git_clone/thorvg
1
0
Fork 0
mirror of https://github.com/thorvg/thorvg.git synced 2025-06-14 12:04:29 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 5
3731 commits 31 branches 105 tags 403 MiB
Commit graph

660 commits

Author SHA1 Message Date
Hermet Park
d6fffd13c2 api: revise the engine initializer for the 1.0 release. 
This change introduces the CanvasEngine::All type to automatically
initialize the engines available on the current system.

These revisions improve the usability of these APIs.

Addtions:
- enum class CanvasEngine::All

Modifications:
- Result Initializer::init(CanvasEngine engine, uint32_t threads) ->
Result Initializer::init(uint32_t threads, CanvasEngine engine = tvg::CanvasEngine::All)
 2023-10-30 11:48:02 +09:00
Mira Grudzinska
25a1321243 common: stroke dash offset support with new apis. 
This change just allows users to use the offset of the stroke dash.
Actually feature enhacement has been introduced by
478e45f9f3.

@APIs:
uint32_t Shape::strokeDash(const float** dashPattern) ->
uint32_t Shape::strokeDash(const float** dashPattern, float* offset = nullptr)

Result Shape::strokeDash(const float* dashPattern, uint32_t cnt) ->
Result Shape::strokeDash(const float* dashPattern, uint32_t cnt, float offset = 0.0f)

Tvg_Result tvg_shape_set_stroke_dash(Tvg_Paint* paint, const float* dashPattern, uint32_t cnt) ->
Tvg_Result tvg_shape_set_stroke_dash(Tvg_Paint* paint, const float* dashPattern, uint32_t cnt, float offset)

Tvg_Result tvg_shape_get_stroke_dash(const Tvg_Paint* paint, const float** dashPattern, uint32_t* cnt) ->
Tvg_Result tvg_shape_get_stroke_dash(const Tvg_Paint* paint, const float** dashPattern, uint32_t* cnt, float* offset)

@Issue: https://github.com/thorvg/thorvg/issues/1372
 2023-10-30 11:47:51 +09:00
Hermet Park
f3a2d2a5a6 portability: addressed all compilation warnings from MSVC 2023-10-27 14:20:50 +09:00
Hermet Park
baee5ec767 renderer/paint: added a blend update flag. 
Keep track of the update changes accurately.
We can utilize this value change in the backend engine.
 2023-10-27 12:13:00 +09:00
Hermet Park
db55481e97 renamed stroke apis family. 
float Shape::stroke(float width) -> float Shape::strokeWidth(float width)
Result Shape::stroke(uint8_t r, uint8_t g, uint8_t b, uint8_t a = 255) -> Result Shape::strokeFill(uint8_t r, uint8_t g, uint8_t b, uint8_t a = 255)
Result Shape::stroke(std::unique_ptr<Fill> f) -> Result Shape::strokeFill(std::unique_ptr<Fill> f)
Result Shape::stroke(const float* dashPattern, uint32_t cnt, float offset = 0.0f) ->  Result Shape::strokeDash(const float* dashPattern, uint32_t cnt, float offset = 0.0f)
Result Shape::stroke(StrokeCap cap) -> Result Shape::strokeCap(StrokeCap cap)
Result Shape::stroke(StrokeJoin join) -> Result Shape::strokeJoin(StrokeJoin join)
Result Shape::strokeColor(uint8_t* r, uint8_t* g, uint8_t* b, uint8_t* a = nullptr) const -> Result Shape::strokeFill(uint8_t* r, uint8_t* g, uint8_t* b, uint8_t* a = nullptr) const

@Issue: https://github.com/thorvg/thorvg/issues/1372
 2023-10-27 11:46:51 +09:00
Hermet Park
989b995189 renderer: revise the internal paints structure. 
Get rid of the polymorphism function table,
use the switch directly instead.

We profiled, both binary & performance is better than before.

Tested on a local machine (single thread):
- Lottie: 2ms improved
- Binary: -0.5kb
 2023-10-27 11:24:44 +09:00
Hermet Park
978f85c3ea apis: optimize for compact data size. 
Still it needs to size down of the PathCommand.

