mutter/cogl/winsys/cogl-winsys-glx.c

2701 lines
89 KiB
C
Raw Normal View History

Intial Re-layout of the Cogl source code and introduction of a Cogl Winsys As part of an incremental process to have Cogl be a standalone project we want to re-consider how we organise the Cogl source code. Currently this is the structure I'm aiming for: cogl/ cogl/ <put common source here> winsys/ cogl-glx.c cogl-wgl.c driver/ gl/ gles/ os/ ? utils/ cogl-fixed cogl-matrix-stack? cogl-journal? cogl-primitives? pango/ The new winsys component is a starting point for migrating window system code (i.e. x11,glx,wgl,osx,egl etc) from Clutter to Cogl. The utils/ and pango/ directories aren't added by this commit, but they are noted because I plan to add them soon. Overview of the planned structure: * The winsys/ API is the API that binds OpenGL to a specific window system, be that X11 or win32 etc. Example are glx, wgl and egl. Much of the logic under clutter/{glx,osx,win32 etc} should migrate here. * Note there is also the idea of a winsys-base that may represent a window system for which there are multiple winsys APIs. An example of this is x11, since glx and egl may both be used with x11. (currently only Clutter has the idea of a winsys-base) * The driver/ represents a specific varient of OpenGL. Currently we have "gl" representing OpenGL 1.4-2.1 (mostly fixed function) and "gles" representing GLES 1.1 (fixed funciton) and 2.0 (fully shader based) * Everything under cogl/ should fundamentally be supporting access to the GPU. Essentially Cogl's most basic requirement is to provide a nice GPU Graphics API and drawing a line between this and the utility functionality we add to support Clutter should help keep this lean and maintainable. * Code under utils/ as suggested builds on cogl/ adding more convenient APIs or mechanism to optimize special cases. Broadly speaking you can compare cogl/ to OpenGL and utils/ to GLU. * clutter/pango will be moved to clutter/cogl/pango How some of the internal configure.ac/pkg-config terminology has changed: backendextra -> CLUTTER_WINSYS_BASE # e.g. "x11" backendextralib -> CLUTTER_WINSYS_BASE_LIB # e.g. "x11/libclutter-x11.la" clutterbackend -> {CLUTTER,COGL}_WINSYS # e.g. "glx" CLUTTER_FLAVOUR -> {CLUTTER,COGL}_WINSYS clutterbackendlib -> CLUTTER_WINSYS_LIB CLUTTER_COGL -> COGL_DRIVER # e.g. "gl" Note: The CLUTTER_FLAVOUR and CLUTTER_COGL defines are kept for apps As the first thing to take advantage of the new winsys component in Cogl; cogl_get_proc_address() has been moved from cogl/{gl,gles}/cogl.c into cogl/common/cogl.c and this common implementation first trys _cogl_winsys_get_proc_address() but if that fails then it falls back to gmodule.
2009-07-27 21:02:02 -04:00
/*
* Cogl
*
* An object oriented GL/GLES Abstraction/Utility Layer
*
* Copyright (C) 2007,2008,2009,2010,2011,2013 Intel Corporation.
Intial Re-layout of the Cogl source code and introduction of a Cogl Winsys As part of an incremental process to have Cogl be a standalone project we want to re-consider how we organise the Cogl source code. Currently this is the structure I'm aiming for: cogl/ cogl/ <put common source here> winsys/ cogl-glx.c cogl-wgl.c driver/ gl/ gles/ os/ ? utils/ cogl-fixed cogl-matrix-stack? cogl-journal? cogl-primitives? pango/ The new winsys component is a starting point for migrating window system code (i.e. x11,glx,wgl,osx,egl etc) from Clutter to Cogl. The utils/ and pango/ directories aren't added by this commit, but they are noted because I plan to add them soon. Overview of the planned structure: * The winsys/ API is the API that binds OpenGL to a specific window system, be that X11 or win32 etc. Example are glx, wgl and egl. Much of the logic under clutter/{glx,osx,win32 etc} should migrate here. * Note there is also the idea of a winsys-base that may represent a window system for which there are multiple winsys APIs. An example of this is x11, since glx and egl may both be used with x11. (currently only Clutter has the idea of a winsys-base) * The driver/ represents a specific varient of OpenGL. Currently we have "gl" representing OpenGL 1.4-2.1 (mostly fixed function) and "gles" representing GLES 1.1 (fixed funciton) and 2.0 (fully shader based) * Everything under cogl/ should fundamentally be supporting access to the GPU. Essentially Cogl's most basic requirement is to provide a nice GPU Graphics API and drawing a line between this and the utility functionality we add to support Clutter should help keep this lean and maintainable. * Code under utils/ as suggested builds on cogl/ adding more convenient APIs or mechanism to optimize special cases. Broadly speaking you can compare cogl/ to OpenGL and utils/ to GLU. * clutter/pango will be moved to clutter/cogl/pango How some of the internal configure.ac/pkg-config terminology has changed: backendextra -> CLUTTER_WINSYS_BASE # e.g. "x11" backendextralib -> CLUTTER_WINSYS_BASE_LIB # e.g. "x11/libclutter-x11.la" clutterbackend -> {CLUTTER,COGL}_WINSYS # e.g. "glx" CLUTTER_FLAVOUR -> {CLUTTER,COGL}_WINSYS clutterbackendlib -> CLUTTER_WINSYS_LIB CLUTTER_COGL -> COGL_DRIVER # e.g. "gl" Note: The CLUTTER_FLAVOUR and CLUTTER_COGL defines are kept for apps As the first thing to take advantage of the new winsys component in Cogl; cogl_get_proc_address() has been moved from cogl/{gl,gles}/cogl.c into cogl/common/cogl.c and this common implementation first trys _cogl_winsys_get_proc_address() but if that fails then it falls back to gmodule.
2009-07-27 21:02:02 -04:00
*
* 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 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library. If not, see
* <http://www.gnu.org/licenses/>.
*
*
* Authors:
* Robert Bragg <robert@linux.intel.com>
Intial Re-layout of the Cogl source code and introduction of a Cogl Winsys As part of an incremental process to have Cogl be a standalone project we want to re-consider how we organise the Cogl source code. Currently this is the structure I'm aiming for: cogl/ cogl/ <put common source here> winsys/ cogl-glx.c cogl-wgl.c driver/ gl/ gles/ os/ ? utils/ cogl-fixed cogl-matrix-stack? cogl-journal? cogl-primitives? pango/ The new winsys component is a starting point for migrating window system code (i.e. x11,glx,wgl,osx,egl etc) from Clutter to Cogl. The utils/ and pango/ directories aren't added by this commit, but they are noted because I plan to add them soon. Overview of the planned structure: * The winsys/ API is the API that binds OpenGL to a specific window system, be that X11 or win32 etc. Example are glx, wgl and egl. Much of the logic under clutter/{glx,osx,win32 etc} should migrate here. * Note there is also the idea of a winsys-base that may represent a window system for which there are multiple winsys APIs. An example of this is x11, since glx and egl may both be used with x11. (currently only Clutter has the idea of a winsys-base) * The driver/ represents a specific varient of OpenGL. Currently we have "gl" representing OpenGL 1.4-2.1 (mostly fixed function) and "gles" representing GLES 1.1 (fixed funciton) and 2.0 (fully shader based) * Everything under cogl/ should fundamentally be supporting access to the GPU. Essentially Cogl's most basic requirement is to provide a nice GPU Graphics API and drawing a line between this and the utility functionality we add to support Clutter should help keep this lean and maintainable. * Code under utils/ as suggested builds on cogl/ adding more convenient APIs or mechanism to optimize special cases. Broadly speaking you can compare cogl/ to OpenGL and utils/ to GLU. * clutter/pango will be moved to clutter/cogl/pango How some of the internal configure.ac/pkg-config terminology has changed: backendextra -> CLUTTER_WINSYS_BASE # e.g. "x11" backendextralib -> CLUTTER_WINSYS_BASE_LIB # e.g. "x11/libclutter-x11.la" clutterbackend -> {CLUTTER,COGL}_WINSYS # e.g. "glx" CLUTTER_FLAVOUR -> {CLUTTER,COGL}_WINSYS clutterbackendlib -> CLUTTER_WINSYS_LIB CLUTTER_COGL -> COGL_DRIVER # e.g. "gl" Note: The CLUTTER_FLAVOUR and CLUTTER_COGL defines are kept for apps As the first thing to take advantage of the new winsys component in Cogl; cogl_get_proc_address() has been moved from cogl/{gl,gles}/cogl.c into cogl/common/cogl.c and this common implementation first trys _cogl_winsys_get_proc_address() but if that fails then it falls back to gmodule.
2009-07-27 21:02:02 -04:00
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "cogl-i18n-private.h"
#include "cogl-util.h"
#include "cogl-winsys-private.h"
#include "cogl-feature-private.h"
#include "cogl-context-private.h"
#include "cogl-framebuffer.h"
#include "cogl-swap-chain-private.h"
#include "cogl-renderer-private.h"
#include "cogl-glx-renderer-private.h"
#include "cogl-onscreen-template-private.h"
#include "cogl-glx-display-private.h"
#include "cogl-private.h"
#include "cogl-texture-2d-private.h"
#include "cogl-texture-rectangle-private.h"
#include "cogl-pipeline-opengl-private.h"
#include "cogl-frame-info-private.h"
#include "cogl-framebuffer-private.h"
#include "cogl-onscreen-private.h"
#include "cogl-swap-chain-private.h"
#include "cogl-xlib-renderer.h"
#include "cogl-util.h"
Add -Wmissing-declarations to maintainer flags and fix problems This option to GCC makes it give a warning whenever a global function is defined without a declaration. This should catch cases were we've defined a function but forgot to put it in a header. In that case it is either only used within one file so we should make it static or we should declare it in a header. The following changes where made to fix problems: • Some functions were made static • cogl-path.h (the one containing the 1.0 API) was split into two files, one defining the functions and one defining the enums so that cogl-path.c can include the enum and function declarations from the 2.0 API as well as the function declarations from the 1.0 API. • cogl2-clip-state has been removed. This only had one experimental function called cogl_clip_push_from_path but as this is unstable we might as well remove it favour of the equivalent cogl_framebuffer_* API. • The GLX, SDL and WGL winsys's now have a private header to define their get_vtable function instead of directly declaring in the C file where it is called. • All places that were calling COGL_OBJECT_DEFINE need to have the cogl_is_whatever function declared so these have been added either as a public function or in a private header. • Some files that were not including the header containing their function declarations have been fixed to do so. • Any unused error quark functions have been removed. If we later want them we should add them back one by one and add a declaration for them in a header. • _cogl_is_framebuffer has been renamed to cogl_is_framebuffer and made a public function with a declaration in cogl-framebuffer.h • Similarly for CoglOnscreen. • cogl_vdraw_indexed_attributes is called cogl_framebuffer_vdraw_indexed_attributes in the header. The definition has been changed to match the header. • cogl_index_buffer_allocate has been removed. This had no declaration and I'm not sure what it's supposed to do. • CoglJournal has been changed to use the internal CoglObject macro so that it won't define an exported cogl_is_journal symbol. • The _cogl_blah_pointer_from_handle functions have been removed. CoglHandle isn't used much anymore anyway and in the few places where it is used I think it's safe to just use the implicit cast from void* to the right type. • The test-utils.h header for the conformance tests explicitly disables the -Wmissing-declaration option using a pragma because all of the tests declare their main function without a header. Any mistakes relating to missing declarations aren't really important for the tests. • cogl_quaternion_init_from_quaternion and init_from_matrix have been given declarations in cogl-quaternion.h Reviewed-by: Robert Bragg <robert@linux.intel.com>
2012-03-06 13:21:28 -05:00
#include "cogl-winsys-glx-private.h"
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
#include "cogl-error-private.h"
#include "cogl-poll-private.h"
#include "cogl-version.h"
#include <stdlib.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/time.h>
#include <fcntl.h>
#include <time.h>
Intial Re-layout of the Cogl source code and introduction of a Cogl Winsys As part of an incremental process to have Cogl be a standalone project we want to re-consider how we organise the Cogl source code. Currently this is the structure I'm aiming for: cogl/ cogl/ <put common source here> winsys/ cogl-glx.c cogl-wgl.c driver/ gl/ gles/ os/ ? utils/ cogl-fixed cogl-matrix-stack? cogl-journal? cogl-primitives? pango/ The new winsys component is a starting point for migrating window system code (i.e. x11,glx,wgl,osx,egl etc) from Clutter to Cogl. The utils/ and pango/ directories aren't added by this commit, but they are noted because I plan to add them soon. Overview of the planned structure: * The winsys/ API is the API that binds OpenGL to a specific window system, be that X11 or win32 etc. Example are glx, wgl and egl. Much of the logic under clutter/{glx,osx,win32 etc} should migrate here. * Note there is also the idea of a winsys-base that may represent a window system for which there are multiple winsys APIs. An example of this is x11, since glx and egl may both be used with x11. (currently only Clutter has the idea of a winsys-base) * The driver/ represents a specific varient of OpenGL. Currently we have "gl" representing OpenGL 1.4-2.1 (mostly fixed function) and "gles" representing GLES 1.1 (fixed funciton) and 2.0 (fully shader based) * Everything under cogl/ should fundamentally be supporting access to the GPU. Essentially Cogl's most basic requirement is to provide a nice GPU Graphics API and drawing a line between this and the utility functionality we add to support Clutter should help keep this lean and maintainable. * Code under utils/ as suggested builds on cogl/ adding more convenient APIs or mechanism to optimize special cases. Broadly speaking you can compare cogl/ to OpenGL and utils/ to GLU. * clutter/pango will be moved to clutter/cogl/pango How some of the internal configure.ac/pkg-config terminology has changed: backendextra -> CLUTTER_WINSYS_BASE # e.g. "x11" backendextralib -> CLUTTER_WINSYS_BASE_LIB # e.g. "x11/libclutter-x11.la" clutterbackend -> {CLUTTER,COGL}_WINSYS # e.g. "glx" CLUTTER_FLAVOUR -> {CLUTTER,COGL}_WINSYS clutterbackendlib -> CLUTTER_WINSYS_LIB CLUTTER_COGL -> COGL_DRIVER # e.g. "gl" Note: The CLUTTER_FLAVOUR and CLUTTER_COGL defines are kept for apps As the first thing to take advantage of the new winsys component in Cogl; cogl_get_proc_address() has been moved from cogl/{gl,gles}/cogl.c into cogl/common/cogl.c and this common implementation first trys _cogl_winsys_get_proc_address() but if that fails then it falls back to gmodule.
2009-07-27 21:02:02 -04:00
#include <GL/glx.h>
#include <X11/Xlib.h>
Intial Re-layout of the Cogl source code and introduction of a Cogl Winsys As part of an incremental process to have Cogl be a standalone project we want to re-consider how we organise the Cogl source code. Currently this is the structure I'm aiming for: cogl/ cogl/ <put common source here> winsys/ cogl-glx.c cogl-wgl.c driver/ gl/ gles/ os/ ? utils/ cogl-fixed cogl-matrix-stack? cogl-journal? cogl-primitives? pango/ The new winsys component is a starting point for migrating window system code (i.e. x11,glx,wgl,osx,egl etc) from Clutter to Cogl. The utils/ and pango/ directories aren't added by this commit, but they are noted because I plan to add them soon. Overview of the planned structure: * The winsys/ API is the API that binds OpenGL to a specific window system, be that X11 or win32 etc. Example are glx, wgl and egl. Much of the logic under clutter/{glx,osx,win32 etc} should migrate here. * Note there is also the idea of a winsys-base that may represent a window system for which there are multiple winsys APIs. An example of this is x11, since glx and egl may both be used with x11. (currently only Clutter has the idea of a winsys-base) * The driver/ represents a specific varient of OpenGL. Currently we have "gl" representing OpenGL 1.4-2.1 (mostly fixed function) and "gles" representing GLES 1.1 (fixed funciton) and 2.0 (fully shader based) * Everything under cogl/ should fundamentally be supporting access to the GPU. Essentially Cogl's most basic requirement is to provide a nice GPU Graphics API and drawing a line between this and the utility functionality we add to support Clutter should help keep this lean and maintainable. * Code under utils/ as suggested builds on cogl/ adding more convenient APIs or mechanism to optimize special cases. Broadly speaking you can compare cogl/ to OpenGL and utils/ to GLU. * clutter/pango will be moved to clutter/cogl/pango How some of the internal configure.ac/pkg-config terminology has changed: backendextra -> CLUTTER_WINSYS_BASE # e.g. "x11" backendextralib -> CLUTTER_WINSYS_BASE_LIB # e.g. "x11/libclutter-x11.la" clutterbackend -> {CLUTTER,COGL}_WINSYS # e.g. "glx" CLUTTER_FLAVOUR -> {CLUTTER,COGL}_WINSYS clutterbackendlib -> CLUTTER_WINSYS_LIB CLUTTER_COGL -> COGL_DRIVER # e.g. "gl" Note: The CLUTTER_FLAVOUR and CLUTTER_COGL defines are kept for apps As the first thing to take advantage of the new winsys component in Cogl; cogl_get_proc_address() has been moved from cogl/{gl,gles}/cogl.c into cogl/common/cogl.c and this common implementation first trys _cogl_winsys_get_proc_address() but if that fails then it falls back to gmodule.
