view engine/core/view/camera.cpp @ 422:9d94f4676d17

Moved the management of cameras into the Map class. The View class no longer exists since it now serves no purpose. The engine class itself holds the renderers and these get passed to each map that is loaded and then passed to each camera on each map. This change makes it possible for multiple maps to be loaded at the same time with the same camera id without a name clash. fixes[t:342]
author vtchill@33b003aa-7bff-0310-803a-e67f0ece8222
date Fri, 12 Feb 2010 06:34:50 +0000
parents 219b8036708c
children 4f2752d45458
line wrap: on
line source

/***************************************************************************
 *   Copyright (C) 2005-2008 by the FIFE team                              *
 *   http://www.fifengine.de                                               *
 *   This file is part of FIFE.                                            *
 *                                                                         *
 *   FIFE is free software; you can redistribute it and/or                 *
 *   modify it under the terms of the GNU Lesser General Public            *
 *   License as published by the Free Software Foundation; either          *
 *   version 2.1 of the License, or (at your option) any later version.    *
 *                                                                         *
 *   This library is distributed in the hope that it will be useful,       *
 *   but WITHOUT ANY WARRANTY; without even the implied warranty of        *
 *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU     *
 *   Lesser General Public License for more details.                       *
 *                                                                         *
 *   You should have received a copy of the GNU Lesser General Public      *
 *   License along with this library; if not, write to the                 *
 *   Free Software Foundation, Inc.,                                       *
 *   51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA          *
 ***************************************************************************/

// Standard C++ library includes

// 3rd party library includes
#include <SDL.h>

// FIFE includes
// These includes are split up in two parts, separated by one empty line
// First block: files included from the FIFE root src directory
// Second block: files included from the same folder

#include "model/metamodel/grids/cellgrid.h"
#include "model/metamodel/action.h"
#include "model/metamodel/timeprovider.h"
#include "model/structures/map.h"
#include "model/structures/layer.h"
#include "model/structures/instancetree.h"
#include "model/structures/instance.h"
#include "model/structures/location.h"
#include "util/log/logger.h"
#include "util/math/fife_math.h"
#include "util/math/angles.h"
#include "util/time/timemanager.h"
#include "video/renderbackend.h"
#include "video/image.h"
#include "video/imagepool.h"
#include "video/animation.h"
#include "video/animationpool.h"

#include "camera.h"
#include "visual.h"


namespace FIFE {
	static Logger _log(LM_CAMERA);

	Camera::Camera(const std::string& id,
		Layer *layer,
		const Rect& viewport,
		const ExactModelCoordinate& emc,
		RenderBackend* renderbackend,
		ImagePool* ipool,
		AnimationPool* apool):
			m_id(id),
			m_matrix(),
			m_inverse_matrix(),
			m_tilt(0),
			m_rotation(0),
			m_zoom(1),
			m_location(),
			m_prev_origo(ScreenPoint(0,0,0)),
			m_cur_origo(ScreenPoint(0,0,0)),
			m_viewport(),
			m_screen_cell_width(1),
			m_screen_cell_height(1),
			m_reference_scale(1),
			m_enabled(true),
			m_attachedto(NULL),
			m_image_dimensions(),
			m_iswarped(false),
			m_renderers(),
			m_pipeline(),
			m_updated(false),
			m_renderbackend(renderbackend),
			m_ipool(ipool),
			m_apool(apool),
			m_layer_to_instances() {
		m_location.setLayer(layer);
		m_location.setExactLayerCoordinates(emc);
		m_viewport = viewport;
	}

