AuroraMiddleware/gtk
1
0
Fork 1
mirror of https://gitlab.gnome.org/GNOME/gtk.git synced 2024-12-27 14:10:30 +00:00
Code Issues Projects Releases Wiki Activity
76caf2c812
gtk/gsk/gl/gskglglyphlibrary.c

438 lines
13 KiB
C
Raw Normal View History

Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
/* gskglglyphlibrary.c
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
*
* Copyright 2020 Christian Hergert <chergert@redhat.com>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this program. If not, see <http://www.gnu.org/licenses/>.
*
* SPDX-License-Identifier: LGPL-2.1-or-later
*/
#include "config.h"
#include <gdk/gdkglcontextprivate.h>
memorytexture: Split out GdkMemoryFormat handling Also, now make gdk_memory_convert() the only conversion functions and allow conversions between any 2 formats by going via a float[4]. This could be optimized via fast-paths, but so far it isn't.
2021-09-22 00:01:41 +00:00
#include <gdk/gdkmemoryformatprivate.h>
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
#include <gdk/gdkprofilerprivate.h>
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
#include "gskglcommandqueueprivate.h"
#include "gskgldriverprivate.h"
#include "gskglglyphlibraryprivate.h"
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
glyph cache: Add debug output Show the size of data uploaded per glyph.
2023-07-02 02:31:28 +00:00
#include "gskdebugprivate.h"
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
#define MAX_GLYPH_SIZE 128
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
G_DEFINE_TYPE (GskGLGlyphLibrary, gsk_gl_glyph_library, GSK_TYPE_GL_TEXTURE_LIBRARY)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
GskGLGlyphLibrary *
gsk_gl_glyph_library_new (GskGLDriver *driver)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
g_return_val_if_fail (GSK_IS_GL_DRIVER (driver), NULL);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
return g_object_new (GSK_TYPE_GL_GLYPH_LIBRARY,
"driver", driver,
NULL);
}
static guint
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
gsk_gl_glyph_key_hash (gconstpointer data)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
const GskGLGlyphKey *key = data;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
/* We do not store the hash within the key because GHashTable will already
* store the hash value for us and so this is called only a single time per
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
* cached item. This saves an extra 4 bytes per GskGLGlyphKey which means on
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
* 64-bit, we fit nicely within 2 pointers (the smallest allocation size
* for GSlice).
*/
return GPOINTER_TO_UINT (key->font) ^
key->glyph ^
(key->xshift << 24) ^
(key->yshift << 26) ^
key->scale;
}
static gboolean
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
gsk_gl_glyph_key_equal (gconstpointer v1,
gconstpointer v2)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
return memcmp (v1, v2, sizeof (GskGLGlyphKey)) == 0;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
}
static void
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
gsk_gl_glyph_key_free (gpointer data)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
GskGLGlyphKey *key = data;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
g_clear_object (&key->font);
gsk: Stop using g_slice
2023-03-03 11:48:49 +00:00
g_free (key);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
}
static void
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
gsk_gl_glyph_value_free (gpointer data)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
gsk: Stop using g_slice
2023-03-03 11:48:49 +00:00
g_free (data);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
}
ngl: clear Glyph front cache at the beginning of a frame We don't want to be responsible for duplicating the effort of the hash table, we just want to speed up subsequent lookups. Otherwise, we risk not marking glyph usage when tracking usage for compaction.
2021-03-19 00:54:48 +00:00
static void
gsk/gl: only clear glyph cache durign reclaimation We don't need to clear the front cache on every frame as we can clear it specifically when we do reclaimation to avoid unnecessary memset() calls.
2022-03-15 22:55:30 +00:00
gsk_gl_glyph_library_clear_cache (GskGLTextureLibrary *library)
ngl: clear Glyph front cache at the beginning of a frame We don't want to be responsible for duplicating the effort of the hash table, we just want to speed up subsequent lookups. Otherwise, we risk not marking glyph usage when tracking usage for compaction.
2021-03-19 00:54:48 +00:00
{
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
GskGLGlyphLibrary *self = (GskGLGlyphLibrary *)library;
ngl: clear Glyph front cache at the beginning of a frame We don't want to be responsible for duplicating the effort of the hash table, we just want to speed up subsequent lookups. Otherwise, we risk not marking glyph usage when tracking usage for compaction.
