Bug 1011 - Fix TFP fallback mechanism (take 2)
* clutter/glx/clutter-glx-texture-pixmap.c:
* tests/test-pixmap.c:
Slightly modified patch from Gwenole Beauchesne applied
to better handle both falling back to non accelerated tfp
and improvements to the test case.
new layout code
* clutter/clutter-texture.c (clutter_texture_new_from_actor): Now
parents the source actor if it doesn't already have a parent so
that it will get an allocation during layout.
* tests/test-fbo.c: One of the tests tries to ensure that the
ClutterTexture clone keeps the source actor alive by derefing
it. However as actors have a floating reference then test-fbo
doesn't have its own reference once the source is parented so
unrefing just steals the parent's reference and causes
badness. The test now claims the floating reference before cloning
the source so that it can safely be unref'd later.
This is now do-what-I-mean like clutter_actor_get_size so that if
the allocation box is available it will use that, otherwise it
will use the preferred size. clutter_actor_transform_vertices has
been converted to clutter_actor_transform_and_project_box so that
it can be used by both functions. Based on a patch by Emmanuele
Bassi.
(clutter_actor_get_abs_allocation_vertices)
(clutter_actor_get_allocation_vertices): These two functions now
force a relayout if the allocation box is not available.
* tests/test-fbo.c: Fixed the shader to use texture2D instead of
texture2DRect now that GL_EXT_texture_rectangle is no longer used.
Add a note to the clutter_stage_read_pixels doc that the alpha channel
isn't guaranteed to contain sensible data
* tests/test-stage-read-pixels.c: (update_snapshot):
Overwrite the alpha data when using clutter_stage_read_pixels; fixes
this test on non-nvidia drivers
* clutter/glx/clutter-stage-glx.c:
* clutter/x11/clutter-backend-x11.c:
* clutter/x11/clutter-backend-x11.h:
* clutter/x11/clutter-event-x11.c:
* clutter/x11/clutter-x11.h:
* tests/test-devices.c:
Disable use of XInput and add an explicit clutter_x11_enable_xinput
to enable it.
Also fix up the x11 pre-init calls to not need g_type_init.
Correct documentation about accepted pixel formats
* tests/test-stage-read-pixels.c: (main):
Don't try to free a ClutterTexture after main, fixes segfault on quit
Bug 997 - automatic updates not working for named TFP pixmaps,
at least in x11
* clutter/glx/clutter-glx-texture-pixmap.c:
* clutter/glx/clutter-glx-texture-pixmap.h:
* clutter/x11/clutter-backend-x11.c:
* clutter/x11/clutter-x11-texture-pixmap.c:
* clutter/x11/clutter-x11-texture-pixmap.h:
* clutter/x11/clutter-x11.h:
* configure.ac:
* tests/test-pixmap.c:
Rework Andy Wingos patch a little adding more safety for now also
handling redirect Windows (as well as pixmaps)
* clutter/clutter-texture.c:
(clutter_texture_get_preferred_width),
(clutter_texture_get_preferred_height),
(clutter_texture_set_property),
(clutter_texture_get_property),
(clutter_texture_class_init),
(clutter_texture_init): Add the new :keep-aspect-ratio property
to ClutterTexture; when set to TRUE the texture will return a
preferred width maintaining the aspect ratio with the given height
and a preferred height maintaining the aspect ratio with the
given width. This allows to set the width or the height and have
the texture automatically request the height or the width respectively
while maintaining the aspect ratio of the original image.
* tests/test-script.json: Update to test the new :keep-aspect-ratio
property.
* clutter/clutter-actor.c:
(clutter_actor_set_parent): Document and maintain the invariant
that after setting a realized parent on an actor, the actor is
also going to be realized.
(clutter_actor_unparent): Change the invariant that an unparented
actor is also unrealized: the paint is fast enough to avoid
unrealizing, since it also causes more problems that what it's
worth.
* tests/test-invariants.c (test_show_on_set_parent): Update the
invariants test because we changed the invariants.
* clutter/clutter-actor.c (clutter_actor_unparent): Reset the
:show-on-set-parent property to TRUE when unparenting.
* tests/Makefile.am: Add test-invariant to the build.
* tests/test-invariants.c: Test the invariants that we are going
to honour (and document, at some point).
