By default NSWindow does not listen to mousemoved events and hence the
default behaviour for Actors using the "motion-event" signal differs
from backend to backend.
Using setAcceptsMouseMovedEvents seems to fix it; unfortunately, I
cannot verify it, but since nobody is currently working on the Quartz
backend I guess it cannot get more broken than how currently is.
Thanks to: Michael <michael@f3k.org>
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1687
It would be useful inside a custom actor's paint function to be able to
tell if this is a primary paint call, or if we are in fact painting on
behalf of a clone.
In Mutter we have an optimization not to paint occluded windows; this is
desirable for the windows per se, to conserve bandwith to the card, but
if something like an application switcher is using clones of these windows,
they will not get painted either; currently we have no way of
differentiating between the two.
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1685
The CLUTTER_TEXTURE_IN_CLONE_PAINT was used with the old CloneTexture
actor; now that we have ClutterClone nothing sets the private flag
anymore, and the flag itself is not needed.
Updating the WM hints on the stage window shortcircuits if the stage
is in WITHDRAWN state, so we need to move the update_wm_hints() call
after the flag has been unset.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
If we manually wait for the VBLANK with:
- SGI_video_sync
- Direct usage of the DRM ioctl
Then we should call glFinish() first, or otherwise the swap-buffers
may be delayed by pending drawing and cause a tear.
http://bugzilla.openedhand.com/show_bug.cgi?id=1636
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
commit e2c4a2a9f8 fixed one thing but broke many others things :-/
hopfully this fixes that.
It turned out that the journal was mistakenly setting the OVERRIDE_LAYER0
flush option for all entries, but some other logic errors were also
uncovered in _cogl_material_equal.
Added an internal clutter function, _clutter_master_clock_ensure_next_iteration
that ensures another iteration of the master clock, can be called from repaint
functions as well as other threads.
To help us handle sliced textures; When flushing materials there is an
override option that can be given to replace the texture name for layer0
so we may iterate the slices without needing to modify the material
in use.
Since improving the journal's ability to batch state changes we added a
_cogl_material_equals function that is used by the journal to compare
materials and identify when a state change is required, but this wasn't
correctly considering the layer0 override resulting in false positives that
meant the journal wouldn't update the GL state and the first texture name
was used for all slices.
The cost of glGetFloatv with Mesa is still representing a majority of our
time in OpenGL for some applications, and the last thing left using this is
the current-matrix API when getting the projection matrix.
This adds a matrix stack for the projection matrix, so all getting, setting
and modification of the projection matrix is now managed by Cogl and it's only
when we come to draw that we flush changes to the matrix to OpenGL.
This also brings us closer to being able to drop internal use of the
deprecated OpenGL matrix functions, re: commit 54159f5a1d
The clutter_actor_get_allocation_coords() is not used, and since
the switch to floats in the Actor's API, it returns exactly what
the get_allocation_box() returns.
Currently, the transformation matrix for an actor is constructed
from scenegraph-related accessors. An actor, though, can call COGL
API to add new transformations inside the paint() implementation,
for instance:
static void
my_foo_paint (ClutterActor *a)
{
...
cogl_translate (-scroll_x, -scroll_y, 0);
...
}
Unfortunately these transformations will be completely ignored by
the scenegraph machinery; for instance, getting the actor-relative
coordinates from event coordinates is going to break badly because
of this.
In order to make the scenegraph aware of the potential of additional
transformations, we need a ::apply_transform() virtual function. This
vfunc will pass a CoglMatrix which can be used to apply additional
operations:
static void
my_foo_apply_transform (ClutterActor *a, CoglMatrix *m)
{
CLUTTER_ACTOR_CLASS (my_foo_parent_class)->apply_transform (a, m);
...
cogl_matrix_translate (m, -scroll_x, -scroll_y, 0);
...
}
The ::paint() implementation will be called with the actor already
using the newly applied transformation matrix, as expected:
static void
my_foo_paint (ClutterActor *a)
{
...
}
The ::apply_transform() implementations *must* chain up, so that the
various transformations of each class are preserved. The default
implementation inside ClutterActor applies all the transformations
defined by the scenegraph-related accessors.
Actors performing transformations inside the paint() function will
continue to work as previously.
Scanners like gtk-doc and g-ir-scanner get confused by:
typedef struct _Foo {
...
} Foo;
And expect instead:
typedef struct _Foo Foo;
struct _Foo {
...
};
CoglMatrix definition should be changed to avoid the former type.
The race we were experiencing in the X11 backends is apparently
back after the fix in commit 00a3c698.
This time, just delaying the setting of the SYNC_MATRICES flag
is not enough, so we can resume the use of a STAGE_IN_RESIZE
private flag.
This should also fix bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1668
In order to validate the sequence of:
XResizeWindow
ConfigureNotify
glViewport
that should happen on X11 we need to add debug annotations to the
calls to glViewport() done through COGL.
This avoids some calls to glGetFloatv, which have at least proven to be very
in-efficient in mesa at this point in time, since it always updates all derived
state even when it may not relate to the state being requested.
Fixes and adds a unit test for creating and drawing using materials with
COGL_INVALID_HANDLE texture layers.
This may be valid if for example the user has set a texture combine string
that only references a constant color.
_cogl_material_flush_layers_gl_state will bind the fallback texture for any
COGL_INVALID_HANDLE layer, later though we could explicitly check when the
current blend mode does't actually reference a texture source in which case
binding the fallback texture is redundant.
This tests drawing using cogl_rectangle, cogl_polygon and
cogl_vertex_buffer_draw.
Although we wouldn't recommend developers try and interleve OpenGL drawing
with Cogl drawing - we would prefer patches that improve Cogl to avoid this
if possible - we are providing a simple mechanism that will at least give
developers a fighting chance if they find it necissary.
Note: we aren't helping developers change OpenGL state to modify the
behaviour of Cogl drawing functions - it's unlikley that can ever be
reliably supported - but if they are trying to do something like:
- setup some OpenGL state.
- draw using OpenGL (e.g. glDrawArrays() )
- reset modified OpenGL state.
- continue using Cogl to draw
They should surround their blocks of raw OpenGL with cogl_begin_gl() and
cogl_end_gl():
cogl_begin_gl ();
- setup some OpenGL state.
- draw using OpenGL (e.g. glDrawArrays() )
- reset modified OpenGL state.
cogl_end_gl ();
- continue using Cogl to draw
Again; we aren't supporting code like this:
- setup some OpenGL state.
- use Cogl to draw
- reset modified OpenGL state.
When the internals of Cogl evolves, this is very liable to break.
cogl_begin_gl() will flush all internally batched Cogl primitives, and emit
all internal Cogl state to OpenGL as if it were going to draw something
itself.
The result is that the OpenGL modelview matrix will be setup; the state
corresponding to the current source material will be setup and other world
state such as backface culling, depth and fogging enabledness will be also
be sent to OpenGL.
Note: no special material state is flushed, so if developers want Cogl to setup
a simplified material state it is the their responsibility to set a simple
source material before calling cogl_begin_gl. E.g. by calling
cogl_set_source_color4ub().
Note: It is the developers responsibility to restore any OpenGL state that they
modify to how it was after calling cogl_begin_gl() if they don't do this then
the result of further Cogl calls is undefined.
This function should only need to be called in exceptional circumstances
since Cogl can normally determine internally when a flush is necessary.
As an optimization Cogl drawing functions may batch up primitives
internally, so if you are trying to use raw GL outside of Cogl you stand a
better chance of being successful if you ask Cogl to flush any batched
geometry before making your state changes.
cogl_flush() ensures that the underlying driver is issued all the commands
necessary to draw the batched primitives. It provides no guarantees about
when the driver will complete the rendering.
This provides no guarantees about the GL state upon returning and to avoid
confusing Cogl you should aim to restore any changes you make before
resuming use of Cogl.
If you are making state changes with the intention of affecting Cogl drawing
primitives you are 100% on your own since you stand a good chance of
conflicting with Cogl internals. For example clutter-gst which currently
uses direct GL calls to bind ARBfp programs will very likely break when Cogl
starts to use ARBfb programs internally for the material API, but for now it
can use cogl_flush() to at least ensure that the ARBfp program isn't applied
to additional primitives.
This does not provide a robust generalized solution supporting safe use of
raw GL, its use is very much discouraged.
Previously we would call _cogl_material_pre_change_notify unconditionally, but
now we wait until we really know we are removing a layer before notifying the
change, which will require a journal flush.
Since the convenience functions cogl_set_source_color4ub and
cogl_set_source_texture share a single material, cogl_set_source_color4ub
always calls cogl_material_remove_layer. Often this is a NOP though and
shouldn't require a journal flush.
This gets performance back to where it was before reverting the per-actor
material commits.
This reverts commit 8cf42ea8ac5c05f6b443c453f9c6c2a3cd75acfa.
Since the journal puts material colors in the vertex array accumulated for
drawing we don't need to flush the journal simply due to color changes which
means using cogl_set_source_color4ub is no longer a concern.
This reverts commit 85243da382025bd516937c76a61b8381f6e74689.
Since the journal puts material colors in the vertex array accumulated for
drawing we don't need to flush the journal simply due to color changes
which means using cogl_set_source_color4ub is no longer a concern.
Before any cogl vertex buffer drawing we call
enable_state_for_drawing_buffer which sets up the GL state, but we weren't
disabling unsed client texture coord arrays.
This simplifies the vertex data uploading in the journal, and could improve
performance. Modifying a VBO mid-scene could reqire synchronizing with the
GPU or some form of shadowing/copying to avoid modifying data that the GPU
is currently processing; the buffer was also being marked as GL_STATIC_DRAW
which could have made things worse.
Now we simply create a GL_STATIC_DRAW VBO for each flush and and delete it
when we are finished.
Using cogl_rectangle (and thus the journal) in
_cogl_add_path_to_stencil_buffer means we have to consider all the state
that the journal may change in case it may interfer with the direct GL calls
used. This has proven to be error prone and in this case the journal is an
unnecissary overhead. We now simply call glRectf instead of using
cogl_rectangle.
For small runs of text like icon labels, we can get better performance
going through the Cogl journal since text may then be batched together
with other geometry.
