zcy
/
solvespace
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
solvespace/src/render/render.h

399 lines
14 KiB
C
Raw Normal View History

Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
//-----------------------------------------------------------------------------
// Backend-agnostic rendering interface, and various backends we use.
//
// Copyright 2016 whitequark
//-----------------------------------------------------------------------------
#ifndef SOLVESPACE_RENDER_H
#define SOLVESPACE_RENDER_H
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
//-----------------------------------------------------------------------------
Implement a platform abstraction for timers. This commit temporarily disables tooltip functionality; it will be handled natively in a platform abstraction for windows using much simpler code.
2018-07-11 05:35:31 +00:00
// Interfaces common for all renderers
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
//-----------------------------------------------------------------------------
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
enum class StipplePattern : uint32_t;
// A mapping from 3d sketch coordinates to 2d screen coordinates, using
// an axonometric projection.
class Camera {
public:
Implement a platform abstraction for windows. This commit removes a large amount of code partially duplicated between the text and the graphics windows, and opens the path to having more than one model window on screen at any given time, as well as simplifies platform work. This commit also adds complete support for High-DPI device pixel ratio. It adds support for font scale factor (a fractional factor on top of integral device pixel ratio) on the platform side, but not on the application side. This commit also adds error checking to all Windows API calls (within the abstracted code) and fixes a significant number of misuses and non-future-proof uses of Windows API. This commit also makes uses of Windows API idiomatic, e.g. using the built-in vertical scroll bar, native tooltips, control subclassing instead of hooks in the global dispatch loop, and so on. It reinstates tooltip support and removes menu-related hacks.
2018-07-12 19:29:44 +00:00
double width;
double height;
double pixelRatio;
bool gridFit;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
Vector offset;
Vector projRight;
Vector projUp;
double scale;
double tangent;
bool IsPerspective() const { return tangent != 0.0; }
Point2d ProjectPoint(Vector p) const;
Vector ProjectPoint3(Vector p) const;
Vector ProjectPoint4(Vector p, double *w) const;
Vector UnProjectPoint(Point2d p) const;
Vector UnProjectPoint3(Vector p) const;
Vector VectorFromProjs(Vector rightUpForward) const;
Vector AlignToPixelGrid(Vector v) const;
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
SBezier ProjectBezier(SBezier b) const;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
void LoadIdentity();
void NormalizeProjectionVectors();
};
// A description of scene lighting.
class Lighting {
public:
RgbaColor backgroundColor;
double ambientIntensity;
double lightIntensity[2];
Vector lightDirection[2];
};
Add an interface for view-independent rendering. To actually achieve improved performance with the OpenGL 2 renderer, we have to cache geometry that doesn't change when the viewport does (note that the rendered pixels can change quite dramatically because we can reconfigure shaders; e.g. stippling can be drawn in screen coordinates). This commit adds a BatchCanvas interface that can be implemented by renderers, and uses it for drawing entities such as lines and points.
2016-06-30 15:54:35 +00:00
class BatchCanvas;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
// An interface for populating a drawing area with geometry.
class Canvas {
public:
// Stroke and fill styles are addressed with handles to be able to quickly
// group geometry into indexed draw calls.
class hStroke {
public:
uint32_t v;
};
class hFill {
public:
uint32_t v;
};
// The layer of a geometry describes how it occludes other geometry.
// Within a layer, geometry with higher z-index occludes geometry with lower z-index,
// or geometry drawn earlier if z-indexes match.
enum class Layer {
NORMAL, // Occluded by geometry with lower Z coordinate
OCCLUDED, // Only drawn over geometry with lower Z coordinate
DEPTH_ONLY, // Like NORMAL, but only affects future occlusion, not color
BACK, // Always drawn below all other geometry
FRONT, // Always drawn above all other geometry
Implement BatchCanvas for GL2.
2016-06-30 15:54:35 +00:00
LAST = FRONT
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
};
Unify displayEdges and displayOutlines. This has the following benefits: * Less geometry to generate; we can do both in one pass; * Less geometry to draw; * Eliminate overdraw of outlines on top of emphasized edges; * In future, being able to seamlessly stitch stippled lines. The contour edges are now also drawn before emphasized edges; this makes intersections of contour and emphasized edges look better as the thinner emphasized edge doesn't clobber the depth buffer.
