GeomVertexArrayFormat

class GeomVertexArrayFormat

Bases: TypedWritableReferenceCount, GeomEnums

This describes the structure of a single array within a Geom data. See GeomVertexFormat for the parent class which collects together all of the individual GeomVertexArrayFormat objects.

A particular array may include any number of standard or user-defined columns. All columns consist of a sequence of one or more numeric values, packed in any of a variety of formats; the semantic meaning of each column is defined in general with its contents member, and in particular by its name. The standard array types used most often are named “vertex”, “normal”, “texcoord”, and “color”; other kinds of data may be piggybacked into the data record simply by choosing a unique name.

Inheritance diagram

GeomVertexArrayFormat(void)

GeomVertexArrayFormat(GeomVertexArrayFormat const &copy)

GeomVertexArrayFormat(CPT_InternalName name0, int num_components0, GeomEnums::NumericType numeric_type0, GeomEnums::Contents contents0)

GeomVertexArrayFormat(CPT_InternalName name0, int num_components0, GeomEnums::NumericType numeric_type0, GeomEnums::Contents contents0, CPT_InternalName name1, int num_components1, GeomEnums::NumericType numeric_type1, GeomEnums::Contents contents1)

GeomVertexArrayFormat(CPT_InternalName name0, int num_components0, GeomEnums::NumericType numeric_type0, GeomEnums::Contents contents0, CPT_InternalName name1, int num_components1, GeomEnums::NumericType numeric_type1, GeomEnums::Contents contents1, CPT_InternalName name2, int num_components2, GeomEnums::NumericType numeric_type2, GeomEnums::Contents contents2)

GeomVertexArrayFormat(CPT_InternalName name0, int num_components0, GeomEnums::NumericType numeric_type0, GeomEnums::Contents contents0, CPT_InternalName name1, int num_components1, GeomEnums::NumericType numeric_type1, GeomEnums::Contents contents1, CPT_InternalName name2, int num_components2, GeomEnums::NumericType numeric_type2, GeomEnums::Contents contents2, CPT_InternalName name3, int num_components3, GeomEnums::NumericType numeric_type3, GeomEnums::Contents contents3)

int add_column(CPT_InternalName name, int num_components, GeomEnums::NumericType numeric_type, GeomEnums::Contents contents, int start = -1, int column_alignment = 0)

int add_column(GeomVertexColumn const &column)

Adds a new column to the specification. This is a table of per-vertex floating-point numbers such as “vertex” or “normal”; you must specify where in each record the table starts, and how many components (dimensions) exist per vertex.

The return value is the index number of the new data type.

Adds a new column to the specification. This is a table of per-vertex floating-point numbers such as “vertex” or “normal”; you must specify where in each record the table starts, and how many components (dimensions) exist per vertex.

Adding a column with the same name as a previous type, or that overlaps with one or more previous types, quietly removes the previous type(s).

The return value is the index number of the new data type.

void align_columns_for_animation(void)

Reprocesses the columns in the format to align the C_point and C_vector columns to 16-byte boundaries to allow for the more efficient SSE2 operations (assuming SSE2 is enabled in the build).

The caller is responsible for testing vertex_animation_align_16 to decide whether to call this method.

void clear_columns(void)

Removes all columns previously added, sets the stride to zero, and prepares to start over.

int count_unused_space(void) const

Returns the number of bytes per row that are not assigned to any column.

static TypeHandle get_class_type(void)

GeomVertexColumn const *get_column(int i) const

GeomVertexColumn const *get_column(InternalName const *name) const

GeomVertexColumn const *get_column(int start_byte, int num_bytes) const

Returns the ith column of the array.

Returns the specification with the indicated name, or NULL if the name is not used.

Returns the first specification that overlaps with any of the indicated bytes in the range, or NULL if none do.

int get_divisor(void) const

Returns the divisor attribute for the data in this array. If 0, it contains per-vertex data. If 1, it contains per-instance data. If higher than 1, the read row is advanced for each n instances.

std::string get_format_string(bool pad = true) const

Returns a string with format codes representing the exact memory layout of the columns in memory, as understood by Python’s struct module. If pad is true, extra padding bytes are added to the end as ‘x’ characters as needed.

int get_num_columns(void) const

Returns the number of different columns in the array.

int get_pad_to(void) const

Returns the byte divisor to which the data record must be padded to meet hardware limitations. For instance, if this is 4, the stride will be automatically rounded up to the next multiple of 4 bytes. This value is automatically increased as needed to ensure the individual numeric components in the array are word-aligned.

int get_stride(void) const

Returns the total number of bytes reserved in the array for each vertex.

int get_total_bytes(void) const

Returns the total number of bytes used by the data types within the format, including gaps between elements.

bool has_column(InternalName const *name) const

Returns true if the array has the named column, false otherwise.

bool is_data_subset_of(GeomVertexArrayFormat const &other) const

Returns true if all of the fields in this array format are also present and equivalent in the other array format, and in the same byte positions, and the stride is the same. That is, true if this format can share the same data pointer as the other format (with possibly some unused gaps).

bool is_registered(void) const

Returns true if this format has been registered, false if it has not. It may not be used for a Geom until it has been registered, but once registered, it may no longer be modified.

void output(std::ostream &out) const

void pack_columns(void)

Removes wasted space between columns.

ConstPointerTo<GeomVertexArrayFormat> register_format(GeomVertexArrayFormat const *format)

Adds the indicated format to the registry, if there is not an equivalent format already there; in either case, returns the pointer to the equivalent format now in the registry.

This is similar to GeomVertexFormat::register_format(), except that you generally need not call it explicitly. Calling GeomVertexFormat::register_format() automatically registers all of the nested array formats.

void remove_column(InternalName const *name)

Removes the column with the indicated name, if any. This leaves a gap in the byte structure.

void set_divisor(int divisor)

Set this to 0 to indicate that this array contains per-vertex data, or to 1 to indicate that it contains per-instance data. If higher than 1, the read row is advanced for each n instances.

void set_pad_to(int pad_to)

Explicitly sets the byte divisor to which the data record must be padded to meet hardware limitations. See get_pad_to(). Normally it is not necessary to call this unless you have some specific requirements for row-to-row data alignment. Note that this value may be automatically increased at each subsequent call to add_column().

void set_stride(int stride)

Changes the total number of bytes reserved in the array for each vertex. You may not reduce this below get_total_bytes(), but you may increase it arbitrarily.

virtual bool unref(void) const

This method overrides ReferenceCount::unref() to unregister the object when its reference count goes to zero.

void write(std::ostream &out, int indent_level = 0) const

void write_with_data(std::ostream &out, int indent_level, GeomVertexArrayData const *array_data) const