2021-03-19 00:54:48 +00:00
gsk/gl: only clear glyph cache durign reclaimation We don't need to clear the front cache on every frame as we can clear it specifically when we do reclaimation to avoid unnecessary memset() calls.
2022-03-15 22:55:30 +00:00
g_assert (GSK_IS_GL_GLYPH_LIBRARY (self));
ngl: clear Glyph front cache at the beginning of a frame We don't want to be responsible for duplicating the effort of the hash table, we just want to speed up subsequent lookups. Otherwise, we risk not marking glyph usage when tracking usage for compaction.
2021-03-19 00:54:48 +00:00
memset (self->front, 0, sizeof self->front);
}
gsk/gl: make texture libraries more autonomous This moves a lot of the texture atlas control out of the driver and into the various texture libraries through their base GskGLTextureLibrary class. Additionally, this gives more control to libraries on allocating which can be necessary for some tooling such as future Glyphy integration. As part of this, the 1x1 pixel initialization is moved to the Glyph library which is the only place where it is actually needed. The compact vfunc now is responsible for compaction and it allows for us to iterate the atlas hashtable a single time instead of twice as we were doing previously. The init_atlas vfunc is used to do per-library initialization such as adding a 1x1 pixel in the Glyph cache used for coloring lines. The allocate vfunc purely allocates but does no upload. This can be useful for situations where a library wants to reuse the allocator from the base class but does not want to actually insert a key/value entry. The glyph library uses this for it's 1x1 pixel. In the future, we will also likely want to decouple the rectangle packing implementation from the atlas structure, or at least move it into a union so that we do not allocate unused memory for alternate allocators.
2022-03-18 21:51:41 +00:00
static void
gsk_gl_glyph_library_init_atlas (GskGLTextureLibrary *self,
GskGLTextureAtlas *atlas)
{
gboolean packed G_GNUC_UNUSED;
int x, y;
guint gl_format;
guint gl_type;
guint8 pixel_data[4 * 3 * 3];
g_assert (GSK_IS_GL_GLYPH_LIBRARY (self));
g_assert (atlas != NULL);
/* Insert a single pixel at 0,0 for use in coloring */
gdk_gl_context_push_debug_group_printf (gdk_gl_context_get_current (),
"Initializing Atlas");
packed = gsk_gl_texture_library_allocate (self, atlas, 3, 3, &x, &y);
g_assert (packed);
g_assert (x == 0 && y == 0);
memset (pixel_data, 255, sizeof pixel_data);
gl: Fix text carets going missing on NVidia Use the same logic for uploading the 'corner pixel' that we use for uploading glyphs, since that works. Fixes: #6348
2024-01-21 13:22:47 +00:00
if (gdk_gl_context_get_use_es (gdk_gl_context_get_current ()))
gsk/gl: make texture libraries more autonomous This moves a lot of the texture atlas control out of the driver and into the various texture libraries through their base GskGLTextureLibrary class. Additionally, this gives more control to libraries on allocating which can be necessary for some tooling such as future Glyphy integration. As part of this, the 1x1 pixel initialization is moved to the Glyph library which is the only place where it is actually needed. The compact vfunc now is responsible for compaction and it allows for us to iterate the atlas hashtable a single time instead of twice as we were doing previously. The init_atlas vfunc is used to do per-library initialization such as adding a 1x1 pixel in the Glyph cache used for coloring lines. The allocate vfunc purely allocates but does no upload. This can be useful for situations where a library wants to reuse the allocator from the base class but does not want to actually insert a key/value entry. The glyph library uses this for it's 1x1 pixel. In the future, we will also likely want to decouple the rectangle packing implementation from the atlas structure, or at least move it into a union so that we do not allocate unused memory for alternate allocators.