2009-07-27 21:02:02 -04:00
#define COGL_ONSCREEN_X11_EVENT_MASK (StructureNotifyMask | ExposureMask)
#define MAX_GLX_CONFIG_ATTRIBS 30
typedef struct _CoglContextGLX
{
GLXDrawable current_drawable;
} CoglContextGLX;
typedef struct _CoglOnscreenXlib
{
Window xwin;
int x, y;
CoglBool is_foreign_xwin;
CoglOutput *output;
} CoglOnscreenXlib;
typedef struct _CoglOnscreenGLX
{
CoglOnscreenXlib _parent;
GLXDrawable glxwin;
uint32_t last_swap_vsync_counter;
CoglBool pending_sync_notify;
CoglBool pending_complete_notify;
CoglBool pending_resize_notify;
} CoglOnscreenGLX;
typedef struct _CoglTexturePixmapGLX
{
GLXPixmap glx_pixmap;
CoglBool has_mipmap_space;
CoglBool can_mipmap;
CoglTexture *glx_tex;
CoglBool bind_tex_image_queued;
CoglBool pixmap_bound;
} CoglTexturePixmapGLX;
/* Define a set of arrays containing the functions required from GL
for each winsys feature */
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
#define COGL_WINSYS_FEATURE_BEGIN(major_version, minor_version, \
name, namespaces, extension_names, \
feature_flags, \
winsys_feature) \
static const CoglFeatureFunction \
cogl_glx_feature_ ## name ## _funcs[] = {
#define COGL_WINSYS_FEATURE_FUNCTION(ret, name, args) \
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
{ G_STRINGIFY (name), G_STRUCT_OFFSET (CoglGLXRenderer, name) },
#define COGL_WINSYS_FEATURE_END() \
{ NULL, 0 }, \
};
#include "cogl-winsys-glx-feature-functions.h"
/* Define an array of features */
#undef COGL_WINSYS_FEATURE_BEGIN
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
#define COGL_WINSYS_FEATURE_BEGIN(major_version, minor_version, \
name, namespaces, extension_names, \
feature_flags, \
winsys_feature) \
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
{ major_version, minor_version, \
0, namespaces, extension_names, \
feature_flags, \
0, \
winsys_feature, \
cogl_glx_feature_ ## name ## _funcs },
#undef COGL_WINSYS_FEATURE_FUNCTION
#define COGL_WINSYS_FEATURE_FUNCTION(ret, name, args)
#undef COGL_WINSYS_FEATURE_END
#define COGL_WINSYS_FEATURE_END()
static const CoglFeatureData winsys_feature_data[] =
{
#include "cogl-winsys-glx-feature-functions.h"
};
Intial Re-layout of the Cogl source code and introduction of a Cogl Winsys As part of an incremental process to have Cogl be a standalone project we want to re-consider how we organise the Cogl source code. Currently this is the structure I'm aiming for: cogl/ cogl/ <put common source here> winsys/ cogl-glx.c cogl-wgl.c driver/ gl/ gles/ os/ ? utils/ cogl-fixed cogl-matrix-stack? cogl-journal? cogl-primitives? pango/ The new winsys component is a starting point for migrating window system code (i.e. x11,glx,wgl,osx,egl etc) from Clutter to Cogl. The utils/ and pango/ directories aren't added by this commit, but they are noted because I plan to add them soon. Overview of the planned structure: * The winsys/ API is the API that binds OpenGL to a specific window system, be that X11 or win32 etc. Example are glx, wgl and egl. Much of the logic under clutter/{glx,osx,win32 etc} should migrate here. * Note there is also the idea of a winsys-base that may represent a window system for which there are multiple winsys APIs. An example of this is x11, since glx and egl may both be used with x11. (currently only Clutter has the idea of a winsys-base) * The driver/ represents a specific varient of OpenGL. Currently we have "gl" representing OpenGL 1.4-2.1 (mostly fixed function) and "gles" representing GLES 1.1 (fixed funciton) and 2.0 (fully shader based) * Everything under cogl/ should fundamentally be supporting access to the GPU. Essentially Cogl's most basic requirement is to provide a nice GPU Graphics API and drawing a line between this and the utility functionality we add to support Clutter should help keep this lean and maintainable. * Code under utils/ as suggested builds on cogl/ adding more convenient APIs or mechanism to optimize special cases. Broadly speaking you can compare cogl/ to OpenGL and utils/ to GLU. * clutter/pango will be moved to clutter/cogl/pango How some of the internal configure.ac/pkg-config terminology has changed: backendextra -> CLUTTER_WINSYS_BASE # e.g. "x11" backendextralib -> CLUTTER_WINSYS_BASE_LIB # e.g. "x11/libclutter-x11.la" clutterbackend -> {CLUTTER,COGL}_WINSYS # e.g. "glx" CLUTTER_FLAVOUR -> {CLUTTER,COGL}_WINSYS clutterbackendlib -> CLUTTER_WINSYS_LIB CLUTTER_COGL -> COGL_DRIVER # e.g. "gl" Note: The CLUTTER_FLAVOUR and CLUTTER_COGL defines are kept for apps As the first thing to take advantage of the new winsys component in Cogl; cogl_get_proc_address() has been moved from cogl/{gl,gles}/cogl.c into cogl/common/cogl.c and this common implementation first trys _cogl_winsys_get_proc_address() but if that fails then it falls back to gmodule.
2009-07-27 21:02:02 -04:00
static CoglFuncPtr
_cogl_winsys_renderer_get_proc_address (CoglRenderer *renderer,
Don't use eglGetProcAddress to retrieve core functions According to the EGL spec, eglGetProcAddress should only be used to retrieve extension functions. It also says that returning non-NULL does not mean the extension is available so you could interpret this as saying that the function is allowed to return garbage for core functions. This seems to happen at least for the Android implementation of EGL. To workaround this the winsys's are now passed down a flag to say whether the function is from the core API. This information is already in the gl-prototypes headers as the minimum core GL version and as a pair of flags to specify whether it is available in core GLES1 and GLES2. If the function is in core the EGL winsys will now avoid using eglGetProcAddress and always fallback to querying the library directly with the GModule API. The GLX winsys is left alone because glXGetProcAddress apparently supports querying core API and extension functions. The WGL winsys could ideally be changed because wglGetProcAddress should also only be used for extension functions but the situation is slightly different because WGL considers anything from GL > 1.1 to be an extension so it would need a bit more information to determine whether to query the function directly from the library. The SDL winsys is also left alone because it's not as easy to portably determine which GL library SDL has chosen to load in order to resolve the symbols directly. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 72089730ad06ccdd38a344279a893965ae68cec1) Since we aren't able to break API on the 1.12 branch cogl_get_proc_address is still supported but isn't easily able to determine whether the given name corresponds to a core symbol or not. For now we just assume the symbol being queried isn't part of the core GL api and update the documentation accordingly.
2012-06-20 07:42:31 -04:00
const char *name,
CoglBool in_core)
Intial Re-layout of the Cogl source code and introduction of a Cogl Winsys As part of an incremental process to have Cogl be a standalone project we want to re-consider how we organise the Cogl source code. Currently this is the structure I'm aiming for: cogl/ cogl/ <put common source here> winsys/ cogl-glx.c cogl-wgl.c driver/ gl/ gles/ os/ ? utils/ cogl-fixed cogl-matrix-stack? cogl-journal? cogl-primitives? pango/ The new winsys component is a starting point for migrating window system code (i.e. x11,glx,wgl,osx,egl etc) from Clutter to Cogl. The utils/ and pango/ directories aren't added by this commit, but they are noted because I plan to add them soon. Overview of the planned structure: * The winsys/ API is the API that binds OpenGL to a specific window system, be that X11 or win32 etc. Example are glx, wgl and egl. Much of the logic under clutter/{glx,osx,win32 etc} should migrate here. * Note there is also the idea of a winsys-base that may represent a window system for which there are multiple winsys APIs. An example of this is x11, since glx and egl may both be used with x11. (currently only Clutter has the idea of a winsys-base) * The driver/ represents a specific varient of OpenGL. Currently we have "gl" representing OpenGL 1.4-2.1 (mostly fixed function) and "gles" representing GLES 1.1 (fixed funciton) and 2.0 (fully shader based) * Everything under cogl/ should fundamentally be supporting access to the GPU. Essentially Cogl's most basic requirement is to provide a nice GPU Graphics API and drawing a line between this and the utility functionality we add to support Clutter should help keep this lean and maintainable. * Code under utils/ as suggested builds on cogl/ adding more convenient APIs or mechanism to optimize special cases. Broadly speaking you can compare cogl/ to OpenGL and utils/ to GLU. * clutter/pango will be moved to clutter/cogl/pango How some of the internal configure.ac/pkg-config terminology has changed: backendextra -> CLUTTER_WINSYS_BASE # e.g. "x11" backendextralib -> CLUTTER_WINSYS_BASE_LIB # e.g. "x11/libclutter-x11.la" clutterbackend -> {CLUTTER,COGL}_WINSYS # e.g. "glx" CLUTTER_FLAVOUR -> {CLUTTER,COGL}_WINSYS clutterbackendlib -> CLUTTER_WINSYS_LIB CLUTTER_COGL -> COGL_DRIVER # e.g. "gl" Note: The CLUTTER_FLAVOUR and CLUTTER_COGL defines are kept for apps As the first thing to take advantage of the new winsys component in Cogl; cogl_get_proc_address() has been moved from cogl/{gl,gles}/cogl.c into cogl/common/cogl.c and this common implementation first trys _cogl_winsys_get_proc_address() but if that fails then it falls back to gmodule.
2009-07-27 21:02:02 -04:00
{
CoglGLXRenderer *glx_renderer = renderer->winsys;
Intial Re-layout of the Cogl source code and introduction of a Cogl Winsys As part of an incremental process to have Cogl be a standalone project we want to re-consider how we organise the Cogl source code. Currently this is the structure I'm aiming for: cogl/ cogl/ <put common source here> winsys/ cogl-glx.c cogl-wgl.c driver/ gl/ gles/ os/ ? utils/ cogl-fixed cogl-matrix-stack? cogl-journal? cogl-primitives? pango/ The new winsys component is a starting point for migrating window system code (i.e. x11,glx,wgl,osx,egl etc) from Clutter to Cogl. The utils/ and pango/ directories aren't added by this commit, but they are noted because I plan to add them soon. Overview of the planned structure: * The winsys/ API is the API that binds OpenGL to a specific window system, be that X11 or win32 etc. Example are glx, wgl and egl. Much of the logic under clutter/{glx,osx,win32 etc} should migrate here. * Note there is also the idea of a winsys-base that may represent a window system for which there are multiple winsys APIs. An example of this is x11, since glx and egl may both be used with x11. (currently only Clutter has the idea of a winsys-base) * The driver/ represents a specific varient of OpenGL. Currently we have "gl" representing OpenGL 1.4-2.1 (mostly fixed function) and "gles" representing GLES 1.1 (fixed funciton) and 2.0 (fully shader based) * Everything under cogl/ should fundamentally be supporting access to the GPU. Essentially Cogl's most basic requirement is to provide a nice GPU Graphics API and drawing a line between this and the utility functionality we add to support Clutter should help keep this lean and maintainable. * Code under utils/ as suggested builds on cogl/ adding more convenient APIs or mechanism to optimize special cases. Broadly speaking you can compare cogl/ to OpenGL and utils/ to GLU. * clutter/pango will be moved to clutter/cogl/pango How some of the internal configure.ac/pkg-config terminology has changed: backendextra -> CLUTTER_WINSYS_BASE # e.g. "x11" backendextralib -> CLUTTER_WINSYS_BASE_LIB # e.g. "x11/libclutter-x11.la" clutterbackend -> {CLUTTER,COGL}_WINSYS # e.g. "glx" CLUTTER_FLAVOUR -> {CLUTTER,COGL}_WINSYS clutterbackendlib -> CLUTTER_WINSYS_LIB CLUTTER_COGL -> COGL_DRIVER # e.g. "gl" Note: The CLUTTER_FLAVOUR and CLUTTER_COGL defines are kept for apps As the first thing to take advantage of the new winsys component in Cogl; cogl_get_proc_address() has been moved from cogl/{gl,gles}/cogl.c into cogl/common/cogl.c and this common implementation first trys _cogl_winsys_get_proc_address() but if that fails then it falls back to gmodule.
2009-07-27 21:02:02 -04:00
Don't use eglGetProcAddress to retrieve core functions According to the EGL spec, eglGetProcAddress should only be used to retrieve extension functions. It also says that returning non-NULL does not mean the extension is available so you could interpret this as saying that the function is allowed to return garbage for core functions. This seems to happen at least for the Android implementation of EGL. To workaround this the winsys's are now passed down a flag to say whether the function is from the core API. This information is already in the gl-prototypes headers as the minimum core GL version and as a pair of flags to specify whether it is available in core GLES1 and GLES2. If the function is in core the EGL winsys will now avoid using eglGetProcAddress and always fallback to querying the library directly with the GModule API. The GLX winsys is left alone because glXGetProcAddress apparently supports querying core API and extension functions. The WGL winsys could ideally be changed because wglGetProcAddress should also only be used for extension functions but the situation is slightly different because WGL considers anything from GL > 1.1 to be an extension so it would need a bit more information to determine whether to query the function directly from the library. The SDL winsys is also left alone because it's not as easy to portably determine which GL library SDL has chosen to load in order to resolve the symbols directly. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 72089730ad06ccdd38a344279a893965ae68cec1) Since we aren't able to break API on the 1.12 branch cogl_get_proc_address is still supported but isn't easily able to determine whether the given name corresponds to a core symbol or not. For now we just assume the symbol being queried isn't part of the core GL api and update the documentation accordingly.
2012-06-20 07:42:31 -04:00
/* The GLX_ARB_get_proc_address extension documents that this should
* work for core functions too so we don't need to do anything
* special with in_core */
return glx_renderer->glXGetProcAddress ((const GLubyte *) name);
Intial Re-layout of the Cogl source code and introduction of a Cogl Winsys As part of an incremental process to have Cogl be a standalone project we want to re-consider how we organise the Cogl source code. Currently this is the structure I'm aiming for: cogl/ cogl/ <put common source here> winsys/ cogl-glx.c cogl-wgl.c driver/ gl/ gles/ os/ ? utils/ cogl-fixed cogl-matrix-stack? cogl-journal? cogl-primitives? pango/ The new winsys component is a starting point for migrating window system code (i.e. x11,glx,wgl,osx,egl etc) from Clutter to Cogl. The utils/ and pango/ directories aren't added by this commit, but they are noted because I plan to add them soon. Overview of the planned structure: * The winsys/ API is the API that binds OpenGL to a specific window system, be that X11 or win32 etc. Example are glx, wgl and egl. Much of the logic under clutter/{glx,osx,win32 etc} should migrate here. * Note there is also the idea of a winsys-base that may represent a window system for which there are multiple winsys APIs. An example of this is x11, since glx and egl may both be used with x11. (currently only Clutter has the idea of a winsys-base) * The driver/ represents a specific varient of OpenGL. Currently we have "gl" representing OpenGL 1.4-2.1 (mostly fixed function) and "gles" representing GLES 1.1 (fixed funciton) and 2.0 (fully shader based) * Everything under cogl/ should fundamentally be supporting access to the GPU. Essentially Cogl's most basic requirement is to provide a nice GPU Graphics API and drawing a line between this and the utility functionality we add to support Clutter should help keep this lean and maintainable. * Code under utils/ as suggested builds on cogl/ adding more convenient APIs or mechanism to optimize special cases. Broadly speaking you can compare cogl/ to OpenGL and utils/ to GLU. * clutter/pango will be moved to clutter/cogl/pango How some of the internal configure.ac/pkg-config terminology has changed: backendextra -> CLUTTER_WINSYS_BASE # e.g. "x11" backendextralib -> CLUTTER_WINSYS_BASE_LIB # e.g. "x11/libclutter-x11.la" clutterbackend -> {CLUTTER,COGL}_WINSYS # e.g. "glx" CLUTTER_FLAVOUR -> {CLUTTER,COGL}_WINSYS clutterbackendlib -> CLUTTER_WINSYS_LIB CLUTTER_COGL -> COGL_DRIVER # e.g. "gl" Note: The CLUTTER_FLAVOUR and CLUTTER_COGL defines are kept for apps As the first thing to take advantage of the new winsys component in Cogl; cogl_get_proc_address() has been moved from cogl/{gl,gles}/cogl.c into cogl/common/cogl.c and this common implementation first trys _cogl_winsys_get_proc_address() but if that fails then it falls back to gmodule.
