the wire cannot have tolerance larger than *MaxTol*. The value *EpsGeom*UnitFactor* is passed into *ShapeFix_Wire::SetPrecision* and *MaxTol* is passed into *ShapeFix_Wire::MaxTolerance*. To find out how these parameters affect the resulting tolerance changes, please, refer to class *ShapeFix_Wire*.
* *IGESToBRep_TopoCurve::TransferCurveOnFace* and *IGESToBRep_TopoCurve::TransferBoundaryOnFace* build a wire from 3D and 2D representations of a curve on surface. Edges and vertices of the wire cannot have tolerance larger than *MaxTol*. The value *EpsGeom*UnitFactor* is passed into *ShapeFix_Wire::SetPrecision* and *MaxTol* is passed into *ShapeFix_Wire::MaxTolerance*. To find out how these parameters affect the resulting tolerance changes, refer to class *ShapeFix_Wire*.
* *IGESToBRep_TopoCurve::TransferTopoBasicCurve* and *IGESToBRep_TopoCurve::Transfer2dTopoBasicCurve* -- the boundary vertices of an edge (or a wire if a curve was of C0 continuity) translated from a basis IGES curve (*BSplineCurve, CopiousData, Line,* etc.) are built with tolerance *EpsGeom*UnitFactor*, the edge tolerance is *Precision::Confusion*. If a curve was divided into several edges, the common vertices of such adjacent edges have tolerance *Precision::Confusion*.
@@ -710,7 +710,7 @@ if (!Interface_Static::SetRVal(;write.precision.val;,0.01))
~~~~~
Default value is 0.0001.
<h4>write.iges.resource.name</h4> and <h4>write.iges.sequence</h4> are the same as the corresponding read.iges.\* parameters, please, see above. Note that the default sequence for writing contains *DirectFaces* operator, which converts elementary surfaces based on left-hand axes (valid in CASCADE) to right-hand axes (which are valid only in IGES).
<h4>write.iges.resource.name</h4> and <h4>write.iges.sequence</h4> are the same as the corresponding read.iges.\* parameters. Note that the default sequence for writing contains *DirectFaces* operator, which converts elementary surfaces based on left-hand axes (valid in CASCADE) to right-hand axes (which are valid only in IGES).
Default values :
~~~~~
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