2022-03-18 21:51:41 +00:00
{
gl_format = GL_RGBA;
gl_type = GL_UNSIGNED_BYTE;
}
else
{
gl_format = GL_BGRA;
changes to fix the memorytexture regression
2024-04-05 08:30:10 +00:00
gl_type = GL_UNSIGNED_INT_8_8_8_8_REV;
gsk/gl: make texture libraries more autonomous This moves a lot of the texture atlas control out of the driver and into the various texture libraries through their base GskGLTextureLibrary class. Additionally, this gives more control to libraries on allocating which can be necessary for some tooling such as future Glyphy integration. As part of this, the 1x1 pixel initialization is moved to the Glyph library which is the only place where it is actually needed. The compact vfunc now is responsible for compaction and it allows for us to iterate the atlas hashtable a single time instead of twice as we were doing previously. The init_atlas vfunc is used to do per-library initialization such as adding a 1x1 pixel in the Glyph cache used for coloring lines. The allocate vfunc purely allocates but does no upload. This can be useful for situations where a library wants to reuse the allocator from the base class but does not want to actually insert a key/value entry. The glyph library uses this for it's 1x1 pixel. In the future, we will also likely want to decouple the rectangle packing implementation from the atlas structure, or at least move it into a union so that we do not allocate unused memory for alternate allocators.
2022-03-18 21:51:41 +00:00
}
glBindTexture (GL_TEXTURE_2D, atlas->texture_id);
glTexSubImage2D (GL_TEXTURE_2D, 0,
0, 0,
3, 3,
gl_format, gl_type,
pixel_data);
gdk_gl_context_pop_debug_group (gdk_gl_context_get_current ());
self->driver->command_queue->n_uploads++;
}
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
static void
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
gsk_gl_glyph_library_finalize (GObject *object)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
GskGLGlyphLibrary *self = (GskGLGlyphLibrary *)object;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
g_clear_pointer (&self->surface_data, g_free);
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
G_OBJECT_CLASS (gsk_gl_glyph_library_parent_class)->finalize (object);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
}
static void
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
gsk_gl_glyph_library_class_init (GskGLGlyphLibraryClass *klass)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
GObjectClass *object_class = G_OBJECT_CLASS (klass);
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
GskGLTextureLibraryClass *library_class = GSK_GL_TEXTURE_LIBRARY_CLASS (klass);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
object_class->finalize = gsk_gl_glyph_library_finalize;
ngl: clear Glyph front cache at the beginning of a frame We don't want to be responsible for duplicating the effort of the hash table, we just want to speed up subsequent lookups. Otherwise, we risk not marking glyph usage when tracking usage for compaction.
2021-03-19 00:54:48 +00:00
gsk/gl: only clear glyph cache durign reclaimation We don't need to clear the front cache on every frame as we can clear it specifically when we do reclaimation to avoid unnecessary memset() calls.
2022-03-15 22:55:30 +00:00
library_class->clear_cache = gsk_gl_glyph_library_clear_cache;
gsk/gl: make texture libraries more autonomous This moves a lot of the texture atlas control out of the driver and into the various texture libraries through their base GskGLTextureLibrary class. Additionally, this gives more control to libraries on allocating which can be necessary for some tooling such as future Glyphy integration. As part of this, the 1x1 pixel initialization is moved to the Glyph library which is the only place where it is actually needed. The compact vfunc now is responsible for compaction and it allows for us to iterate the atlas hashtable a single time instead of twice as we were doing previously. The init_atlas vfunc is used to do per-library initialization such as adding a 1x1 pixel in the Glyph cache used for coloring lines. The allocate vfunc purely allocates but does no upload. This can be useful for situations where a library wants to reuse the allocator from the base class but does not want to actually insert a key/value entry. The glyph library uses this for it's 1x1 pixel. In the future, we will also likely want to decouple the rectangle packing implementation from the atlas structure, or at least move it into a union so that we do not allocate unused memory for alternate allocators.
2022-03-18 21:51:41 +00:00
library_class->init_atlas = gsk_gl_glyph_library_init_atlas;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
}
static void
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
gsk_gl_glyph_library_init (GskGLGlyphLibrary *self)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
GskGLTextureLibrary *tl = (GskGLTextureLibrary *)self;
ngl: Reduce type checking overhead No need to be very careful here, these are entirely internal types.