For larger runs of text though we still use VBOs since the cost of logging
the quads becomes too expensive, including the software transform which
isn't at all optimized at this point. VBOs also have the further advantage
of avoiding repeated validation of vertices by the driver and repeated
mapping of data into the GPU so long as the text doesn't change.
Currently the threshold is 100 vertices/25 quads. This number was plucked
out of thin air and should be tuned later.
With this change I see ~180% fps improvment for test-text. (x61s + i965 +
Mesa 7.6-devel)
We were missing the simplest test of all: are the two CoglHandles equal and
are the flush option flags for each material equal? This should improve
batching for some common cases.
Whenever we modify a material we call _cogl_material_pre_change_notify which
checks to see if the material is referenced by the journal and if so flushes
if before we modify the material.
Since the journal logs material colors directly into a vertex array (to
avoid us repeatedly calling glColor) then we know we never need to flush
the journal when material colors change.
Since most Clutter actors aren't much more than textured quads; flushing the
journal typically involves lots of 'change modelview; draw quad' sequences.
The amount of overhead involved in uploading a new modelview and queuing
that primitive is huge in comparison to simply transforming 4 vertices by
the current modelview when logging quads. (Note if your GPU supports HW
vertex transform, then it still does the projective and viewport transforms)
At the same time a --cogl-debug=disable-software-transform option has been
added for comparison and debugging.
This change allows typical pick scenes to be batched into a single draw call
and I'm seeing test-pick run over 200% faster with this. (i965 + Mesa
7.6-devel)
Enabling this option makes Cogl trace how the journal is managing to batch
your rectangles. The journal staggers how it emmits state to the GL driver
and the batches will normally get smaller for each stage, but ideally you
don't want to be in a situation where Cogl is only able to draw one quad per
modelview change and draw call.
E.g. this is a fairly ideal example:
BATCHING: journal len = 101
BATCHING: vbo offset batch len = 101
BATCHING: material batch len = 101
BATCHING: modelview batch len = 101
This isn't:
BATCHING: journal len = 1
BATCHING: vbo offset batch len = 1
BATCHING: material batch len = 1
BATCHING: modelview batch len = 1
BATCHING: journal len = 1
BATCHING: vbo offset batch len = 1
BATCHING: material batch len = 1
BATCHING: modelview batch len = 1
<repeat>
When this option is used Cogl will print a trace of all quads that get
logged into the journal, and a trace of quads as they get flushed.
If you are seeing a bug with the geometry being drawn by Cogl this may give
some clues by letting you sanity check the numbers being logged vs the
numbers being emitted.
For testing the VBO fallback paths it helps to be able to disable the
COGL_FEATURE_VBOS feature flag. When VBOs aren't available Cogl should use
client side malloc()'d buffers instead.
Previously we only used the Cogl matrix stack API for indirect contexts, but
it's too costly to keep on requesting modelview matrices from GL (for
logging in the journal) even for direct rendering.
I also experimented with a patch for mesa to improve performance and
discussed this with upstream, but we agreed to consider the GL matrix API
essentially deprecated. (For reference the GLES 2 and GL 3 specs have
removed the matrix APIs)
CoglColors shouldn't be compared using memcmp since they may contain
uninitialized padding bytes.
The prototype is also suitable for passing to g_hash_table_new as the
key_equal_func.
_cogl_pango_display_list_add_texture now uses this instead of memcmp.
We now put the color of materials into the vertex array used by the journal
instead of calling glColor() but the number of requests for the material
color were quite expensive so we have changed the material color to
internally be byte components instead of floats to avoid repeat conversions
and added _cogl_material_get_colorubv as a fast-path for the journal to
copy data into the vertex array.
To improve batching of geometry in the Cogl journal we need to avoid modifying
materials midscene.
Currently cogl_set_source_color and cogl_set_source_texture simply modify a
single shared material. In the future we can improve this so they use a pool
of materials that gets recycled as the journal is flushed, but for now we
give all ClutterRectangles their own private materials for painting with.
Currently cogl_set_source_color uses a single shared material which means
each actor that uses it causes the journal to flush if the color changes.
Until we improve cogl_set_source_color to use a pool of materials that can
be recycled as the journal is flushed we avoid mid-scene material changes by
giving all actors a private material instead.
The number of material layers enabled when logging a quad in the journal
determines the stride of the corresponding vertex data (since we need a set
of texture coordinates for each layer.) By padding data in the case where we
have only one layer we can avoid a change in stride if we are mixing single
and double layer primitives in a scene (e.g. relevent for a composite
manager that may use 2 layers for all shaped windows) Avoiding stride
changes means we can minimize calls to gl{Vertex,Color}Pointer when flushing
the journal.
Since we need to update the texcoord pointers when the actual number of
layers changes, this adds another batch_and_call() stage to deal with
glTexCoordPointer and enabling/disabling the client arrays.
Previously the journal was always flushed at the end of
_cogl_rectangles_with_multitexture_coords, (i.e. the end of any
cogl_rectangle* calls) but now we have broadened the potential for batching
geometry. In ideal circumstances we will only flush once per scene.
In summary the journal works like this:
When you use any of the cogl_rectangle* APIs then nothing is emitted to the
GPU at this point, we just log one or more quads into the journal. A
journal entry consists of the quad coordinates, an associated material
reference, and a modelview matrix. Ideally the journal only gets flushed
once at the end of a scene, but in fact there are things to consider that
may cause unwanted flushing, including:
- modifying materials mid-scene
This is because each quad in the journal has an associated material
reference (i.e. not copy), so if you try and modify a material that is
already referenced in the journal we force a flush first)
NOTE: For now this means you should avoid using cogl_set_source_color()
since that currently uses a single shared material. Later we
should change it to use a pool of materials that is recycled
when the journal is flushed.
- modifying any state that isn't currently logged, such as depth, fog and
backface culling enables.
The first thing that happens when flushing, is to upload all the vertex data
associated with the journal into a single VBO.
We then go through a process of splitting up the journal into batches that
have compatible state so they can be emitted to the GPU together. This is
currently broken up into 3 levels so we can stagger the state changes:
1) we break the journal up according to changes in the number of material layers
associated with logged quads. The number of layers in a material determines
the stride of the associated vertices, so we have to update our vertex
array offsets at this level. (i.e. calling gl{Vertex,Color},Pointer etc)
2) we further split batches up according to material compatability. (e.g.
materials with different textures) We flush material state at this level.
3) Finally we split batches up according to modelview changes. At this level
we update the modelview matrix and actually emit the actual draw command.
This commit is largely about putting the initial design in-place; this will be
followed by other changes that take advantage of the extended batching.
Removed the G_PARAM_CONSTRUCT from the property registration of
"load-async" and "load-data-async". It made it impossible to use only
load-data-async, as the async loading state would be unset when
load-async got set it's default FALSE value.
Use signed integers while combining window space clip rectangles, so we avoid
arithmatic errors later resulting in glScissor getting negative width and
height arguments.
Previously this was RGBA_8888. It souldn't really make a difference but for
consistency we expect almost all textures in use to have an internaly
premultiplied pixel format.
_cogl_texture_download_from_gl needs to create transient CoglBitmaps when
downloading sliced textures from GL, and then copies these as subregions
into the final target_bitmap. _cogl_texture_download_from_gl also supports
target_bitmaps with a different format to the source CoglTexture being
downloaded.
The problem was that in the case of slice textures we were always looking
at the format of the CoglTexture, not of the target_bitmap when setting
up the transient slice bitmap.
To allow for flushing of batched geometry within Cogl we can't support users
directly calling glReadPixels. glReadPixels is also awkward, not least
because it returns upside down image data.
All the unit tests have been swithed over and clutter_stage_read_pixels now
sits on top of this too.
I've found this is something I do quite often when debugging rendering
problems since its a simple way to wipe out lots of geometry and removes a
lot of unpredictable noise when logging geometry passing through the Cogl
journal.
We were calculating our vertex stride and allocating our vertex array
differently depending on whether the user passed TRUE for use_color or not.
The problem was that we were always writting color data to the array
regardless of use_color.
There was also a bug with _cogl_texture_sliced_polygon in that it was
writing byte color components but we were expecting float components. We
now use byte components in _cogl_multitexture_unsliced_polygon too and pass
GL_UNSIGNED_BYTE to glColorPointer.
Cogl already add similar defines but with the CLUTTER namespace
(CLUTTER_COGL_HAS_GL and CLUTTER_COGL_HAS_GLES). Let's just add two
similar defines with the COGL namespace. Removing the CLUTTER_COGL ones
could break applications silently for no real good reason.
While grepping through the public headers looking for invalid use of
private HAVE_* defines, I stumbled upon two out of sync comments. Yes
it's a very minor trivial change.
JSON-GLib provides simple accessors for basic types so that we
can avoid getting the JsonNode out of a complex type. This makes
the code simpler to understand.
Currently, Clutter depends on the internal copy of JSON-GLib for
the ClutterScript parser. This is done to allow building Clutter
on platforms that do not have the library installed on the system.
Just like we use the internal PNG/JPEG loader as a fallback in
case we don't have GdkPixbuf or CoreGraphics available, we should
use the internal copy of JSON-GLib only in case the system copy
is not present.
The change is simply to move the default for the --with-json
configure switch from "internal" to "check".
In order to allow stricter compliance, a third setting should
be present: "system", which fails if the system copy is not
available.
We should also change the introspection generation to avoid
breaking in case we require the installed Json-1.0.gir instead
of the generated ClutterJson.gir
The clutter_actor_pick() function just emits the ::pick signal
on the actor. Nobody should be using it, since the paint() method
is already context sensitive and will result in a ::pick emission
by itself. The clutter_actor_pick() is just confusing things.
In the int-to-float switch for actor properties, the ::size-change signal
was moved to use floats instead of integers. Sub-pixel precision for image
size is meaningless, though, so we should revert it back to ints.
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1659
Instead of using a specific function to check whether the X
server supports the XInput extension we can use the generic
Xlib function XQueryExtension(). This cuts down the extra
checks inside the configure.ac and simplifies the code inside
clutter_x11_register_xinput().
Currently, XInput support requires a function call. In order to
make it easier for people to test it, we can also add a command
line switch that moves the pointer device detection and handling
to XInput. This should ensure that, at least for people building
Clutter with --enable-xinput, applications can be easily migrated
and regressions can be caught.
