This repository has been archived on 2023-10-09. You can view files and clone it, but cannot push or open issues or pull requests.
Files
blender-archive/source/blender/python/mathutils/mathutils.h
Campbell Barton 5270ac5ed8 Fix GC tracking error for instances of mathutils types
Mathutils types were always GC tracked even when it wasn't intended.
Not having to track objects speeds up Python execution.

In an isolated benchmark created to stress test the GC
creating 4-million vectors (re-assigning them 100 times), this gives
an overall ~2.5x speedup, see: P3221.

Details:

Since [0] (which added support for sub-classed mathutils types)
tp_alloc was called which defaults to PyType_GenericAlloc which always
GC tracked the resulting object when Py_TPFLAGS_HAVE_GC was set.

Avoid using PyType_GenericAlloc unless the type is sub-classed,
in that case the object is un-tracked.

Add asserts that the tracked state is as expected before tracking &
un-tracking, to ensure changes to object creation don't cause objects
to be tracked unintentionally.

Also assign the PyTypeObject.tp_is_gc callback so types optionally GC
track objects only do so when an object is referenced.

[0]: fbd9364944
2022-09-28 17:53:30 +10:00

231 lines
8.4 KiB
C++

/* SPDX-License-Identifier: GPL-2.0-or-later */
#pragma once
/** \file
* \ingroup pymathutils
*/
/* Can cast different mathutils types to this, use for generic functions. */
#include "BLI_compiler_attrs.h"
struct DynStr;
extern char BaseMathObject_is_wrapped_doc[];
extern char BaseMathObject_is_frozen_doc[];
extern char BaseMathObject_is_valid_doc[];
extern char BaseMathObject_owner_doc[];
PyObject *_BaseMathObject_new_impl(PyTypeObject *root_type, PyTypeObject *base_type);
#define BASE_MATH_NEW(struct_name, root_type, base_type) \
((struct_name *)_BaseMathObject_new_impl(&root_type, base_type))
/** #BaseMathObject.flag */
enum {
/**
* Do not own the memory used in this vector,
* \note This is error prone if the memory may be freed while this vector is in use.
* Prefer using callbacks where possible, see: #Mathutils_RegisterCallback
*/
BASE_MATH_FLAG_IS_WRAP = (1 << 0),
/**
* Prevent changes to the vector so it can be used as a set or dictionary key for example.
* (typical use cases for tuple).
*/
BASE_MATH_FLAG_IS_FROZEN = (1 << 1),
};
#define BASE_MATH_FLAG_DEFAULT 0
#define BASE_MATH_MEMBERS(_data) \
/** Array of data (alias), wrapped status depends on wrapped status. */ \
PyObject_VAR_HEAD \
float *_data; \
/** If this vector references another object, otherwise NULL, *Note* this owns its reference */ \
PyObject *cb_user; \
/** Which user functions do we adhere to, RNA, etc */ \
unsigned char cb_type; \
/** Sub-type: location, rotation... \
* to avoid defining many new functions for every attribute of the same type */ \
unsigned char cb_subtype; \
/** Wrapped data type. */ \
unsigned char flag
typedef struct {
BASE_MATH_MEMBERS(data);
} BaseMathObject;
/* types */
#include "mathutils_Color.h"
#include "mathutils_Euler.h"
#include "mathutils_Matrix.h"
#include "mathutils_Quaternion.h"
#include "mathutils_Vector.h"
/* avoid checking all types */
#define BaseMathObject_CheckExact(v) (Py_TYPE(v)->tp_dealloc == (destructor)BaseMathObject_dealloc)
PyObject *BaseMathObject_owner_get(BaseMathObject *self, void *);
PyObject *BaseMathObject_is_wrapped_get(BaseMathObject *self, void *);
PyObject *BaseMathObject_is_frozen_get(BaseMathObject *self, void *);
PyObject *BaseMathObject_is_valid_get(BaseMathObject *self, void *);
extern char BaseMathObject_freeze_doc[];
PyObject *BaseMathObject_freeze(BaseMathObject *self);
int BaseMathObject_traverse(BaseMathObject *self, visitproc visit, void *arg);
int BaseMathObject_clear(BaseMathObject *self);
void BaseMathObject_dealloc(BaseMathObject *self);
int BaseMathObject_is_gc(BaseMathObject *self);
PyMODINIT_FUNC PyInit_mathutils(void);
int EXPP_FloatsAreEqual(float A, float B, int maxDiff);
int EXPP_VectorsAreEqual(const float *vecA, const float *vecB, int size, int floatSteps);
typedef struct Mathutils_Callback Mathutils_Callback;
/** Checks the user is still valid. */
typedef int (*BaseMathCheckFunc)(BaseMathObject *);
/** Gets the vector from the user. */
typedef int (*BaseMathGetFunc)(BaseMathObject *, int);
/** Sets the users vector values once its modified. */
typedef int (*BaseMathSetFunc)(BaseMathObject *, int);
/** Same as above but only for an index. */
typedef int (*BaseMathGetIndexFunc)(BaseMathObject *, int, int);
/** Same as above but only for an index. */
typedef int (*BaseMathSetIndexFunc)(BaseMathObject *, int, int);
struct Mathutils_Callback {