* tests/test-layout.c: Take into account the origin of the
parent, if it has changed, and queue a relayout if we are
using the transformed box when computing the layout.
Bug #815 - Split up request, allocation, and paint box
* clutter/clutter-actor.[ch]: Rework the size allocation,
request and paint area. Now ::request_coords() is called
::allocate(), and ::query_coords() has been split into
::get_preferred_width() and ::get_preferred_height(). See
the documentation and the layout test on how to implement
a container and layout manager with the new API. (#915,
based on a patch by Havoc Pennington, Lucas Rocha and Johan
Bilien)
* clutter/clutter-clone-texture.c: Port CloneTexture to
the new size negotiation API; it just means forwarding
the requests to the parent texture.
* clutter/clutter-deprecated.h: Add deprecated and replaced
API.
* clutter/clutter-entry.c: Port Entry to the new size
negotiation API.
* clutter/clutter-group.c: Port Group to the new size
negotiation API; the semantics of the Group actor do not
change.
* clutter/clutter-label.c: Port Label to the new size
negotiation API, and vastly simplify the code.
* clutter/clutter-main.[ch]: Add API for executing a
relayout when needed.
* clutter/clutter-private.h: Add new Stage private API.
* clutter/clutter-rectangle.c: Update the get_abs_opacity()
call to get_paint_opacity().
* clutter/clutter-stage.c:
(clutter_stage_get_preferred_width),
(clutter_stage_get_preferred_height),
(clutter_stage_allocate),
(clutter_stage_class_init): Port Stage to the new size
negotiation API.
* clutter/clutter-texture.c: Port Texture to the new size
negotiation API.
* clutter/clutter-types.h: Add ClutterRequestMode enumeration.
* clutter/x11/clutter-stage-x11.c: Port the X11 stage
implementation to the new size negotiation API.
* tests/Makefile.am: Add the layout manager test case.
* tests/test-opacity.c: Update.
* tests/test-project.c: Update.
* tests/test-layout.c: Test case for a layout manager implemented
using the new size negotiation API; the layout manager handles
both transformed and untransformed children.
Bug #950 - AltGr not handled
* clutter/osx/clutter-event-osx.c: (clutter_event_osx_translate):
* clutter/x11/clutter-event-x11.c: (translate_key_event):
* tests/test-events.c: (fill_keybuf), (input_cb):
Apply patch from Tommi Komulainen, fill the unicode_value attribute of
the ClutterKeyEvent struct. Also use XKeycodeToKeysym, as suggested in
bug #950, comment #2
Bug#959 - Multiple minor improvements
* configure.ac: pick up GDKPIXBUF_PREFIX from gdk-pixbuf-2.0
module, not pango
* clutter/clutter-keysyms-table.h: cosmetic fix for ifdef guard
name in the comment (leftover from earlier commit)
* tests/test-textures.c: don't depend on GdkPixbuf as we're not
using any actual features from it, just plain pixel buffer
manipulation
* clutter/clutter-timeline.c
* clutter/clutter-timeline.h (clutter_timeline_list_markers): Use
gsize* to return number of items, not guint*
* tests/test-paint-wrapper.c: use correct type (guint*) in call
to clutter_actor_get_size()
* tests/test-depth.c (janus_group): properly take height2 into
account when calculating needed height for the rectangle
* tests/test-cogl-tex-getset.c: use rowstride, not width*4, when
calculating pixel offsets
* tests/test-cogl-tex-getset.c: Don't assume/force RGBA format,
also support ARGB format (needed with quartz imageloader.)
control the use of mipmapping.
* clutter/clutter-texture.h: added ClutterTextureQuality enum.
* tests/test-texture-quality.c: new test.
* tests/Makefile.am: added test-texture-quality
* clutter/clutter-texture.c:
(clutter_texture_class_init): Do not unrealize on hide(), and
do not realize() on show. By default, clutter_actor_show() will
realize ourselves, and we don't want to unrealize when hidden
to avoid paying the penalty of reading back the texture data
from the video memory.
* tests/test-actors.c: Show all textures again when pressing
the 'r' key.
sampler2DRect now that GL_TEXTURE_RECTANGLE support is disabled in
revision 2834. The ClutterTexture actor now has the
'disable-slicing' property set. The distance in texture
coordinates between pixels is passed in as a uniform so that the
box-blur shader can still work.