The StageManager singleton instance is already kept around
by the clutter_stage_manager_get_default() function; there is
no need to have it inside the main Clutter context as well.
Instead of using _clutter_context_get_default() and checking the
is_initialized flag, we should use the newly added private function
that does not cause side effects, especially for functions that have
to be called before any other Clutter function.
The _clutter_context_get_default() function will automatically
create the main Clutter context; if we just want to check whether
Clutter has been initialized this will complicate matters, by
requiring a call to g_type_init() inside the client code.
Instead, we should simply provide an internal API that checks
whether the main Clutter context exists and if it has been
initialized, without any side effect.
The input device API is split halfway thorugh the backends in a very
weird way. The data structures are private, as they should, but most
of the information should be available in the main API since it's
generic enough.
The device type enumeration, for instance, should be common across
every backend; the accessors for device type and id should live in the
core API. The internal API should always use ClutterInputDevice and
not the private X11 implementation when dealing with public structures
like ClutterEvent.
By adding accessors for the device type and id, and by moving the
device type enumeration into the core API we can cut down the amount
of symbols private and/or visible only to the X11 backends; this way
when other backends start implementing multi-pointer support we can
share the same API across the code.
HAVE_COGL_GLES2 is defined in config.h through the configure script and
should not be used in public headers.
The patch makes configure generate the right define that can be used
later in the header.
The clutter_context_get_default() function is private, but shared
across Clutter. For this reason, it should be prefixed by '_' so
that the symbol is hidden from the shared object.
ActorBox should have methods for easily extracting the X and Y
coordinates of the origin, and the width and height separately.
These methods will make it easier for high-level language bindings
to manipulate ActorBox instances and avoid the Geometry type.
The Vertex and ActorBox boxed types are meant to be used across
the API, but are fairly difficult to bind. Their memory management
is also unclear, and has to go through the indirection of
g_boxed_copy() and g_boxed_free().
Cairo stores the image data in ARGB native byte order so we need to
upload this as BGRA on little endian architectures and ARGB on big
endian. ClutterTexture doesn't currently expose any flags to describe
ARGB format so until we can fix the Clutter API it now uses the Cogl
API directly.
In order to chain up animations using clutter_actor_animate() and
friends you have to use an idle handler that guarantees that the
main loop spins at least once after the animation pointer has been
detached from the actor.
This has several drawbacks, first and foremost the fact that the
slice of the main loop for the idle handler might be starved by
other operations, like redrawing. This inevitably leads to tricks
with priorities and the like, contributing to the overall complexity.
Instead, we should guarantee that the animation instance created by
clutter_actor_animate() is valid for the ::completed signal until
it reaches its default handler; after that, the animation is detached
from the actor and destroyed. This means that it's possible to
create a new animation after the first is complete by simply using
g_signal_connect_after().
This unfortunately makes it impossible to keep a reference to the
animation pointer attached to the actor by using g_object_ref(); a
way to "fix" this would be to have a clutter_animation_attach()
and a clutter_animation_detach() pair of methods that allow attaching
any animation to an actor. This might overcomplicate what it is
the simple animation API, though, so it's currently not implemented
and left for future versions.
The test-easing interactive demo has been modified to show how
the animation queuing works by adding a command line switch that
recenters the animated actor once the first animation has ended.
Continuing in the tradition on making clutter_actor_animate() the
next g_object_connect(), here's the addition of the signal-after::
and signal-swapped:: modifiers for the automagic signal connection
arguments.
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1646
In order to be ready for the next major version of GLib we need to
disable single header inclusion by using the G_DISABLE_SINGLE_INCLUDES
define in the build process.
We can't short-circuit the emission of ::queue-redraw for not-visible
actors, since ClutterClone uses that signal to know when things need
to be redrawn.
Calling clutter_actor_queue_redraw() out of clutter_actor_real_map() /
clutter_actor_real_unmap() was causing the flag state to get set
incorrectly from _clutter_actor_set_enable_paint_unmapped(), because
a paint queueing a redraw was not expected.
Moving queuing the redraw to clutter_actor_hide()/show() fixes this, and
also fixes a problem where showing a child of a cloned actor wouldn't
cause the clone to be repainted.
http://bugzilla.openedhand.com/show_bug.cgi?id=1484
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
If we have an not-visible texture pixmap, we need to:
- Still update it if it is realized, since it won't be
updated when shown. And it might be also be cloned.
- Queue a redraw if even if not visible, since it
it might be cloned.
http://bugzilla.openedhand.com/show_bug.cgi?id=1647
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
My patch to choose a premultiplied format when the user gives
COGL_PIXEL_FORMAT_ANY for the internal_format broke the case where the data
in question doesn't have and alpha channel.
This was accidentally missed when merging the premultiplication branch
since I merged a local version of the branch that missed this commit.
Although the underlying materials should allow layers with INVALID_HANDLES
it shouldn't be necissary to expose that via cogl_set_source_texture() and
it's easier to resolve a warning/crash here than odd artefacts/crashes later
in the pipeline.
Merge branch 'premultiplication'
[cogl-texture docs] Improves the documentation of the internal_format args
[test-premult] Adds a unit test for texture upload premultiplication semantics
[fog] Document that fogging only works with opaque or unmultipled colors
[test-blend-strings] Explicitly request RGBA_888 tex format for test textures
[premultiplication] Be more conservative with what data gets premultiplied
[bitmap] Fixes _cogl_bitmap_fallback_unpremult
[cogl-bitmap] Fix minor copy and paste error in _cogl_bitmap_fallback_premult
Avoid unnecesary unpremultiplication when saving to local data
Don't unpremultiply Cairo data
Default to a blend function that expects premultiplied colors
Implement premultiplication for CoglBitmap
Use correct texture format for pixmap textures and FBO's
Add cogl_color_premultiply()
Clarifies that if you give COGL_PIXEL_FORMAT_ANY as the internal format for
cogl_texture_new_from_file or cogl_texture_new_from_data then Cogl will
choose a premultiplied internal format.
The fixed function fogging provided by OpenGL only works with unmultiplied
colors (or if the color has an alpha of 1.0) so since we now premultiply
textures and colors by default a note to this affect has been added to
clutter_stage_set_fog and cogl_set_fog.
test-depth.c no longer uses clutter_stage_set_fog for this reason.
In the future when we can depend on fragment shaders we should also be
able to support fogging of premultiplied primitives.
We don't want to force texture data to be premultipled if the user
explicitly specifies a non premultiplied internal_format such as
COGL_PIXEL_FORMAT_RGBA_8888. So now Cogl will only automatically
premultiply data when COGL_PIXEL_FORMAT_ANY is given for the
internal_format, or a premultiplied internal format such as
COGL_PIXEL_FORMAT_RGBA_8888_PRE is requested but non-premultiplied source
data is given.
This approach is consistent with OpenVG image formats which have already
influenced Cogl's pixel format semantics.
The _cogl_unpremult_alpha_{first,last} functions which
_cogl_bitmap_fallback_unpremult depends on were incorrectly casting each
of the byte components of a texel to a gulong and performing shifts as
if it were dealing with the whole texel.
It now just uses array indexing to access the byte components without
needing to cast or manually shift any bits around.
Even though we used to depend on unpremult whenever we used a
ClutterCairoTexture, clutter_cairo_texture_context_destroy had it's own
unpremult code which worked which is why this bug wouldn't have been noticed
before.
Now that we typically have premultiplied data stored in Cogl
textures, when fetching a texture into local data for temporary
storage, use a premultiplied format to avoid an unpremultiply/
premultiply roundtrip.
http://bugzilla.openedhand.com/show_bug.cgi?id=1406
Signed-off-by: Robert Bragg <robert@linux.intel.com>
Instead of unpremultiplying the Cairo data, pass it directly to Cogl
in premultiplied form; we now *prefer* premultiplied data.
http://bugzilla.openedhand.com/show_bug.cgi?id=1406
Signed-off-by: Robert Bragg <robert@linux.intel.com>
Many operations, like mixing two textures together or alpha-blending
onto a destination with alpha, are done most logically if texture data
is in premultiplied form. We also have many sources of premultiplied
texture data, like X pixmaps, FBOs, cairo surfaces. Rather than trying
to work with two different types of texture data, simplify things by
always premultiplying texture data before uploading to GL.
Because the default blend function is changed to accommodate this,
uses of pure-color CoglMaterial need to be adapted to add
premultiplication.
gl/cogl-texture.c gles/cogl-texture.c: Always premultiply
non-premultiplied texture data before uploading to GL.
cogl-material.c cogl-material.h: Switch the default blend functions
to ONE, ONE_MINUS_SRC_ALPHA so they work correctly with premultiplied
data.
cogl.c: Make cogl_set_source_color() premultiply the color.
cogl.h.in color-material.h: Add some documentation about
premultiplication and its interaction with color values.
cogl-pango-render.c clutter-texture.c tests/interactive/test-cogl-offscreen.c:
Use premultiplied colors.
http://bugzilla.openedhand.com/show_bug.cgi?id=1406
Signed-off-by: Robert Bragg <robert@linux.intel.com>
cogl-bitmap.c cogl-bitmap-pixbuf.c cogl-bitmap-fallback.c cogl-bitmap-private.h:
Add _cogl_bitmap_can_premult(), _cogl_bitmap_premult() and implement
a reasonably fast implementation in the "fallback" code.
http://bugzilla.openedhand.com/show_bug.cgi?id=1406
Signed-off-by: Robert Bragg <robert@linux.intel.com>
RGBA data in X pixmaps and in FBOs is already premultiplied; use
the right format when creating cogl textures.
http://bugzilla.openedhand.com/show_bug.cgi?id=1406
Signed-off-by: Robert Bragg <robert@linux.intel.com>
Since commit d743aeaa updated the internal copy of JSON-GLib and
added a new private header file, we need to fix the build to avoid
a distcheck failure.
