ModelPool
-
class ModelPool
This class unifies all references to the same filename, so that multiple attempts to load the same model will return the same pointer. Note that the default behavior is thus to make instances: use with caution. Use the copy_subgraph() method on Node (or use
NodePath::copy_to()) to make modifiable copies of the node.Unlike
TexturePool, this class does not automatically resolve the model filenames before loading, so a relative path and an absolute path to the same model will appear to be different filenames.However, see the Loader class, which is now the preferred interface for loading models. The Loader class can resolve filenames, supports threaded loading, and can automatically consult the ModelPool, according to the supplied
LoaderOptions.Inheritance diagram
-
void add_model(ModelRoot *model)
Adds the indicated already-loaded model to the pool. The model will replace any previously-loaded model in the pool that had the same filename.
Deprecated: Use the one-parameter add_model(model) instead.
Adds the indicated already-loaded model to the pool. The model will replace any previously-loaded model in the pool that had the same filename.
-
int garbage_collect(void)
Releases only those models in the pool that have a reference count of exactly 1; i.e. only those models that are not being used outside of the pool. Returns the number of models released.
-
ModelRoot *get_model(Filename const &filename, bool verify)
Returns the model that has already been previously loaded, or NULL otherwise. If verify is true, it will check if the file is still up-to- date (and hasn’t been modified in the meantime), and if not, will still return NULL.
-
bool has_model(Filename const &filename)
Returns true if the model has ever been loaded, false otherwise. Note that this does not guarantee that the model is still up-to-date.
-
void list_contents(std::ostream &out)
-
void list_contents(void)
Lists the contents of the model pool to the indicated output stream.
Lists the contents of the model pool to cout.
-
ModelRoot *load_model(Filename const &filename, LoaderOptions const &options = LoaderOptions())
Loads the given filename up as a model, if it has not already been loaded, and returns the new model. If a model with the same filename was previously loaded, returns that one instead (unless cache-check-timestamps is true and the file has recently changed). If the model file cannot be found, or cannot be loaded for some reason, returns NULL.
-
void release_all_models(void)
Releases all models in the pool and restores the pool to the empty state.
-
void release_model(ModelRoot *model)
Removes the indicated model from the pool, indicating it will never be loaded again; the model may then be freed. If this function is never called, a reference count will be maintained on every model every loaded, and models will never be freed.
Deprecated: Use release_model(model) instead.
Removes the indicated model from the pool, indicating it will never be loaded again; the model may then be freed. If this function (and
garbage_collect()) is never called, a reference count will be maintained on every model every loaded, and models will never be freed.The model’s get_fullpath() value should not have been changed during its lifetime, or this function may fail to locate it in the pool.
-
bool verify_model(Filename const &filename)
Loads the given filename up as a model, if it has not already been loaded, and returns true to indicate success, or false to indicate failure. If this returns true, it is probable that a subsequent call to
load_model()with the same model name will return a validPandaNode.However, even if this returns true, it is still possible for a subsequent call to
load_model()to fail. This can happen if cache-check-timestamps is true, and the on-disk file is subsequently modified to replace it with an invalid model.
-
static void write(std::ostream &out)
Lists the contents of the model pool to the indicated output stream. Helps with debugging.
-
void add_model(ModelRoot *model)