@Issue: https://github.com/thorvg/thorvg/issues/1372
 2023-10-26 15:11:31 +09:00
Hermet Park
21911fa1c8 apis: remove deprecated 
- Result Picture::bounds(float* x, float* y, float* w, float* h) const
- Result Picture::load(const char* data, uint32_t size, bool copy = false)
- Result Canvas::reserve(uint32_t size)
- Result Scene::reserve(uint32_t size)

@Issue: https://github.com/thorvg/thorvg/issues/1372
 2023-10-26 15:03:13 +09:00
Hermet Park
b2499739ee renderer/shape: enable returning count values only 2023-10-26 13:02:54 +09:00
Hermet Park
6e3674b904 canvas/paint: ++exception handling 
enhanced reference count verification
to prevent unintentional deletion of used composition targets.
 2023-10-26 11:22:39 +09:00
SergeyLebedkin
8200bc5fc3
wg_engine: Added a feature to draw multiple radial gradient filled shapes 2023-10-24 18:24:16 +09:00
Hermet Park
e570064eba animation/lottie: updated the frame count unit. 
replace the frame count unit from the int32_t to float
since animations could smoothly interpolate key-frames.

This notificably improve the animation smoothness in Lottie

Beta API changes:
Result Animation::frame(uint32_t no) -> Result Animation::frame(float no)
uint32_t Animation::curFrame() const -> float Animation::curFrame() const
uint32_t Animation::totalFrame() const -> float Animation::totalFrame() const
 2023-10-24 11:49:57 +09:00
RuiwenTang
fea0d1bb77 gl_engine: use raw pointer to pass and hold GlRenderTask 2023-10-23 18:08:14 +09:00
Sergii Liebodkin
94eabc609c wg_engine: Added ability to draw multiple linear gradient filled shapes 
[issues 1479: LinearGradient](thorvg#1479)

In order to build you need third party libraries. Before you start please read this: [LearnWebGPU](https://eliemichel.github.io/LearnWebGPU/getting-started/hello-webgpu.html)

Usage example:

    // init glfw
    glfwInit();

    // create a windowed mode window and its opengl context
    glfwWindowHint(GLFW_CLIENT_API, GLFW_NO_API);
    GLFWwindow* window = glfwCreateWindow(800, 800, "WebGPU base app", nullptr, nullptr);

    // get window size
    int width{}, height{};
    glfwGetWindowSize(window, &width, &height);

    // init engine webgpu
    tvg::Initializer::init(tvg::CanvasEngine::Wg, 0);

    // create wg canvas
    auto canvasWg = tvg::WgCanvas::gen();
    canvas_wg->target(glfwGetWin32Window(window), width, height);

    // gradient color stops
    tvg::Fill::ColorStop colorStops[2];
    colorStops[0] = {0, 0, 0, 0, 255};
    colorStops[1] = {1, 255, 255, 255, 255};
    // linear gradient
    auto fill = tvg::LinearGradient::gen();
    fill->linear(0, 0, 400, 400);
    fill->colorStops(colorStops, 2);
    // prepare rectangle
    auto shape1 = tvg::Shape::gen();
    shape1->appendRect(0, 0, 400, 400); //x, y, w, h
    shape1->fill(std::move(fill));
    canvas_wg->push(std::move(shape1));

    // gradient color stops
    tvg::Fill::ColorStop colorStops2[3];
    colorStops2[0] = { 0, 255, 0, 0, 255 };
    colorStops2[1] = { 0.5, 255, 255, 0, 255 };
    colorStops2[2] = { 1, 255, 255, 255, 255 };
    // linear gradient
    auto fill2 = tvg::LinearGradient::gen();
    fill2->linear(400, 200, 400, 600);
    fill2->colorStops(colorStops2, 3);
    // prepare circle
    auto shape2 = tvg::Shape::gen();
    shape2->appendCircle(400, 400, 200, 200); //cx, cy, radiusW, radiusH
    shape2->fill(std::move(fill2));
    canvas_wg->push(std::move(shape2));