2021-07-24 13:50:43 +00:00
tl->max_entry_size = MAX_GLYPH_SIZE;
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
gsk_gl_texture_library_set_funcs (tl,
gsk_gl_glyph_key_hash,
gsk_gl_glyph_key_equal,
gsk_gl_glyph_key_free,
gsk_gl_glyph_value_free);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
}
static cairo_surface_t *
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
gsk_gl_glyph_library_create_surface (GskGLGlyphLibrary *self,
int stride,
int width,
int height,
int uwidth,
int uheight)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
cairo_surface_t *surface;
gsize n_bytes;
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
g_assert (GSK_IS_GL_GLYPH_LIBRARY (self));
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
g_assert (width > 0);
g_assert (height > 0);
n_bytes = stride * height;
if G_LIKELY (n_bytes > self->surface_data_len)
{
self->surface_data = g_realloc (self->surface_data, n_bytes);
self->surface_data_len = n_bytes;
}
memset (self->surface_data, 0, n_bytes);
surface = cairo_image_surface_create_for_data (self->surface_data,
CAIRO_FORMAT_ARGB32,
width, height, stride);
ngl: Use exact device scales for glyphs We are rendering the glyphs on a larger surface, and we should avoid introducing unnecessary rounding errors here. Also, I've found that we always need to enlarge the surface by one pixels in each direction to avoid cutting off the tops of large glyphs.
2021-09-17 23:07:55 +00:00
cairo_surface_set_device_scale (surface, width / (double)uwidth, height / (double)uheight);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
return surface;
}
static void
render_glyph (cairo_surface_t *surface,
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
const GskGLGlyphKey *key,
const GskGLGlyphValue *value)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
cairo_t *cr;
Go back to using pango for glyph rendering Using just cairo makes us lose hexboxes. So, until we implement that ourselves, go back to using pango.
2021-11-20 16:13:52 +00:00
PangoGlyphString glyph_string;
PangoGlyphInfo glyph_info;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
g_assert (surface != NULL);
cr = cairo_create (surface);
cairo_set_source_rgba (cr, 1, 1, 1, 1);
Go back to using pango for glyph rendering Using just cairo makes us lose hexboxes. So, until we implement that ourselves, go back to using pango.
2021-11-20 16:13:52 +00:00
glyph_info.glyph = key->glyph;
glyph_info.geometry.width = value->ink_rect.width * 1024;
Fix kerning of GL rendered glyphs Multiple scaled shifts by 1024 too. Fixes: #4602
2022-01-31 14:41:18 +00:00
glyph_info.geometry.x_offset = (0.25 * key->xshift - value->ink_rect.x) * 1024;
glyph_info.geometry.y_offset = (0.25 * key->yshift - value->ink_rect.y) * 1024;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
Go back to using pango for glyph rendering Using just cairo makes us lose hexboxes. So, until we implement that ourselves, go back to using pango.
2021-11-20 16:13:52 +00:00
glyph_string.num_glyphs = 1;
glyph_string.glyphs = &glyph_info;
pango_cairo_show_glyph_string (cr, key->font, &glyph_string);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
cairo_destroy (cr);
cairo_surface_flush (surface);
}
static void
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
gsk_gl_glyph_library_upload_glyph (GskGLGlyphLibrary *self,
const GskGLGlyphKey *key,
const GskGLGlyphValue *value,
gsk: Make glyph upload more similar to icons When uploading glyphs to the atlas, clear the padding bytes so that no dirt remains there.
2021-10-08 08:14:19 +00:00
int packed_x,
int packed_y,
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
int width,
int height,
int uwidth,
int uheight)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
GskGLTextureLibrary *tl = (GskGLTextureLibrary *)self;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
G_GNUC_UNUSED gint64 start_time = GDK_PROFILER_CURRENT_TIME;
cairo_surface_t *surface;
guchar *pixel_data;
guchar *free_data = NULL;
guint gl_format;
guint gl_type;
guint texture_id;
gsize stride;
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
g_assert (GSK_IS_GL_GLYPH_LIBRARY (self));
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
g_assert (key != NULL);
g_assert (value != NULL);
stride = cairo_format_stride_for_width (CAIRO_FORMAT_ARGB32, width);
gdk_gl_context_push_debug_group_printf (gdk_gl_context_get_current (),
"Uploading glyph %d",
key->glyph);
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
surface = gsk_gl_glyph_library_create_surface (self, stride, width, height, uwidth, uheight);
Go back to using pango for glyph rendering Using just cairo makes us lose hexboxes. So, until we implement that ourselves, go back to using pango.