2009-07-27 21:02:02 -04:00
}
static CoglOnscreen *
find_onscreen_for_xid (CoglContext *context, uint32_t xid)
{
GList *l;
for (l = context->framebuffers; l; l = l->next)
{
CoglFramebuffer *framebuffer = l->data;
CoglOnscreenXlib *xlib_onscreen;
if (framebuffer->type != COGL_FRAMEBUFFER_TYPE_ONSCREEN)
continue;
/* Does the GLXEvent have the GLXDrawable or the X Window? */
xlib_onscreen = COGL_ONSCREEN (framebuffer)->winsys;
if (xlib_onscreen->xwin == (Window)xid)
return COGL_ONSCREEN (framebuffer);
}
return NULL;
}
static void
ensure_ust_type (CoglRenderer *renderer,
GLXDrawable drawable)
{
CoglGLXRenderer *glx_renderer = renderer->winsys;
CoglXlibRenderer *xlib_renderer = _cogl_xlib_renderer_get_data (renderer);
int64_t ust;
int64_t msc;
int64_t sbc;
struct timeval tv;
struct timespec ts;
int64_t current_system_time;
int64_t current_monotonic_time;
if (glx_renderer->ust_type != COGL_GLX_UST_IS_UNKNOWN)
return;
glx_renderer->ust_type = COGL_GLX_UST_IS_OTHER;
if (glx_renderer->glXGetSyncValues == NULL)
goto out;
if (!glx_renderer->glXGetSyncValues (xlib_renderer->xdpy, drawable,
&ust, &msc, &sbc))
goto out;
/* This is the time source that existing (buggy) linux drm drivers
* use */
gettimeofday (&tv, NULL);
current_system_time = (tv.tv_sec * G_GINT64_CONSTANT (1000000)) + tv.tv_usec;
if (current_system_time > ust - 1000000 &&
current_system_time < ust + 1000000)
{
glx_renderer->ust_type = COGL_GLX_UST_IS_GETTIMEOFDAY;
goto out;
}
/* This is the time source that the newer (fixed) linux drm
* drivers use (Linux >= 3.8) */
clock_gettime (CLOCK_MONOTONIC, &ts);
current_monotonic_time = (ts.tv_sec * G_GINT64_CONSTANT (1000000)) +
(ts.tv_nsec / G_GINT64_CONSTANT (1000));
if (current_monotonic_time > ust - 1000000 &&
current_monotonic_time < ust + 1000000)
{
glx_renderer->ust_type = COGL_GLX_UST_IS_MONOTONIC_TIME;
goto out;
}
out:
COGL_NOTE (WINSYS, "Classified OML system time as: %s",
glx_renderer->ust_type == COGL_GLX_UST_IS_GETTIMEOFDAY ? "gettimeofday" :
(glx_renderer->ust_type == COGL_GLX_UST_IS_MONOTONIC_TIME ? "monotonic" :
"other"));
return;
}
static int64_t
ust_to_nanoseconds (CoglRenderer *renderer,
GLXDrawable drawable,
int64_t ust)
{
CoglGLXRenderer *glx_renderer = renderer->winsys;
ensure_ust_type (renderer, drawable);
switch (glx_renderer->ust_type)
{
case COGL_GLX_UST_IS_UNKNOWN:
g_assert_not_reached ();
break;
case COGL_GLX_UST_IS_GETTIMEOFDAY:
case COGL_GLX_UST_IS_MONOTONIC_TIME:
return 1000 * ust;
case COGL_GLX_UST_IS_OTHER:
/* In this case the scale of UST is undefined so we can't easily
* scale to nanoseconds.
*
* For example the driver may be reporting the rdtsc CPU counter
* as UST values and so the scale would need to be determined
* empirically.
*
* Potentially we could block for a known duration within
* ensure_ust_type() to measure the timescale of UST but for now
* we just ignore unknown time sources */
return 0;
}
return 0;
}
static int64_t
_cogl_winsys_get_clock_time (CoglContext *context)
{
CoglGLXRenderer *glx_renderer = context->display->renderer->winsys;
/* We don't call ensure_ust_type() because we don't have a drawable
* to work with. cogl_get_clock_time() is documented to only work
* once a valid, non-zero, timestamp has been retrieved from Cogl.
*/
switch (glx_renderer->ust_type)
{
case COGL_GLX_UST_IS_UNKNOWN:
case COGL_GLX_UST_IS_OTHER:
return 0;
case COGL_GLX_UST_IS_GETTIMEOFDAY:
{
struct timeval tv;
gettimeofday(&tv, NULL);
return tv.tv_sec * G_GINT64_CONSTANT (1000000000) +
tv.tv_usec * G_GINT64_CONSTANT (1000);
}
case COGL_GLX_UST_IS_MONOTONIC_TIME:
{
struct timespec ts;
clock_gettime (CLOCK_MONOTONIC, &ts);
return ts.tv_sec * G_GINT64_CONSTANT (1000000000) + ts.tv_nsec;
}
}
g_assert_not_reached();
return 0;
}
static void
flush_pending_notifications_cb (void *data,
void *user_data)
{
CoglFramebuffer *framebuffer = data;
if (framebuffer->type == COGL_FRAMEBUFFER_TYPE_ONSCREEN)
{
CoglOnscreen *onscreen = COGL_ONSCREEN (framebuffer);
CoglOnscreenGLX *glx_onscreen = onscreen->winsys;
CoglBool pending_sync_notify = glx_onscreen->pending_sync_notify;
CoglBool pending_complete_notify = glx_onscreen->pending_complete_notify;
/* If swap_region is called then notifying the sync event could
* potentially immediately queue a subsequent pending notify so
* we need to clear the flag before invoking the callback */
glx_onscreen->pending_sync_notify = FALSE;
glx_onscreen->pending_complete_notify = FALSE;
if (pending_sync_notify)
{
CoglFrameInfo *info = g_queue_peek_head (&onscreen->pending_frame_infos);
_cogl_onscreen_notify_frame_sync (onscreen, info);
}
if (pending_complete_notify)
{
CoglFrameInfo *info = g_queue_pop_head (&onscreen->pending_frame_infos);
_cogl_onscreen_notify_complete (onscreen, info);
cogl_object_unref (info);
}
if (glx_onscreen->pending_resize_notify)
{
_cogl_onscreen_notify_resize (onscreen);
glx_onscreen->pending_resize_notify = FALSE;
}
}
}
static void
flush_pending_notifications_idle (void *user_data)
{
CoglContext *context = user_data;
CoglRenderer *renderer = context->display->renderer;
CoglGLXRenderer *glx_renderer = renderer->winsys;
/* This needs to be disconnected before invoking the callbacks in
* case the callbacks cause it to be queued again */
_cogl_closure_disconnect (glx_renderer->flush_notifications_idle);
glx_renderer->flush_notifications_idle = NULL;
g_list_foreach (context->framebuffers,
flush_pending_notifications_cb,
NULL);
}
static void
set_sync_pending (CoglOnscreen *onscreen)
{
CoglOnscreenGLX *glx_onscreen = onscreen->winsys;
CoglContext *context = COGL_FRAMEBUFFER (onscreen)->context;
CoglRenderer *renderer = context->display->renderer;
CoglGLXRenderer *glx_renderer = renderer->winsys;
/* We only want to dispatch sync events when the application calls
* cogl_context_dispatch so instead of immediately notifying we
* queue an idle callback */
if (!glx_renderer->flush_notifications_idle)
{
glx_renderer->flush_notifications_idle =
_cogl_poll_renderer_add_idle (renderer,
flush_pending_notifications_idle,
context,
NULL);
}
glx_onscreen->pending_sync_notify = TRUE;
}
static void
set_complete_pending (CoglOnscreen *onscreen)
{
CoglOnscreenGLX *glx_onscreen = onscreen->winsys;
CoglContext *context = COGL_FRAMEBUFFER (onscreen)->context;
CoglRenderer *renderer = context->display->renderer;
CoglGLXRenderer *glx_renderer = renderer->winsys;
/* We only want to notify swap completion when the application calls
* cogl_context_dispatch so instead of immediately notifying we
* queue an idle callback */
if (!glx_renderer->flush_notifications_idle)
{
glx_renderer->flush_notifications_idle =
_cogl_poll_renderer_add_idle (renderer,
flush_pending_notifications_idle,
context,
NULL);
}
glx_onscreen->pending_complete_notify = TRUE;
}
static void
notify_swap_buffers (CoglContext *context, GLXBufferSwapComplete *swap_event)
{
CoglOnscreen *onscreen = find_onscreen_for_xid (context, (uint32_t)swap_event->drawable);
CoglOnscreenGLX *glx_onscreen;
if (!onscreen)
return;
glx_onscreen = onscreen->winsys;
/* We only want to notify that the swap is complete when the
application calls cogl_context_dispatch so instead of immediately
notifying we'll set a flag to remember to notify later */
set_sync_pending (onscreen);
if (swap_event->ust != 0)
{
CoglFrameInfo *info = g_queue_peek_head (&onscreen->pending_frame_infos);
info->presentation_time =
ust_to_nanoseconds (context->display->renderer,
glx_onscreen->glxwin,
swap_event->ust);
}
set_complete_pending (onscreen);
}
static void
update_output (CoglOnscreen *onscreen)
{
CoglOnscreenXlib *xlib_onscreen = onscreen->winsys;
CoglFramebuffer *framebuffer = COGL_FRAMEBUFFER (onscreen);
CoglContext *context = framebuffer->context;
CoglDisplay *display = context->display;
CoglOutput *output;
int width, height;
width = cogl_framebuffer_get_width (framebuffer);
height = cogl_framebuffer_get_height (framebuffer);
output = _cogl_xlib_renderer_output_for_rectangle (display->renderer,
xlib_onscreen->x,
xlib_onscreen->y,
width, height);
if (xlib_onscreen->output != output)
{
if (xlib_onscreen->output)
cogl_object_unref (xlib_onscreen->output);
xlib_onscreen->output = output;
if (output)
cogl_object_ref (xlib_onscreen->output);
}
}
static void
notify_resize (CoglContext *context,
XConfigureEvent *configure_event)
{
CoglOnscreen *onscreen = find_onscreen_for_xid (context,
configure_event->window);
CoglFramebuffer *framebuffer = COGL_FRAMEBUFFER (onscreen);
CoglRenderer *renderer = context->display->renderer;
CoglGLXRenderer *glx_renderer = renderer->winsys;
CoglOnscreenGLX *glx_onscreen;
CoglOnscreenXlib *xlib_onscreen;
if (!onscreen)
return;
glx_onscreen = onscreen->winsys;
xlib_onscreen = onscreen->winsys;
_cogl_framebuffer_winsys_update_size (framebuffer,
configure_event->width,
configure_event->height);
/* We only want to notify that a resize happened when the
* application calls cogl_context_dispatch so instead of immediately
* notifying we queue an idle callback */
if (!glx_renderer->flush_notifications_idle)
{
glx_renderer->flush_notifications_idle =
_cogl_poll_renderer_add_idle (renderer,
flush_pending_notifications_idle,
context,
NULL);
}
glx_onscreen->pending_resize_notify = TRUE;
if (!xlib_onscreen->is_foreign_xwin)
{
int x, y;
if (configure_event->send_event)
{
x = configure_event->x;
y = configure_event->y;
}
else
{
Window child;
XTranslateCoordinates (configure_event->display,
configure_event->window,
DefaultRootWindow (configure_event->display),
0, 0, &x, &y, &child);
}
xlib_onscreen->x = x;
xlib_onscreen->y = y;
update_output (onscreen);
}
}
static CoglFilterReturn
glx_event_filter_cb (XEvent *xevent, void *data)
{
CoglContext *context = data;
2011-04-12 16:02:14 -04:00
#ifdef GLX_INTEL_swap_event
CoglGLXRenderer *glx_renderer;
2011-04-12 16:02:14 -04:00
#endif
if (xevent->type == ConfigureNotify)
{
notify_resize (context,
&xevent->xconfigure);
/* we let ConfigureNotify pass through */
return COGL_FILTER_CONTINUE;
}
#ifdef GLX_INTEL_swap_event
2011-04-12 16:02:14 -04:00
glx_renderer = context->display->renderer->winsys;
if (xevent->type == (glx_renderer->glx_event_base + GLX_BufferSwapComplete))
{
GLXBufferSwapComplete *swap_event = (GLXBufferSwapComplete *) xevent;
notify_swap_buffers (context, swap_event);
/* remove SwapComplete events from the queue */
return COGL_FILTER_REMOVE;
}
#endif /* GLX_INTEL_swap_event */
if (xevent->type == Expose)
{
CoglOnscreen *onscreen =
find_onscreen_for_xid (context, xevent->xexpose.window);
if (onscreen)
{
CoglOnscreenDirtyInfo info;
info.x = xevent->xexpose.x;
info.y = xevent->xexpose.y;
info.width = xevent->xexpose.width;
info.height = xevent->xexpose.height;
_cogl_onscreen_queue_dirty (onscreen, &info);
}
return COGL_FILTER_CONTINUE;
}
return COGL_FILTER_CONTINUE;
}
static void
_cogl_winsys_renderer_disconnect (CoglRenderer *renderer)
{
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
CoglGLXRenderer *glx_renderer = renderer->winsys;
_cogl_xlib_renderer_disconnect (renderer);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
if (glx_renderer->libgl_module)
g_module_close (glx_renderer->libgl_module);
g_slice_free (CoglGLXRenderer, renderer->winsys);
}
static CoglBool
update_all_outputs (CoglRenderer *renderer)
{
GList *l;
_COGL_GET_CONTEXT (context, FALSE);
if (context->display == NULL) /* during connection */
return FALSE;
if (context->display->renderer != renderer)
return FALSE;
for (l = context->framebuffers; l; l = l->next)
{
CoglFramebuffer *framebuffer = l->data;
if (framebuffer->type != COGL_FRAMEBUFFER_TYPE_ONSCREEN)
continue;
update_output (COGL_ONSCREEN (framebuffer));
}
return TRUE;
}
static void
_cogl_winsys_renderer_outputs_changed (CoglRenderer *renderer)
{
update_all_outputs (renderer);
}
static CoglBool
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
resolve_core_glx_functions (CoglRenderer *renderer,
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
CoglError **error)
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
{
CoglGLXRenderer *glx_renderer;
glx_renderer = renderer->winsys;
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
if (!g_module_symbol (glx_renderer->libgl_module, "glXQueryExtension",
(void **) &glx_renderer->glXQueryExtension) ||
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
!g_module_symbol (glx_renderer->libgl_module, "glXQueryVersion",
(void **) &glx_renderer->glXQueryVersion) ||
!g_module_symbol (glx_renderer->libgl_module, "glXQueryExtensionsString",
(void **) &glx_renderer->glXQueryExtensionsString) ||
(!g_module_symbol (glx_renderer->libgl_module, "glXGetProcAddress",
(void **) &glx_renderer->glXGetProcAddress) &&
!g_module_symbol (glx_renderer->libgl_module, "glXGetProcAddressARB",
(void **) &glx_renderer->glXGetProcAddress)) ||
!g_module_symbol (glx_renderer->libgl_module, "glXQueryDrawable",
(void **) &glx_renderer->glXQueryDrawable))
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
{
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
COGL_WINSYS_ERROR_INIT,
"Failed to resolve required GLX symbol");
return FALSE;
}
return TRUE;
}
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
static void
update_base_winsys_features (CoglRenderer *renderer)
{
CoglGLXRenderer *glx_renderer = renderer->winsys;
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (renderer);
const char *glx_extensions;
int default_screen;
char **split_extensions;
int i;
default_screen = DefaultScreen (xlib_renderer->xdpy);
glx_extensions =
glx_renderer->glXQueryExtensionsString (xlib_renderer->xdpy,
default_screen);
COGL_NOTE (WINSYS, " GLX Extensions: %s", glx_extensions);
split_extensions = g_strsplit (glx_extensions, " ", 0 /* max_tokens */);
for (i = 0; i < G_N_ELEMENTS (winsys_feature_data); i++)
if (_cogl_feature_check (renderer,
"GLX", winsys_feature_data + i,
glx_renderer->glx_major,
glx_renderer->glx_minor,
COGL_DRIVER_GL, /* the driver isn't used */
split_extensions,
glx_renderer))
{
glx_renderer->legacy_feature_flags |=
winsys_feature_data[i].feature_flags;
if (winsys_feature_data[i].winsys_feature)
COGL_FLAGS_SET (glx_renderer->base_winsys_features,
winsys_feature_data[i].winsys_feature,
TRUE);
}
g_strfreev (split_extensions);
/* Note: the GLX_SGI_video_sync spec explicitly states this extension
* only works for direct contexts. */
if (!glx_renderer->is_direct)
{
glx_renderer->glXGetVideoSync = NULL;
glx_renderer->glXWaitVideoSync = NULL;
COGL_FLAGS_SET (glx_renderer->base_winsys_features,
COGL_WINSYS_FEATURE_VBLANK_COUNTER,
FALSE);
}
COGL_FLAGS_SET (glx_renderer->base_winsys_features,
COGL_WINSYS_FEATURE_MULTIPLE_ONSCREEN,
TRUE);
if (glx_renderer->glXWaitVideoSync ||
glx_renderer->glXWaitForMsc)
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
COGL_FLAGS_SET (glx_renderer->base_winsys_features,
COGL_WINSYS_FEATURE_VBLANK_WAIT,
TRUE);
}
static CoglBool
_cogl_winsys_renderer_connect (CoglRenderer *renderer,
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
CoglError **error)
{
CoglGLXRenderer *glx_renderer;
CoglXlibRenderer *xlib_renderer;
renderer->winsys = g_slice_new0 (CoglGLXRenderer);
glx_renderer = renderer->winsys;
xlib_renderer = _cogl_xlib_renderer_get_data (renderer);
if (!_cogl_xlib_renderer_connect (renderer, error))
goto error;