	Camera::~Camera() {
		std::map<std::string, RendererBase*>::iterator r_it = m_renderers.begin();
		for(; r_it != m_renderers.end(); ++r_it) {
			delete r_it->second;
		}
		m_renderers.clear();
	}

	void Camera::setTilt(double tilt) {
		m_tilt = tilt;
		updateReferenceScale();
		updateMatrices();
		m_iswarped = true;
	}

	double Camera::getTilt() const {
		return m_tilt;
	}

	void Camera::setRotation(double rotation) {
		m_rotation = rotation;
		updateReferenceScale();
		updateMatrices();
		m_iswarped = true;
	}

	double Camera::getRotation() const {
		return m_rotation;
	}

	void Camera::setZoom(double zoom) {
		m_zoom = zoom;
		if (m_zoom < 0.001) {
			m_zoom = 0.001;
		}
		updateMatrices();
		m_iswarped = true;
	}

	double Camera::getZoom() const {
		return m_zoom;
	}

	void Camera::setCellImageDimensions(unsigned int width, unsigned int height) {
		m_screen_cell_width = width;
		m_screen_cell_height = height;
		updateReferenceScale();
		updateMatrices();
		m_iswarped = true;
	}

	void Camera::setLocation(const Location& location) {
		// initialize first set properly
		if ((m_prev_origo == m_cur_origo) && (m_prev_origo == ScreenPoint(0,0,0))) {
			m_cur_origo = toScreenCoordinates(ExactModelCoordinate(0,0,0));
			m_prev_origo = m_cur_origo;
		}
		m_location = location;

		CellGrid* cg = NULL;
		if (m_location.getLayer()) {
			cg = m_location.getLayer()->getCellGrid();
		} else {
			throw Exception("Location without layer given to Camera::setLocation");
		}
		if (!cg) {
			throw Exception("Camera layer has no cellgrid specified");
		}

		updateMatrices();

		m_cur_origo = toScreenCoordinates(ExactModelCoordinate(0,0,0));
	}

	Point Camera::getCellImageDimensions() {
		return getCellImageDimensions(m_location.getLayer());
	}

	Point Camera::getCellImageDimensions(Layer* layer) {
		if (layer == m_location.getLayer()) {
			return Point( m_screen_cell_width, m_screen_cell_height );
		}
		std::map<Layer*, Point>::iterator it = m_image_dimensions.find(layer);
		if (it != m_image_dimensions.end()) {
			return it->second;
		}
		Point p;
		CellGrid* cg = layer->getCellGrid();
		assert(cg);
		DoublePoint dimensions = getLogicalCellDimensions(layer);
		p.x = static_cast<int>(round(m_reference_scale * dimensions.x));
		p.y = static_cast<int>(round(m_reference_scale * dimensions.y));
		m_image_dimensions[layer] = p;
		return p;
	}

	Location Camera::getLocation() const {
		return m_location;
	}

	Location& Camera::getLocationRef() {
		return m_location;
	}

	void Camera::setViewPort(const Rect& viewport) {
		m_viewport = viewport;
	}

	const Rect& Camera::getViewPort() const {
		return m_viewport;
	}

	void Camera::setEnabled(bool enabled) {
		m_enabled = enabled;
	}

	bool Camera::isEnabled() {
		return m_enabled;
	}

	void Camera::updateMatrices() {
		double scale = m_reference_scale;
		m_matrix.loadScale(scale, scale, scale);
		if (m_location.getLayer()) {
			CellGrid* cg = m_location.getLayer()->getCellGrid();
			if (cg) {
				ExactModelCoordinate pt = m_location.getMapCoordinates();
				m_matrix.applyTranslate( -pt.x *m_reference_scale,-pt.y *m_reference_scale, 0);
			}
		}
		scale = m_zoom;
		m_matrix.applyScale(scale, scale, scale);
		m_matrix.applyRotate(-m_rotation, 0.0, 0.0, 1.0);
		m_matrix.applyRotate(-m_tilt, 1.0, 0.0, 0.0);
		m_inverse_matrix = m_matrix.inverse();
	}

	void Camera::calculateZValue(ScreenPoint& screen_coords) {
		int dy = -(screen_coords.y - toScreenCoordinates(m_location.getMapCoordinates()).y);
		screen_coords.z = static_cast<int>(tan(m_tilt * (M_PI / 180.0)) * static_cast<double>(dy));
	}