2021-11-20 16:13:52 +00:00
render_glyph (surface, key, value);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
texture_id = GSK_GL_TEXTURE_ATLAS_ENTRY_TEXTURE (value);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
g_assert (texture_id > 0);
gsk: Fix icon and glyph upload with GLES We can't avoid a copy here, unfortunately. Fixes: #6216
2023-12-18 02:52:05 +00:00
if (gdk_gl_context_get_use_es (gdk_gl_context_get_current ()))
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
pixel_data = free_data = g_malloc (width * height * 4);
gsk: Make glyph upload more similar to icons When uploading glyphs to the atlas, clear the padding bytes so that no dirt remains there.
2021-10-08 08:14:19 +00:00
gdk_memory_convert (pixel_data, width * 4,
memorytexture: Split out GdkMemoryFormat handling Also, now make gdk_memory_convert() the only conversion functions and allow conversions between any 2 formats by going via a float[4]. This could be optimized via fast-paths, but so far it isn't.
2021-09-22 00:01:41 +00:00
GDK_MEMORY_R8G8B8A8_PREMULTIPLIED,
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
cairo_image_surface_get_data (surface),
gsk: Make glyph upload more similar to icons When uploading glyphs to the atlas, clear the padding bytes so that no dirt remains there.
2021-10-08 08:14:19 +00:00
stride,
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
GDK_MEMORY_DEFAULT,
width, height);
gsk: Make glyph upload more similar to icons When uploading glyphs to the atlas, clear the padding bytes so that no dirt remains there.
2021-10-08 08:14:19 +00:00
stride = width * 4;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
gl_format = GL_RGBA;
gl_type = GL_UNSIGNED_BYTE;
}
else
{
pixel_data = cairo_image_surface_get_data (surface);
gl_format = GL_BGRA;
changes to fix the memorytexture regression
2024-04-05 08:30:10 +00:00
gl_type = GL_UNSIGNED_INT_8_8_8_8_REV;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
}
gsk: Make glyph upload more similar to icons When uploading glyphs to the atlas, clear the padding bytes so that no dirt remains there.
2021-10-08 08:14:19 +00:00
glPixelStorei (GL_UNPACK_ROW_LENGTH, stride / 4);
glBindTexture (GL_TEXTURE_2D, texture_id);
glTexSubImage2D (GL_TEXTURE_2D, 0,
packed_x + 1, packed_y + 1,
width, height,
gl_format, gl_type,
pixel_data);
/* Padding top */
glTexSubImage2D (GL_TEXTURE_2D, 0,
packed_x + 1, packed_y,
width, 1,
gl_format, gl_type,
pixel_data);
/* Padding left */
glTexSubImage2D (GL_TEXTURE_2D, 0,
packed_x, packed_y + 1,
1, height,
gl_format, gl_type,
pixel_data);
/* Padding top left */
glTexSubImage2D (GL_TEXTURE_2D, 0,
packed_x, packed_y,
1, 1,
gl_format, gl_type,
pixel_data);
/* Padding right */
glPixelStorei (GL_UNPACK_ROW_LENGTH, width);
glPixelStorei (GL_UNPACK_SKIP_PIXELS, width - 1);
glTexSubImage2D (GL_TEXTURE_2D, 0,
packed_x + width + 1, packed_y + 1,
1, height,
gl_format, gl_type,
pixel_data);
/* Padding top right */
glTexSubImage2D (GL_TEXTURE_2D, 0,
packed_x + width + 1, packed_y,
1, 1,
gl_format, gl_type,
pixel_data);
/* Padding bottom */
glPixelStorei (GL_UNPACK_SKIP_PIXELS, 0);
glPixelStorei (GL_UNPACK_ROW_LENGTH, 0);
glPixelStorei (GL_UNPACK_SKIP_ROWS, height - 1);
glTexSubImage2D (GL_TEXTURE_2D, 0,
packed_x + 1, packed_y + 1 + height,
width, 1,
gl_format, gl_type,
pixel_data);
/* Padding bottom left */
glTexSubImage2D (GL_TEXTURE_2D, 0,
packed_x, packed_y + 1 + height,
1, 1,
gl_format, gl_type,
pixel_data);
/* Padding bottom right */
glPixelStorei (GL_UNPACK_ROW_LENGTH, width);
glPixelStorei (GL_UNPACK_SKIP_PIXELS, width - 1);
glTexSubImage2D (GL_TEXTURE_2D, 0,
packed_x + 1 + width, packed_y + 1 + height,
1, 1,
gl_format, gl_type,
pixel_data);
/* Reset this */
glPixelStorei (GL_UNPACK_SKIP_PIXELS, 0);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
glPixelStorei (GL_UNPACK_ROW_LENGTH, 0);
gsk: Make glyph upload more similar to icons When uploading glyphs to the atlas, clear the padding bytes so that no dirt remains there.