Add a GL 3 driver This adds a new CoglDriver for GL 3 called COGL_DRIVER_GL3. When requested, the GLX, EGL and SDL2 winsyss will set the necessary attributes to request a forward-compatible core profile 3.1 context. That means it will have no deprecated features. To simplify the explosion of checks for specific combinations of context->driver, many of these conditionals have now been replaced with private feature flags that are checked instead. The GL and GLES drivers now initialise these private feature flags depending on which driver is used. The fixed function backends now explicitly check whether the fixed function private feature is available which means the GL3 driver will fall back to always using the GLSL progend. Since Rob's latest patches the GLSL progend no longer uses any fixed function API anyway so it should just work. The driver is currently lower priority than COGL_DRIVER_GL so it will not be used unless it is specificly requested. We may want to change this priority at some point because apparently Mesa can make some memory savings if a core profile context is used. In GL 3, getting the combined extensions string with glGetString is deprecated so this patch changes it to use glGetStringi to build up an array of extensions instead. _cogl_context_get_gl_extensions now returns this array instead of trying to return a const string. The caller is expected to free the array. Some issues with this patch: • GL 3 does not support GL_ALPHA format textures. We should probably make this a feature flag or something. Cogl uses this to render text which currently just throws a GL error and breaks so it's pretty important to do something about this before considering the GL3 driver to be stable. • GL 3 doesn't support client side vertex buffers. This probably doesn't matter because CoglBuffer won't normally use malloc'd buffers if VBOs are available, but it might but worth making malloc'd buffers a private feature and forcing it not to use them. • GL 3 doesn't support the default vertex array object. This patch just makes it create and bind a single non-default vertex array object which gets used just like the normal default object. Ideally it would be good to use vertex array objects properly and attach them to a CoglPrimitive to cache the state. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 66c9db993595b3a22e63f4c201ea468bc9b88cb6)
2012-09-26 15:32:36 -04:00
if (renderer->driver != COGL_DRIVER_GL &&
renderer->driver != COGL_DRIVER_GL3)
{
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_INIT,
"GLX Backend can only be used in conjunction with OpenGL");
goto error;
}
glx_renderer->libgl_module = g_module_open (COGL_GL_LIBNAME,
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
G_MODULE_BIND_LAZY);
if (glx_renderer->libgl_module == NULL)
{
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
COGL_WINSYS_ERROR_INIT,
"Failed to dynamically open the OpenGL library");
goto error;
}
if (!resolve_core_glx_functions (renderer, error))
goto error;
if (!glx_renderer->glXQueryExtension (xlib_renderer->xdpy,
&glx_renderer->glx_error_base,
&glx_renderer->glx_event_base))
{
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_INIT,
"XServer appears to lack required GLX support");
goto error;
}
/* XXX: Note: For a long time Mesa exported a hybrid GLX, exporting
* extensions specified to require GLX 1.3, but still reporting 1.2
* via glXQueryVersion. */
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
if (!glx_renderer->glXQueryVersion (xlib_renderer->xdpy,
&glx_renderer->glx_major,
&glx_renderer->glx_minor)
|| !(glx_renderer->glx_major == 1 && glx_renderer->glx_minor >= 2))
{
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_INIT,
"XServer appears to lack required GLX 1.2 support");
goto error;
}
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
update_base_winsys_features (renderer);
glx_renderer->dri_fd = -1;
return TRUE;
error:
_cogl_winsys_renderer_disconnect (renderer);
return FALSE;
}
static CoglBool
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
update_winsys_features (CoglContext *context, CoglError **error)
{
CoglGLXDisplay *glx_display = context->display->winsys;
CoglGLXRenderer *glx_renderer = context->display->renderer->winsys;
_COGL_RETURN_VAL_IF_FAIL (glx_display->glx_context, FALSE);
if (!_cogl_context_update_features (context, error))
return FALSE;
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
memcpy (context->winsys_features,
glx_renderer->base_winsys_features,
sizeof (context->winsys_features));
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
context->feature_flags |= glx_renderer->legacy_feature_flags;
context->feature_flags |= COGL_FEATURE_ONSCREEN_MULTIPLE;
COGL_FLAGS_SET (context->features,
COGL_FEATURE_ID_ONSCREEN_MULTIPLE, TRUE);
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
if (glx_renderer->glXCopySubBuffer || context->glBlitFramebuffer)
{
CoglGpuInfo *info = &context->gpu;
CoglGpuInfoArchitecture arch = info->architecture;
COGL_FLAGS_SET (context->winsys_features, COGL_WINSYS_FEATURE_SWAP_REGION, TRUE);
/*
* "The "drisw" binding in Mesa for loading sofware renderers is
* broken, and neither glBlitFramebuffer nor glXCopySubBuffer
* work correctly."
* - ajax
* - https://bugzilla.gnome.org/show_bug.cgi?id=674208
*
* This is broken in software Mesa at least as of 7.10 and got
* fixed in Mesa 10.1
*/
if (info->driver_package == COGL_GPU_INFO_DRIVER_PACKAGE_MESA &&
info->driver_package_version < COGL_VERSION_ENCODE (10, 1, 0) &&
(arch == COGL_GPU_INFO_ARCHITECTURE_LLVMPIPE ||
arch == COGL_GPU_INFO_ARCHITECTURE_SOFTPIPE ||
arch == COGL_GPU_INFO_ARCHITECTURE_SWRAST))
{
COGL_FLAGS_SET (context->winsys_features,
COGL_WINSYS_FEATURE_SWAP_REGION, FALSE);
}
}
/* Note: glXCopySubBuffer and glBlitFramebuffer won't be throttled
* by the SwapInterval so we have to throttle swap_region requests
* manually... */
if (_cogl_winsys_has_feature (COGL_WINSYS_FEATURE_SWAP_REGION) &&
_cogl_winsys_has_feature (COGL_WINSYS_FEATURE_VBLANK_WAIT))
COGL_FLAGS_SET (context->winsys_features,
COGL_WINSYS_FEATURE_SWAP_REGION_THROTTLE, TRUE);
if (_cogl_winsys_has_feature (COGL_WINSYS_FEATURE_SYNC_AND_COMPLETE_EVENT))
{
/* TODO: remove this deprecated feature */
COGL_FLAGS_SET (context->features,
COGL_FEATURE_ID_SWAP_BUFFERS_EVENT,
TRUE);
COGL_FLAGS_SET (context->features,
COGL_FEATURE_ID_PRESENTATION_TIME,
TRUE);
}
/* We'll manually handle queueing dirty events in response to
* Expose events from X */
COGL_FLAGS_SET (context->private_features,
COGL_PRIVATE_FEATURE_DIRTY_EVENTS,
TRUE);
return TRUE;
}
static void
glx_attributes_from_framebuffer_config (CoglDisplay *display,
CoglFramebufferConfig *config,
int *attributes)
{
CoglGLXRenderer *glx_renderer = display->renderer->winsys;
int i = 0;
attributes[i++] = GLX_DRAWABLE_TYPE;
attributes[i++] = GLX_WINDOW_BIT;
attributes[i++] = GLX_RENDER_TYPE;
attributes[i++] = GLX_RGBA_BIT;
attributes[i++] = GLX_DOUBLEBUFFER;
attributes[i++] = GL_TRUE;
attributes[i++] = GLX_RED_SIZE;
attributes[i++] = 1;
attributes[i++] = GLX_GREEN_SIZE;
attributes[i++] = 1;
attributes[i++] = GLX_BLUE_SIZE;
attributes[i++] = 1;
attributes[i++] = GLX_ALPHA_SIZE;
attributes[i++] = config->swap_chain->has_alpha ? 1 : GLX_DONT_CARE;
attributes[i++] = GLX_DEPTH_SIZE;
attributes[i++] = 1;
attributes[i++] = GLX_STENCIL_SIZE;
attributes[i++] = config->need_stencil ? 1: GLX_DONT_CARE;
if (glx_renderer->glx_major == 1 && glx_renderer->glx_minor >= 4 &&
config->samples_per_pixel)
{
attributes[i++] = GLX_SAMPLE_BUFFERS;
attributes[i++] = 1;
attributes[i++] = GLX_SAMPLES;
attributes[i++] = config->samples_per_pixel;
}
attributes[i++] = None;
g_assert (i < MAX_GLX_CONFIG_ATTRIBS);
}
/* It seems the GLX spec never defined an invalid GLXFBConfig that
* we could overload as an indication of error, so we have to return
* an explicit boolean status. */
static CoglBool
find_fbconfig (CoglDisplay *display,
CoglFramebufferConfig *config,
GLXFBConfig *config_ret,
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
CoglError **error)
{
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (display->renderer);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
CoglGLXRenderer *glx_renderer = display->renderer->winsys;
GLXFBConfig *configs = NULL;
int n_configs;
static int attributes[MAX_GLX_CONFIG_ATTRIBS];
CoglBool ret = TRUE;
int xscreen_num = DefaultScreen (xlib_renderer->xdpy);
glx_attributes_from_framebuffer_config (display, config, attributes);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
configs = glx_renderer->glXChooseFBConfig (xlib_renderer->xdpy,
xscreen_num,
attributes,
&n_configs);
if (!configs || n_configs == 0)
{
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_CREATE_CONTEXT,
"Failed to find any compatible fbconfigs");
ret = FALSE;
goto done;
}
if (config->swap_chain->has_alpha)
{
int i;
for (i = 0; i < n_configs; i++)
{
XVisualInfo *vinfo;
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
vinfo = glx_renderer->glXGetVisualFromFBConfig (xlib_renderer->xdpy,
configs[i]);
if (vinfo == NULL)
continue;
if (vinfo->depth == 32 &&
(vinfo->red_mask | vinfo->green_mask | vinfo->blue_mask)
!= 0xffffffff)
{
COGL_NOTE (WINSYS, "Found an ARGB FBConfig [index:%d]", i);
*config_ret = configs[i];
goto done;
}
}
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_CREATE_CONTEXT,
"Unable to find fbconfig with rgba visual");
ret = FALSE;
goto done;
}
else
{
COGL_NOTE (WINSYS, "Using the first available FBConfig");
*config_ret = configs[0];
}
done:
XFree (configs);
return ret;
}
Add a GL 3 driver This adds a new CoglDriver for GL 3 called COGL_DRIVER_GL3. When requested, the GLX, EGL and SDL2 winsyss will set the necessary attributes to request a forward-compatible core profile 3.1 context. That means it will have no deprecated features. To simplify the explosion of checks for specific combinations of context->driver, many of these conditionals have now been replaced with private feature flags that are checked instead. The GL and GLES drivers now initialise these private feature flags depending on which driver is used. The fixed function backends now explicitly check whether the fixed function private feature is available which means the GL3 driver will fall back to always using the GLSL progend. Since Rob's latest patches the GLSL progend no longer uses any fixed function API anyway so it should just work. The driver is currently lower priority than COGL_DRIVER_GL so it will not be used unless it is specificly requested. We may want to change this priority at some point because apparently Mesa can make some memory savings if a core profile context is used. In GL 3, getting the combined extensions string with glGetString is deprecated so this patch changes it to use glGetStringi to build up an array of extensions instead. _cogl_context_get_gl_extensions now returns this array instead of trying to return a const string. The caller is expected to free the array. Some issues with this patch: • GL 3 does not support GL_ALPHA format textures. We should probably make this a feature flag or something. Cogl uses this to render text which currently just throws a GL error and breaks so it's pretty important to do something about this before considering the GL3 driver to be stable. • GL 3 doesn't support client side vertex buffers. This probably doesn't matter because CoglBuffer won't normally use malloc'd buffers if VBOs are available, but it might but worth making malloc'd buffers a private feature and forcing it not to use them. • GL 3 doesn't support the default vertex array object. This patch just makes it create and bind a single non-default vertex array object which gets used just like the normal default object. Ideally it would be good to use vertex array objects properly and attach them to a CoglPrimitive to cache the state. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 66c9db993595b3a22e63f4c201ea468bc9b88cb6)
2012-09-26 15:32:36 -04:00
static GLXContext
create_gl3_context (CoglDisplay *display,
GLXFBConfig fb_config)
{
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (display->renderer);
CoglGLXRenderer *glx_renderer = display->renderer->winsys;
/* We want a core profile 3.1 context with no deprecated features */
static const int attrib_list[] =
{
GLX_CONTEXT_MAJOR_VERSION_ARB, 3,
GLX_CONTEXT_MINOR_VERSION_ARB, 1,
GLX_CONTEXT_PROFILE_MASK_ARB, GLX_CONTEXT_CORE_PROFILE_BIT_ARB,
GLX_CONTEXT_FLAGS_ARB, GLX_CONTEXT_FORWARD_COMPATIBLE_BIT_ARB,
None
};
/* Make sure that the display supports the GLX_ARB_create_context
extension */
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
if (glx_renderer->glXCreateContextAttribs == NULL)
Add a GL 3 driver This adds a new CoglDriver for GL 3 called COGL_DRIVER_GL3. When requested, the GLX, EGL and SDL2 winsyss will set the necessary attributes to request a forward-compatible core profile 3.1 context. That means it will have no deprecated features. To simplify the explosion of checks for specific combinations of context->driver, many of these conditionals have now been replaced with private feature flags that are checked instead. The GL and GLES drivers now initialise these private feature flags depending on which driver is used. The fixed function backends now explicitly check whether the fixed function private feature is available which means the GL3 driver will fall back to always using the GLSL progend. Since Rob's latest patches the GLSL progend no longer uses any fixed function API anyway so it should just work. The driver is currently lower priority than COGL_DRIVER_GL so it will not be used unless it is specificly requested. We may want to change this priority at some point because apparently Mesa can make some memory savings if a core profile context is used. In GL 3, getting the combined extensions string with glGetString is deprecated so this patch changes it to use glGetStringi to build up an array of extensions instead. _cogl_context_get_gl_extensions now returns this array instead of trying to return a const string. The caller is expected to free the array. Some issues with this patch: • GL 3 does not support GL_ALPHA format textures. We should probably make this a feature flag or something. Cogl uses this to render text which currently just throws a GL error and breaks so it's pretty important to do something about this before considering the GL3 driver to be stable. • GL 3 doesn't support client side vertex buffers. This probably doesn't matter because CoglBuffer won't normally use malloc'd buffers if VBOs are available, but it might but worth making malloc'd buffers a private feature and forcing it not to use them. • GL 3 doesn't support the default vertex array object. This patch just makes it create and bind a single non-default vertex array object which gets used just like the normal default object. Ideally it would be good to use vertex array objects properly and attach them to a CoglPrimitive to cache the state. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 66c9db993595b3a22e63f4c201ea468bc9b88cb6)
2012-09-26 15:32:36 -04:00
return NULL;
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
return glx_renderer->glXCreateContextAttribs (xlib_renderer->xdpy,
fb_config,
NULL /* share_context */,
True, /* direct */
attrib_list);