	ExactModelCoordinate Camera::toMapCoordinates(ScreenPoint screen_coords, bool z_calculated) {
		if (!z_calculated) {
			calculateZValue(screen_coords);
		}
		screen_coords.x -= m_viewport.w / 2;
		screen_coords.y -= m_viewport.h / 2;

		return m_inverse_matrix  * intPt2doublePt(screen_coords);
	}

	ScreenPoint Camera::toScreenCoordinates(ExactModelCoordinate elevation_coords) {
		ExactModelCoordinate p = elevation_coords;
		ScreenPoint pt = doublePt2intPt( m_matrix* p  );
		pt.x += m_viewport.w / 2;
		pt.y += m_viewport.h / 2;
		return pt;
	}

	DoublePoint Camera::getLogicalCellDimensions(Layer* layer) {
		CellGrid* cg = NULL;
		if (layer) {
			cg = layer->getCellGrid();
		}
		assert(cg);

		ModelCoordinate cell(0,0);
		std::vector<ExactModelCoordinate> vertices;
		cg->getVertices(vertices, cell);

		DoubleMatrix mtx;
		mtx.loadRotate(m_rotation, 0.0, 0.0, 1.0);
		mtx.applyRotate(m_tilt, 1.0, 0.0, 0.0);
		double x1, x2, y1, y2;
		for (unsigned int i = 0; i < vertices.size(); i++) {
			vertices[i] = cg->toMapCoordinates(vertices[i]);
			vertices[i] = mtx * vertices[i];
			if (i == 0) {
				x1 = x2 = vertices[0].x;
				y1 = y2 = vertices[0].y;
			} else {
				x1 = std::min(vertices[i].x, x1);
				x2 = std::max(vertices[i].x, x2);
				y1 = std::min(vertices[i].y, y1);
				y2 = std::max(vertices[i].y, y2);
			}
		}
		return DoublePoint( x2 - x1, y2 - y1 );
	}

	void Camera::updateReferenceScale() {
		DoublePoint dim = getLogicalCellDimensions(m_location.getLayer());
		m_reference_scale = static_cast<double>(m_screen_cell_width) / dim.x;

		FL_DBG(_log, "Updating reference scale");
		FL_DBG(_log, LMsg("   tilt=") << m_tilt << " rot=" << m_rotation);
		FL_DBG(_log, LMsg("   m_screen_cell_width=") << m_screen_cell_width);
	}

	void Camera::getMatchingInstances(ScreenPoint screen_coords, Layer& layer, std::list<Instance*>& instances) {
		instances.clear();
		const std::vector<Instance*>& layer_instances = m_layer_to_instances[&layer];
		std::vector<Instance*>::const_iterator instance_it = layer_instances.end();
		while (instance_it != layer_instances.begin()) {
			--instance_it;
			Instance* i = (*instance_it);
			InstanceVisual* visual = i->getVisual<InstanceVisual>();
			InstanceVisualCacheItem& vc = visual->getCacheItem(this);
			if ((vc.dimensions.contains(Point(screen_coords.x, screen_coords.y)))) {
				assert(vc.image);
				Uint8 r, g, b, a;
				int x = screen_coords.x - vc.dimensions.x;
				int y = screen_coords.y - vc.dimensions.y;
				if (m_zoom != 1.0) {
					double fx = static_cast<double>(x);
					double fy = static_cast<double>(y);
					double fow = static_cast<double>(vc.image->getWidth());
					double foh = static_cast<double>(vc.image->getHeight());
					double fsw = static_cast<double>(vc.dimensions.w);
					double fsh = static_cast<double>(vc.dimensions.h);
					x = static_cast<int>(round(fx / fsw * fow));
					y = static_cast<int>(round(fy / fsh * foh));
				}
				vc.image->getPixelRGBA(x, y, &r, &g, &b, &a);
				// instance is hit with mouse if not totally transparent
				if (a != 0) {
					instances.push_back(i);
				}
			}
		}
	}