    // gradient color stops
    tvg::Fill::ColorStop colorStops3[4];
    colorStops3[0] = { 0, 0, 127, 0, 127 };
    colorStops3[1] = { 0.25, 0, 170, 170, 170 };
    colorStops3[2] = { 0.5, 200, 0, 200, 200 };
    colorStops3[3] = { 1, 255, 255, 255, 255 };
    // linear gradient
    auto fill3 = tvg::LinearGradient::gen();
    fill3->linear(450, 600, 750, 600);
    fill3->colorStops(colorStops3, 4);
    // prepare ellipse
    auto shape3 = tvg::Shape::gen();
    shape3->appendCircle(600, 600, 150, 100); //cx, cy, radiusW, radiusH
    shape3->fill(std::move(fill3));
    canvas_wg->push(std::move(shape3));

    while (!glfwWindowShouldClose(window)) {
        // webgpu
        canvas_wg->draw();
        canvas_wg->sync();

        // pull events
        glfwPollEvents();
    }

    // terminate engine and window
    tvg::Initializer::term(tvg::CanvasEngine::Wg);
    glfwDestroyWindow(window);
    glfwTerminate();
 2023-10-23 18:05:56 +09:00
Hermet Park
6a18e694d1 sw_engine/raster: optimized the scaled image rasterization 
Unified common logic for scaled image raster operations,
Avoid on-spot pixel computation as possible.

Tested on local machine (single thread)

Lottie: 0.057s -> 0.053s (-0.004s)
 2023-10-23 11:02:10 +09:00
Sergii Liebodkin
14b2508cd1 wg_engine: Added ability to draw multiple solid color filled shapes 
[issues 1479: Shape](https://github.com/thorvg/thorvg/issues/1479)

In order to build you need third party libraries. Before you start please read this: [LearnWebGPU](https://eliemichel.github.io/LearnWebGPU/getting-started/hello-webgpu.html)

Usage example:

    // init glfw
    glfwInit();

    // create a windowed mode window and its opengl context
    glfwWindowHint(GLFW_CLIENT_API, GLFW_NO_API);
    GLFWwindow* window = glfwCreateWindow(800, 800, "WebGPU base app", nullptr, nullptr);

    // get window size
    int width{}, height{};
    glfwGetWindowSize(window, &width, &height);

    // init engine webgpu
    tvg::Initializer::init(tvg::CanvasEngine::Wg, 0);

    // create wg canvas
    auto canvasWg = tvg::WgCanvas::gen();
    canvas_wg->target(glfwGetWin32Window(window), width, height);

    // prepare a shape (Rectangle + Rectangle + Circle + Circle)
    auto shape1 = tvg::Shape::gen();
    shape1->appendRect(0, 0, 200, 200);                //x, y, w, h
    shape1->appendRect(100, 100, 300, 300, 100, 100);  //x, y, w, h, rx, ry
    shape1->appendCircle(400, 400, 100, 100);          //cx, cy, radiusW, radiusH
    shape1->appendCircle(400, 500, 170, 100);          //cx, cy, radiusW, radiusH
    shape1->fill(255, 255, 0);                         //r, g, b

    canvas_wg->push(std::move(shape1));

    while (!glfwWindowShouldClose(window)) {
        // webgpu
        canvas_wg->draw();
        canvas_wg->sync();

        // pull events
        glfwPollEvents();
    }

    // terminate engine and window
    tvg::Initializer::term(tvg::CanvasEngine::Wg);
    glfwDestroyWindow(window);
    glfwTerminate();
 2023-10-20 18:37:52 +09:00
Hermet Park
61081c02af sw_engine raster: fixed a default alpha blending bug. 
alpha value has been missed by a mistake,
a regression by c50d2fd

Issue: https://github.com/thorvg/thorvg/issues/1716
 2023-10-20 13:27:17 +09:00
Hermet Park
d81f5d29fb sw_engine/math: fine-tuning optimization 
Try to minimize the use of sqrt() and arctan() calls
when possible. These calls can be relatively expensive
when accumulated within a single frame.

Also repalce the division with shift operation.
since split cubic function is one of the significant hot-spots
in the data processing, we could earn a noticable enhancement.

