Commit Graph

76051 Commits

Author SHA1 Message Date
Benjamin Otte
3b5a2d60c7 Merge branch 'wip/otte/gl-version' into 'main'
Make sure highest possible version GLContext is created

See merge request GNOME/gtk!5870
2023-04-27 00:39:19 +00:00
Benjamin Otte
9254ab8503 wgl: Improve error messages when GL init fails
In particular, we want to get the GL version, when the Windows box/VM
has an unsuitable GL implementation.

This is somewhat helpful in analyzing failures to bring up GL on
machines where users claim GL does work.
2023-04-27 02:19:25 +02:00
Benjamin Otte
d6afcee1e4 wgl: Create context during WGL initialization
This way, we can realize it and either print success information about
it or return NULL if that fails.

This makes it more likely that we fail early, which means we can then
initialize EGL.
2023-04-27 02:16:46 +02:00
Benjamin Otte
c71ca481c1 win32: Get rid of display->gl_version
We can just query the display's GL context.

And it's not used otherwise anymore.
2023-04-27 02:13:33 +02:00
Benjamin Otte
8ef38c46b5 win32: Refactor context creation
This refactor achieves the following:

 * check GL version against proper matching context version
   In particular, for legacy contexts, we now actually check
 * make sure the actual version is set, even for legacy contexts
 * make sure set_is_legacy() is set properly
2023-04-27 02:13:33 +02:00
Benjamin Otte
4333d754b8 gdk: Add workaround for Visual Studio
apparently casting something to itself makes it not constant.
2023-04-27 02:13:33 +02:00
Benjamin Otte
9f82d537b0 mac: Properly set GLContext.is_legacy() 2023-04-27 02:13:32 +02:00
Benjamin Otte
7c7a3d67ca glcontext: assert all contexts set the version on realize()
Now that all contexts do that, insist that they keep doing it.

And because they keep doing it, we can support querying the GL version
from gdk_gl_context_get_version() without requiring the context to be
made current.
2023-04-27 02:13:32 +02:00
Benjamin Otte
d33b82249b mac: Try all different OpenGL profiles
.. and pick the best one that is supported.

Also make sure that the resulting context has at least the desired
version, otherwise bail.
2023-04-27 02:13:32 +02:00
Benjamin Otte
5f833f1d31 glcontext: Compute matching version the simple way
Do it all in one function instead of requiring two different ones.
2023-04-27 02:13:32 +02:00
Benjamin Otte
b8958419e6 win32: Make sure highest possible GL version is created
Mirror EGL here.
2023-04-27 02:13:32 +02:00
Benjamin Otte
919c90182f win32: Pass the version properly through the creation stack
We were using major/minor instead of GdkGLVersion.

And we were resetting back to 0 and ignoring the required min version
which we should not do.
2023-04-27 02:13:32 +02:00
Benjamin Otte
ab2a548479 glx: Make sure highest possible GL version is created
Mirror EGL here.
2023-04-27 02:13:32 +02:00
Benjamin Otte
34662fc4b0 egl: Make sure highest possible GL version is created
The EGL spec states:

    The context returned must be the specified version, or a later
    version which is backwards compatible with that version.
    Even if a later version is returned, the specified version
    must correspond to a defined version of the client API.

GTK has so far been relying on EGL implementations returning a
later version, because that is what Mesa does.
But ANGLE does not do that and only provides the minimum version, which
means Windows EGL has been forced to use a lower EGL version for no
reason.

So fix this and try versions in order from highest to lowest.
2023-04-27 02:13:32 +02:00
Benjamin Otte
5b376cedcf gdk: Move GdkGLAPI enum into gdkenums.h 2023-04-27 02:13:32 +02:00
Benjamin Otte
3aefed39b1 glcontext: Use GdkGLVersion elsewhere
... and add a convenience API to generate GL versions from strings to
make the gdk_gl_context_check() API nicer.
2023-04-27 02:13:32 +02:00
Benjamin Otte
f86429177a gdk: Introduce GdkGLVersion
... and use it.

Makes the code simpler.
2023-04-27 02:13:32 +02:00
Benjamin Otte
dacfed3e11 win32: Remove an outdated check
We require GL 3.0, so checking for less than 2.0 makes no sense anymore.
2023-04-27 02:13:32 +02:00
Matthias Clasen
8fb2ee7d67 Fix a crash
Fix an oversight from c87b193d2a.
2023-04-27 00:09:18 +02:00
Benjamin Otte
29e7186829 Merge branch 'wip/otte/treelistmodel-madness' into 'main'
treelistmodel: Be safer during collapsing

See merge request GNOME/gtk!5875
2023-04-26 22:06:40 +00:00
Matthias Clasen
951e3b3d6b Fix the build with clang
This broke our macos ci.
2023-04-26 23:54:43 +02:00
Benjamin Otte
3c76f3fb58 treelistmodel: Delay notifies from TreeListRow
Don't notify during destruction, notify afterwards.
This way we don't call into user code from a half-destructed node.

Note that this changes the order in which those notifies happen when
collapsing a large tree: From parent node before child nodes to child
nodes before parent node.