2016-06-26 06:39:27 +00:00
// The outlines are the collection of all edges that may be drawn.
// Outlines can be classified as emphasized or not; emphasized outlines indicate an abrupt
// change in the surface curvature. These are indicated by the SOutline tag.
// Outlines can also be classified as contour or not; contour outlines indicate the boundary
// of the filled mesh. Whether an outline is a part of contour or not depends on point of view.
enum class DrawOutlinesAs {
Implement an OpenGL 2 renderer. There are two main reasons to desire an OpenGL 2 renderer: 1. Compatibility. The compatibility profile, ironically, does not offer a lot of compatibility, and our OpenGL 1 renderer will not run on Android, iOS, or WebGL. 2. Performance. The immediate mode does not scale, and in fact becomes very slow with only a moderate amount of lines on screen, and only a somewhat large amount of triangles. This commit implements a basic OpenGL 2 renderer that uses only features from the (OpenGL 3.2) core profile. It is not yet faster than the OpenGL 1 renderer, primarily because it uses a lot of small draw calls. This commit uses OpenGL 2 on Linux and Mac OS X directly (i.e. links to the GL symbols from version 2+); on Windows this is impossible with the default drivers, so for now OpenGL 1 is still used there.
2016-06-30 15:54:35 +00:00
EMPHASIZED_AND_CONTOUR = 0, // Both emphasized and contour outlines
EMPHASIZED_WITHOUT_CONTOUR = 1, // Emphasized outlines except those also belonging to contour
CONTOUR_ONLY = 2 // Contour outlines only
Unify displayEdges and displayOutlines. This has the following benefits: * Less geometry to generate; we can do both in one pass; * Less geometry to draw; * Eliminate overdraw of outlines on top of emphasized edges; * In future, being able to seamlessly stitch stippled lines. The contour edges are now also drawn before emphasized edges; this makes intersections of contour and emphasized edges look better as the thinner emphasized edge doesn't clobber the depth buffer.
2016-06-26 06:39:27 +00:00
};
Allow selecting unit (px/mm) in Canvas::Stroke. By directly specifying the desired end result to the renderer, we can avoid regeneration of geometry.
2016-06-30 15:54:35 +00:00
// Stroke widths, etc, can be scale-invariant (in pixels) or scale-dependent (in millimeters).
enum class Unit {
MM,
PX
};
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
class Stroke {
public:
hStroke h;
Layer layer;
int zIndex;
RgbaColor color;
double width;
Allow selecting unit (px/mm) in Canvas::Stroke. By directly specifying the desired end result to the renderer, we can avoid regeneration of geometry.
2016-06-30 15:54:35 +00:00
Unit unit;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
StipplePattern stipplePattern;
double stippleScale;
Eliminate several memory leaks. All leaks found with valgrind while running the test suite. This commit also clears the Cairo cache to improve SNR of Valgrind output.
2016-08-01 03:52:12 +00:00
void Clear() { *this = {}; }
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
bool Equals(const Stroke &other) const;
Allow selecting unit (px/mm) in Canvas::Stroke. By directly specifying the desired end result to the renderer, we can avoid regeneration of geometry.
2016-06-30 15:54:35 +00:00
double WidthMm(const Camera &camera) const;
double WidthPx(const Camera &camera) const;
double StippleScaleMm(const Camera &camera) const;
double StippleScalePx(const Camera &camera) const;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
};
enum class FillPattern {
SOLID, CHECKERED_A, CHECKERED_B
};
class Fill {
public:
hFill h;
Layer layer;
int zIndex;
RgbaColor color;
FillPattern pattern;
Implement an OpenGL 2 renderer. There are two main reasons to desire an OpenGL 2 renderer: 1. Compatibility. The compatibility profile, ironically, does not offer a lot of compatibility, and our OpenGL 1 renderer will not run on Android, iOS, or WebGL. 2. Performance. The immediate mode does not scale, and in fact becomes very slow with only a moderate amount of lines on screen, and only a somewhat large amount of triangles. This commit implements a basic OpenGL 2 renderer that uses only features from the (OpenGL 3.2) core profile. It is not yet faster than the OpenGL 1 renderer, primarily because it uses a lot of small draw calls. This commit uses OpenGL 2 on Linux and Mac OS X directly (i.e. links to the GL symbols from version 2+); on Windows this is impossible with the default drivers, so for now OpenGL 1 is still used there.