Tested on single thread local machine:

Lottie: 0.080 -> 0.052s (-0.028s)
Performance: 0.023 -> 0.022 (-0.001s)
Binary: +34
 2023-10-20 13:18:28 +09:00
Hermet Park
3ef740145f gl_engine: fix a compiler warning. 
../src/renderer/gl_engine/tvgGlRenderPass.cpp:37:29: warning: macro expands to multiple statements [-Wmultistatement-macros]
   37 |     if (mColorTex) GL_CHECK(glDeleteTextures(1, &mColorTex));
 2023-10-20 11:26:10 +09:00
RuiwenTang
6a2b34df73 gl_engine: support basic alpha composite 
* introduce a new class GlRenderPass to hold off-screen rendering result
* add basic alpha composite support with begin/end render-pass
 2023-10-20 10:51:04 +09:00
Hermet Park
4de2b953d1 sw_engine/math: Enhance trigonometric functions. 
Streamlining computations with floating-point operations in rotation
thereby improving 'thorvg' speed.

Also use the well-optimized posix math functions instead of
custom math.

Test on my local machine.

Lottie: -0.008s (0.073 -> 0.065)
Performance: -0.0013s (0.0154 -> 0.0141)
Binary: -323
 2023-10-19 10:39:50 +09:00
JunsuChoi
fdb2a17504 renderer/initializer: Support for initializing SW and GL engines together 
Initialize both renderers to allow init(SW | GL);
 2023-10-16 11:04:45 +09:00
SergeyLebedkin
8d5c728119
wg_engine: introduce a webgpu canvas(engine) 
WebGPU is a Render Hardware Interface built on top of the various APIs 
provided by the driver/OS depending on your platform. 

WebGPU exposes an API for performing operations, 
such as rendering and computation, on a Graphics Processing Unit.

WebGPU official documentation: https://www.w3.org/TR/webgpu/

The new engine type introduced: tvg::CanvasEngine::Wg

The new canvas type introduced: tvg::WgCanvas

Example:
$meson setup build -Dengines=wg_beta

`
    // init engine webgpu
    tvg::Initializer::init(tvg::CanvasEngine::Wg, 0);

    // create wg canvas
    auto canvasWg = tvg::WgCanvas::gen();
    canvas_wg->target(glfwGetWin32Window(window), width, height);

    // ...

    // terminate engine and window
    tvg::Initializer::term(tvg::CanvasEngine::Wg);
`
Still this feature is under the beta

Issue: https://github.com/thorvg/thorvg/issues/1479
 2023-10-13 22:59:32 +09:00
RuiwenTang
4722550e06 gl_engine: support clip by using scissor and stencil 2023-10-12 23:18:12 +09:00
Hermet Park
0cc6cfffef sw_engine/stroke: enhanced the quality of the dash line corners. 
Previously, the engine didn't properly cover the dash line corner styles
because it considered a new line to start at the corner.

This update modifies the logic to recognize curved lines
as a single line, including the corners.

There may still be some quality issues,
but it's an improvement over the previous version.

@Issue: https://github.com/thorvg/thorvg/issues/121
 2023-10-05 14:30:10 +09:00
RuiwenTang
6a3a03f29f gl_engine: support render image 2023-10-04 12:41:22 +09:00
Hermet Park
12260198d1 release: bump up version v0.11.0 2023-09-28 08:30:34 +09:00
Hermet Park
5ecd3fb479 sw_engine: Correct the color conversion condition. 
The color conversion is supposed to take into account the differences between
straight alpha premultiplied color and pre-multiplied alpha color.

The previous logic does not perfectly cover these conditions.

The problem was occured in the thorvg viewer with a jpeg bgra format.
 2023-09-26 18:57:34 +09:00
Hermet Park
51c3a8912c sw_engine: fix an invalid memory access. 
the surface and mesh data can be missed by an invalid condition.
this fixes an invalid memory access problem.

Issue: https://github.com/thorvg/thorvg/issues/1671
 2023-09-26 17:41:08 +09:00
Hermet Park
e4ade98e8c sw_engine: fix a regression stroke bug 
reverted changes that occurred this issue from d683d2e

Issue: https://github.com/thorvg/thorvg/issues/1670
 2023-09-26 14:26:57 +09:00
Hermet Park
74b67919e0 tvg: support radial gradient focal properties 
properly store/restore the radial gradient focal properties
from the tvg loader and saver
 2023-09-26 13:05:27 +09:00
Hermet Park
ed23b432bb tvg: support dash offset property 
properly store/restore the dash offset property
from the tvg loader and saver

Issue: https://github.com/thorvg/thorvg/issues/1617
 2023-09-26 13:05:27 +09:00
Hermet Park
1819fed033 renderer/paint: fixed a mismatched reference count. 
This correction ensures a consistent use of 'ref' and 'unref' for paints to release memory properly.
The memory leak occurred when a picture was not pushed to a valid canvas.

