Clang-tidy applying rule for cppcoreguidelines-pro-type-member-init.
Updated: TKernel and TKMath
Update constructor in some classes instead of direct initialization
Refactor Bnd_BoundSortBox and Bnd_Box constructors to initialize member variables directly
Update empty method guards to new style with regex (see PR).
Used clang-format 18.1.8.
New actions to validate code formatting is added.
Update .clang-format with disabling of include sorting.
It is temporary changes, then include will be sorted.
Apply formatting for /src and /tools folder.
The files with .hxx,.cxx,.lxx,.h,.pxx,.hpp,*.cpp extensions.
Updated multiple classes to utilize the new NCollection_MapAlgo
for union, intersection, and other set operations,
improving code structure and maintainability.
Deprecated older methods in NCollection_Map in favor of the new algorithmic approaches.
Reorganized style for param to the next templates:
- "@param theParameter description ..."
- "@param[in] theParameter description ..."
- "@param[out] theParameter description ..."
- "@param[in][out] theParameter description ..."
The replacement was with keeping spacing, no removing of extra spaces.
In some files '/' was used instead of '@', that was not updated yet.
NCollection containers update:
- NCollection_Array1 - updated functionality
- NCollection_Array2 - NCollection_Array1 as a wrapper for 2array
- NCollection_Vector -> NCollection_DynamicArray was renamed and reworked.
TCollection:
- Use static empty string to avoid allocations on empty string
NCollection allocators update:
- NCollection_Allocator - allocator that used Standard::Allocate
- NCollection_OccAllocator - allocator-wrapper that used OCC BaseAllocator objects
- NCollection_IncAllocator - rework to increase performance
Standard:
- Rework functionality to use different allocation libs
- Implement basic of new way to wrap allocations tools
- Define 4 ways to allocation (defines in configure stage)
Additional changes:
- Hash function uses std::hash functionality
- size_t as a hash value
- New HashUtils with Murmur and FVN hash algo for x32 and x64
- Deprecated _0.cxx and .gxx DE classes reorganized
- Create own utility for std memory
- Update Standard_Transient to be more platform-independent
Math TK changes:
- math_Vector -> match_BaseVector<>
- Buffer decreased to cash 32 elements instead of 512
Microsoft Visual Studio Professional 2019
Version 16.11.11
std=c++20
Get rid of warning C26440: "If your function may not throw, declare it noexcept"
"If code is not supposed to cause any exceptions,
it should be marked as such by using the 'noexcept' specifier.
This would help to simplify error handling on the client code side,
as well as enable compiler to do additional optimizations."
Microsoft Visual Studio Professional 2019
Version 16.11.11
std=c++20
Get rid of warnings about uninitialized class fields in constructor initialization block.
Removed usage of std::auto_ptr.
opencascade::std workarounds for VS2008 have been dropped, when possible.
OCCT_NO_RVALUE_REFERENCE workaround for VS2008 has been removed.
Added transform persistence property to Graphic3d_Group and Select3D_SensitiveEntity.
SelectMgr_ViewerSelector, Graphic3d_Layer and OpenGl_Structure have been updated
to process per-group transform persistence within picking, ZFit and rendering.
Added zoomable state to Prs3d_ArrowAspect supported by PrsDim_Dimension.
Added gp_GTrsf::SetMat4(), opposite to gp_GTrsf::GetMat4().
Make preprocessor expressions consistent with respect to guard expressions.
Fixed usage of macros __QNX__, IRIX, OCCT_DEBUG, DO_INVERSE, DRAW, CHFI3D_DEB by value.
Removed obsolete hack for Sun Workshop 5.0 compiler.
Introduced NCollection_Mat3 class similar to NCollection_Mat4.
Added NCollection_Mat4::operator() alias to NCollection_Mat4::Value().
Added NCollection_Mat4::Negated(), ::Subtracted(), ::Added(), ::Divided() operations.
Added a note to NCollection_Mat4::GetData() that matrix values are stored in column-major order.
Removed duplicated operator== operators from NCollection_Mat4/NCollection_Vec4/NCollection_Vec3/NCollection_Vec2 classes.
Removed obsolete and no more used matrix state fields from OpenGl_Workspace.
Removed obsoiete structure OpenGl_Matrix (replaced by OpenGl_Mat4).
OpenGl_ShaderProgram - duplicated methods have been replaced by templates.
Introduced NCollection_AliasedArray collection defining a general byte array
to be aliased as array of elements of mutable type.
Poly_ArrayOfNodes specializes this interface for collecting 2d/3d points with single/double precision.
Poly_Triangulation relies on this new class for definition of 3D nodes and UV coordinates
using gp_Pnt and gp_Pnt2d by default as before.
RWGltf_CafReader now fills in triangulation using single precision by default (as stored in the file).
The workaround for a bug occurred in MS Visual Studio 2019 / Win32 / Release configuration with disabled Whole Program Optimization is implemented.
The problem was at the line "std::stable_sort(aPairList.begin(), aPairList.end(), BRepExtrema_CheckPair_Comparator);" of BRepExtrema_DistShapeShape.cxx source file,
and linked to NCollection_Vector::Differ() method.
This workaround was suggested by MSFT team.
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.
- 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
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
"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.
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.
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.
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.
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().