Add a GL 3 driver This adds a new CoglDriver for GL 3 called COGL_DRIVER_GL3. When requested, the GLX, EGL and SDL2 winsyss will set the necessary attributes to request a forward-compatible core profile 3.1 context. That means it will have no deprecated features. To simplify the explosion of checks for specific combinations of context->driver, many of these conditionals have now been replaced with private feature flags that are checked instead. The GL and GLES drivers now initialise these private feature flags depending on which driver is used. The fixed function backends now explicitly check whether the fixed function private feature is available which means the GL3 driver will fall back to always using the GLSL progend. Since Rob's latest patches the GLSL progend no longer uses any fixed function API anyway so it should just work. The driver is currently lower priority than COGL_DRIVER_GL so it will not be used unless it is specificly requested. We may want to change this priority at some point because apparently Mesa can make some memory savings if a core profile context is used. In GL 3, getting the combined extensions string with glGetString is deprecated so this patch changes it to use glGetStringi to build up an array of extensions instead. _cogl_context_get_gl_extensions now returns this array instead of trying to return a const string. The caller is expected to free the array. Some issues with this patch: • GL 3 does not support GL_ALPHA format textures. We should probably make this a feature flag or something. Cogl uses this to render text which currently just throws a GL error and breaks so it's pretty important to do something about this before considering the GL3 driver to be stable. • GL 3 doesn't support client side vertex buffers. This probably doesn't matter because CoglBuffer won't normally use malloc'd buffers if VBOs are available, but it might but worth making malloc'd buffers a private feature and forcing it not to use them. • GL 3 doesn't support the default vertex array object. This patch just makes it create and bind a single non-default vertex array object which gets used just like the normal default object. Ideally it would be good to use vertex array objects properly and attach them to a CoglPrimitive to cache the state. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 66c9db993595b3a22e63f4c201ea468bc9b88cb6)
2012-09-26 15:32:36 -04:00
}
static CoglBool
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
create_context (CoglDisplay *display, CoglError **error)
{
CoglGLXDisplay *glx_display = display->winsys;
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (display->renderer);
CoglGLXRenderer *glx_renderer = display->renderer->winsys;
CoglBool support_transparent_windows =
display->onscreen_template->config.swap_chain->has_alpha;
GLXFBConfig config;
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
CoglError *fbconfig_error = NULL;
XSetWindowAttributes attrs;
XVisualInfo *xvisinfo;
GLXDrawable dummy_drawable;
CoglXlibTrapState old_state;
_COGL_RETURN_VAL_IF_FAIL (glx_display->glx_context == NULL, TRUE);
glx_display->found_fbconfig =
find_fbconfig (display, &display->onscreen_template->config, &config,
&fbconfig_error);
if (!glx_display->found_fbconfig)
{
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_CREATE_CONTEXT,
"Unable to find suitable fbconfig for the GLX context: %s",
fbconfig_error->message);
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
cogl_error_free (fbconfig_error);
return FALSE;
}
glx_display->fbconfig = config;
glx_display->fbconfig_has_rgba_visual = support_transparent_windows;
COGL_NOTE (WINSYS, "Creating GLX Context (display: %p)",
xlib_renderer->xdpy);
Add a GL 3 driver This adds a new CoglDriver for GL 3 called COGL_DRIVER_GL3. When requested, the GLX, EGL and SDL2 winsyss will set the necessary attributes to request a forward-compatible core profile 3.1 context. That means it will have no deprecated features. To simplify the explosion of checks for specific combinations of context->driver, many of these conditionals have now been replaced with private feature flags that are checked instead. The GL and GLES drivers now initialise these private feature flags depending on which driver is used. The fixed function backends now explicitly check whether the fixed function private feature is available which means the GL3 driver will fall back to always using the GLSL progend. Since Rob's latest patches the GLSL progend no longer uses any fixed function API anyway so it should just work. The driver is currently lower priority than COGL_DRIVER_GL so it will not be used unless it is specificly requested. We may want to change this priority at some point because apparently Mesa can make some memory savings if a core profile context is used. In GL 3, getting the combined extensions string with glGetString is deprecated so this patch changes it to use glGetStringi to build up an array of extensions instead. _cogl_context_get_gl_extensions now returns this array instead of trying to return a const string. The caller is expected to free the array. Some issues with this patch: • GL 3 does not support GL_ALPHA format textures. We should probably make this a feature flag or something. Cogl uses this to render text which currently just throws a GL error and breaks so it's pretty important to do something about this before considering the GL3 driver to be stable. • GL 3 doesn't support client side vertex buffers. This probably doesn't matter because CoglBuffer won't normally use malloc'd buffers if VBOs are available, but it might but worth making malloc'd buffers a private feature and forcing it not to use them. • GL 3 doesn't support the default vertex array object. This patch just makes it create and bind a single non-default vertex array object which gets used just like the normal default object. Ideally it would be good to use vertex array objects properly and attach them to a CoglPrimitive to cache the state. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 66c9db993595b3a22e63f4c201ea468bc9b88cb6)
2012-09-26 15:32:36 -04:00
if (display->renderer->driver == COGL_DRIVER_GL3)
glx_display->glx_context = create_gl3_context (display, config);
else
glx_display->glx_context =
glx_renderer->glXCreateNewContext (xlib_renderer->xdpy,
config,
GLX_RGBA_TYPE,
NULL,
True);
if (glx_display->glx_context == NULL)
{
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_CREATE_CONTEXT,
"Unable to create suitable GL context");
return FALSE;
}
glx_renderer->is_direct =
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXIsDirect (xlib_renderer->xdpy, glx_display->glx_context);
COGL_NOTE (WINSYS, "Setting %s context",
glx_renderer->is_direct ? "direct" : "indirect");
/* XXX: GLX doesn't let us make a context current without a window
* so we create a dummy window that we can use while no CoglOnscreen
* framebuffer is in use.
*/
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
xvisinfo = glx_renderer->glXGetVisualFromFBConfig (xlib_renderer->xdpy,
config);
if (xvisinfo == NULL)
{
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_CREATE_CONTEXT,
"Unable to retrieve the X11 visual");
return FALSE;
}
_cogl_xlib_renderer_trap_errors (display->renderer, &old_state);
attrs.override_redirect = True;
attrs.colormap = XCreateColormap (xlib_renderer->xdpy,
DefaultRootWindow (xlib_renderer->xdpy),
xvisinfo->visual,
AllocNone);
attrs.border_pixel = 0;
glx_display->dummy_xwin =
XCreateWindow (xlib_renderer->xdpy,
DefaultRootWindow (xlib_renderer->xdpy),
-100, -100, 1, 1,
0,
xvisinfo->depth,
CopyFromParent,
xvisinfo->visual,
CWOverrideRedirect | CWColormap | CWBorderPixel,
&attrs);
/* Try and create a GLXWindow to use with extensions dependent on
* GLX versions >= 1.3 that don't accept regular X Windows as GLX
* drawables. */
if (glx_renderer->glx_major == 1 && glx_renderer->glx_minor >= 3)
{
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_display->dummy_glxwin =
glx_renderer->glXCreateWindow (xlib_renderer->xdpy,
config,
glx_display->dummy_xwin,
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
NULL);
}
if (glx_display->dummy_glxwin)
dummy_drawable = glx_display->dummy_glxwin;
else
dummy_drawable = glx_display->dummy_xwin;
COGL_NOTE (WINSYS, "Selecting dummy 0x%x for the GLX context",
(unsigned int) dummy_drawable);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXMakeContextCurrent (xlib_renderer->xdpy,
dummy_drawable,
dummy_drawable,
glx_display->glx_context);
XFree (xvisinfo);
if (_cogl_xlib_renderer_untrap_errors (display->renderer, &old_state))
{
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_CREATE_CONTEXT,
"Unable to select the newly created GLX context");
return FALSE;
}
return TRUE;
}
static void
_cogl_winsys_display_destroy (CoglDisplay *display)
{
CoglGLXDisplay *glx_display = display->winsys;
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (display->renderer);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
CoglGLXRenderer *glx_renderer = display->renderer->winsys;
_COGL_RETURN_IF_FAIL (glx_display != NULL);
if (glx_display->glx_context)
{
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXMakeContextCurrent (xlib_renderer->xdpy,
None, None, NULL);
glx_renderer->glXDestroyContext (xlib_renderer->xdpy,
glx_display->glx_context);
glx_display->glx_context = NULL;
}
if (glx_display->dummy_glxwin)
{
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXDestroyWindow (xlib_renderer->xdpy,
glx_display->dummy_glxwin);
glx_display->dummy_glxwin = None;
}
if (glx_display->dummy_xwin)
{
XDestroyWindow (xlib_renderer->xdpy, glx_display->dummy_xwin);
glx_display->dummy_xwin = None;
}
g_slice_free (CoglGLXDisplay, display->winsys);
display->winsys = NULL;
}
static CoglBool
_cogl_winsys_display_setup (CoglDisplay *display,
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
CoglError **error)
{
CoglGLXDisplay *glx_display;
int i;
_COGL_RETURN_VAL_IF_FAIL (display->winsys == NULL, FALSE);
glx_display = g_slice_new0 (CoglGLXDisplay);
display->winsys = glx_display;
if (!create_context (display, error))
goto error;
for (i = 0; i < COGL_GLX_N_CACHED_CONFIGS; i++)
glx_display->glx_cached_configs[i].depth = -1;
return TRUE;
error:
_cogl_winsys_display_destroy (display);
return FALSE;
}
static CoglBool
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_winsys_context_init (CoglContext *context, CoglError **error)
{
context->winsys = g_new0 (CoglContextGLX, 1);
cogl_xlib_renderer_add_filter (context->display->renderer,
glx_event_filter_cb,
context);
return update_winsys_features (context, error);
}
static void
_cogl_winsys_context_deinit (CoglContext *context)
{
cogl_xlib_renderer_remove_filter (context->display->renderer,
glx_event_filter_cb,
context);
g_free (context->winsys);
}
static CoglBool
_cogl_winsys_onscreen_init (CoglOnscreen *onscreen,
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
CoglError **error)
{
CoglFramebuffer *framebuffer = COGL_FRAMEBUFFER (onscreen);
CoglContext *context = framebuffer->context;
CoglDisplay *display = context->display;
CoglGLXDisplay *glx_display = display->winsys;
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (display->renderer);
CoglGLXRenderer *glx_renderer = display->renderer->winsys;
Window xwin;
CoglOnscreenXlib *xlib_onscreen;
CoglOnscreenGLX *glx_onscreen;
GLXFBConfig fbconfig;
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
CoglError *fbconfig_error = NULL;
_COGL_RETURN_VAL_IF_FAIL (glx_display->glx_context, FALSE);
if (!find_fbconfig (display, &framebuffer->config,
&fbconfig,
&fbconfig_error))
{
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_CREATE_CONTEXT,
"Unable to find suitable fbconfig for the GLX context: %s",
fbconfig_error->message);
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
cogl_error_free (fbconfig_error);
return FALSE;
}
/* Update the real number of samples_per_pixel now that we have
* found an fbconfig... */
if (framebuffer->config.samples_per_pixel)
{
int samples;
int status = glx_renderer->glXGetFBConfigAttrib (xlib_renderer->xdpy,
fbconfig,
GLX_SAMPLES,
&samples);
g_return_val_if_fail (status == Success, TRUE);
framebuffer->samples_per_pixel = samples;
}
/* FIXME: We need to explicitly Select for ConfigureNotify events.
* For foreign windows we need to be careful not to mess up any
* existing event mask.
* We need to document that for windows we create then toolkits
* must be careful not to clear event mask bits that we select.
*/
/* XXX: Note we ignore the user's original width/height when
* given a foreign X window. */
if (onscreen->foreign_xid)
{
Status status;
CoglXlibTrapState state;
XWindowAttributes attr;
int xerror;
xwin = onscreen->foreign_xid;
_cogl_xlib_renderer_trap_errors (display->renderer, &state);
status = XGetWindowAttributes (xlib_renderer->xdpy, xwin, &attr);
XSync (xlib_renderer->xdpy, False);
xerror = _cogl_xlib_renderer_untrap_errors (display->renderer, &state);
if (status == 0 || xerror)
{
char message[1000];
XGetErrorText (xlib_renderer->xdpy, xerror, message, sizeof(message));
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_CREATE_ONSCREEN,
"Unable to query geometry of foreign xid 0x%08lX: %s",
xwin, message);
return FALSE;
}
_cogl_framebuffer_winsys_update_size (framebuffer,
attr.width, attr.height);
/* Make sure the app selects for the events we require... */
onscreen->foreign_update_mask_callback (onscreen,
COGL_ONSCREEN_X11_EVENT_MASK,
onscreen->foreign_update_mask_data);
}
else
{
int width;
int height;
CoglXlibTrapState state;
XVisualInfo *xvisinfo;
XSetWindowAttributes xattr;
unsigned long mask;
int xerror;
width = cogl_framebuffer_get_width (framebuffer);
height = cogl_framebuffer_get_height (framebuffer);
_cogl_xlib_renderer_trap_errors (display->renderer, &state);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
xvisinfo = glx_renderer->glXGetVisualFromFBConfig (xlib_renderer->xdpy,
fbconfig);
if (xvisinfo == NULL)
{
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_CREATE_ONSCREEN,
"Unable to retrieve the X11 visual of context's "
"fbconfig");
return FALSE;
}
/* window attributes */
xattr.background_pixel = WhitePixel (xlib_renderer->xdpy,
DefaultScreen (xlib_renderer->xdpy));
xattr.border_pixel = 0;
/* XXX: is this an X resource that we are leaking‽... */
xattr.colormap = XCreateColormap (xlib_renderer->xdpy,
DefaultRootWindow (xlib_renderer->xdpy),
xvisinfo->visual,
AllocNone);
xattr.event_mask = COGL_ONSCREEN_X11_EVENT_MASK;
mask = CWBorderPixel | CWColormap | CWEventMask;
xwin = XCreateWindow (xlib_renderer->xdpy,
DefaultRootWindow (xlib_renderer->xdpy),
0, 0,
width, height,
0,
xvisinfo->depth,
InputOutput,
xvisinfo->visual,
mask, &xattr);
XFree (xvisinfo);
XSync (xlib_renderer->xdpy, False);
xerror = _cogl_xlib_renderer_untrap_errors (display->renderer, &state);
if (xerror)
{
char message[1000];
XGetErrorText (xlib_renderer->xdpy, xerror,
message, sizeof (message));
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
_cogl_set_error (error, COGL_WINSYS_ERROR,
COGL_WINSYS_ERROR_CREATE_ONSCREEN,
"X error while creating Window for CoglOnscreen: %s",
message);
return FALSE;
}
}
onscreen->winsys = g_slice_new0 (CoglOnscreenGLX);
xlib_onscreen = onscreen->winsys;
glx_onscreen = onscreen->winsys;
xlib_onscreen->xwin = xwin;
xlib_onscreen->is_foreign_xwin = onscreen->foreign_xid ? TRUE : FALSE;
/* Try and create a GLXWindow to use with extensions dependent on
* GLX versions >= 1.3 that don't accept regular X Windows as GLX
* drawables. */
if (glx_renderer->glx_major == 1 && glx_renderer->glx_minor >= 3)
{
glx_onscreen->glxwin =
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXCreateWindow (xlib_renderer->xdpy,
fbconfig,
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
xlib_onscreen->xwin,
NULL);
}
#ifdef GLX_INTEL_swap_event
if (_cogl_winsys_has_feature (COGL_WINSYS_FEATURE_SYNC_AND_COMPLETE_EVENT))
{
GLXDrawable drawable =
glx_onscreen->glxwin ? glx_onscreen->glxwin : xlib_onscreen->xwin;
/* similarly to above, we unconditionally select this event
* because we rely on it to advance the master clock, and
* drive redraw/relayout, animations and event handling.