	void Camera::getMatchingInstances(Rect screen_rect, Layer& layer, std::list<Instance*>& instances) {
		instances.clear();
		const std::vector<Instance*>& layer_instances = m_layer_to_instances[&layer];
		std::vector<Instance*>::const_iterator instance_it = layer_instances.end();
		while (instance_it != layer_instances.begin()) {
			--instance_it;
			Instance* i = (*instance_it);
			InstanceVisual* visual = i->getVisual<InstanceVisual>();
			InstanceVisualCacheItem& vc = visual->getCacheItem(this);
			if ((vc.dimensions.intersects(screen_rect))) {
				assert(vc.image);
				Uint8 r, g, b, a;
				for(int xx = screen_rect.x; xx < screen_rect.x + screen_rect.w; xx++) {
					for(int yy = screen_rect.y; yy < screen_rect.y + screen_rect.h; yy++) {
						if ((vc.dimensions.contains(Point(xx, yy)))) {
							int x = xx - vc.dimensions.x;
							int y = yy - vc.dimensions.y;
							if (m_zoom != 1.0) {
								double fx = static_cast<double>(x);
								double fy = static_cast<double>(y);
								double fow = static_cast<double>(vc.image->getWidth());
								double foh = static_cast<double>(vc.image->getHeight());
								double fsw = static_cast<double>(vc.dimensions.w);
								double fsh = static_cast<double>(vc.dimensions.h);
								x = static_cast<int>(round(fx / fsw * fow));
								y = static_cast<int>(round(fy / fsh * foh));
							}
							vc.image->getPixelRGBA(x, y, &r, &g, &b, &a);
							// instance is hit with mouse if not totally transparent
							if (a != 0) {
								instances.push_back(i);
								goto found_non_transparent_pixel;
							}
						}
					}
				}
				found_non_transparent_pixel:;
			}
		}
	}

	void Camera::getMatchingInstances(Location& loc, std::list<Instance*>& instances, bool use_exactcoordinates) {
		instances.clear();
		const std::vector<Instance*>& layer_instances = m_layer_to_instances[loc.getLayer()];
		std::vector<Instance*>::const_iterator instance_it = layer_instances.end();
		while (instance_it != layer_instances.begin()) {
			--instance_it;
			Instance* i = (*instance_it);
			if (use_exactcoordinates) {
				if (i->getLocationRef().getExactLayerCoordinatesRef() == loc.getExactLayerCoordinatesRef()) {
					instances.push_back(i);
				}
			} else {
				if (i->getLocationRef().getLayerCoordinates() == loc.getLayerCoordinates()) {
					instances.push_back(i);
				}
			}
		}
	}

	void Camera::attach(Instance *instance) {
		// fail if the layers aren't the same
		if (m_location.getLayer()->getId() != instance->getLocation().getLayer()->getId()) {
			FL_WARN(_log, "Tried to attach camera to instance on different layer.");
			return ;
		}
		m_attachedto = instance;
	}

	void Camera::detach() {
		m_attachedto = NULL;
	}

	void Camera::update() {
		if( !m_attachedto ) {
			return;
		}
		Location loc(m_location);
		loc.setExactLayerCoordinates( m_attachedto->getLocationRef().getExactLayerCoordinates(m_location.getLayer()) );
		setLocation(loc);
		updateMatrices();
	}

	void Camera::refresh() {
		updateMatrices();
		m_iswarped = true;
	}

	void Camera::resetUpdates() {
		m_iswarped = false;
		m_prev_origo = m_cur_origo;
	}

	bool pipelineSort(const RendererBase* lhs, const RendererBase* rhs) {
		return (lhs->getPipelinePosition() < rhs->getPipelinePosition());
	}

	void Camera::addRenderer(RendererBase* renderer) {
		renderer->setRendererListener(this);
		m_renderers[renderer->getName()] = renderer;
		if (renderer->isEnabled()) {
			m_pipeline.push_back(renderer);
		}
		m_pipeline.sort(pipelineSort);
	}

	void Camera::onRendererPipelinePositionChanged(RendererBase* renderer) {
		m_pipeline.sort(pipelineSort);
	}

	void Camera::onRendererEnabledChanged(RendererBase* renderer) {
		assert(m_renderers[renderer->getName()]);
		if (renderer->isEnabled()) {
			FL_LOG(_log, LMsg("Enabling renderer ") << renderer->getName());
			m_pipeline.push_back(renderer);
			m_pipeline.sort(pipelineSort);
		} else {
			m_pipeline.remove(renderer);
		}
	}