Merge branch 'master-clock-updates'
* master-clock-updates: (22 commits)
Change the paint forcing on the Text cache text
[timelines] Improve marker hit check and don't fudge the delta
Revert "[timeline] Don't clamp the elapsed time when a looping tl reaches the end"
[tests] Don't add a newline to the end of g_test_message calls
[test-timeline] Add a marker at the beginning of the timeline
[timeline] Don't clamp the elapsed time when a looping tl reaches the end
[master-clock] Throttle if no redraw was performed
[docs] Update Clutter's API reference
Force a paint instead of calling clutter_redraw()
Fix clutter_redraw() to match the redraw cycle
Run the repaint functions inside the redraw cycle
Remove useless manual timeline ticking
Move elapsed-time calculations into ClutterTimeline
Limit the frame rate when not syncing to VBLANK
Decrease the main-loop priority of the frame cycle
Avoid motion-compression in test-picking test
Compress events as part of the frame cycle
Remove stage update idle and do updates from the master clock
Call g_main_context_wakeup() when we start running timelines
Remove unused msecs_delta member
...
Markers added at the start of the timeline need to be special cased
because we effectively check whether the marker time is greater than
the old frame time or less than or equal to the new frame time but we
never actually emit a frame for the start of the timeline.
If the timeline is looping then it adjusts the position to interpolate
a wrapped around position. However we do not emit a new frame after
setting this position so we need to check for markers again there.
clutter_timeline_get_delta should return the actual wall clock time
between emissions of the new-frame signal - even if the elapsed time
has been fudged at the end of the timeline. To make this work it no
longer directly manipulates priv->msecs_delta but instead passes a
delta value to check_markers.
Someone might hide the previously focused actor in the focus-out signal
handler and as key focus still appears to be on that actor we'd get
re-entrant call and get glib critical from g_object_weak_unref
This changes clutter_stage_get_key_focus() to return stage/NULL during
focus-out signal emission. It used to be the actor focus-out was being
emitted on.
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1547
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
This reverts commit 9c5663d671.
The patch was causing problems for applications that expect the
elapsed_time to be at either end of the timeline when the completed
signal is fired. For example test-behave swaps the direction of the
timeline in the completed handler but if the time has overflowed the
end then the timeline would only take a short time to get back the
beginning. This caused the animation to just vibrate around the
beginning.
The new-frame signal of a timeline was previously guaranteed to be
emitted with the elapsed_time set to the end before it emits the
completed signal. This doesn't necessarily make sense for looping
timelines because it would cause the elapsed time to be clamped to a
slightly off value whenever the timeline restarts. This patch makes it
perform the wrap around before emitting the new-frame signal so that
the elapsed time always corresponds to the time elapsed since the
timeline was started.
Additionally it no longer fudges the msecs_delta property to make the
marker check work so clutter_timeline_get_delta will always return the
wall clock time since the last frame.
A flag in the master clock is now set whenever the dispatch caused an
actual redraw of a stage. If this flag is not set during the prepare
and check functions then it will resort to limiting the redraw
attempts to the default frame rate as if vblank syncing was
disabled. Otherwise if a timeline is running that does not cause the
scene to change then it would busy-wait with 100% CPU until the next
frame.
This fix was suggested by Owen Taylor in:
http://bugzilla.openedhand.com/show_bug.cgi?id=1637
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
We do not need the whole redraw machinery inside
clutter_stage_read_pixels(): instead, we want Clutter to drop everything,
paint and call glReadPixels() with the current buffer.
The clutter_redraw() function is used by embedding toolkits to
force a redraw on a stage. Since everything is performed by
toggling a flag inside the Stage itself and then letting the
master clock advance, we need a ClutterStage method to ensure
that we start the master clock and redraw.
Instead of calculating a delta in the master clock, and passing that
into each timeline, make each timeline individually responsible for
remembering the last time and computing the delta.
This:
- Fixes a problem where we could spin infinitely processing
timeline-only frames with < 1msec differences.
- Makes timelines consistently start timing on the first frame;
instead of doing different things for the first started timeline
and other timelines.
- Improves accuracy of elapsed time computations by avoiding
accumulating microsecond => millisecond truncation errors.
http://bugzilla.openedhand.com/show_bug.cgi?id=1637
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
clutter-master-clock.c clutter-master-clock.h: When the
SYNC_TO_VBLANK feature is not available, wait for 1/frame_rate
seconds since the start of the last frame before drawing the next
frame. Add _clutter_master_clock_start_running() to abstract
the usage of g_main_context_wakeup()
clutter-stage.c: Add _clutter_master_clock_start_running()
clutter-main.c: Update docs for clutter_set_default_frame_rate()
clutter_get_default_frame_rate() to no longer talk about timeline
frame rates.
test-text-perf.c test-text.c: Set a frame rate of 1000fps so that
frame-rate limiting doesn't affect the result.
http://bugzilla.openedhand.com/show_bug.cgi?id=1637
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Change CLUTTER_PRIORITY_REDRAW to be lower than the GTK+ resize
and relayout priorities to avoid starving GTK+ when run in the
same process as clutter.
Remove the unused CLUTTER_PRIORITY_TIMELINE
http://bugzilla.openedhand.com/show_bug.cgi?id=1637
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Instead of trying to guess about which motion events are
extraneous, queue up all events until we process a frame.
This allows us to look ahead and reliably compress consecutive
sequence of motion events.
clutter-main.c: Feed received events to the stage for queueing.
Remove old compression code. Remove clutter_get_motion_events_frequency()
clutter_set_motion_events_frequency()
clutter-stage.c: Keep a queue of pending events.
clutter-master-clock.c: Add processng of queued events to the
clock source dispatch function.
http://bugzilla.openedhand.com/show_bug.cgi?id=1637
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
When a redraw is queued on a stage, simply set a flag; then in
the check/prepare functions of the master clock source, check
for stages that need redrawing.
This avoids the complexity of having multiple competing sources
at the same priority and makes the update ordering more reliable and
understandable.
http://bugzilla.openedhand.com/show_bug.cgi?id=1637
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
If a timeline is added from a different thread, we need to
call g_main_context_wakeup() to wake the main thread up to
start updating the timeline.
http://bugzilla.openedhand.com/show_bug.cgi?id=1637
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Instead of keeping a list of all timelines, and connecting to
signals and weak notifies, simply keep a list of running timelines;
this greatly simplifies both the book-keeping, and also determining
if there are any running timelines.
http://bugzilla.openedhand.com/show_bug.cgi?id=1637
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Remove code to advance the master clock after drawing a frame; if
there are any running timelines the master clock will do another
frame by itself, and the clock will be advanced before running
that frame.
With this change, there is no point in queueing an extra frame
redraw after completing a timeline, since we are always advancing
the timeline *before* redrawing, so remove that code as well.
(This does mean that calling clutter_timeline_stop() won't implicitly
cause the stage to be redrawn; this doesn't seem like something
an app should rely on in any case.)
http://bugzilla.openedhand.com/show_bug.cgi?id=1637
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The clutter_stage_fullscreen() and clutter_stage_unfullscreen() are
a GDK-ism. The underlying implementation is already using an accessor
with a boolean parameter.
This should take the amount of collisions between properties, methods
and signals to zero.
The :fullscreen property is very much confusing as it is implemented.
It can be written to a value, but the whole process might fail. If we
set:
g_object_set (stage, "fullscreen", TRUE, NULL);
and the fullscreen process fails or it is not implemented, the value
will be reset to FALSE (if we're lucky) or left TRUE (most of the
times).
The writability is just a shorthand for invoking clutter_stage_fullscreen()
or clutter_stage_unfullscreen() depending on a boolean value without
using an if.
The :fullscreen property also greatly confuses high level languages,
since the same symbol is used:
- for a method name (Clutter.Stage.fullscreen())
- for a property name (Clutter.Stage.fullscreen)
- for a signal (Clutter.Stage::fullscreen)
For these reasons, the :fullscreen should be renamed to :fullscreen-set
and be read-only. Implementations of the Stage should only emit the
StageState event to change from normal to fullscreen, and the Stage
will automatically update the value of the property and emit a notify
signal for it.
There have been changes in JSON-GLib upstream to clean up the
data structures, and facilitate introspection.
We still not use the updated JsonParser with the (private) JsonScanner
code, since it's a fork of GLib's GScanner.
Otherwise if there is an error before the slices are created it will
try to free the first_pixels array and crash.
It now also checks whether first_pixels has been created before using
it to update the mipmaps. This should only happen for
cogl_texture_new_from_foreign and doesn't matter if the FBO extension
is available. It would be better in this case to fetch the first pixel
using glGetTexImage as Owen mentioned in the last commit.
tex->first_pixels was never set for foreign textures, leading
to a crash when the texture object is freed.
As a quick fix, simply set to NULL. A more complete fix would
require remembering if we had ever seen the first pixel uploaded,
and if not, doing a glReadPixel to get it before triggering the
mipmap update.
http://bugzilla.openedhand.com/show_bug.cgi?id=1645
Signed-off-by: Neil Roberts <neil@linux.intel.com>
It's very common that there's no reasonable fallback to do if the
blend or combine string you set isn't supported. So, rather than
requiring everybody to pass in a GError purely to catch syntax erorrs,
automatically g_warning() if a parse error is encountered and @error
is NULL.
http://bugzilla.openedhand.com/show_bug.cgi?id=1642
Signed-off-by: Robert Bragg <robert@linux.intel.com>
When we complete a timeline, we clamp the elapsed_time variable
to the range of the timeline. We need to adjust msecs_delta so that
when we check for hit markers we have the correct interval.
http://bugzilla.openedhand.com/show_bug.cgi?id=1641
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The Animation should be referenced during the notification of the
alpha value, since the callback is invoked depending on the Alpha
and it won't vivify the Animation instance for us.
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1537
ClutterEvent is not really gobject-introspection friendly because
of the whole discriminated union thing. In particular, if you get
a ClutterEvent in a signal handler, you probably can't access the
event-type-specific fields, and you probably can't call methods
like clutter_key_event_symbol() either, because you can't cast the
ClutterEvent to a ClutterKeyEvent.
The cleanest solution is to turn every accessor into ClutterEvent
methods, accepting a ClutterEvent* and internally checking the event
type.
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1585
According to clutter_texture_set_cogl_texture you should unref the handle as
the texture takes its own.
Signed-off-by: Robert Bragg <robert@linux.intel.com>
The OpenGL spec states that if you create a pixmap using glXCreatePixmap you
should use glXDestroyPixmap to destroy it.