No actual use case for this, just thought it would be safer.
2023-04-26 23:40:56 +02:00
Benjamin Otte
8766a6fab2 treelistmodel: Be safer during collapsing
While we are collapsing a subtree, some signal handlers may not be
disconnected while we are doing this. By adding this check and not
giving those nodes no longer access to the model, we can stop it from
modifying it while we are trying to collapse stuff.

Fixes some crashes in gnome-builder.
2023-04-26 23:30:37 +02:00
Benjamin Otte
f52975c220 Merge branch 'wip/otte/surface-surgery' into 'main'
refactor internal surface API

See merge request GNOME/gtk!5846
2023-04-26 21:03:24 +00:00
Daniel Boles
abd4a57031 3to4: Fix we said we replace doubles with doubles,
when what we are replacing are ints!
2023-04-26 21:30:10 +01:00
Benjamin Otte
9d0448756f display: Remove ::create_surface() vfunc
Instead, have a toplevel_type and popup_type in GdkDisplay and
call g_object_new() with those types.
2023-04-26 21:03:34 +02:00
Benjamin Otte
03d7ce3287 wayland: Set default title in toplevel_init() 2023-04-26 21:03:34 +02:00
Benjamin Otte
f2083d36a1 macos: Add surface from ::constructed() 2023-04-26 21:03:34 +02:00
Benjamin Otte
cb642bec25 gdk: Remove GDK_SURFACE_DRAG enum member
I want to remove the whole enum, but this value was used all over the
place. So removing it first ensures I didn't forget anything.
2023-04-26 21:03:34 +02:00
Benjamin Otte
8c530264f7 wayland: Create drag surface directly
... instead of going through create_surface().
2023-04-26 21:03:34 +02:00
Benjamin Otte
d99042dd91 wayland: Remove struct member
It's unused.
2023-04-26 21:03:34 +02:00
Benjamin Otte
486196c979 x11: Add private gdk_x11_drag_surface_new()
... and use it.
2023-04-26 21:03:34 +02:00
Benjamin Otte
83faacabe3 x11: Set frame clock in ::constructed 2023-04-26 21:03:34 +02:00
Benjamin Otte
31aae62f9a x11: Move window construction to ::constructed() 2023-04-26 21:03:34 +02:00
Benjamin Otte
726be8e2b2 broadway: Create gdk_broadway_drag_surface_new()
... and use it.
2023-04-26 21:03:34 +02:00
Benjamin Otte
ab0fa08d8e win32: Create gdk_win32_draw_surface_new()
... and use it.
2023-04-26 21:03:34 +02:00
Benjamin Otte
42cea18f3e macos: Set frame clock in ::constructed 2023-04-26 21:03:34 +02:00
Benjamin Otte
44e54e1b4c broadway: Set frame clock in ::constructed
... instead of passing it to g_object_new().
2023-04-26 21:03:34 +02:00
Benjamin Otte
a79ae95e7e broadway: Move surface init code into ::constructed() 2023-04-26 21:03:34 +02:00
Benjamin Otte
77fe9116af win32: Set frame clock in constructed()
... instead of passing it as construct argument.
2023-04-26 21:03:34 +02:00
Benjamin Otte
808cde74be wayland: Construct the frame clock in the surface
... instead of passing it to g_object_new().
2023-04-26 21:03:34 +02:00
Benjamin Otte
13120ccf9d surface: Make gdk_surface_set_frame_clock() available
... to backends.

That way, frame clocks can be constructed by the backends' surface
implementations and dont need to be passed in as construct arguments.

Also add an assertion that they are indeed constructed.
2023-04-26 21:03:34 +02:00
Benjamin Otte
a306401023 macos: Move native window creation to ::constructed()
That way, it doesn't need a specific init function.

Also chain up last, so that the generic initialization code in
GdkSurface::constructed can access a fully initialized macos surface.
2023-04-26 21:03:34 +02:00
Benjamin Otte
7e18a1cea7 macos: Don't pass sizes to macos_surface_new()
They're 0, 0, 100, 100 always, so just use those values everywhere.
2023-04-26 21:03:34 +02:00
Benjamin Otte
9a6e6be785 macos: Create drag surface directly
Do not go via macos_surface_new().
2023-04-26 21:03:34 +02:00
Benjamin Otte
093a4e83d5 macos: Move construction stuff into ::constructed 2023-04-26 21:03:34 +02:00
Benjamin Otte
428798b53f macos: toplevels have no parent 2023-04-26 21:03:34 +02:00
Benjamin Otte
4fa81cf70f win32: Move surface init code into ::constructed() 2023-04-26 21:03:34 +02:00
Benjamin Otte
9ac31fe0a0 win32: Abort on error
Returning NULL from a function that must not return NULL is not a good
idea.
2023-04-26 21:03:34 +02:00
Benjamin Otte
69fabb3ce9 win32: Remove outdated debug messages
They are not needed anymore as they print stuff that's no decided in
that function anymore.
2023-04-26 21:03:34 +02:00