	RendererBase* Camera::getRenderer(const std::string& name) {
		return m_renderers[name];
	}

	class InstanceDistanceSort {
	public:
		Camera* cam;
		inline bool operator()(const Instance* lhs, const Instance* rhs) {
			InstanceVisual* liv = lhs->getVisual<InstanceVisual>();
			InstanceVisual* riv = rhs->getVisual<InstanceVisual>();
			InstanceVisualCacheItem& lic = liv->getCacheItem(cam);
			InstanceVisualCacheItem& ric = riv->getCacheItem(cam);
			if (lic.screenpoint.z == ric.screenpoint.z) {
				return liv->getStackPosition() < riv->getStackPosition();
			}
			return lic.screenpoint.z < ric.screenpoint.z;
		}
	};

	void Camera::resetRenderers() {
		std::map<std::string, RendererBase*>::iterator r_it = m_renderers.begin();
		for (; r_it != m_renderers.end(); ++r_it) {
			Map* map = m_location.getMap();
			r_it->second->reset();
			r_it->second->activateAllLayers(map);
		}
	}

	void Camera::render() {
		ScreenPoint cammove = m_prev_origo - m_cur_origo;

		Map* map = m_location.getMap();
		if (!map) {
			FL_ERR(_log, "No map for camera found");
			return;
		}
		//if ((!map->isChanged()) && (!m_iswarped) && (cammove == ScreenPoint(0,0,0))) {
		//	return;
		//}

		// update each layer
		m_renderbackend->pushClipArea(getViewPort());

		m_layer_to_instances.clear();

		const std::list<Layer*>& layers = map->getLayers();
		std::list<Layer*>::const_iterator layer_it = layers.begin();
		for (;layer_it != layers.end(); ++layer_it) {

			// sort instances on layer based on stack position + camera distance. done only once
			//  here instead passing it to each renderer.
			// instances are checked first if their image intersects with the viewport.
			// this reduces processing load during sorting later
			const std::vector<Instance*>& allinstances = (*layer_it)->getInstances();
			std::vector<Instance*>::const_iterator instance_it = allinstances.begin();
			std::vector<Instance*>& instances_to_render = m_layer_to_instances[*layer_it];
			for (;instance_it != allinstances.end(); ++instance_it) {
				Instance* instance = *instance_it;
				InstanceVisual* visual = instance->getVisual<InstanceVisual>();
				if(!visual->isVisible())
					continue;
				InstanceVisualCacheItem& vc = visual->getCacheItem(this);

				// use cached values if there is no need to do full recalculation
				ScreenPoint drawpt;
				int angle = 0;
				if (m_updated && (!m_iswarped) && (!(instance->getChangeInfo() & (ICHANGE_LOC | ICHANGE_ROTATION | ICHANGE_FACING_LOC))) && (vc.image)) {
					int pos_estimate_x = vc.screenpoint.x - cammove.x;
					int pos_estimate_y = vc.screenpoint.y - cammove.y;
					int pos_estimate_z = vc.screenpoint.z - cammove.z;
					angle = vc.facing_angle;
					//std::cout << "orig x = " << drawpt.x << ", est x = " << pos_estimate_x << "\n";
					//std::cout << "orig y = " << drawpt.y << ", est y = " << pos_estimate_y << "\n";
					drawpt.x = pos_estimate_x;
					drawpt.y = pos_estimate_y;
					drawpt.z = pos_estimate_z;
					//drawpt.z = toScreenCoordinates( instance->getLocationRef().getMapCoordinates() ).z;
				} else {
					drawpt = toScreenCoordinates( instance->getLocationRef().getMapCoordinates() );
					vc.facing_angle = getAngleBetween(instance->getLocationRef(), instance->getFacingLocation());
					angle = vc.facing_angle;
				}