2016-06-30 15:54:35 +00:00
std::shared_ptr<const Pixmap> texture;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
Eliminate several memory leaks. All leaks found with valgrind while running the test suite. This commit also clears the Cairo cache to improve SNR of Valgrind output.
2016-08-01 03:52:12 +00:00
void Clear() { *this = {}; }
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
bool Equals(const Fill &other) const;
};
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
IdList<Stroke, hStroke> strokes = {};
IdList<Fill, hFill> fills = {};
BitmapFont bitmapFont = {};
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
Eliminate several memory leaks. All leaks found with valgrind while running the test suite. This commit also clears the Cairo cache to improve SNR of Valgrind output.
2016-08-01 03:52:12 +00:00
virtual void Clear();
Fix "virtual function without virtual destructor" warning. NFC.
2019-08-21 16:07:07 +00:00
virtual ~Canvas() = default;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
hStroke GetStroke(const Stroke &stroke);
hFill GetFill(const Fill &fill);
Use a dedicated BitmapFont instance per Canvas. Before this commit, updates to the bitmap font in the graphics window cause missing characters in the text window and vice versa. This is because BitmapFont contains a texture that's mutated, and sharing it would also require sharing display lists between GL contexts, which is not done (and overly complicated for this case anyway).
2017-01-10 02:52:51 +00:00
BitmapFont *GetBitmapFont();
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
virtual const Camera &GetCamera() const = 0;
virtual void DrawLine(const Vector &a, const Vector &b, hStroke hcs) = 0;
virtual void DrawEdges(const SEdgeList &el, hStroke hcs) = 0;
virtual bool DrawBeziers(const SBezierList &bl, hStroke hcs) = 0;
Unify displayEdges and displayOutlines. This has the following benefits: * Less geometry to generate; we can do both in one pass; * Less geometry to draw; * Eliminate overdraw of outlines on top of emphasized edges; * In future, being able to seamlessly stitch stippled lines. The contour edges are now also drawn before emphasized edges; this makes intersections of contour and emphasized edges look better as the thinner emphasized edge doesn't clobber the depth buffer.
2016-06-26 06:39:27 +00:00
virtual void DrawOutlines(const SOutlineList &ol, hStroke hcs, DrawOutlinesAs drawAs) = 0;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
virtual void DrawVectorText(const std::string &text, double height,
const Vector &o, const Vector &u, const Vector &v,
hStroke hcs) = 0;
virtual void DrawQuad(const Vector &a, const Vector &b, const Vector &c, const Vector &d,
hFill hcf) = 0;
Use Canvas::Stroke, not Canvas::Fill, for drawing a point. Our points are more like fat lines than actual quads, in that they are scale-invariant.
2016-07-17 14:46:38 +00:00
virtual void DrawPoint(const Vector &o, hStroke hcs) = 0;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
virtual void DrawPolygon(const SPolygon &p, hFill hcf) = 0;
Reimplement drawing of the mesh wireframe. OpenGL 2 and newer do not have the glPolygonMode(..., GL_LINES) API, so produce the wireframe on our side. It's somewhat slow, and draws every line three times, but it is cached when the OpenGL 2 renderer is used, and this should do for a debugging feature.