This issue was reported by the unit-test memory sanitizer.
 2023-09-26 10:48:33 +09:00
RuiwenTang
453cba7ddd gl_engine: make GlRenderTask generic with uniform block 
* Use uniform block to pack all color informations
* Move the actual gl draw call into GlRenderer::sync function, so all
  data is been uploaded into GPU
* Make GlRenderTask simple and generic for direct gl draw
 2023-09-21 22:24:07 +09:00
Hermet Park
9e3b74bac5 sw_engine fill: fixed radial fill focal issue. 
This might be a workaround to correct the issue.
The threshold value comes from the mathematical inaccuracy.

@Issue:https://github.com/thorvg/thorvg/issues/1555
 2023-09-21 22:19:42 +09:00
Hermet Park
edb156f4fc sw_engine/shape: fix a dash-line infite-loop bug. 
a regresion bug was introduced by d683d2e70d
 2023-09-21 21:55:53 +09:00
Hermet Park
bd9d0cea6d sw_engine: properly implemented the missing image composition. 
The 8-bit masking target is the destination with a 32-bit source image and 8-bit composition.
 2023-09-20 18:42:06 +09:00
Martin Capitanio
5ed7034c45 sw_engine: fix VS 2017 issues with Godot's NOMINMAX 
Fixes #1661
 2023-09-19 21:51:54 +09:00
Hermet Park
8aa12ca468 taskschduler: fix a regression deadlock issue 
This fix introduces a workaround to enforce synchronous tasking on worker threads.
Sometimes, out of threads get stuck in a deadlock condition.

@Issue: https://github.com/thorvg/thorvg/issues/1636
 2023-09-18 21:06:01 +09:00
Hermet Park
d683d2e70d sw_engine: Do not handle exceptions for zero line length. 
While it may represent a dot with a stroke width,
ignoring its length is not a good idea.

this also eliminates 0 size dash dots caused by unexpected empty lengths.
 2023-09-15 11:16:38 +09:00
RuiwenTang
5298f338de gl_engine: fix compilation erros due to redundant header inclusion 2023-09-14 19:49:35 +09:00
Hermet Park
ed1b4a4923 sw_engine: allow 8bit grayscale gradient filling 2023-09-14 01:03:22 +09:00
Hermet Park
e9490e375c sw_engine: corrected a wrong blending equation at 8 bits grayscale rle. 2023-09-14 01:03:22 +09:00
Hermet Park
cea63125a4 sw_engine: fix a wrong compuation. 
corrected a mistake caused by a dead brain.
 2023-09-12 21:40:01 +09:00
Hermet Park
46d831ead7 sw_engine: skip the rendering if possible. 
zero scale would not produce any visible artifacts.
we can skip the further processing sonner.
 2023-09-12 20:52:33 +09:00
Hermet Park
2257211852 sw_engine: fix a regression bug by 0dafe9ff64 
rectified the stroke scale factor acquisition.
 2023-09-12 11:44:27 +09:00
Hermet Park
0dafe9ff64 sw_engine: revised code. 
add a concrete logic for validating the stroke width
 2023-09-05 18:41:20 +09:00
Hermet Park
78cb32794a sw_engine: utilize the memory pool for dash stroking. 2023-09-05 12:38:57 +09:00
Hermet Park
64a444aeb6 loader/lottie: support focal property in the radial gradient 2023-09-05 10:29:12 +09:00
Hermet Park
fa134692b0 Revert "renderer taskscheduler: code refactoring." 
This reverts commit 77de62068c.

This fixes the sanitizer thread report (data race) issue.

@Issue: https://github.com/thorvg/thorvg/issues/1623
 2023-09-05 10:22:53 +09:00