				//generate angle based on camera rotation and instance rotation
				angle = angle + m_rotation + instance->getRotation();

				Image* image = NULL;
				Action* action = instance->getCurrentAction();
				if (action) {
					FL_DBG(_log, "Instance has action");
					int animation_id = action->getVisual<ActionVisual>()->getAnimationIndexByAngle(vc.facing_angle + m_rotation);

					int facing_angle = vc.facing_angle;
					if (facing_angle < 0){
						facing_angle = 360+facing_angle; //make it a positive angle
					}
					facing_angle %= 360;
					instance->setRotation(facing_angle);  //update so the animation has correct rotation

					Animation& animation = m_apool->getAnimation(animation_id);
					unsigned int animtime = instance->getActionRuntime() % animation.getDuration();
					image = animation.getFrameByTimestamp(animtime);
				} else {
					FL_DBG(_log, "No action");
					int imageid = vc.getStaticImageIndexByAngle(angle, instance);
					FL_DBG(_log, LMsg("Instance does not have action, using static image with id ") << imageid);
					if (imageid >= 0) {
						image = &m_ipool->getImage(imageid);
					} else {
						// there was no static image for instance, trying default action
						action = instance->getObject()->getDefaultAction();
						if (action) {
							int animation_id = action->getVisual<ActionVisual>()->getAnimationIndexByAngle(angle);
							Animation& animation = m_apool->getAnimation(animation_id);
							unsigned int animtime = instance->getActionRuntime() % animation.getDuration();
							image = animation.getFrameByTimestamp(animtime);
						}
					}
				}
				if (image) {
					vc.image = image;
					vc.screenpoint = drawpt;

					int w = image->getWidth();
					int h = image->getHeight();
					drawpt.x -= w / 2;
					drawpt.x += image->getXShift();
					drawpt.y -= h / 2;
					drawpt.y += image->getYShift();
					Rect r = Rect(drawpt.x, drawpt.y, w, h);

					vc.dimensions = r;
					if (m_zoom != 1.0) {
						// NOTE: Due to image alignment, there is additional additions and substractions on image dimensions
						//       There's probabaly some better solution for this, but works "good enough" for now.
						//       In case additions / substractions are removed, gaps appear between tiles.
						r.w = static_cast<unsigned int>(ceil(static_cast<double>(vc.dimensions.w) * m_zoom)) + 2;
						r.h = static_cast<unsigned int>(ceil(static_cast<double>(vc.dimensions.h) * m_zoom)) + 2;
						int xo = static_cast<int>(ceil(static_cast<double>(vc.image->getXShift()) * m_zoom)) - vc.image->getXShift();
						int yo = static_cast<int>(ceil(static_cast<double>(vc.image->getYShift()) * m_zoom)) - vc.image->getYShift();
						r.x = vc.dimensions.x - static_cast<unsigned int>(ceil(static_cast<double>(r.w - vc.dimensions.w) / 2)) + xo - 1;
						r.y = vc.dimensions.y - static_cast<unsigned int>(ceil(static_cast<double>(r.h - vc.dimensions.h) / 2)) + yo - 1;
						vc.dimensions = r;
					}

					if (vc.dimensions.intersects(getViewPort())) {
						instances_to_render.push_back(instance);
					}
				}
			}

			InstanceDistanceSort ids;
			ids.cam = this;
			std::stable_sort(instances_to_render.begin(), instances_to_render.end(), ids);

			std::list<RendererBase*>::iterator r_it = m_pipeline.begin();
			for (; r_it != m_pipeline.end(); ++r_it) {
				if ((*r_it)->isActivedLayer(*layer_it)) {
					(*r_it)->render(this, *layer_it, instances_to_render);
				}
			}
		}
		m_renderbackend->popClipArea();
		resetUpdates();
		m_updated = true;
	}

}