* tests/Makefile.am:
* tests/test-paint-wrapper.c: Add a test case (merely, a copy
of test-actors.c) that verifies that handlers to the ::paint
signal are called in the right order.
* clutter/cogl/gles/cogl.c:
* clutter/cogl/gl/cogl.c: The clip planes are now set using the
inverse projection matrix as the modelview matrix so that they can
be specified in screen coordinates.
* clutter/cogl/gles/cogl-context.h (CoglContext):
* clutter/cogl/gl/cogl-context.h (CoglContext): Added a member to
cache the inverse projection matrix
* clutter/clutter-fixed.h: Added a constant for converting from
radians to degrees.
* clutter/clutter-fixed.c (clutter_atani, clutter_atan2i): Added
fixed-point versions of atan and atan2.
* tests/test-clip.c: Added a test for clipping with various
rotations and depths.
* tests/Makefile.am (noinst_PROGRAMS): Added test-clip
* clutter/x11/clutter-backend-x11.c:
* clutter/clutter-event.h:
* clutter/clutter-feature.h:
* clutter/clutter-fixed.c:
* clutter/clutter-model.h: Fix documentation.
* clutter/eglnative/clutter-backend-egl.[ch]:
* clutter/eglnative/clutter-event-egl.c: Add the same solution
used for the SDL backend in order to get the time of an event.
This should fix the motion event throttling and the click count
on button press.
* tests/test-pixmap.c (create_pixmap), (main): Fix preprocessor
directives.
* tests/test-events.c: Clean up a bit, and print out more
messages when pressing one of the rectangles.
* tests/test-threads.c: Be more verbose in the console, and
apply another behaviour on the rectangle, to show that we
are not blocking.
* clutter/x11/clutter-x11-texture-pixmap.c:
* configure.ac:
* tests/Makefile.am:
Remove the XComposite dep from Clutter itself, just use
in test-pixmap (if available)
Rework the stage wrapper/implementation relation: remove
duplicated code and all the bookkeeping from the backends into
ClutterStage whenever possible, to reduce the amount of work a
backend must do (and possibly get wrong). Thanks to Tommi
Komulainen.
* clutter/clutter-main.c:
(clutter_init_with_args), (clutter_init): Realize the default
stage after creation. The default stage is special, because we
use it in the initialization sequence. This removes the burden
from the backends and reduces the things a backend can get
wrong.
* clutter/clutter-stage.c:
(clutter_stage_show): Make sure to realize the implementation if
it hasn't been realized yet.
(clutter_stage_realize): Set the REALIZED flag and call
clutter_stage_ensure_current() if the implementation was
successfully realized.
(clutter_stage_unrealized): Call clutter_stage_ensure_current()
on unrealize.
* clutter/glx/clutter-backend-glx.c:
(clutter_backend_glx_create_stage): Do not realize the stage anymore
when creating it, and let the normal realization sequence take
place.
(clutter_backend_glx_ensure_context): Trap for X11 errors.
* clutter/glx/clutter-stage-glx.c:
(clutter_stage_glx_realize): Chain up to the X11 implementation
so that we can set up the window state (title, cursor visibility)
when we actually have a X window. Also, do not call
clutter_stage_ensure_current(), and rely on the wrapper to do
it for us. This means we can drop setting the REALIZED flag on
the wrapper.
(clutter_stage_glx_unrealize): Do not call
clutter_stage_ensure_current() ourselves, and rely on the wrapper
to do it for us.
* clutter/x11/clutter-stage-x11.c:
(set_wm_title), (set_cursor_visible): Move the WM title and
cursor visibility code inside their own functions.
(clutter_stage_x11_realize): Set the window title and whether the
cursor is visible or not after realizing the stage.
(clutter_stage_x11_set_cursor_visible),
(clutter_stage_x11_set_title): Call set_wm_title() and
set_cursor_visible().
(clutter_stage_x11_finalize): Free the title string.
* clutter/x11/clutter-stage-x11.h: Save more of the stage state,
so that we can set it even when the stage hasn't been realized
yet.
* clutter/eglnative/clutter-backend-egl.c:
(clutter_backend_egl_create_stage):
* clutter/eglnative/clutter-stage-egl.c:
(clutter_stage_egl_unrealize),
(clutter_stage_egl_realize): Update the eglnative backend.
