renderAttrib.h

Go to the documentation of this file.

/**
 * PANDA 3D SOFTWARE
 * Copyright (c) Carnegie Mellon University.  All rights reserved.
 *
 * All use of this software is subject to the terms of the revised BSD
 * license.  You should have received a copy of this license along
 * with this source code in a file named "LICENSE."
 *
 * @file renderAttrib.h
 * @author drose
 * @date 2002-02-21
 */

#ifndef RENDERATTRIB_H
#define RENDERATTRIB_H

#include "pandabase.h"

#include "typedWritableReferenceCount.h"
#include "renderAttribRegistry.h"
#include "pointerTo.h"
#include "simpleHashMap.h"
#include "lightReMutex.h"
#include "pStatCollector.h"

class AttribSlots;
class GraphicsStateGuardianBase;
class CullTraverser;
class CullTraverserData;
class RenderState;

/**
 * This is the base class for a number of render attributes (other than
 * transform) that may be set on scene graph nodes to control the appearance
 * of geometry.  This includes TextureAttrib, ColorAttrib, etc.
 *
 * RenderAttrib represents render attributes that always propagate down to the
 * leaves without regard to the particular node they are assigned to.  A
 * RenderAttrib will have the same effect on a leaf node whether it is
 * assigned to the graph at the leaf or several nodes above.  This is
 * different from RenderEffect, which represents a particular render property
 * that is applied immediately to the node on which it is encountered, like
 * billboarding or decaling.
 *
 * You should not attempt to create or modify a RenderAttrib directly;
 * instead, use the make() method of the appropriate kind of attrib you want.
 * This will allocate and return a new RenderAttrib of the appropriate type,
 * and it may share pointers if possible.  Do not modify the new RenderAttrib
 * if you wish to change its properties; instead, create a new one.
 */
class EXPCL_PANDA_PGRAPH RenderAttrib : public TypedWritableReferenceCount {
protected:
  RenderAttrib();

public:
  RenderAttrib(const RenderAttrib &copy) = delete;
  virtual ~RenderAttrib();

  RenderAttrib &operator = (const RenderAttrib &copy) = delete;

PUBLISHED:
  INLINE CPT(RenderAttrib) compose(const RenderAttrib *other) const;
  INLINE CPT(RenderAttrib) invert_compose(const RenderAttrib *other) const;
  virtual bool lower_attrib_can_override() const;

public:
  virtual bool has_cull_callback() const;
  virtual bool cull_callback(CullTraverser *trav, const CullTraverserData &data) const;

PUBLISHED:
  INLINE int compare_to(const RenderAttrib &other) const;
  INLINE size_t get_hash() const;
  INLINE CPT(RenderAttrib) get_unique() const;

  virtual bool unref() const final;

  virtual void output(std::ostream &out) const;
  virtual void write(std::ostream &out, int indent_level) const;

  static int get_num_attribs();
  static void list_attribs(std::ostream &out);
  static int garbage_collect();
  static bool validate_attribs();

  virtual int get_slot() const=0;
  MAKE_PROPERTY(slot, get_slot);

  enum PandaCompareFunc {   // intentionally defined to match D3DCMPFUNC
    M_none=0,           // alpha-test disabled (always-draw)
    M_never,            // Never draw.
    M_less,             // incoming < reference_alpha
    M_equal,            // incoming == reference_alpha
    M_less_equal,       // incoming <= reference_alpha
    M_greater,          // incoming > reference_alpha
    M_not_equal,        // incoming != reference_alpha
    M_greater_equal,    // incoming >= reference_alpha
    M_always            // Always draw.
  };

  // This is the enumerated type for TexGenAttrib.  It is inherited into
  // TexGenAttrib.  It is defined up at this level only to avoid circular
  // dependencies in the header files.
  enum TexGenMode {
    M_off,

    // In the types below, "eye" means the coordinate space of the observing
    // camera, and "world" means world coordinates, e.g.  the coordinate space
    // of the root of the graph.