Signed-off-by: Robert Bragg <robert@linux.intel.com>
Setting the pixmap for an unrealized ClutterGLXTexturePixmap should
not cause it to be realized, and certainly shouldn't cause the the
REALIZED flag to be set without using clutter_actor_realize().
This patch uses the simple approach that;
- pixmap changes on an unrealized ClutterGLXTexturePixmap
are ignored
- when the ClutterGLXTexturePixmap is realized, we then create
the GLXPixmap and the corresponding texture.
The call to clutter_glx_texture_pixmap_update_area() is moved
from create_cogl_texture() to
clutter_glx_texture_pixmap_create_glx_pixmap() since
create_cogl_texture() is only called from one place, and updating
the area is really something we do *after* creating the texture,
not part of creating the texture.
clutter_glx_texture_pixmap_create_glx_pixmap() is reorganized a
bit to avoid debug-logging confusingly if it's called before a pixmap
has been set, and for readability.
http://bugzilla.openedhand.com/show_bug.cgi?id=1635
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
An implementaton of realize() never needs to set the
CLUTTER_ACTOR_REALIZED flag, though it can unset the flag if
things fail unexpectedly. (Previously, stage backend implementations
had to do this since clutter_actor_realize() wasn't used; this
is no longer the case.)
http://bugzilla.openedhand.com/show_bug.cgi?id=1634
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Due to the accumulation of floating point errors, natural_width
and min_width can diverge significantly even if the math for
computing them is correct. So just clamp natural_width to
min_width instead of warning about it.
http://bugzilla.openedhand.com/show_bug.cgi?id=1632
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
If we use float temporaries when computing the bounds of
a group, then, depending on what variables are kept in registers
and what stored on the stack, the accumulated difference between
natural_width and min_width can be more than FLOAT_EPSILON.
Using double temporaries will eliminate the difference in most
cases, or, very rarely, reduce it to a last-bit error.
http://bugzilla.openedhand.com/show_bug.cgi?id=1632
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
If we are cloning an source actor with an unmapped parent, then when
we temporarily map the source actor:
- We need to skip the check that a mapped actor has a mapped
parent.
- We need to realize the actor's parents before mapping it,
or we'll get an assertion failure in clutter_actor_update_map_state()
because an actor with an unmapped parent is !may_be_realized.
http://bugzilla.openedhand.com/show_bug.cgi?id=1633
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Setting the stage size using clutter_actor_set_size() is almost always
wrong: the X11 stage implementation should save the size and queue a
relayout -- like it does when receiving a ConfigureNotify. The same
should happen when setting it to be full screen.
Since we build the Cogl GIR inside /clutter/cogl we should be looking
there when building the Clutter GIR. Otherwise g-ir-scanner will look
inside the gir directory -- and if you never built Clutter before it
will error out.
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1638
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The load-finished signal has a GError* argument which is meant to
signify whether the loading was successful. However many of the
places in ClutterTexture that emit this signal directly pass their
'error' variable which is a GError** and will be NULL or not
completely independently of whether there was an error. If the
argument was dereferenced it would probably crash.
The test-texture-async interactive test case should also verify
that the ::load-finished signal is correctly emitted.
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1622
Correctly apply De Morgan's laws to the short-circuit test in
clutter_timeline_pause(); it was short-circuiting always and
never actually pausing.
http://bugzilla.openedhand.com/show_bug.cgi?id=1629
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The commit 2c95b378 prevents clutter_animation_setup_property from being
called with fixed:: property names. This patch adds a additional
parameter "is_fixed" to clutter_animation_setup_property instead of
searching for "fixed::" in property_name.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
It should be possible render a single PangoLayout with different
colors without recalculating the layout. This was not working because
the color used at the first edit was being stored in the display
list. This broke changing the opacity on a ClutterText.
Now each node in the display list has a 'color override' flag which
marks whether it should use the base color or not. The base color is
now passed in from _cogl_pango_display_list_render_texture. The alpha
value is always taken from the base color.
The clutter_redraw() function is used by libraries embedding
Clutter inside another toolkit, instead of queueing a redraw
on the embedded stage. This means that clutter_redraw() should
perform the same sequence of actions done by the redraw idle
callback.
Clutter short-circuits painting when an actor's opacity is
zero. However if the actor is being painted from a ClutterClone then
it will be painted using the clone's opacity instead so the test was
broken.
* 1.0-integration: (138 commits)
[x11] Disable XInput by default
[xinput] Invert the XI extension version check
[cogl-primitives] Fix an unused variable warning when building GLES
[clutter-stage-egl] Pass -1,-1 to clutter_stage_x11_fix_window_size
Update the GLES backend to have the layer filters in the material
[gles/cogl-shader] Add a missing semicolon
[cogl] Move the texture filters to be a property of the material layer
[text] Fix Pango unit to pixels conversion
[actor] Force unrealization on destroy only for non-toplevels
[x11] Rework map/unmap and resizing
[xinput] Check for the XInput entry points
[units] Validate units against the ParamSpec
[actor] Add the ::allocation-changed signal
[actor] Use flags to control allocations
[units] Rework Units into logical distance value
Remove a stray g_value_get_int()
Remove usage of Units and macros
[cogl-material] Allow setting a layer with an invalid texture handle
[timeline] Remove the concept of frames from timelines
[gles/cogl-shader] Fix parameter spec for cogl_shader_get_info_log
...
Conflicts:
configure.ac
The XInput support in Clutter is still using XI 1.x. This will never
work correctly, and we are all waiting for XInput 2 anyway. The changes
internally should be minimal, so we can leave everything in place, but
it's better to disable XInput support by default -- at least for the
time being.
The texture filters are now a property of the material layer rather
than the texture object. Whenever a texture is painted with a material
it sets the filters on all of the GL textures in the Cogl texture. The
filter is cached so that it won't be changed unnecessarily.
The automatic mipmap generation has changed so that the mipmaps are
only generated when the texture is painted instead of every time the
data changes. Changing the texture sets a flag to mark that the
mipmaps are dirty. This works better if the FBO extension is available
because we can use glGenerateMipmap. If the extension is not available
it will temporarily enable automatic mipmap generation and reupload
the first pixel of each slice. This requires tracking the data for the
first pixel.
The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with
COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to
auto-mipmapping. The mipmap generation is now effectively free if you
are not using a mipmap filter mode so you would only want to disable
it if you had some special reason to generate your own mipmaps.
ClutterTexture no longer has to store its own copy of the filter
mode. Instead it stores it in the material and the property is
directly set and read from that. This fixes problems with the filters
getting out of sync when a cogl handle is set on the texture
directly. It also avoids the mess of having to rerealize the texture
if the filter quality changes to HIGH because Cogl will take of
generating the mipmaps if needed.
The mapping and unmapping of the X11 stage implementation is
a bit bong. It's asynchronous, for starters, when it really
can avoid it by tracking the state internally.
The ordering of the map/unmap sequence is also broken with
respect to the resizing.
By tracking the state internally into StageX11 we can safely
remove the MapNotify and UnmapNotify X event handling.
In theory, we should use _NET_WM_STATE a lot more, and reuse
the X11 state flags for fullscreening as well.
Apparently, the XInput extension is using the same pkg-config
file ('xi') for both the 1.x and the 2.x API, so we need to
check for both the 1.x XGetExtensionVersion and the 2.x
XQueryInputVersion.
When declaring a property using ClutterParamSpecUnits we pass a
default type to limit the type of units we accept as valid values
for the property.
This means that we need to add the unit type check as part of the
validation process.
Sometimes it is useful to be able to track changes in the allocation
flags, like the absolute origin, inside children of a container.
Using the notify::allocation signal is not enough, in these cases, so
we need a specific signal that gives us both the allocation box and the
allocation flags.
Instead of passing a boolean value, the ::allocate virtual function
should use a bitmask and flags. This gives us room for expansion
without breaking API/ABI, and allows to encode more information to
the allocation process instead of just changes of absolute origin.
Units as they have been implemented since Clutter 0.4 have always been
misdefined as "logical distance unit", while they were just pixels with
fractionary bits.
Units should be reworked to be opaque structures to hold a value and
its unit type, that can be then converted into pixels when Clutter needs
to paint or compute size requisitions and perform allocations.
The previous API should be completely removed to avoid collisions, and
a new type:
ClutterUnits
should be added; the ability to install GObject properties using
ClutterUnits should be maintained.
It was previously possible to create a material layer with no texture
by setting some property on it such as the matrix. However it was not
possible to get back to that state without removing the layer and
recreating it. It is useful to be able to remove the texture to free
resources without forgetting the state of the layer so we can put a
different texture in later.
Timelines no longer work in terms of a frame rate and a number of
frames but instead just have a duration in milliseconds. This better
matches the working of the master clock where if any timelines are
running it will redraw as fast as possible rather than limiting to the
lowest rated timeline.
Most applications will just create animations and expect them to
finish in a certain amount of time without caring about how many
frames are drawn. If a frame is going to be drawn it might as well
update all of the animations to some fraction of the total animation
rather than rounding to the nearest whole frame.
The 'frame_num' parameter of the new-frame signal is now 'msecs' which
is a number of milliseconds progressed along the
timeline. Applications should use clutter_timeline_get_progress
instead of the frame number.
Markers can now only be attached at a time value. The position is
stored in milliseconds rather than at a frame number.
test-timeline-smoothness and test-timeline-dup-frames have been
removed because they no longer make sense.
The clutter_actor_map and unmap functions need to be called to
properly update the mapped state. This matches the changes to the X11
stage in 125bded8.
If the application code calls for destruction of an actor we need
to make sure that the actor is unrealized before running the dispose
sequence; otherwise, we might trigger an assertion failure on composite
actors.
The commit 762873e79e is completely
and utterly wrong and I should have never pushed it.
Serves me well for trying to work on three different branches and
on three different things.
Currently, the clock source spins a redraw every time there is at
least a timeline running. If the timelines were not advanced in
the previous frame, though, because their interval is larger than
the vblanking interval then this will lead to excessive redraws of
the scenegraph even if nothing has changed.
To avoid this a simple guard should be set by the MasterClock::advance
method in case no timeline was effectively advanced, and checked
before dispatching the stage redraws.