* clutter/eglx/clutter-backend-egl.c:
(clutter_backend_egl_ensure_context),
(clutter_backend_egl_create_stage):
* clutter/eglx/clutter-stage-egl.c:
(clutter_stage_egl_unrealize),
(clutter_stage_egl_realize): Update the eglx backend.
* clutter/sdl/clutter-backend-sdl.c:
(clutter_backend_sdl_create_stage):
* clutter/sdl/clutter-stage-sdl.c:
(clutter_stage_sdl_realize): Update the sdl backend.
* clutter/fruity/clutter-backend-fruity.c:
(clutter_backend_fruity_create_stage):
* clutter/sdl/clutter-stage-fruity.c:
(clutter_stage_fruity_realize): Update the fruity backend.
* tests/test-multistage.c (on_button_press): Bail out if
clutter_stage_new() returns NULL.
* HACKING.backends: Update backend writing documentation.
* clutter/clutter-script-private.h: Add a flag for the
default stage.
* clutter/clutter-script.c:
(json_object_end): If the "type" member is "ClutterStage"
and we have a "is-default" member set to true then this
is the default stage.
(clutter_script_construct_object): Special case the default
stage instead of each ClutterStage.
(object_info_free): Ditto as above.
* tests/test-script.json: Test the creation of a non-default
stage and the ::destroy handler to quit.
Bug #902 - Support transform from G_TYPE_INT to ClutterUnit
* clutter/clutter-units.c:
(clutter_value_transform_int_unit),
(clutter_unit_get_type): Add GValue transformation function from
integer values to ClutterUnit. (#902)
(param_unit_validate): Fix validation for ClutterParamSpecUnit;
this allows writable ClutterUnit properties.
* clutter/clutter-fixed.c:
(clutter_value_transform_int_fixed),
(clutter_value_transform_double_fixed),
(clutter_value_transform_float_fixed): Add GValue transformation
functions from native types (int, double, float) to ClutterFixed.
(clutter_fixed_get_type): Register the new transformation functions.
(param_fixed_validate): Fix validation for ClutterParamSpecUnit;
this allows writable ClutterFixed properties.
* clutter/cogl/gl(es)/cogl-texture.h:
* clutter/cogl/gl(es)/cogl-texture.c:
cogl_texture_new_* functions take a gboolean auto_mipmap argument.
If TRUE automatic mipmap generation is enabled during the process
of slice texture object creation.
(cogl_texture_new_from_foreign:) now allows mipmap min filter
flags.
* clutter/clutter-texture.c:
* clutter/glx/clutter-glx-texture-pixmap.c:
* tests/test-cogl-offscreen.c:
* tests/test-cogl-tex-tile.c:
* tests/test-cogl-tex-convert.c:
* tests/test-cogl-tex-polygon.c:
* tests/test-cogl-tex-getset.c:
Pass FALSE for auto_mipmap to cogl_texture_new_*.
* clutter/pango/pangoclutter-render.c:
(tc_get:) Pass TRUE to cogl_texture_new_with_size and use mipmap
min filter for nicer glyphs at small scales. As a result test-text
has gone all beautiful now.
* clutter/cogl/common/cogl-primitives.c: api review touch ups.
* clutter/cogl/gl/cogl-primitives.c: (cogl_path_fill),
(cogl_path_stroke): indentation.
* clutter/cogl/gles/cogl-primitives.c:
(_cogl_path_fill_nodes): free allocated resources.
* tests/test-cogl-primitives.c: updated to new API, added rotation to
test to show that cogl renders paths correct under perspective
distortion.
* clutter/glx/clutter-glx-texture-pixmap.c:
* tests/test-pixmap.c:
Fix up texture-pixmap, now works.
Thanks to Neil for spotting we were actually using
wrong texture type in fbconfig (COGL change).
primitives and path API a bit smaller and more resembling cairo.
* clutter/cogl/gl/cogl-primitives.c:
* clutter/cogl/gles/cogl-primitives.c:
* clutter/cogl/gles/cogl.c:
* clutter/clutter-actor.c:
* clutter/clutter-rectangle.c:
* tests/test-cogl-offscreen.c:
* tests/test-cogl-primitives.c:
* tests/test-cogl-tex-convert.c:
* tests/test-cogl-tex-foreign.c:
* tests/test-cogl-tex-getset.c:
* tests/test-cogl-tex-tile.c: updated according to changes in cogl.