2016-11-17 16:47:45 +00:00
virtual void DrawMesh(const SMesh &m, hFill hcfFront, hFill hcfBack = {}) = 0;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
virtual void DrawFaces(const SMesh &m, const std::vector<uint32_t> &faces, hFill hcf) = 0;
virtual void DrawPixmap(std::shared_ptr<const Pixmap> pm,
const Vector &o, const Vector &u, const Vector &v,
const Point2d &ta, const Point2d &tb, hFill hcf) = 0;
virtual void InvalidatePixmap(std::shared_ptr<const Pixmap> pm) = 0;
Add an interface for view-independent rendering. To actually achieve improved performance with the OpenGL 2 renderer, we have to cache geometry that doesn't change when the viewport does (note that the rendered pixels can change quite dramatically because we can reconfigure shaders; e.g. stippling can be drawn in screen coordinates). This commit adds a BatchCanvas interface that can be implemented by renderers, and uses it for drawing entities such as lines and points.
2016-06-30 15:54:35 +00:00
virtual std::shared_ptr<BatchCanvas> CreateBatch();
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
};
Traits work for handles to permit sharing functionality and operators. NFC.
2019-07-09 14:41:46 +00:00
template<>
struct IsHandleOracle<Canvas::hStroke> : std::true_type {};
template<>
struct IsHandleOracle<Canvas::hFill> : std::true_type {};
Add an interface for view-independent rendering. To actually achieve improved performance with the OpenGL 2 renderer, we have to cache geometry that doesn't change when the viewport does (note that the rendered pixels can change quite dramatically because we can reconfigure shaders; e.g. stippling can be drawn in screen coordinates). This commit adds a BatchCanvas interface that can be implemented by renderers, and uses it for drawing entities such as lines and points.
2016-06-30 15:54:35 +00:00
// An interface for view-dependent visualization.
Hide OpenGL implementation details. Abstract the exact details of the OpenGL renderer in the render.h header; this allows us to use GL-specific types in the renderer class and functions without including OpenGL (and Windows, where applicable) headers in every source file.
2016-06-28 08:48:06 +00:00
class ViewportCanvas : public Canvas {
public:
GTK: remove GlWidget, use standard Gtk::GLArea. This removes a large number of hacks from the codebase, speeds up the rendering, and removes tearing when dragging entities.
2017-01-02 16:21:05 +00:00
virtual void SetCamera(const Camera &camera) = 0;
Hide OpenGL implementation details. Abstract the exact details of the OpenGL renderer in the render.h header; this allows us to use GL-specific types in the renderer class and functions without including OpenGL (and Windows, where applicable) headers in every source file.
2016-06-28 08:48:06 +00:00
virtual void SetLighting(const Lighting &lighting) = 0;
Refactor the renderer frame flush functionality. This commit does three things: * Recognizes that BeginFrame()/EndFrame() are badly named, since BeginFrame() sets up framebuffer, and EndFrame() flushes a frame, and they do not have to be called in pairs; and so renames them to NewFrame()/FlushFrame(). * Reduces the amount of frame flushes in GraphicsWindow::Paint() to two, which is the minimum since we use two different cameras for geometry and UI; * Changes the FPS measurement code to only take into account the time spent rendering our main geometry, and not the UI rendering or window system interaction time.
2016-11-16 06:14:22 +00:00
virtual void NewFrame() = 0;
virtual void FlushFrame() = 0;
Hide OpenGL implementation details. Abstract the exact details of the OpenGL renderer in the render.h header; this allows us to use GL-specific types in the renderer class and functions without including OpenGL (and Windows, where applicable) headers in every source file.
2016-06-28 08:48:06 +00:00
virtual std::shared_ptr<Pixmap> ReadFrame() = 0;
virtual void GetIdent(const char **vendor, const char **renderer, const char **version) = 0;
};
Add an interface for view-independent rendering. To actually achieve improved performance with the OpenGL 2 renderer, we have to cache geometry that doesn't change when the viewport does (note that the rendered pixels can change quite dramatically because we can reconfigure shaders; e.g. stippling can be drawn in screen coordinates). This commit adds a BatchCanvas interface that can be implemented by renderers, and uses it for drawing entities such as lines and points.