When creating a Cogl texture from a Cogl bitmap it would steal the
data by setting the bitmap_owner flag and clearing the data pointer
from the bitmap. The data would be freed by the time the
new_from_bitmap is finished. There is no reason to do this because the
data will be freed when the Cogl bitmap is unref'd and it is confusing
not to be able to reuse the bitmap for creating multiple textures.
clutter_color_from_string() only supported the "#rrggbbaa" format with
alpha channel, this patch adds support for "#rgba".
Colors in "#rrggbb" format were parsed manually, this is now left to
the pango color parsing fallback, since that's handling it just fine.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The cogl_shader_get_info_log() function is very inconvenient for
language bindings and for regular use, as it requires a static
buffer to be filled -- basically just providing a wrapper around
glGetInfoLogARB().
Since COGL aims to be a more convenient API than raw GL we should
just make cogl_shader_get_info_log() return an allocated string
with the GLSL compiler log.
Instead of using GL_TRIANGLES and uploading the indices every time, it
now uses GL_QUADS instead on OpenGL. Under GLES it still uses indices
but it uses the new cogl_vertex_buffer_indices_get_for_quads function
to avoid uploading the vertices every time.
This requires the _cogl_vertex_buffer_indices_pointer_from_handle
function to be exposed privately to the rest of Cogl.
The static_indices array has been removed from the Cogl context.
The GIR file for Clutter still contains symbols from COGL, even
though we provide a Cogl GIR as well. The Clutter GIR should
depend on the Cogl GIR instead.
All the underlying implementation and the public entry points have
been switched to floats; the only missing bits are the Actor properties
that deal with positioning and sizing.
This usually means a major pain when dealing with GValues and varargs
functions. While GValue will warn you when dealing with the wrong
conversions, varags will simply die an horrible (and hard to debug)
death via segfault. Nothing much to do here, except warn people in the
release notes and hope for the best.
The documentation for ClutterTexture's set_from_rgb_data() and
set_from_yuv_data() says:
Note: This function is likely to change in future versions.
This is not true, since they'll remain for the whole 1.x API cycle.
Now that CoglVertexBuffers support indices we can use them with GLES
to avoid duplicating vertices. Regular GL still uses GL_QUADS because
it is shown to still have a performance benefit over indices with the
Intel drivers.
This function can be used as an efficient way of drawing groups of
quads without using GL_QUADS. It generates a VBO containing the
indices needed to render using pairs of GL_TRIANGLES. The VBO is
globally cached so that it only needs to be uploaded whenever more
indices are requested than ever before.
The allocate_available_size() method is a convenience method in
the same spirit as allocate_preferred_size(). While the latter
will allocate the preferred size of an actor regardless of the
available size provided by the actor's parent -- and thus it's
suitable for simple fixed layout managers like ClutterGroup -- the
former will take into account the available size provided by the
parent and never allocate more than that; it is, thus, suitable
for simple fluid layout managers.
The cogl-enum-types.h file is created by glib-mkenums under
/clutter/cogl/common, and then copied in /clutter/cogl in order
to make the inclusion of that file work inside cogl.h.
Since we're copying it in a different location, the Makefile
for that location has to clean up the copy.
Notifications should be fired off from both the internal timeline and
the wrapping animation here, so notifiers should be frozen around these
property setters.
Signed-off-by: Jonas Bonn <jonas@southpole.se>
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Just a couple of final cleanups after the reimplementation of the
Animation model.
i) _set_mode does not need to set the timeline on the alpha
ii) freeze notifications around the setting of a new alpha
Signed-off-by: Jonas Bonn <jonas@southpole.se>
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The "started" signal is sent first after the timeline has been set to the
'running' state. For this reason, checking if the clock has any running
timelines running will always return true in the "started" signal handler:
the timeline that sent the signal is running.
What needs to be checked in the signal handler is if there are any
timelines running other than the one that emitted the ::started signal,
which we know is running anyway.
This prevents frames from being lost at the beginning of an animation when
a timeline is started after a quiescent period.
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1617
Signed-off-by: Jonas Bonn <jonas@southpole.se>
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
We avoid rebuilding cogl-enum-types.h and cogl-enum-types.c by
using a "guard" -- a stamp file that will block Makefile. Since
we need cogl-enum-types.h into /clutter/cogl as well for the
cogl.h include to work, if we copy the cogl-enum-types.h
unconditionally it will cause a rebuild of the whole COGL; which
will cause a full rebuild.
To solve this, we can copy the header file when generating it
under the stamp file.
The libclutter-cogl internal object should be the only dependency
for Clutter, since we are already copying it inside clutter/cogl
for the introspection scanner. For this reason, the backend-specific,
real internal object should be built with the backend encoded into
the file name, like libclutter-common. This makes the build output
a little bit more clear: instead of having two:
LINK libclutter-cogl-common.la
...
LINK libclutter-cogl.la
LINK libclutter-cogl.la
We'll have:
LINK libclutter-cogl-common.la
...
LINK libclutter-cogl-gl.la
LINK libclutter-cogl.la
Same applies for the GLES backend.
Just like we do with GObject types and G_DEFINE_TYPE, we should
use the g_once_init_enter/g_once_init_leave mechanism to make the
GType registration of enumeration types thread safe.
The setup_viewport() function should only be used by Clutter and
not by application code.
It can be emulated by changing the Stage size and perspective and
requeueing a redraw after calling clutter_stage_ensure_viewport().
The backface culling enabling function was split and renamed, just
like the depth testing one, so we need to add the macro to the
cogl-deprecated.h header.
Previously indices were tightly bound to a particular Cogl vertex buffer
but we would like to be able to share indices so now we have
cogl_vertex_buffer_indices_new () which returns a CoglHandle.
In particular we could like to have a shared set of indices for drawing
lists of quads that can be shared between the pango renderer and the
Cogl journal.
At the moment Cogl doesn't do much batching of quads so most of the time we
are flushing a single quad at a time. This patch simplifies how we submit
those quads to OpenGL by using glDrawArrays with GL_TRIANGLE_FAN mode
instead of sending indexed vertices using GL_TRIANGLES mode.
Note: I hope to follow up soon with changes that improve our batching and
also move the indices into a VBO so they don't need to be re-validated every
time we call glDrawElements.
To assist people porting code from 0.8, the cogl_texture_* functions that
have been replaced now have defines that give some hint as to how they
should be replaced.
cogl_enable_depth_test and cogl_enable_backface_culling have been renamed
and now have corresponding getters, the new functions are:
cogl_set_depth_test_enabled
cogl_get_depth_test_enabled
cogl_set_backface_culling_enabled
cogl_get_backface_culling_enabled
This adds cogl_matrix api for multiplying matrices either by a perspective
or ortho projective transform. The internal matrix stack and current-matrix
APIs also have corresponding support added.
New public API:
cogl_matrix_perspective
cogl_matrix_ortho
cogl_ortho
cogl_set_modelview_matrix
cogl_set_projection_matrix
cogl_create_context is dealt with internally when _cogl_get_default context
is called, and cogl_destroy_context is currently never called.
It might be nicer later to get an object back when creating a context so
Cogl can support multiple contexts, so these functions are being removed
from the API until we get a chance to address context management properly.
For now cogl_destroy_context is still exported as _cogl_destroy_context so
Clutter could at least install a library deinit handler to call it.
Originally cogl_vertex_buffer_add_indices let the user pass in their own unique
ID for the indices; now the Id is generated internally and returned to the
caller.
It's now possible to add arrays of indices to a Cogl vertex buffer and
they will be put into an OpenGL vertex buffer object. Since it's quite
common for index arrays to be static it saves the OpenGL driver from
having to validate them repeatedly.
This changes the cogl_vertex_buffer_draw_elements API: It's no longer
possible to provide a pointer to an index array at draw time. So
cogl_vertex_buffer_draw_elements now takes an indices identifier that
should correspond to an idendifier returned when calling
cogl_vertex_buffer_add_indices ()
This is being removed before we release Clutter 1.0 since the implementation
wasn't complete, and so we assume no one is using this yet. Util we have
someone with a good usecase, we can't pretend to support breaking out into
raw OpenGL.
There were a number of functions intended to support creating of new
primitives using materials, but at this point they aren't used outside of
Cogl so until someone has a usecase and we can get feedback on this
API, it's being removed before we release Clutter 1.0.
This removes the following API:
cogl_material_set_blend_factors
cogl_material_set_layer_combine_function
cogl_material_set_layer_combine_arg_src
cogl_material_set_layer_combine_arg_op
These were rather awkward to use, so since it's expected very few people are
using them at this point and it should be straight forward to switch over
to blend strings, the API is being removed before we release Clutter 1.0.
Setting up layer combine functions and blend modes is very awkward to do
programatically. This adds a parser for string based descriptions which are
more consise and readable.
E.g. a material layer combine function could now be given as:
"RGBA = ADD (TEXTURE[A], PREVIOUS[RGB])"
or
"RGB = REPLACE (PREVIOUS)"
"A = MODULATE (PREVIOUS, TEXTURE)"
The simple syntax and grammar are only designed to expose standard fixed
function hardware, more advanced combining must be done with shaders.
This includes standalone documentation of blend strings covering the aspects
that are common to blending and texture combining, and adds documentation
with examples specific to the new cogl_material_set_blend() and
cogl_material_layer_set_combine() functions.
Note: The hope is to remove the now redundant bits of the material API
before 1.0
After long deliberation, the Animation class handling of the
:mode, :duration and :loop properties, as well as the conditions
for creating the Alpha and Timeline instances, came out as far too
complicated for their own good.
This is a rework of the API/parameters matrix and behaviour:
- :mode accessors will create an Alpha, if needed
- :duration and :loop accessors will create an Alpha and a Timeline
if needed
- :alpha will set or unset the Alpha
- :timeline will set or unset the Timeline
Plus, more documentation on the Animation class itself.
Many thanks to Jonas Bonn <jonas@southpole.se> for the feedback
and the ideas.
The Animatable interface implementation will always have the computed
value applied, whilst the non-Animatable objects go through the
interval validation first to avoid incurring in assertions and
warnings.
The Animatable::animate_property() should also be able to validate the
property it's supposed to interpolate, and eventually discard it. This
requires adding a return value to the virtual function (and its wrapper
function).