*/
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXSelectEvent (xlib_renderer->xdpy,
drawable,
GLX_BUFFER_SWAP_COMPLETE_INTEL_MASK);
}
#endif /* GLX_INTEL_swap_event */
return TRUE;
}
static void
_cogl_winsys_onscreen_deinit (CoglOnscreen *onscreen)
{
CoglFramebuffer *framebuffer = COGL_FRAMEBUFFER (onscreen);
CoglContext *context = framebuffer->context;
CoglContextGLX *glx_context = context->winsys;
CoglGLXDisplay *glx_display = context->display->winsys;
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (context->display->renderer);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
CoglGLXRenderer *glx_renderer = context->display->renderer->winsys;
CoglXlibTrapState old_state;
CoglOnscreenXlib *xlib_onscreen = onscreen->winsys;
CoglOnscreenGLX *glx_onscreen = onscreen->winsys;
GLXDrawable drawable;
/* If we never successfully allocated then there's nothing to do */
if (glx_onscreen == NULL)
return;
if (xlib_onscreen->output != NULL)
{
cogl_object_unref (xlib_onscreen->output);
xlib_onscreen->output = NULL;
}
_cogl_xlib_renderer_trap_errors (context->display->renderer, &old_state);
drawable =
glx_onscreen->glxwin == None ? xlib_onscreen->xwin : glx_onscreen->glxwin;
/* Cogl always needs a valid context bound to something so if we are
* destroying the onscreen that is currently bound we'll switch back
* to the dummy drawable. Although the documentation for
* glXDestroyWindow states that a currently bound window won't
* actually be destroyed until it is unbound, it looks like this
* doesn't work if the X window itself is destroyed */
if (drawable == glx_context->current_drawable)
{
GLXDrawable dummy_drawable = (glx_display->dummy_glxwin == None ?
glx_display->dummy_xwin :
glx_display->dummy_glxwin);
glx_renderer->glXMakeContextCurrent (xlib_renderer->xdpy,
dummy_drawable,
dummy_drawable,
glx_display->glx_context);
glx_context->current_drawable = dummy_drawable;
}
if (glx_onscreen->glxwin != None)
{
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXDestroyWindow (xlib_renderer->xdpy,
glx_onscreen->glxwin);
glx_onscreen->glxwin = None;
}
if (!xlib_onscreen->is_foreign_xwin && xlib_onscreen->xwin != None)
{
XDestroyWindow (xlib_renderer->xdpy, xlib_onscreen->xwin);
xlib_onscreen->xwin = None;
}
else
xlib_onscreen->xwin = None;
XSync (xlib_renderer->xdpy, False);
_cogl_xlib_renderer_untrap_errors (context->display->renderer, &old_state);
g_slice_free (CoglOnscreenGLX, onscreen->winsys);
onscreen->winsys = NULL;
}
static void
_cogl_winsys_onscreen_bind (CoglOnscreen *onscreen)
{
CoglContext *context = COGL_FRAMEBUFFER (onscreen)->context;
CoglContextGLX *glx_context = context->winsys;
CoglGLXDisplay *glx_display = context->display->winsys;
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (context->display->renderer);
CoglGLXRenderer *glx_renderer = context->display->renderer->winsys;
CoglOnscreenXlib *xlib_onscreen = onscreen->winsys;
CoglOnscreenGLX *glx_onscreen = onscreen->winsys;
CoglXlibTrapState old_state;
GLXDrawable drawable;
drawable =
glx_onscreen->glxwin ? glx_onscreen->glxwin : xlib_onscreen->xwin;
if (glx_context->current_drawable == drawable)
return;
_cogl_xlib_renderer_trap_errors (context->display->renderer, &old_state);
COGL_NOTE (WINSYS,
"MakeContextCurrent dpy: %p, window: 0x%x (%s), context: %p",
xlib_renderer->xdpy,
(unsigned int) drawable,
xlib_onscreen->is_foreign_xwin ? "foreign" : "native",
glx_display->glx_context);
glx_renderer->glXMakeContextCurrent (xlib_renderer->xdpy,
drawable,
drawable,
glx_display->glx_context);
/* In case we are using GLX_SGI_swap_control for vblank syncing
* we need call glXSwapIntervalSGI here to make sure that it
* affects the current drawable.
*
* Note: we explicitly set to 0 when we aren't using the swap
* interval to synchronize since some drivers have a default
* swap interval of 1. Sadly some drivers even ignore requests
* to disable the swap interval.
*
* NB: glXSwapIntervalSGI applies to the context not the
* drawable which is why we can't just do this once when the
* framebuffer is allocated.
*
* FIXME: We should check for GLX_EXT_swap_control which allows
* per framebuffer swap intervals. GLX_MESA_swap_control also
* allows per-framebuffer swap intervals but the semantics tend
* to be more muddled since Mesa drivers tend to expose both the
* MESA and SGI extensions which should technically be mutually
* exclusive.
*/
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
if (glx_renderer->glXSwapInterval)
{
CoglFramebuffer *fb = COGL_FRAMEBUFFER (onscreen);
if (fb->config.swap_throttled)
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
glx_renderer->glXSwapInterval (1);
else
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
glx_renderer->glXSwapInterval (0);
}
XSync (xlib_renderer->xdpy, False);
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
/* FIXME: We should be reporting a CoglError here
*/
if (_cogl_xlib_renderer_untrap_errors (context->display->renderer,
&old_state))
{
g_warning ("X Error received while making drawable 0x%08lX current",
drawable);
return;
}
glx_context->current_drawable = drawable;
}
static void
_cogl_winsys_wait_for_gpu (CoglOnscreen *onscreen)
{
CoglFramebuffer *framebuffer = COGL_FRAMEBUFFER (onscreen);
CoglContext *ctx = framebuffer->context;
ctx->glFinish ();
}
static void
_cogl_winsys_wait_for_vblank (CoglOnscreen *onscreen)
{
CoglFramebuffer *framebuffer = COGL_FRAMEBUFFER (onscreen);
CoglContext *ctx = framebuffer->context;
CoglGLXRenderer *glx_renderer;
CoglXlibRenderer *xlib_renderer;
glx_renderer = ctx->display->renderer->winsys;
xlib_renderer = _cogl_xlib_renderer_get_data (ctx->display->renderer);
if (glx_renderer->glXWaitForMsc ||
glx_renderer->glXGetVideoSync)
{
CoglFrameInfo *info = g_queue_peek_tail (&onscreen->pending_frame_infos);
if (glx_renderer->glXWaitForMsc)
{
CoglOnscreenGLX *glx_onscreen = onscreen->winsys;
Drawable drawable = glx_onscreen->glxwin;
int64_t ust;
int64_t msc;
int64_t sbc;
glx_renderer->glXGetSyncValues (xlib_renderer->xdpy, drawable,
&ust, &msc, &sbc);
glx_renderer->glXWaitForMsc (xlib_renderer->xdpy, drawable,
0, 2, (msc + 1) % 2,
&ust, &msc, &sbc);
info->presentation_time = ust_to_nanoseconds (ctx->display->renderer,
drawable,
ust);
}
else
{
uint32_t current_count;
struct timespec ts;
glx_renderer->glXGetVideoSync (&current_count);
glx_renderer->glXWaitVideoSync (2,
(current_count + 1) % 2,
&current_count);
clock_gettime (CLOCK_MONOTONIC, &ts);
info->presentation_time =
ts.tv_sec * G_GINT64_CONSTANT (1000000000) + ts.tv_nsec;
}
}
}
static uint32_t
_cogl_winsys_get_vsync_counter (CoglContext *ctx)
{
uint32_t video_sync_count;
CoglGLXRenderer *glx_renderer;
glx_renderer = ctx->display->renderer->winsys;
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
glx_renderer->glXGetVideoSync (&video_sync_count);
return video_sync_count;
}
#ifndef GLX_BACK_BUFFER_AGE_EXT
#define GLX_BACK_BUFFER_AGE_EXT 0x20F4
#endif
static int
_cogl_winsys_onscreen_get_buffer_age (CoglOnscreen *onscreen)
{
CoglFramebuffer *framebuffer = COGL_FRAMEBUFFER (onscreen);
CoglContext *context = framebuffer->context;
CoglXlibRenderer *xlib_renderer = _cogl_xlib_renderer_get_data (context->display->renderer);
CoglGLXRenderer *glx_renderer = context->display->renderer->winsys;
CoglOnscreenGLX *glx_onscreen = onscreen->winsys;
CoglOnscreenXlib *xlib_onscreen = onscreen->winsys;
GLXDrawable drawable = glx_onscreen->glxwin ? glx_onscreen->glxwin : xlib_onscreen->xwin;
unsigned int age;
if (!_cogl_winsys_has_feature (COGL_WINSYS_FEATURE_BUFFER_AGE))
return 0;
glx_renderer->glXQueryDrawable (xlib_renderer->xdpy, drawable, GLX_BACK_BUFFER_AGE_EXT, &age);
return age;
}
static void
set_frame_info_output (CoglOnscreen *onscreen,
CoglOutput *output)
{
CoglFrameInfo *info = g_queue_peek_tail (&onscreen->pending_frame_infos);
info->output = output;
if (output)
{
float refresh_rate = cogl_output_get_refresh_rate (output);
if (refresh_rate != 0.0)
info->refresh_rate = refresh_rate;
}
}
static void
_cogl_winsys_onscreen_swap_region (CoglOnscreen *onscreen,
const int *user_rectangles,
int n_rectangles)
{
CoglFramebuffer *framebuffer = COGL_FRAMEBUFFER (onscreen);
CoglContext *context = framebuffer->context;
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (context->display->renderer);
CoglGLXRenderer *glx_renderer = context->display->renderer->winsys;
CoglOnscreenXlib *xlib_onscreen = onscreen->winsys;
CoglOnscreenGLX *glx_onscreen = onscreen->winsys;
GLXDrawable drawable =
glx_onscreen->glxwin ? glx_onscreen->glxwin : xlib_onscreen->xwin;
uint32_t end_frame_vsync_counter = 0;
CoglBool have_counter;
CoglBool can_wait;
int x_min = 0, x_max = 0, y_min = 0, y_max = 0;
/*
* We assume that glXCopySubBuffer is synchronized which means it won't prevent multiple
* blits per retrace if they can all be performed in the blanking period. If that's the
* case then we still want to use the vblank sync menchanism but
* we only need it to throttle redraws.
*/
CoglBool blit_sub_buffer_is_synchronized =
_cogl_winsys_has_feature (COGL_WINSYS_FEATURE_SWAP_REGION_SYNCHRONIZED);
int framebuffer_width = cogl_framebuffer_get_width (framebuffer);
int framebuffer_height = cogl_framebuffer_get_height (framebuffer);
int *rectangles = g_alloca (sizeof (int) * n_rectangles * 4);
int i;
/* glXCopySubBuffer expects rectangles relative to the bottom left corner but
* we are given rectangles relative to the top left so we need to flip
* them... */
memcpy (rectangles, user_rectangles, sizeof (int) * n_rectangles * 4);
for (i = 0; i < n_rectangles; i++)
{
int *rect = &rectangles[4 * i];
if (i == 0)
{
x_min = rect[0];
x_max = rect[0] + rect[2];
y_min = rect[1];
y_max = rect[1] + rect[3];
}
else
{
x_min = MIN (x_min, rect[0]);
x_max = MAX (x_max, rect[0] + rect[2]);
y_min = MIN (y_min, rect[1]);
y_max = MAX (y_max, rect[1] + rect[3]);
}
rect[1] = framebuffer_height - rect[1] - rect[3];
}
_cogl_framebuffer_flush_state (framebuffer,
framebuffer,
COGL_FRAMEBUFFER_STATE_BIND);
if (framebuffer->config.swap_throttled)
{
have_counter =
_cogl_winsys_has_feature (COGL_WINSYS_FEATURE_VBLANK_COUNTER);
can_wait = _cogl_winsys_has_feature (COGL_WINSYS_FEATURE_VBLANK_WAIT);
}
else
{
have_counter = FALSE;
can_wait = FALSE;
}
/* We need to ensure that all the rendering is done, otherwise
* redraw operations that are slower than the framerate can
* queue up in the pipeline during a heavy animation, causing a
* larger and larger backlog of rendering visible as lag to the
* user.
*
* For an exaggerated example consider rendering at 60fps (so 16ms
* per frame) and you have a really slow frame that takes 160ms to
* render, even though painting the scene and issuing the commands
* to the GPU takes no time at all. If all we did was use the
* video_sync extension to throttle the painting done by the CPU
* then every 16ms we would have another frame queued up even though
* the GPU has only rendered one tenth of the current frame. By the
* time the GPU would get to the 2nd frame there would be 9 frames
* waiting to be rendered.
*
* The problem is that we don't currently have a good way to throttle
* the GPU, only the CPU so we have to resort to synchronizing the
* GPU with the CPU to throttle it.
*
* Note: since calling glFinish() and synchronizing the CPU with
* the GPU is far from ideal, we hope that this is only a short
* term solution.
* - One idea is to using sync objects to track render
* completion so we can throttle the backlog (ideally with an
* additional extension that lets us get notifications in our
* mainloop instead of having to busy wait for the
* completion.)
* - Another option is to support clipped redraws by reusing the
* contents of old back buffers such that we can flip instead
* of using a blit and then we can use GLX_INTEL_swap_events
* to throttle. For this though we would still probably want an
* additional extension so we can report the limited region of
* the window damage to X/compositors.
*/
_cogl_winsys_wait_for_gpu (onscreen);
if (blit_sub_buffer_is_synchronized && have_counter && can_wait)
{
end_frame_vsync_counter = _cogl_winsys_get_vsync_counter (context);
/* If we have the GLX_SGI_video_sync extension then we can
* be a bit smarter about how we throttle blits by avoiding
* any waits if we can see that the video sync count has
* already progressed. */
if (glx_onscreen->last_swap_vsync_counter == end_frame_vsync_counter)
_cogl_winsys_wait_for_vblank (onscreen);
}
else if (can_wait)
_cogl_winsys_wait_for_vblank (onscreen);
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
if (glx_renderer->glXCopySubBuffer)
{
Display *xdpy = xlib_renderer->xdpy;
int i;
for (i = 0; i < n_rectangles; i++)
{
int *rect = &rectangles[4 * i];
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
glx_renderer->glXCopySubBuffer (xdpy, drawable,
rect[0], rect[1], rect[2], rect[3]);
}
}
else if (context->glBlitFramebuffer)
{
int i;
/* XXX: checkout how this state interacts with the code to use
* glBlitFramebuffer in Neil's texture atlasing branch */
/* glBlitFramebuffer is affected by the scissor so we need to
* ensure we have flushed an empty clip stack to get rid of it.
* We also mark that the clip state is dirty so that it will be
* flushed to the correct state the next time something is
* drawn */
_cogl_clip_stack_flush (NULL, framebuffer);
context->current_draw_buffer_changes |= COGL_FRAMEBUFFER_STATE_CLIP;
context->glDrawBuffer (GL_FRONT);
for (i = 0; i < n_rectangles; i++)
{
int *rect = &rectangles[4 * i];
int x2 = rect[0] + rect[2];
int y2 = rect[1] + rect[3];
context->glBlitFramebuffer (rect[0], rect[1], x2, y2,
rect[0], rect[1], x2, y2,
GL_COLOR_BUFFER_BIT, GL_NEAREST);
}
context->glDrawBuffer (GL_BACK);
}
/* NB: unlike glXSwapBuffers, glXCopySubBuffer and
* glBlitFramebuffer don't issue an implicit glFlush() so we
* have to flush ourselves if we want the request to complete in
* a finite amount of time since otherwise the driver can batch
* the command indefinitely. */
context->glFlush ();
/* NB: It's important we save the counter we read before acting on
* the swap request since if we are mixing and matching different
* swap methods between frames we don't want to read the timer e.g.