2016-06-30 15:54:35 +00:00
// An interface for view-independent visualization.
class BatchCanvas : public Canvas {
public:
const Camera &GetCamera() const override;
virtual void Finalize() = 0;
virtual void Draw() = 0;
};
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
// A wrapper around Canvas that simplifies drawing UI in screen coordinates.
class UiCanvas {
public:
Hide OpenGL implementation details. Abstract the exact details of the OpenGL renderer in the render.h header; this allows us to use GL-specific types in the renderer class and functions without including OpenGL (and Windows, where applicable) headers in every source file.
2016-06-28 08:48:06 +00:00
std::shared_ptr<Canvas> canvas;
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
bool flip = false;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
Allow using z-index in UiCanvas, and use it for tooltips. Before this commit, tooltips in the text window are drawn under the red "X" indicating a disabled button. After this commit, they are moved on top of that. This commit also alters the OpenGL renderers' SetCamera() method to clear the depth buffer, as that would interfere with drawing the UI; the toolbar would get occluded by geometry.
2016-11-15 21:35:06 +00:00
void DrawLine(int x1, int y1, int x2, int y2, RgbaColor color, int width = 1,
int zIndex = 0);
void DrawRect(int l, int r, int t, int b, RgbaColor fillColor, RgbaColor outlineColor,
int zIndex = 0);
void DrawPixmap(std::shared_ptr<const Pixmap> pm, int x, int y,
int zIndex = 0);
void DrawBitmapChar(char32_t codepoint, int x, int y, RgbaColor color,
int zIndex = 0);
void DrawBitmapText(const std::string &str, int x, int y, RgbaColor color,
int zIndex = 0);
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
int Flip(int y) const { return flip ? (int)canvas->GetCamera().height - y : y; }
};
// A canvas that performs picking against drawn geometry.
class ObjectPicker : public Canvas {
public:
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
Camera camera = {};
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
// Configuration.
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
Point2d point = {};
double selRadius = 0.0;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
// Picking state.
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
double minDistance = 0.0;
int maxZIndex = 0;
uint32_t position = 0;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
const Camera &GetCamera() const override { return camera; }
void DrawLine(const Vector &a, const Vector &b, hStroke hcs) override;
void DrawEdges(const SEdgeList &el, hStroke hcs) override;
bool DrawBeziers(const SBezierList &bl, hStroke hcs) override { return false; }
Unify displayEdges and displayOutlines. This has the following benefits: * Less geometry to generate; we can do both in one pass; * Less geometry to draw; * Eliminate overdraw of outlines on top of emphasized edges; * In future, being able to seamlessly stitch stippled lines. The contour edges are now also drawn before emphasized edges; this makes intersections of contour and emphasized edges look better as the thinner emphasized edge doesn't clobber the depth buffer.
2016-06-26 06:39:27 +00:00
void DrawOutlines(const SOutlineList &ol, hStroke hcs, DrawOutlinesAs drawAs) override;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
void DrawVectorText(const std::string &text, double height,
const Vector &o, const Vector &u, const Vector &v,
hStroke hcs) override;
void DrawQuad(const Vector &a, const Vector &b, const Vector &c, const Vector &d,
hFill hcf) override;
Use Canvas::Stroke, not Canvas::Fill, for drawing a point. Our points are more like fat lines than actual quads, in that they are scale-invariant.
2016-07-17 14:46:38 +00:00
void DrawPoint(const Vector &o, hStroke hcs) override;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
void DrawPolygon(const SPolygon &p, hFill hcf) override;
Reimplement drawing of the mesh wireframe. OpenGL 2 and newer do not have the glPolygonMode(..., GL_LINES) API, so produce the wireframe on our side. It's somewhat slow, and draws every line three times, but it is cached when the OpenGL 2 renderer is used, and this should do for a debugging feature.
2016-11-17 16:47:45 +00:00
void DrawMesh(const SMesh &m, hFill hcfFront, hFill hcfBack) override;
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
void DrawFaces(const SMesh &m, const std::vector<uint32_t> &faces, hFill hcf) override;
void DrawPixmap(std::shared_ptr<const Pixmap> pm,
const Vector &o, const Vector &u, const Vector &v,
const Point2d &ta, const Point2d &tb, hFill hcf) override;
void InvalidatePixmap(std::shared_ptr<const Pixmap> pm) override {}
void DoCompare(double distance, int zIndex, int comparePosition = 0);
void DoQuad(const Vector &a, const Vector &b, const Vector &c, const Vector &d,
int zIndex, int comparePosition = 0);
bool Pick(std::function<void()> drawFn);
};
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
// A canvas that renders onto a 2d surface, performing z-index sorting, occlusion testing, etc,
// on the CPU.