The Animation code will then apply the computed value only if the
animate_property() returns TRUE -- unifying the code path with the
non-Animatable objects.
The Animation class should proxy the :mode, :duration and :loop
properties whenever possible, to avoid them going out of sync when
changed using the Alpha and Timeline instances directly.
Currently, if Timeline:duration is changed, querying Animation:duration
will yield the old value, but the animation itself (being driven by
the Timeline) will use the Timeline's :duration new value. This holds
for the :loop and :mode properties as well.
Instead, the getters for the Animation's :duration, :loop and
:mode properties should ask the relevant object -- if any. The
loop, duration and mode values inside AnimationPrivate should only
be used if no Timeline or no Alpha instances are available, or
when creating new instances.
The Animation should not directly manipulate a Timeline instance,
but it should defer to the Alpha all handling of the timeline.
This means that:
- set_duration() and set_loop() will either create a Timeline or
will set the :duration and :loop properties on the Timeline; if
the Timeline must be created, and no Alpha instance is available,
then a new Alpha instance will be created as well and the newly
create Timeline will be assigned to the Alpha
- if set_mode() on an Animation instance without an Alpha, the
Alpha will be created; a Timeline will also be created
- set_alpha() will replace the Alpha; if the new Alpha does not
have a Timeline associated then a Timeline will be created using
the current :duration and :loop properties of Animation; otherwise,
if the replaced Alpha had a timeline, the timeline will be
transferred to the new one
The CoglTexture constructors expose the "max-waste" argument for
controlling the maximum amount of wasted areas for slicing or,
if set to -1, disables slicing.
Slicing is really relevant only for large images that are never
repeated, so it's a useful feature only in controlled use cases.
Specifying the amount of wasted area is, on the other hand, just
a way to mess up this feature; 99% the times, you either pull this
number out of thin air, hoping it's right, or you try to do the
right thing and you choose the wrong number anyway.
Instead, we can use the CoglTextureFlags to control whether the
texture should not be sliced (useful for Clutter-GST and for the
texture-from-pixmap actors) and provide a reasonable value for
enabling the slicing ourself. At some point, we might even
provide a way to change the default at compile time or at run time,
for particular platforms.
Since max_waste is gone, the :tile-waste property of ClutterTexture
becomes read-only, and it proxies the cogl_texture_get_max_waste()
function.
Inside Clutter, the only cases where the max_waste argument was
not set to -1 are in the Pango glyph cache (which is a POT texture
anyway) and inside the test cases where we want to force slicing;
for the latter we can create larger textures that will be bigger than
the threshold we set.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Signed-off-by: Robert Bragg <robert@linux.intel.com>
Signed-off-by: Neil Roberts <neil@linux.intel.com>
Sometimes it is necessary for third party code to have a
function called during the redraw process, so that you can
update the scenegraph before it is painted.
If we are short-circuiting the paint when the opacity is zero we still
need to clear the queued_redraw flag otherwise it won't be possible to
queue another redraw of the actor until something else has caused a
paint first.
* master:
[cogl-vertex-buffer] Ensure the clip state before rendering
[test-text-perf] Small fix-ups
Add a test for text performance
[build] Ensure that cogl-debug is disabled by default
[build] The cogl GE macro wasn't passing an int according to the format string
Use the right internal format for GL_ARB_texture_rectangle
[actor_paint] Ensure painting is a NOP for actors with opacity = 0
Make backface culling work with vertex buffers
Before any rendering is done by Cogl it needs to ensure the clip stack
is set up correctly by calling cogl_clip_ensure. This was not being
done for the Cogl vertex buffer so it would still use the clip from
the previous render.
Now that everything is float, the marsharlling function of the
size-change signal should reflect that fact.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
GLES doesn't support GL_QUADS. This patch makes it use GL_TRIANGLES
instead in that case. Unfortunately this means submitting two extra
vertices per quad. It could be better to use indexed elements once
CoglVertexBuffers gains support for that.
When ClutterGLXTexturePixmap uses GL_ARB_texture_rectangle,
it needs to pass the right internal format (GL_RGB or GL_RGBA)
when it initializes the texture with glTexImage2D() or later
handling won't recognize the alpha channel.
http://bugzilla.openedhand.com/show_bug.cgi?id=1586
Signed-off-by: Robert Bragg <robert@linux.intel.com>
Since it is convenient to use geometry with an opacity of 0 for input only
purposes it's a worthwhile optimization to avoid submitting anything
for such actors while painting.
Backface culling is enabled as part of cogl_enable so the different
rendering functions in Cogl need to explicitly opt-in to have backface
culling enabled. Cogl vertex buffers should allow backface culling so
they should check whether it is enabled and then set the appropriate
cogl_enable flag.
In order to cope with the situation where an application renders with
a PangoLayout, makes some changes and then renders again with the same
layout, CoglPangoRenderer needs to detect that the changes have
occured so that it can recreate the display list. This is acheived by
keeping a reference to the first line of the layout. When the layout
is changed Pango will clear the layout pointer in the first line and
create a new line. So if the layout pointer in the line becomes NULL
then we know the layout has changed. This trick was suggested by
Behdad Esfahbod in this email:
http://mail.gnome.org/archives/gtk-i18n-list/2009-May/msg00019.html
When a position is given to cogl_pango_render_layout_subpixel it
translates the GL matrix by the coordinates. However it was not
dividing by PANGO_SCALE so the coordinates were completely wrong.
Most of the operations involving the texture's allocated area require
floats -- either for computations or for setting the geometry into
COGL. So it doesn't make any sense to use get_allocation_coords() and
cast everything to floats.
Currently, COGL depends on defining debug symbols by manually
modifying the source code. When it's done, it will forcefully
print stuff to the console.
Since COGL has also a pretty, runtime selectable debugging API
we might as well switch everything to it.
In order for this to happen, configure needs a new:
--enable-cogl-debug
command line switch; this will enable COGL debugging, the
CoglHandle debugging and will also turn on the error checking
for each GL operation.
The default setting for the COGL debug defines is off, since
it slows down the GL operations; enabling it for a particular
debug build is trivial, though.
COGL has a debug message system like Clutter's own. In parallel,
it also uses a coupld of #defines. Spread around there are also
calls to printf() instead to the more correct g_log* wrappers.
This commit tries to unify and clean up the macros and the
debug message handling inside COGL to be more consistent.
ClutterTexture has many properties that can only be accessed using
the GObject API. This is fairly inefficient and makes binding the
class overly complicated.
The Texture class should have accessor methods for all its properties,
properly documented.
The code for the conversion of the GL error enumeration code
into a string is not following the code style and conventions
we follow in Clutter and COGL.
The GE() macro is also using fprintf(stderr) directly instead
of using g_warning() -- which is redirectable to an alternative
logging system using the g_log* API.
We use math routines inside Cogl, so it's correct to have it in
the LIBADD line. In normal usage something else was pulling in
-lm, but the introspection is relying on linking against the
convenience library.
Based on a patch by: Colin Walters <walters@verbum.org>
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The timeline created when calling set_timeline(NULL) is referenced
even though we implicitly own it. When the Animation is destroyed,
the timeline is then leaked.
Thanks to: Richard Heatley <richard.heatley@starleaf.com>
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1548
Add a method for deleting the current selection inside a Text actor.
This is useful for subclasses.
See bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1521
Based on a patch by: Raymond Liu <raymond.liu@intel.com>
ClutterAnimation currently inherits the initial floating reference
semantics from GInitiallyUnowned. An Animation is, though, meant to
be used as a top-level object, like a Timeline or a Behaviour, and
not "owned" by another object. For this reason, the initial floating
reference does not make any sense.
Document that repeated calls to clutter_cairo_texture_create()
continue drawing on the same cairo_surface_t. Add
clutter_cairo_texture_clear() for when you don't want that behavior.
http://bugzilla.openedhand.com/show_bug.cgi?id=1599
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The cursor x position is already translated, so we do not need to take the
actors allocation into account when calculating scrolling.
Additionally, we need to update the text_x value before running
clutter_text_ensure_cursor_position.
Add any scrolling offset to the x value when in single line mode.
Now that the offset is taken into account in the position_to_coords
function, we do not need to adjust the cursor x manually in
clutter_text_paint.
If the cursor is already at the end of the Text contents then we
need to maintain its position when deleting the previous character
using the relative key binding.
The required "fake" libclutter-cogl.la upon with the main clutter
shared object depends is only built with introspection enabled
instead of being built unconditionally.
Passing:
--library=clutter-@CLUTTER_FLAVOUR@-@CLUTTER_API_VERSION@
to g-ir-scanner, when building Cogl was causing g-ir-scanner to
link the introspection program against the installed clutter library,
if it existed or fail otherwise. Instead copy the handling from
the json/ directory where we link against the convenience library
to scan, and do the generation of the typelib later in the
main clutter/directory.
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1594
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
If text is set, ClutterText should never return less than the layout
height for minimum and preferred heights.
This holds unless ellipsize and wrap are enabled, in which case the
minimum height should be the height of the first line -- which is
the height needed to at the very least show the ellipsization.
Based on a patch by: Thomas Wood <thomas@openedhand.com>
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1598
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
This is the another step into abstracting the backend operations
that are currently spread all across the board back into the
backend implementations where they belong.
The GL context creation, for instance, is demanded to the stage
realization which makes it a critical path for every operation
that is GL-context bound. This usually does not make any difference
since we realize the default stage, but at some point we might
start looking into avoiding the default stage realization in order
to make the Clutter startup faster.
It also makes the code maintainable because every part is self
contained and can be reworked with the minimum amount of pain.
The master clock is using the redraw priority to create the source
that will be used to spin the paint sequence if something is being
animated using a timeline.
Unfortunately, the priority is too high and this causes starvation
when embedding into other toolkits -- like gtk+.
Thanks to Havoc Pennington for catching this.
The XVisualInfo for GL is created when a stage is being realized.
When embedding Clutter inside another toolkit we might not want to
realize a stage to extract the XVisualInfo, then set the stage
window using a foreign X Window -- which will cause a re-realization.
Instead, we should abstract as much as possible into the X11 backend.
Unfortunately, the XVisualInfo for GL is requested using GLX API; for
this reason we have to create a ClutterBackendX11 method that we
override inside the ClutterBackendGLX implementation.