* after calling glFinish() some times and not for others.
*
* In other words; this way we consistently save the time at the end
* of the applications frame such that the counter isn't muddled by
* the varying costs of different swap methods.
*/
if (have_counter)
glx_onscreen->last_swap_vsync_counter = end_frame_vsync_counter;
if (!xlib_onscreen->is_foreign_xwin)
{
CoglOutput *output;
x_min = CLAMP (x_min, 0, framebuffer_width);
x_max = CLAMP (x_max, 0, framebuffer_width);
y_min = CLAMP (y_min, 0, framebuffer_width);
y_max = CLAMP (y_max, 0, framebuffer_height);
output =
_cogl_xlib_renderer_output_for_rectangle (context->display->renderer,
xlib_onscreen->x + x_min,
xlib_onscreen->y + y_min,
x_max - x_min,
y_max - y_min);
set_frame_info_output (onscreen, output);
}
/* XXX: we don't get SwapComplete events based on how we implement
* the _swap_region() API but if cogl-onscreen.c knows we are
* handling _SYNC and _COMPLETE events in the winsys then we need to
* send fake events in this case.
*/
if (_cogl_winsys_has_feature (COGL_WINSYS_FEATURE_SYNC_AND_COMPLETE_EVENT))
{
set_sync_pending (onscreen);
set_complete_pending (onscreen);
}
}
static void
_cogl_winsys_onscreen_swap_buffers_with_damage (CoglOnscreen *onscreen,
const int *rectangles,
int n_rectangles)
{
CoglFramebuffer *framebuffer = COGL_FRAMEBUFFER (onscreen);
CoglContext *context = framebuffer->context;
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (context->display->renderer);
CoglGLXRenderer *glx_renderer = context->display->renderer->winsys;
CoglOnscreenXlib *xlib_onscreen = onscreen->winsys;
CoglOnscreenGLX *glx_onscreen = onscreen->winsys;
CoglBool have_counter;
GLXDrawable drawable;
/* XXX: theoretically this shouldn't be necessary but at least with
* the Intel drivers we have see that if we don't call
* glXMakeContextCurrent for the drawable we are swapping then
* we get a BadDrawable error from the X server. */
_cogl_framebuffer_flush_state (framebuffer,
framebuffer,
COGL_FRAMEBUFFER_STATE_BIND);
drawable = glx_onscreen->glxwin ? glx_onscreen->glxwin : xlib_onscreen->xwin;
if (framebuffer->config.swap_throttled)
{
uint32_t end_frame_vsync_counter = 0;
have_counter =
_cogl_winsys_has_feature (COGL_WINSYS_FEATURE_VBLANK_COUNTER);
/* If the swap_region API is also being used then we need to track
* the vsync counter for each swap request so we can manually
* throttle swap_region requests. */
if (have_counter)
end_frame_vsync_counter = _cogl_winsys_get_vsync_counter (context);
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
if (!glx_renderer->glXSwapInterval)
{
CoglBool can_wait =
_cogl_winsys_has_feature (COGL_WINSYS_FEATURE_VBLANK_WAIT);
/* If we are going to wait for VBLANK manually, we not only
* need to flush out pending drawing to the GPU before we
* sleep, we need to wait for it to finish. Otherwise, we
* may end up with the situation:
*
* - We finish drawing - GPU drawing continues
* - We go to sleep - GPU drawing continues
* VBLANK - We call glXSwapBuffers - GPU drawing continues
* - GPU drawing continues
* - Swap buffers happens
*
* Producing a tear. Calling glFinish() first will cause us
* to properly wait for the next VBLANK before we swap. This
* obviously does not happen when we use _GLX_SWAP and let
* the driver do the right thing
*/
_cogl_winsys_wait_for_gpu (onscreen);
if (have_counter && can_wait)
{
if (glx_onscreen->last_swap_vsync_counter ==
end_frame_vsync_counter)
_cogl_winsys_wait_for_vblank (onscreen);
}
else if (can_wait)
_cogl_winsys_wait_for_vblank (onscreen);
}
}
else
have_counter = FALSE;
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXSwapBuffers (xlib_renderer->xdpy, drawable);
if (have_counter)
glx_onscreen->last_swap_vsync_counter =
_cogl_winsys_get_vsync_counter (context);
set_frame_info_output (onscreen, xlib_onscreen->output);
}
static uint32_t
_cogl_winsys_onscreen_x11_get_window_xid (CoglOnscreen *onscreen)
{
CoglOnscreenXlib *xlib_onscreen = onscreen->winsys;
return xlib_onscreen->xwin;
}
static void
_cogl_winsys_onscreen_update_swap_throttled (CoglOnscreen *onscreen)
{
CoglContext *context = COGL_FRAMEBUFFER (onscreen)->context;
CoglContextGLX *glx_context = context->winsys;
CoglOnscreenGLX *glx_onscreen = onscreen->winsys;
CoglOnscreenXlib *xlib_onscreen = onscreen->winsys;
GLXDrawable drawable =
glx_onscreen->glxwin ? glx_onscreen->glxwin : xlib_onscreen->xwin;
if (glx_context->current_drawable != drawable)
return;
glx_context->current_drawable = 0;
_cogl_winsys_onscreen_bind (onscreen);
}
static void
_cogl_winsys_onscreen_set_visibility (CoglOnscreen *onscreen,
CoglBool visibility)
{
CoglContext *context = COGL_FRAMEBUFFER (onscreen)->context;
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (context->display->renderer);
CoglOnscreenXlib *xlib_onscreen = onscreen->winsys;
if (visibility)
XMapWindow (xlib_renderer->xdpy, xlib_onscreen->xwin);
else
XUnmapWindow (xlib_renderer->xdpy, xlib_onscreen->xwin);
}
static void
_cogl_winsys_onscreen_set_resizable (CoglOnscreen *onscreen,
CoglBool resizable)
{
CoglFramebuffer *framebuffer = COGL_FRAMEBUFFER (onscreen);
CoglContext *context = framebuffer->context;
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (context->display->renderer);
CoglOnscreenXlib *xlib_onscreen = onscreen->winsys;
XSizeHints *size_hints = XAllocSizeHints ();
if (resizable)
{
/* TODO: Add cogl_onscreen_request_minimum_size () */
size_hints->min_width = 1;
size_hints->min_height = 1;
size_hints->max_width = INT_MAX;
size_hints->max_height = INT_MAX;
}
else
{
int width = cogl_framebuffer_get_width (framebuffer);
int height = cogl_framebuffer_get_height (framebuffer);
size_hints->min_width = width;
size_hints->min_height = height;
size_hints->max_width = width;
size_hints->max_height = height;
}
XSetWMNormalHints (xlib_renderer->xdpy, xlib_onscreen->xwin, size_hints);
XFree (size_hints);
}
/* XXX: This is a particularly hacky _cogl_winsys interface... */
static XVisualInfo *
_cogl_winsys_xlib_get_visual_info (void)
{
CoglGLXDisplay *glx_display;
CoglXlibRenderer *xlib_renderer;
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
CoglGLXRenderer *glx_renderer;
_COGL_GET_CONTEXT (ctx, NULL);
_COGL_RETURN_VAL_IF_FAIL (ctx->display->winsys, FALSE);
glx_display = ctx->display->winsys;
xlib_renderer = _cogl_xlib_renderer_get_data (ctx->display->renderer);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer = ctx->display->renderer->winsys;
if (!glx_display->found_fbconfig)
return NULL;
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
return glx_renderer->glXGetVisualFromFBConfig (xlib_renderer->xdpy,
glx_display->fbconfig);
}
static CoglBool
get_fbconfig_for_depth (CoglContext *context,
unsigned int depth,
GLXFBConfig *fbconfig_ret,
CoglBool *can_mipmap_ret)
{
CoglXlibRenderer *xlib_renderer;
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
CoglGLXRenderer *glx_renderer;
CoglGLXDisplay *glx_display;
Display *dpy;
GLXFBConfig *fbconfigs;
int n_elements, i;
int db, stencil, alpha, mipmap, rgba, value;
int spare_cache_slot = 0;
CoglBool found = FALSE;
xlib_renderer = _cogl_xlib_renderer_get_data (context->display->renderer);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer = context->display->renderer->winsys;
glx_display = context->display->winsys;
/* Check if we've already got a cached config for this depth */
for (i = 0; i < COGL_GLX_N_CACHED_CONFIGS; i++)
if (glx_display->glx_cached_configs[i].depth == -1)
spare_cache_slot = i;
else if (glx_display->glx_cached_configs[i].depth == depth)
{
*fbconfig_ret = glx_display->glx_cached_configs[i].fb_config;
*can_mipmap_ret = glx_display->glx_cached_configs[i].can_mipmap;
return glx_display->glx_cached_configs[i].found;
}
dpy = xlib_renderer->xdpy;
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
fbconfigs = glx_renderer->glXGetFBConfigs (dpy, DefaultScreen (dpy),
&n_elements);
db = G_MAXSHORT;
stencil = G_MAXSHORT;
mipmap = 0;
rgba = 0;
for (i = 0; i < n_elements; i++)
{
XVisualInfo *vi;
int visual_depth;
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
vi = glx_renderer->glXGetVisualFromFBConfig (dpy, fbconfigs[i]);
if (vi == NULL)
continue;
visual_depth = vi->depth;
XFree (vi);
if (visual_depth != depth)
continue;
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_ALPHA_SIZE,
&alpha);
glx_renderer->glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_BUFFER_SIZE,
&value);
if (value != depth && (value - alpha) != depth)
continue;
if (glx_renderer->glx_major == 1 && glx_renderer->glx_minor >= 4)
{
glx_renderer->glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_SAMPLES,
&value);
if (value > 1)
continue;
}
value = 0;
if (depth == 32)
{
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_BIND_TO_TEXTURE_RGBA_EXT,
&value);
if (value)
rgba = 1;
}
if (!value)
{
if (rgba)
continue;
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_BIND_TO_TEXTURE_RGB_EXT,
&value);
if (!value)
continue;
}
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_DOUBLEBUFFER,
&value);
if (value > db)
continue;
db = value;
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_STENCIL_SIZE,
&value);
if (value > stencil)
continue;
stencil = value;
/* glGenerateMipmap is defined in the offscreen extension */
if (cogl_has_feature (context, COGL_FEATURE_ID_OFFSCREEN))
{
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_BIND_TO_MIPMAP_TEXTURE_EXT,
&value);
if (value < mipmap)
continue;
mipmap = value;
}
*fbconfig_ret = fbconfigs[i];
*can_mipmap_ret = mipmap;
found = TRUE;
}
if (n_elements)
XFree (fbconfigs);
glx_display->glx_cached_configs[spare_cache_slot].depth = depth;
glx_display->glx_cached_configs[spare_cache_slot].found = found;
glx_display->glx_cached_configs[spare_cache_slot].fb_config = *fbconfig_ret;
glx_display->glx_cached_configs[spare_cache_slot].can_mipmap = mipmap;
return found;
}
static CoglBool
should_use_rectangle (CoglContext *context)
{
if (context->rectangle_state == COGL_WINSYS_RECTANGLE_STATE_UNKNOWN)
{
if (cogl_has_feature (context, COGL_FEATURE_ID_TEXTURE_RECTANGLE))
{
const char *rect_env;
/* Use the rectangle only if it is available and either:
the COGL_PIXMAP_TEXTURE_RECTANGLE environment variable is
set to 'force'
*or*
the env var is set to 'allow' or not set and NPOTs textures
are not available */
context->rectangle_state =
cogl_has_feature (context, COGL_FEATURE_ID_TEXTURE_NPOT) ?
COGL_WINSYS_RECTANGLE_STATE_DISABLE :
COGL_WINSYS_RECTANGLE_STATE_ENABLE;
if ((rect_env = g_getenv ("COGL_PIXMAP_TEXTURE_RECTANGLE")) ||
/* For compatibility, we'll also look at the old Clutter
environment variable */
(rect_env = g_getenv ("CLUTTER_PIXMAP_TEXTURE_RECTANGLE")))
{
if (g_ascii_strcasecmp (rect_env, "force") == 0)
context->rectangle_state =
COGL_WINSYS_RECTANGLE_STATE_ENABLE;
else if (g_ascii_strcasecmp (rect_env, "disable") == 0)
context->rectangle_state =
COGL_WINSYS_RECTANGLE_STATE_DISABLE;
else if (g_ascii_strcasecmp (rect_env, "allow"))
g_warning ("Unknown value for COGL_PIXMAP_TEXTURE_RECTANGLE, "
"should be 'force' or 'disable'");
}
}
else
context->rectangle_state = COGL_WINSYS_RECTANGLE_STATE_DISABLE;
}
return context->rectangle_state == COGL_WINSYS_RECTANGLE_STATE_ENABLE;
}
static CoglBool
try_create_glx_pixmap (CoglContext *context,
CoglTexturePixmapX11 *tex_pixmap,
CoglBool mipmap)
{
CoglTexturePixmapGLX *glx_tex_pixmap = tex_pixmap->winsys;
CoglRenderer *renderer;
CoglXlibRenderer *xlib_renderer;
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
CoglGLXRenderer *glx_renderer;
Display *dpy;
/* We have to initialize this *opaque* variable because gcc tries to
* be too smart for its own good and warns that the variable may be
* used uninitialized otherwise. */
GLXFBConfig fb_config = (GLXFBConfig)0;
int attribs[7];
int i = 0;
GLenum target;
CoglXlibTrapState trap_state;
unsigned int depth = tex_pixmap->depth;
Visual* visual = tex_pixmap->visual;
renderer = context->display->renderer;
xlib_renderer = _cogl_xlib_renderer_get_data (renderer);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer = renderer->winsys;
dpy = xlib_renderer->xdpy;
if (!get_fbconfig_for_depth (context, depth, &fb_config,
&glx_tex_pixmap->can_mipmap))
{
COGL_NOTE (TEXTURE_PIXMAP, "No suitable FBConfig found for depth %i",
depth);
return FALSE;
}
if (should_use_rectangle (context))
{
target = GLX_TEXTURE_RECTANGLE_EXT;
glx_tex_pixmap->can_mipmap = FALSE;
}
else
target = GLX_TEXTURE_2D_EXT;
if (!glx_tex_pixmap->can_mipmap)
mipmap = FALSE;
attribs[i++] = GLX_TEXTURE_FORMAT_EXT;
/* Check whether an alpha channel is used by comparing the total
* number of 1-bits in color masks against the color depth requested
* by the client.
*/
if (_cogl_util_popcountl (visual->red_mask |
visual->green_mask |
visual->blue_mask) == depth)
attribs[i++] = GLX_TEXTURE_FORMAT_RGB_EXT;
else
attribs[i++] = GLX_TEXTURE_FORMAT_RGBA_EXT;
attribs[i++] = GLX_MIPMAP_TEXTURE_EXT;
attribs[i++] = mipmap;
attribs[i++] = GLX_TEXTURE_TARGET_EXT;
attribs[i++] = target;
attribs[i++] = None;
/* We need to trap errors from glXCreatePixmap because it can
* sometimes fail during normal usage. For example on NVidia it gets
* upset if you try to create two GLXPixmaps for the same drawable.