Implement a platform abstraction for windows. This commit removes a large amount of code partially duplicated between the text and the graphics windows, and opens the path to having more than one model window on screen at any given time, as well as simplifies platform work. This commit also adds complete support for High-DPI device pixel ratio. It adds support for font scale factor (a fractional factor on top of integral device pixel ratio) on the platform side, but not on the application side. This commit also adds error checking to all Windows API calls (within the abstracted code) and fixes a significant number of misuses and non-future-proof uses of Windows API. This commit also makes uses of Windows API idiomatic, e.g. using the built-in vertical scroll bar, native tooltips, control subclassing instead of hooks in the global dispatch loop, and so on. It reinstates tooltip support and removes menu-related hacks.
2018-07-12 19:29:44 +00:00
class SurfaceRenderer : public ViewportCanvas {
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
public:
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
Camera camera = {};
Lighting lighting = {};
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
// Chord tolerance, for converting beziers to pwl.
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
double chordTolerance = 0.0;
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
// Render lists.
handle_map<hStroke, SEdgeList> edges;
handle_map<hStroke, SBezierList> beziers;
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
SMesh mesh = {};
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
// State.
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
BBox bbox = {};
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
Eliminate several memory leaks. All leaks found with valgrind while running the test suite. This commit also clears the Cairo cache to improve SNR of Valgrind output.
2016-08-01 03:52:12 +00:00
void Clear() override;
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
// Canvas interface.
const Camera &GetCamera() const override { return camera; }
Implement a platform abstraction for windows. This commit removes a large amount of code partially duplicated between the text and the graphics windows, and opens the path to having more than one model window on screen at any given time, as well as simplifies platform work. This commit also adds complete support for High-DPI device pixel ratio. It adds support for font scale factor (a fractional factor on top of integral device pixel ratio) on the platform side, but not on the application side. This commit also adds error checking to all Windows API calls (within the abstracted code) and fixes a significant number of misuses and non-future-proof uses of Windows API. This commit also makes uses of Windows API idiomatic, e.g. using the built-in vertical scroll bar, native tooltips, control subclassing instead of hooks in the global dispatch loop, and so on. It reinstates tooltip support and removes menu-related hacks.
2018-07-12 19:29:44 +00:00
// ViewportCanvas interface.
void SetCamera(const Camera &camera) override { this->camera = camera; };
void SetLighting(const Lighting &lighting) override { this->lighting = lighting; }
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
void DrawLine(const Vector &a, const Vector &b, hStroke hcs) override;
void DrawEdges(const SEdgeList &el, hStroke hcs) override;
bool DrawBeziers(const SBezierList &bl, hStroke hcs) override;
void DrawOutlines(const SOutlineList &ol, hStroke hcs, DrawOutlinesAs drawAs) override;
void DrawVectorText(const std::string &text, double height,
const Vector &o, const Vector &u, const Vector &v,
hStroke hcs) override;
void DrawQuad(const Vector &a, const Vector &b, const Vector &c, const Vector &d,
hFill hcf) override;
Use Canvas::Stroke, not Canvas::Fill, for drawing a point. Our points are more like fat lines than actual quads, in that they are scale-invariant.
2016-07-17 14:46:38 +00:00
void DrawPoint(const Vector &o, hStroke hcs) override;
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
void DrawPolygon(const SPolygon &p, hFill hcf) override;
Reimplement drawing of the mesh wireframe. OpenGL 2 and newer do not have the glPolygonMode(..., GL_LINES) API, so produce the wireframe on our side. It's somewhat slow, and draws every line three times, but it is cached when the OpenGL 2 renderer is used, and this should do for a debugging feature.