    // Sphere maps are classic static reflection maps.  They are supported on
    // just about any hardware, and require a precomputed 360-degree fisheye
    // image.  Sphere maps only make sense in eye coordinate space.
    M_eye_sphere_map,

/*
 * Cube maps are a modern improvement on the sphere map; they don't suffer
 * from any polar singularities, but they require six texture images.  They
 * can also be generated dynamically for real-time reflections (see
 * GraphicsOutput::make_cube_map()). Typically, a statically-generated cube
 * map will be in eye space, while a dynamically-generated map will be in
 * world space.  Cube mapping is not supported on all hardware.
 */
    M_world_cube_map,
    M_eye_cube_map,

    // Normal maps are most useful for applying diffuse lighting effects via a
    // pregenerated cube map.
    M_world_normal,
    M_eye_normal,

    // Position maps convert XYZ coordinates directly to texture coordinates.
    // This is particularly useful for implementing projective texturing (see
    // NodePath::project_texture()).
    M_world_position,
    M_unused,  // formerly M_object_position, now deprecated.
    M_eye_position,

/*
 * With M_point_sprite, texture coordinates will be generated for large points
 * in the range (0,0) - (1,1) from upper-left to lower-right across the
 * point's face.  Without this, each point will have just a single uniform
 * texture coordinate value across its face.  Unfortunately, the generated
 * texture coordinates are inverted (upside-down) from Panda's usual
 * convention, but this is what the graphics card manufacturers decided to
 * use.  You could use a texture matrix to re-invert the texture, but that
 * will probably force the sprites' vertices to be computed in the CPU. You'll
 * have to paint your textures upside-down if you want true hardware sprites.
 */
    M_point_sprite,

    // M_light_vector generated special 3-d texture coordinates that
    // represented the vector to a particular Light in the scene graph,
    // expressed in each vertex's tangent space.  This has now been removed.
    // We need to reserve the slot in the enum, though, to make sure the
    // following enum value still has the same value.
    M_unused2,

    // M_constant generates the same fixed texture coordinates at each vertex.
    // Not terribly useful, of course, except for certain special effects
    // involving moving a flat color over an object.
    M_constant,
  };

protected:
  INLINE void calc_hash();

  static CPT(RenderAttrib) return_new(RenderAttrib *attrib);
  static CPT(RenderAttrib) return_unique(RenderAttrib *attrib);
  virtual int compare_to_impl(const RenderAttrib *other) const;
  virtual size_t get_hash_impl() const;
  virtual CPT(RenderAttrib) compose_impl(const RenderAttrib *other) const;
  virtual CPT(RenderAttrib) invert_compose_impl(const RenderAttrib *other) const;
  void output_comparefunc(std::ostream &out, PandaCompareFunc fn) const;

public:
  INLINE static int register_slot(TypeHandle type_handle, int sort,
                                  RenderAttrib *default_attrib);

private:
  void release_new();

public:
  static void init_attribs();

private:
  // This mutex protects _attribs.
  static LightReMutex *_attribs_lock;
  typedef SimpleHashMap<const RenderAttrib *, std::nullptr_t, indirect_compare_to_hash<const RenderAttrib *> > Attribs;
  static Attribs *_attribs;

  int _saved_entry;
  size_t _hash;

  // This keeps track of our current position through the garbage collection
  // cycle.
  static size_t _garbage_index;

  static PStatCollector _garbage_collect_pcollector;

  friend class RenderAttribRegistry;

public:
  virtual void write_datagram(BamWriter *manager, Datagram &dg);
  static TypedWritable *change_this(TypedWritable *old_ptr, BamReader *manager);
  virtual void finalize(BamReader *manager);

protected:
  static TypedWritable *new_from_bam(RenderAttrib *attrib, BamReader *manager);
  void fillin(DatagramIterator &scan, BamReader *manager);

public:
  static TypeHandle get_class_type() {
    return _type_handle;
  }
  static void init_type() {
    TypedWritableReferenceCount::init_type();
    register_type(_type_handle, "RenderAttrib",
                  TypedWritableReferenceCount::get_class_type());
  }
  virtual TypeHandle get_type() const {
    return get_class_type();
  }
  virtual TypeHandle force_init_type() {init_type(); return get_class_type();}

private:
  static TypeHandle _type_handle;
};

INLINE std::ostream &operator << (std::ostream &out, const RenderAttrib &attrib) {
  attrib.output(out);
  return out;
}

#include "renderAttrib.I"

#endif