2021-10-08 08:14:19 +00:00
glPixelStorei (GL_UNPACK_SKIP_ROWS, 0);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
cairo_surface_destroy (surface);
g_free (free_data);
gdk_gl_context_pop_debug_group (gdk_gl_context_get_current ());
ngl: Reduce type checking overhead No need to be very careful here, these are entirely internal types.
2021-07-24 13:50:43 +00:00
tl->driver->command_queue->n_uploads++;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
if (gdk_profiler_is_running ())
{
char message[64];
g_snprintf (message, sizeof message, "Size %dx%d", width, height);
Tweak profiling strings Capitalize our mark names, and use GTK as category. The justification is: it looks better in sysprof.
2024-01-21 18:58:09 +00:00
gdk_profiler_add_mark (start_time, GDK_PROFILER_CURRENT_TIME-start_time, "Upload glyph", message);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
}
}
gboolean
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
gsk_gl_glyph_library_add (GskGLGlyphLibrary *self,
GskGLGlyphKey *key,
const GskGLGlyphValue **out_value)
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
{
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
GskGLTextureLibrary *tl = (GskGLTextureLibrary *)self;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
PangoRectangle ink_rect;
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
GskGLGlyphValue *value;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
int width;
int height;
guint packed_x;
guint packed_y;
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
g_assert (GSK_IS_GL_GLYPH_LIBRARY (self));
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
g_assert (key != NULL);
g_assert (out_value != NULL);
pango_font_get_glyph_extents (key->font, key->glyph, &ink_rect, NULL);
pango_extents_to_pixels (&ink_rect, NULL);
ngl: Use exact device scales for glyphs We are rendering the glyphs on a larger surface, and we should avoid introducing unnecessary rounding errors here. Also, I've found that we always need to enlarge the surface by one pixels in each direction to avoid cutting off the tops of large glyphs.
2021-09-17 23:07:55 +00:00
ink_rect.x -= 1;
ink_rect.width += 2;
ink_rect.y -= 1;
ink_rect.height += 2;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
ngl: Reuse texture coordinates When uploading glyphs, reuse the position that we get from gsk_ngl_texture_library_pack, instead of recomputing it.
2021-05-15 13:00:50 +00:00
width = (int) ceil (ink_rect.width * key->scale / 1024.0);
height = (int) ceil (ink_rect.height * key->scale / 1024.0);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
glyph cache: Add debug output Show the size of data uploaded per glyph.
2023-07-02 02:31:28 +00:00
GSK_DEBUG (GLYPH_CACHE, "font %p glyph %u: %u x %u pixels", key->font, key->glyph, width, height);
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
value = gsk_gl_texture_library_pack (tl,
key,
sizeof *value,
width,
height,
1,
&packed_x, &packed_y);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
memcpy (&value->ink_rect, &ink_rect, sizeof ink_rect);
if (key->scale > 0 && width > 0 && height > 0)
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
gsk_gl_glyph_library_upload_glyph (self,
key,
value,
gsk: Make glyph upload more similar to icons When uploading glyphs to the atlas, clear the padding bytes so that no dirt remains there.
2021-10-08 08:14:19 +00:00
packed_x,
packed_y,
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
width,
height,
ink_rect.width,
ink_rect.height);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
*out_value = value;
Rename ngl to gl We have only one gl renderer now, and it is a bit odd for it not be called gl.
2021-10-07 03:15:25 +00:00
return GSK_GL_TEXTURE_ATLAS_ENTRY_TEXTURE (value) != 0;
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
}
Reference in New Issue Copy Permalink