2016-11-17 16:47:45 +00:00
void DrawMesh(const SMesh &m, hFill hcfFront, hFill hcfBack) override;
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
void DrawFaces(const SMesh &m, const std::vector<uint32_t> &faces, hFill hcf) override;
void DrawPixmap(std::shared_ptr<const Pixmap> pm,
const Vector &o, const Vector &u, const Vector &v,
const Point2d &ta, const Point2d &tb, hFill hcf) override;
void InvalidatePixmap(std::shared_ptr<const Pixmap> pm) override;
// Geometry manipulation.
void CalculateBBox();
void ConvertBeziersToEdges();
void CullOccludedStrokes();
// Renderer operations.
void OutputInPaintOrder();
virtual bool CanOutputCurves() const = 0;
virtual bool CanOutputTriangles() const = 0;
virtual void OutputStart() = 0;
virtual void OutputBezier(const SBezier &b, hStroke hcs) = 0;
virtual void OutputTriangle(const STriangle &tr) = 0;
virtual void OutputEnd() = 0;
void OutputBezierAsNonrationalCubic(const SBezier &b, hStroke hcs);
};
//-----------------------------------------------------------------------------
Implement a platform abstraction for timers. This commit temporarily disables tooltip functionality; it will be handled natively in a platform abstraction for windows using much simpler code.
2018-07-11 05:35:31 +00:00
// 2d renderers
Add a Cairo rendering backend. This backend will be used for reproducible offscreen rendering in integration tests.
2016-07-24 13:17:52 +00:00
//-----------------------------------------------------------------------------
class CairoRenderer : public SurfaceRenderer {
public:
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
cairo_t *context = NULL;
Add a simple harness for automated, headless testing. This commit alters the build system substantially; it adds another platform, `headless`, that provides stubs in place of all GUI functions, and provides a library `solvespace_headless` alongside the main executable. To cut down build times, only the few files that have #if defined(HEADLESS) are built twice for the executable and the library; the rest is grouped into a new `solvespace_cad` library. It is not usable on its own and just serves for grouping. This commit also gates the tests behind a -DENABLE_TESTS=ON CMake option, ON by default (but suggested as OFF in the README so that people don't ever have to install cairo to build the executable.) The tests introduced in this commit are (so far) rudimentary, although functional, and they serve as a stepping point towards introducing coverage analysis.
2016-07-25 19:37:48 +00:00
// Renderer configuration.
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
bool antialias = false;
Add a Cairo rendering backend. This backend will be used for reproducible offscreen rendering in integration tests.
2016-07-24 13:17:52 +00:00
// Renderer state.
struct {
hStroke hcs;
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
} current = {};
Add a Cairo rendering backend. This backend will be used for reproducible offscreen rendering in integration tests.
2016-07-24 13:17:52 +00:00
Implement a platform abstraction for windows. This commit removes a large amount of code partially duplicated between the text and the graphics windows, and opens the path to having more than one model window on screen at any given time, as well as simplifies platform work. This commit also adds complete support for High-DPI device pixel ratio. It adds support for font scale factor (a fractional factor on top of integral device pixel ratio) on the platform side, but not on the application side. This commit also adds error checking to all Windows API calls (within the abstracted code) and fixes a significant number of misuses and non-future-proof uses of Windows API. This commit also makes uses of Windows API idiomatic, e.g. using the built-in vertical scroll bar, native tooltips, control subclassing instead of hooks in the global dispatch loop, and so on. It reinstates tooltip support and removes menu-related hacks.
2018-07-12 19:29:44 +00:00
void Clear() override;
void NewFrame() override {}
void FlushFrame() override;
std::shared_ptr<Pixmap> ReadFrame() override;
void GetIdent(const char **vendor, const char **renderer, const char **version) override;
Add a Cairo rendering backend. This backend will be used for reproducible offscreen rendering in integration tests.
2016-07-24 13:17:52 +00:00
void SelectStroke(hStroke hcs);
void MoveTo(Vector p);
void FinishPath();
bool CanOutputCurves() const override { return true; }
bool CanOutputTriangles() const override { return true; }
void OutputStart() override;
void OutputBezier(const SBezier &b, hStroke hcs) override;
void OutputTriangle(const STriangle &tr) override;
void OutputEnd() override;
};
Fix a warning about a virtual destructor.