*/
_cogl_xlib_renderer_trap_errors (renderer, &trap_state);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_tex_pixmap->glx_pixmap =
glx_renderer->glXCreatePixmap (dpy,
fb_config,
tex_pixmap->pixmap,
attribs);
glx_tex_pixmap->has_mipmap_space = mipmap;
XSync (dpy, False);
if (_cogl_xlib_renderer_untrap_errors (renderer, &trap_state))
{
COGL_NOTE (TEXTURE_PIXMAP, "Failed to create pixmap for %p", tex_pixmap);
_cogl_xlib_renderer_trap_errors (renderer, &trap_state);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXDestroyPixmap (dpy, glx_tex_pixmap->glx_pixmap);
XSync (dpy, False);
_cogl_xlib_renderer_untrap_errors (renderer, &trap_state);
glx_tex_pixmap->glx_pixmap = None;
return FALSE;
}
return TRUE;
}
static CoglBool
_cogl_winsys_texture_pixmap_x11_create (CoglTexturePixmapX11 *tex_pixmap)
{
CoglTexturePixmapGLX *glx_tex_pixmap;
CoglContext *ctx = COGL_TEXTURE (tex_pixmap)->context;
if (!_cogl_winsys_has_feature (COGL_WINSYS_FEATURE_TEXTURE_FROM_PIXMAP))
{
tex_pixmap->winsys = NULL;
return FALSE;
}
glx_tex_pixmap = g_new0 (CoglTexturePixmapGLX, 1);
glx_tex_pixmap->glx_pixmap = None;
glx_tex_pixmap->can_mipmap = FALSE;
glx_tex_pixmap->has_mipmap_space = FALSE;
glx_tex_pixmap->glx_tex = NULL;
glx_tex_pixmap->bind_tex_image_queued = TRUE;
glx_tex_pixmap->pixmap_bound = FALSE;
tex_pixmap->winsys = glx_tex_pixmap;
if (!try_create_glx_pixmap (ctx, tex_pixmap, FALSE))
{
tex_pixmap->winsys = NULL;
g_free (glx_tex_pixmap);
return FALSE;
}
return TRUE;
}
static void
free_glx_pixmap (CoglContext *context,
CoglTexturePixmapGLX *glx_tex_pixmap)
{
CoglXlibTrapState trap_state;
CoglRenderer *renderer;
CoglXlibRenderer *xlib_renderer;
CoglGLXRenderer *glx_renderer;
renderer = context->display->renderer;
xlib_renderer = _cogl_xlib_renderer_get_data (renderer);
glx_renderer = renderer->winsys;
if (glx_tex_pixmap->pixmap_bound)
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
glx_renderer->glXReleaseTexImage (xlib_renderer->xdpy,
glx_tex_pixmap->glx_pixmap,
GLX_FRONT_LEFT_EXT);
/* FIXME - we need to trap errors and synchronize here because
* of ordering issues between the XPixmap destruction and the
* GLXPixmap destruction.
*
* If the X pixmap is destroyed, the GLX pixmap is destroyed as
* well immediately, and thus, when Cogl calls glXDestroyPixmap()
* it'll cause a BadDrawable error.
*
* this is technically a bug in the X server, which should not
* destroy either pixmaps until the call to glXDestroyPixmap(); so
* at some point we should revisit this code and remove the
* trap+sync after verifying that the destruction is indeed safe.
*
* for reference, see:
* http://bugzilla.clutter-project.org/show_bug.cgi?id=2324
*/
_cogl_xlib_renderer_trap_errors (renderer, &trap_state);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
glx_renderer->glXDestroyPixmap (xlib_renderer->xdpy,
glx_tex_pixmap->glx_pixmap);
XSync (xlib_renderer->xdpy, False);
_cogl_xlib_renderer_untrap_errors (renderer, &trap_state);
glx_tex_pixmap->glx_pixmap = None;
glx_tex_pixmap->pixmap_bound = FALSE;
}
static void
_cogl_winsys_texture_pixmap_x11_free (CoglTexturePixmapX11 *tex_pixmap)
{
CoglTexturePixmapGLX *glx_tex_pixmap;
if (!tex_pixmap->winsys)
return;
glx_tex_pixmap = tex_pixmap->winsys;
free_glx_pixmap (COGL_TEXTURE (tex_pixmap)->context, glx_tex_pixmap);
if (glx_tex_pixmap->glx_tex)
cogl_object_unref (glx_tex_pixmap->glx_tex);
tex_pixmap->winsys = NULL;
g_free (glx_tex_pixmap);
}
static CoglBool
_cogl_winsys_texture_pixmap_x11_update (CoglTexturePixmapX11 *tex_pixmap,
CoglBool needs_mipmap)
{
CoglTexture *tex = COGL_TEXTURE (tex_pixmap);
CoglContext *ctx = COGL_TEXTURE (tex_pixmap)->context;
CoglTexturePixmapGLX *glx_tex_pixmap = tex_pixmap->winsys;
CoglGLXRenderer *glx_renderer;
/* If we don't have a GLX pixmap then fallback */
if (glx_tex_pixmap->glx_pixmap == None)
return FALSE;
glx_renderer = ctx->display->renderer->winsys;
/* Lazily create a texture to hold the pixmap */
if (glx_tex_pixmap->glx_tex == NULL)
{
CoglPixelFormat texture_format;
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
CoglError *error = NULL;
texture_format = (tex_pixmap->depth >= 32 ?
COGL_PIXEL_FORMAT_RGBA_8888_PRE :
COGL_PIXEL_FORMAT_RGB_888);
if (should_use_rectangle (ctx))
{
glx_tex_pixmap->glx_tex = COGL_TEXTURE (
cogl_texture_rectangle_new_with_size (ctx,
tex->width,
tex->height,
texture_format,
&error));
_cogl_texture_set_internal_format (tex, texture_format);
if (glx_tex_pixmap->glx_tex)
COGL_NOTE (TEXTURE_PIXMAP, "Created a texture rectangle for %p",
tex_pixmap);
else
{
COGL_NOTE (TEXTURE_PIXMAP, "Falling back for %p because a "
"texture rectangle could not be created: %s",
tex_pixmap, error->message);
Adds CoglError api Although we use GLib internally in Cogl we would rather not leak GLib api through Cogl's own api, except through explicitly namespaced cogl_glib_ / cogl_gtype_ feature apis. One of the benefits we see to not leaking GLib through Cogl's public API is that documentation for Cogl won't need to first introduce the Glib API to newcomers, thus hopefully lowering the barrier to learning Cogl. This patch provides a Cogl specific typedef for reporting runtime errors which by no coincidence matches the typedef for GError exactly. If Cogl is built with --enable-glib (default) then developers can even safely assume that a CoglError is a GError under the hood. This patch also enforces a consistent policy for when NULL is passed as an error argument and an error is thrown. In this case we log the error and abort the application, instead of silently ignoring it. In common cases where nothing has been implemented to handle a particular error and/or where applications are just printing the error and aborting themselves then this saves some typing. This also seems more consistent with language based exceptions which usually cause a program to abort if they are not explicitly caught (which passing a non-NULL error signifies in this case) Since this policy for NULL error pointers is stricter than the standard GError convention, there is a clear note in the documentation to warn developers that are used to using the GError api. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit b068d5ea09ab32c37e8c965fc8582c85d1b2db46) Note: Since we can't change the Cogl 1.x api the patch was changed to not rename _error_quark() functions to be _error_domain() functions and although it's a bit ugly, instead of providing our own CoglError type that's compatible with GError we simply #define CoglError to GError unless Cogl is built with glib disabled. Note: this patch does technically introduce an API break since it drops the cogl_error_get_type() symbol generated by glib-mkenum (Since the CoglError enum was replaced by a CoglSystemError enum) but for now we are assuming that this will not affect anyone currently using the Cogl API. If this does turn out to be a problem in practice then we would be able to fix this my manually copying an implementation of cogl_error_get_type() generated by glib-mkenum into a compatibility source file and we could also define the original COGL_ERROR_ enums for compatibility too. Note: another minor concern with cherry-picking this patch to the 1.14 branch is that an api scanner would be lead to believe that some APIs have changed, and for example the gobject-introspection parser which understands the semantics of GError will not understand the semantics of CoglError. We expect most people that have tried to use gobject-introspection with Cogl already understand though that it is not well suited to generating bindings of the Cogl api anyway and we aren't aware or anyone depending on such bindings for apis involving GErrors. (GnomeShell only makes very-very minimal use of Cogl via the gjs bindings for the cogl_rectangle and cogl_color apis.) The main reason we have cherry-picked this patch to the 1.14 branch even given the above concerns is that without it it would become very awkward for us to cherry-pick other beneficial patches from master.
2012-08-31 14:28:27 -04:00
cogl_error_free (error);
free_glx_pixmap (ctx, glx_tex_pixmap);
return FALSE;
}
}
else
{
glx_tex_pixmap->glx_tex = COGL_TEXTURE (
cogl_texture_2d_new_with_size (ctx,
tex->width,
tex->height));
_cogl_texture_set_internal_format (tex, texture_format);
if (cogl_texture_allocate (glx_tex_pixmap->glx_tex, &error))
COGL_NOTE (TEXTURE_PIXMAP, "Created a texture 2d for %p",
tex_pixmap);
else
{
COGL_NOTE (TEXTURE_PIXMAP, "Falling back for %p because a "
"texture 2d could not be created: %s",
tex_pixmap, error->message);
cogl_error_free (error);
free_glx_pixmap (ctx, glx_tex_pixmap);
return FALSE;
}
}
}
if (needs_mipmap)
{
/* If we can't support mipmapping then temporarily fallback */
if (!glx_tex_pixmap->can_mipmap)
return FALSE;
/* Recreate the GLXPixmap if it wasn't previously created with a
* mipmap tree */
if (!glx_tex_pixmap->has_mipmap_space)
{
free_glx_pixmap (ctx, glx_tex_pixmap);
COGL_NOTE (TEXTURE_PIXMAP, "Recreating GLXPixmap with mipmap "
"support for %p", tex_pixmap);
if (!try_create_glx_pixmap (ctx, tex_pixmap, TRUE))
{
/* If the pixmap failed then we'll permanently fallback
* to using XImage. This shouldn't happen. */
COGL_NOTE (TEXTURE_PIXMAP, "Falling back to XGetImage "
"updates for %p because creating the GLXPixmap "
"with mipmap support failed", tex_pixmap);
if (glx_tex_pixmap->glx_tex)
cogl_object_unref (glx_tex_pixmap->glx_tex);
return FALSE;
}
glx_tex_pixmap->bind_tex_image_queued = TRUE;
}
}
if (glx_tex_pixmap->bind_tex_image_queued)
{
GLuint gl_handle, gl_target;
CoglXlibRenderer *xlib_renderer =
_cogl_xlib_renderer_get_data (ctx->display->renderer);
cogl_texture_get_gl_texture (glx_tex_pixmap->glx_tex,
&gl_handle, &gl_target);
COGL_NOTE (TEXTURE_PIXMAP, "Rebinding GLXPixmap for %p", tex_pixmap);
_cogl_bind_gl_texture_transient (gl_target, gl_handle, FALSE);
if (glx_tex_pixmap->pixmap_bound)
Query glX* functions before getting the context to fix GL3 driver The GL3 context is created using the glXCreateContextAttribs function which is part of the GLX_ARB_create_context extension. However previously the function pointers from GLX extensions were only retrieved once the GL context is created. That meant that the GL3 context creation function would always assume that the extension is not supported so it would always fail. This patch changes it to query the functions when the renderer is set up instead. The base winsys feature flags that are determined while querying the functions are stored in a member of CoglGLXRenderer. These are then copied to the CoglContext when it is initialised. The spec for glXGetProcAddress says that the functions returned are context-independent. That implies that it is safe to call it without binding a context although that is not explicitly stated as far as I can tell. A big of googling finds this DRI documentation which says it can be used without a context: http://dri.freedesktop.org/wiki/glXGetProcAddressNeverReturnsNULL And also this code sample: http://www.opengl.org/wiki/Tutorial:_OpenGL_3.0_Context_Creation_%28GLX%29 One point that makes me concerned that this might not always work in practice is that the code in SDL2 to create a GL3 context first creates a dummy GL2 context in order to have something bound before it calls glXGetProcAddress. I think this may just be a misunderstanding based on how wglGetProcAddress works however. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 04a7aca9a98e84e43ac5559305a1358112902e30)
2012-12-12 14:58:35 -05:00
glx_renderer->glXReleaseTexImage (xlib_renderer->xdpy,
glx_tex_pixmap->glx_pixmap,
GLX_FRONT_LEFT_EXT);
glx_renderer->glXBindTexImage (xlib_renderer->xdpy,
glx_tex_pixmap->glx_pixmap,
GLX_FRONT_LEFT_EXT,
NULL);
/* According to the recommended usage in the spec for
* GLX_EXT_texture_pixmap we should release the texture after
* we've finished drawing with it and it is undefined what
* happens if you render to a pixmap that is bound to a texture.
* However that would require the texture backend to know when
* Cogl has finished painting and it may be more expensive to
* keep unbinding the texture. Leaving it bound appears to work
* on Mesa and NVidia drivers and it is also what Compiz does so
* it is probably ok */
glx_tex_pixmap->bind_tex_image_queued = FALSE;
glx_tex_pixmap->pixmap_bound = TRUE;
_cogl_texture_2d_externally_modified (glx_tex_pixmap->glx_tex);
}
return TRUE;
}
static void
_cogl_winsys_texture_pixmap_x11_damage_notify (CoglTexturePixmapX11 *tex_pixmap)
{
CoglTexturePixmapGLX *glx_tex_pixmap = tex_pixmap->winsys;
glx_tex_pixmap->bind_tex_image_queued = TRUE;
}
static CoglTexture *
_cogl_winsys_texture_pixmap_x11_get_texture (CoglTexturePixmapX11 *tex_pixmap)
{
CoglTexturePixmapGLX *glx_tex_pixmap = tex_pixmap->winsys;
return glx_tex_pixmap->glx_tex;
}
static CoglWinsysVtable _cogl_winsys_vtable =
{
.id = COGL_WINSYS_ID_GLX,
.name = "GLX",
.constraints = (COGL_RENDERER_CONSTRAINT_USES_X11 |
COGL_RENDERER_CONSTRAINT_USES_XLIB),
.renderer_get_proc_address = _cogl_winsys_renderer_get_proc_address,
.renderer_connect = _cogl_winsys_renderer_connect,
.renderer_disconnect = _cogl_winsys_renderer_disconnect,
.renderer_outputs_changed = _cogl_winsys_renderer_outputs_changed,
.display_setup = _cogl_winsys_display_setup,
.display_destroy = _cogl_winsys_display_destroy,
.context_init = _cogl_winsys_context_init,
.context_deinit = _cogl_winsys_context_deinit,
.context_get_clock_time = _cogl_winsys_get_clock_time,
.xlib_get_visual_info = _cogl_winsys_xlib_get_visual_info,
.onscreen_init = _cogl_winsys_onscreen_init,
.onscreen_deinit = _cogl_winsys_onscreen_deinit,
.onscreen_bind = _cogl_winsys_onscreen_bind,
.onscreen_swap_buffers_with_damage =
_cogl_winsys_onscreen_swap_buffers_with_damage,
.onscreen_swap_region = _cogl_winsys_onscreen_swap_region,
.onscreen_get_buffer_age = _cogl_winsys_onscreen_get_buffer_age,
.onscreen_update_swap_throttled =
_cogl_winsys_onscreen_update_swap_throttled,
.onscreen_x11_get_window_xid =
_cogl_winsys_onscreen_x11_get_window_xid,
.onscreen_set_visibility = _cogl_winsys_onscreen_set_visibility,
.onscreen_set_resizable =
_cogl_winsys_onscreen_set_resizable,
/* X11 tfp support... */
/* XXX: instead of having a rather monolithic winsys vtable we could
* perhaps look for a way to separate these... */
.texture_pixmap_x11_create =
_cogl_winsys_texture_pixmap_x11_create,
.texture_pixmap_x11_free =
_cogl_winsys_texture_pixmap_x11_free,
.texture_pixmap_x11_update =
_cogl_winsys_texture_pixmap_x11_update,
.texture_pixmap_x11_damage_notify =
_cogl_winsys_texture_pixmap_x11_damage_notify,
.texture_pixmap_x11_get_texture =
_cogl_winsys_texture_pixmap_x11_get_texture,
};
/* XXX: we use a function because no doubt someone will complain
* about using c99 member initializers because they aren't portable
* to windows. We want to avoid having to rigidly follow the real
* order of members since some members are #ifdefd and we'd have
* to mirror the #ifdefing to add padding etc. For any winsys that
* can assume the platform has a sane compiler then we can just use
* c99 initializers for insane platforms they can initialize
* the members by name in a function.
*/
const CoglWinsysVtable *
_cogl_winsys_glx_get_vtable (void)
{
return &_cogl_winsys_vtable;
}