BaseMathCheckFunc check;
BaseMathGetFunc get;
BaseMathSetFunc set;
BaseMathGetIndexFunc get_index;
BaseMathSetIndexFunc set_index;
};
unsigned char Mathutils_RegisterCallback(Mathutils_Callback *cb);
int _BaseMathObject_CheckCallback(BaseMathObject *self);
int _BaseMathObject_ReadCallback(BaseMathObject *self);
int _BaseMathObject_WriteCallback(BaseMathObject *self);
int _BaseMathObject_ReadIndexCallback(BaseMathObject *self, int index);
int _BaseMathObject_WriteIndexCallback(BaseMathObject *self, int index);
void _BaseMathObject_RaiseFrozenExc(const BaseMathObject *self);
void _BaseMathObject_RaiseNotFrozenExc(const BaseMathObject *self);
/* since this is called so often avoid where possible */
#define BaseMath_CheckCallback(_self) \
(((_self)->cb_user ? _BaseMathObject_CheckCallback((BaseMathObject *)_self) : 0))
#define BaseMath_ReadCallback(_self) \
(((_self)->cb_user ? _BaseMathObject_ReadCallback((BaseMathObject *)_self) : 0))
#define BaseMath_WriteCallback(_self) \
(((_self)->cb_user ? _BaseMathObject_WriteCallback((BaseMathObject *)_self) : 0))
#define BaseMath_ReadIndexCallback(_self, _index) \
(((_self)->cb_user ? _BaseMathObject_ReadIndexCallback((BaseMathObject *)_self, _index) : 0))
#define BaseMath_WriteIndexCallback(_self, _index) \
(((_self)->cb_user ? _BaseMathObject_WriteIndexCallback((BaseMathObject *)_self, _index) : 0))
/* support BASE_MATH_FLAG_IS_FROZEN */
#define BaseMath_ReadCallback_ForWrite(_self) \
(UNLIKELY((_self)->flag & BASE_MATH_FLAG_IS_FROZEN) ? \
(_BaseMathObject_RaiseFrozenExc((BaseMathObject *)_self), -1) : \
(BaseMath_ReadCallback(_self)))
#define BaseMath_ReadIndexCallback_ForWrite(_self, _index) \
(UNLIKELY((_self)->flag & BASE_MATH_FLAG_IS_FROZEN) ? \
(_BaseMathObject_RaiseFrozenExc((BaseMathObject *)_self), -1) : \
(BaseMath_ReadIndexCallback(_self, _index)))
#define BaseMath_Prepare_ForWrite(_self) \
(UNLIKELY((_self)->flag & BASE_MATH_FLAG_IS_FROZEN) ? \
(_BaseMathObject_RaiseFrozenExc((BaseMathObject *)_self), -1) : \
0)
#define BaseMathObject_Prepare_ForHash(_self) \
(UNLIKELY(((_self)->flag & BASE_MATH_FLAG_IS_FROZEN) == 0) ? \
(_BaseMathObject_RaiseNotFrozenExc((BaseMathObject *)_self), -1) : \
0)
/* utility func */
/**
* Helper function.
* \return length of `value`, -1 on error.
*/
int mathutils_array_parse(
float *array, int array_num_min, int array_num_max, PyObject *value, const char *error_prefix);
/**
* \return -1 is returned on error and no allocation is made.
*/
int mathutils_array_parse_alloc(float **array,
int array_num_min,
PyObject *value,
const char *error_prefix);
/**
* Parse an array of vectors.
*/
int mathutils_array_parse_alloc_v(float **array,
int array_dim,
PyObject *value,
const char *error_prefix);
/**
* Parse an sequence array_dim integers into array.
*/
int mathutils_int_array_parse(int *array,
int array_dim,
PyObject *value,
const char *error_prefix);
/**
* Parse sequence of array_dim sequences of integers and return allocated result.
*/
int mathutils_array_parse_alloc_vi(int **array,
int array_dim,
PyObject *value,
const char *error_prefix);
/**
* Parse sequence of variable-length sequences of int and return allocated
* triple of arrays to represent the result:
* The flattened sequences are put into *array.
* The start index of each sequence goes into start_table.
* The length of each index goes into len_table.
*/
int mathutils_array_parse_alloc_viseq(
int **array, int **start_table, int **len_table, PyObject *value, const char *error_prefix);
int mathutils_any_to_rotmat(float rmat[3][3], PyObject *value, const char *error_prefix);
/**
* helper function that returns a Python `__hash__`.
*
* \note consistent with the equivalent tuple of floats (CPython's 'tuplehash')
*/
Py_hash_t mathutils_array_hash(const float *float_array, size_t array_len);
/* zero remaining unused elements of the array */
#define MU_ARRAY_ZERO (1u << 30)
/* ignore larger py sequences than requested (just use first elements),
* handy when using 3d vectors as 2d */
#define MU_ARRAY_SPILL (1u << 31)
#define MU_ARRAY_FLAGS (MU_ARRAY_ZERO | MU_ARRAY_SPILL)
/**
* Column vector multiplication (Matrix * Vector).
* <pre>
* [1][4][7] [a]
* [2][5][8] * [b]
* [3][6][9] [c]
* </pre>
*
* \note Vector/Matrix multiplication is not commutative.
* \note Assume read callbacks have been done first.
*/
int column_vector_multiplication(float r_vec[4], VectorObject *vec, MatrixObject *mat);
#ifndef MATH_STANDALONE
/* dynstr as python string utility functions */
/* dynstr as python string utility functions, frees 'ds'! */
PyObject *mathutils_dynstr_to_py(struct DynStr *ds);
#endif