2019-09-10 03:33:05 +00:00
class CairoPixmapRenderer final : public CairoRenderer {
Implement a platform abstraction for windows. This commit removes a large amount of code partially duplicated between the text and the graphics windows, and opens the path to having more than one model window on screen at any given time, as well as simplifies platform work. This commit also adds complete support for High-DPI device pixel ratio. It adds support for font scale factor (a fractional factor on top of integral device pixel ratio) on the platform side, but not on the application side. This commit also adds error checking to all Windows API calls (within the abstracted code) and fixes a significant number of misuses and non-future-proof uses of Windows API. This commit also makes uses of Windows API idiomatic, e.g. using the built-in vertical scroll bar, native tooltips, control subclassing instead of hooks in the global dispatch loop, and so on. It reinstates tooltip support and removes menu-related hacks.
2018-07-12 19:29:44 +00:00
public:
std::shared_ptr<Pixmap> pixmap;
cairo_surface_t *surface = NULL;
void Init();
void Clear() override;
std::shared_ptr<Pixmap> ReadFrame() override;
};
Add a Cairo rendering backend. This backend will be used for reproducible offscreen rendering in integration tests.
2016-07-24 13:17:52 +00:00
//-----------------------------------------------------------------------------
Implement a platform abstraction for timers. This commit temporarily disables tooltip functionality; it will be handled natively in a platform abstraction for windows using much simpler code.
2018-07-11 05:35:31 +00:00
// 3d renderers
Add a new renderer that prepares geometry for 2d backends. SurfaceRenderer is a new renderer implementing the Canvas interface running entirely on the CPU; it projects strokes and triangles in the exact same way as OpenGL would, and it can be used for rendering into raster or vector 2d surfaces.
2016-07-21 20:27:53 +00:00
//-----------------------------------------------------------------------------
Refactor GlOffscreen; remove the GLEW dependency. It was never really needed, since both Linux and OS X, where GlOffscreen is used, guarantee that the API we need is present, on all OS versions we're interested in. Also, reorganize GlOffscreen consistently with the rest of our codebase, and don't use RAII for OpenGL resource management because of its requirement for an active context.
2016-07-24 14:42:44 +00:00
// An offscreen renderer based on OpenGL framebuffers.
class GlOffscreen {
public:
Start using C++ in-place member initialization. We should make good use of in-place member initialization. Many new classes have constructors that effectively do nothing but default-initialize POD members, and when adding new members, it is very easy to miss initializing them. With in-place initialization, the code is more compact, the diffs are nicer, and it's harder to miss them. This commit only converts render/ and platform/ to use in-place member initialization, since there was a bug in CairoRenderer, but we should convert the entire codebase.
2018-07-16 11:05:06 +00:00
unsigned int framebuffer = 0;
unsigned int colorRenderbuffer = 0;
unsigned int depthRenderbuffer = 0;
Refactor GlOffscreen; remove the GLEW dependency. It was never really needed, since both Linux and OS X, where GlOffscreen is used, guarantee that the API we need is present, on all OS versions we're interested in. Also, reorganize GlOffscreen consistently with the rest of our codebase, and don't use RAII for OpenGL resource management because of its requirement for an active context.
2016-07-24 14:42:44 +00:00
std::vector<uint8_t> data;
bool Render(int width, int height, std::function<void()> renderFn);
void Clear();
};
Implement a platform abstraction for timers. This commit temporarily disables tooltip functionality; it will be handled natively in a platform abstraction for windows using much simpler code.
2018-07-11 05:35:31 +00:00
//-----------------------------------------------------------------------------
// Factories
//-----------------------------------------------------------------------------
Hide OpenGL implementation details. Abstract the exact details of the OpenGL renderer in the render.h header; this allows us to use GL-specific types in the renderer class and functions without including OpenGL (and Windows, where applicable) headers in every source file.
2016-06-28 08:48:06 +00:00
std::shared_ptr<ViewportCanvas> CreateRenderer();
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
#endif