This also allows us to move a little bit of complexity from out of
the stage realization, which is currently a very delicate and hard
to debug section.
I was seeing clutter_text_get_selection trying to malloc up to 4Gb due to
unexpected negative arithmetic for the start/end offsets which resulted
in a crash.
This just tests for positions of -1 before deciding if the start/end
positions need to be swapped. The conversion from position to byte offset
already works with -1.
cogl_clip_push_window_rect is implemented using GPU scissoring which allows
the GPU to cull anything that falls outside a given rectangle. Since in the
case of picking we only ever care about a single pixel we can get the GPU to
ignore all geometry that doesn't intersect that pixel and only rasterize for
one pixel.
Previously clipping could only be specified in object coordinates, now
rectangles can also be pushed in window coordinates.
Internally rectangles pushed this way are intersected and then clipped using
scissoring. We also transparently try to convert rectangles pushed in
object coordinates into window coordinates as we anticipate the scissoring
path will be faster then the clip planes and undoubtably it will be faster
than using the stencil buffer.
The stencil buffer is always cleared the first time a clip is used
that needs it and the stencil test is disabled otherwise so there is
no need to clear before a paint.
COGLenum, COGLint and COGLuint which were simply typedefs for GL{enum,int,uint}
have been removed from the API and replaced with specialised enum typedefs, int
and unsigned int. These were causing problems for generating bindings and also
considered poor style.
The cogl texture filter defines CGL_NEAREST and CGL_LINEAR etc are now replaced
by a namespaced typedef 'CoglTextureFilter' so they should be replaced with
COGL_TEXTURE_FILTER_NEAREST and COGL_TEXTURE_FILTER_LINEAR etc.
The shader type defines CGL_VERTEX_SHADER and CGL_FRAGMENT_SHADER are handled by
a CoglShaderType typedef and should be replaced with COGL_SHADER_TYPE_VERTEX and
COGL_SHADER_TYPE_FRAGMENT.
cogl_shader_get_parameteriv has been replaced by cogl_shader_get_type and
cogl_shader_is_compiled. More getters can be added later if desired.
Calling glReadPixels is bad enough in forcing us to synchronize the CPU with
the GPU, but glFinish has even stronger synchonization semantics than
glReadPixels which may negate some driver optimizations possible in
glReadPixels.
Commit 43fa38fcf5 broke out-of-tree builds by removing some of the
builddir directories from the include path. builddir/clutter/cogl and
builddir/clutter are needed because cogl.h and cogl-defines-gl.h are
automatically generated by the configure script. The main clutter
headers are in the srcdir so this needs to be in the path too.
When the stage state changes between active/deactive, send out
key-focus-in/key-focus-out signal for the current key focused
actor on the stage.
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1503
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Setting the stage window using the set_stage_foreign() method will
lead to a re-realization. We need to make sure that the Drawable
currently associated to the GL context is set to None, to avoid a
BadDrawable error or, if we're unlucky, a segfault in the X server.
This commit reverts part of commit 5bcde25c - specifically the
part that forced a realization of the stage if we are ensuring
the GL context with it. This makes Clutter behave like it did
prior to commit 5bcde25c: if we are asked to ensure the GL context
with an unrealized stage we simply pass NULL to the backend
implementation.
When showing a Stage for the first time we end up realizing the stage
implementation before realizing the wrapper. This leads to segmentation
faults or errors coming from the backend because we're fumbling the
state and realization sequence.
Since we are destroying any previously set VisualInfo we keep we know
for sure that stage->xvisinfo is going to be None; hence, no reason to
check this condition.
The verify_map_state() internal method is conditionally compiled
if we have CLUTTER_ENABLE_DEBUG set; for this reason, all calls to
that method should be made conditional.
When building Clutter with introspection enabled everything stops
at Cogl GIR generation because it depends on the installed library
to work. Since we still require some changes in the API to be able
to build the GIR and the typelib for Cogl we should disable the
generation of the GIR as well.
The fix for bug 1138 broke multi-stage support on GLX, causing
X11 to segfault with the following stack trace:
Backtrace:
0: /usr/X11R6/bin/X(xf86SigHandler+0x7e) [0x80c91fe]
1: [0xb7eea400]
2: /usr/lib/xorg/modules/extensions//libglx.so [0xb7ae880c]
3: /usr/lib/xorg/modules/extensions//libglx.so [0xb7aec0d6]
4: /usr/X11R6/bin/X [0x8154c24]
5: /usr/X11R6/bin/X(Dispatch+0x314) [0x808de54]
6: /usr/X11R6/bin/X(main+0x4b5) [0x8074795]
7: /lib/i686/cmov/libc.so.6(__libc_start_main+0xe5) [0xb7c75775]
8: /usr/X11R6/bin/X(FontFileCompleteXLFD+0x21d) [0x8073a81]
which I can only track down to clutter_backend_glx_ensure_current()
being passed a NULL stage -- something that happens when a stage
is not correct realized. That should lead to a glXMakeCurrent(None)
and not to a segmentation fault, though.
When destroying a top-level actor we can actually relax the verification
of the map state, since it might be fully asynchronous and we might not
re-enter inside the mainloop in time to receive the unmap notification.
If the filter means that the there should be no rows left in the model,
clutter_model_get_iter_at_row (model, 0) should return NULL.
Howevever the currene implementation misbehaves and returns a bad iterator.
This change resolves the issue by tracking if we actually found any
non-filtered rows in our pass through the sequence.
OH Bugzilla: 1591
The stage is chaining up to the ClutterGroup::paint instead of
the ClutterGroup::pick method. This works anyway because we
detect the stage by default, but it's not a reliable solution
in case we decide to change the picking further on.
The master clock is currently advanced using a frame source driven
by the default frame rate. This breaks the sync to vblank because
the vblanking rate could be different than 60 Hz -- or it might be
completely disabled (e.g. with CLUTTER_VBLANK=none).
We should be using the main loop to check if we have timelines
playing, and if so queue a redraw on the stages we own.
We should also prepare the subsequent frame at the end of the redraw
process, so if there are new redraw we will have the scene already
in place.
This makes Clutter redraw at the maximum frame rate, which is
limited by the vblanking frequency.
Currently, picking in ClutterGroup pollutes the CLUTTER_DEBUG=paint
logs since it just calls the paint function. Reimplementing the pick
doesn't make us lose anything -- it might even be slightly faster
since we don't have to do a (typed) cast and a class dereference.
The timeline created when calling set_timeline(NULL) is referenced
even though we implicitly own it. When the Animation is destroyed,
the timeline is then leaked.
Thanks to: Richard Heatley <richard.heatley@starleaf.com>
Fixes bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1548
Currently, the conversion from em to units is done by using the
default font name inside the backend. For actors using their own
font/text layout we need a way to specify the font name along
with the quantity we wish to transform.
Commit 515350a7 renamed ::focus-in and ::focus-out to ::key-focus-in
and ::key-focus-out respectively. One signal emission for ::focus-out
escaped the renaming in ClutterStage.
Currently, the default screen guard value is 0, which is a valid
screen number on X11, and it might not be the default.
Patch suggested by: Owen W. Taylor <otaylor@redhat.com>
Currently, the introspection data for Cogl is built right into
Clutter's own typelib. This makes functions like:
cogl_path_round_rectangle()
Appear as:
Clutter.cogl_path_round_rectangle()
It should be possible, instead, to have a Cogl namespace and:
Cogl.path_round_rectangle()
This means building introspection data for Cogl alone. Unfortunately,
there are three types defined in Cogl that confuse the introspection
scanner, and make it impossible to build a typelib:
COGLint
COGLuint
COGLenum
These three types should go away before 1.0, substituted by int,
unsigned int and proper enumeration types. For this reason, we can
just set up the GIR build and wait until the last moment to create
the typelib. Once that has been done, we will be able to safely
remove the Cogl API from the Clutter GIR and typelib and let
people import Cogl if they want to use the Cogl API via introspection.
For consistency, and since those signals are key-related, the
::focus-in signal is not ::key-focus-in and the ::focus-out
signal is now ::key-focus-out.
Add a method for deleting the current selection inside a Text actor.
This is useful for subclasses.
See bug:
http://bugzilla.openedhand.com/show_bug.cgi?id=1521
Based on a patch by: Raymond Liu <raymond.liu@intel.com>
ClutterAnimation currently inherits the initial floating reference
semantics from GInitiallyUnowned. An Animation is, though, meant to
be used as a top-level object, like a Timeline or a Behaviour, and
not "owned" by another object. For this reason, the initial floating
reference does not make any sense.
With the recent change to internal floating point values, ClutterUnit
has become a redundant type, defined to be a float. All integer entry
points are being internally converted to floating point values to be
passed to the GL pipeline with the least amount of conversion.
ClutterUnit is thus exposed as just a "pixel with fractionary bits",
and not -- as users might think -- as generic, resolution and device
independent units. not that it was the case, but a definitive amount
of people was convinced it did provide this "feature", and was flummoxed
about the mere existence of this type.
So, having ClutterUnit exposed in the public API doubles the entry
points and has the following disadvantages:
- we have to maintain twice the amount of entry points in ClutterActor
- we still do an integer-to-float implicit conversion
- we introduce a weird impedance between pixels and "pixels with
fractionary bits"
- language bindings will have to choose what to bind, and resort
to manually overriding the API
+ *except* for language bindings based on GObject-Introspection, as
they cannot do manual overrides, thus will replicate the entire
set of entry points
For these reason, we should coalesces every Actor entry point for
pixels and for ClutterUnit into a single entry point taking a float,
like:
void clutter_actor_set_x (ClutterActor *self,
gfloat x);
void clutter_actor_get_size (ClutterActor *self,
gfloat *width,
gfloat *height);
gfloat clutter_actor_get_height (ClutterActor *self);
etc.
The issues I have identified are:
- we'll have a two cases of compiler warnings:
- printf() format of the return values from %d to %f
- clutter_actor_get_size() taking floats instead of unsigned ints
- we'll have a problem with varargs when passing an integer instead
of a floating point value, except on 64bit platforms where the
size of a float is the same as the size of an int
To be clear: the *intent* of the API should not change -- we still use
pixels everywhere -- but:
- we remove ambiguity in the API with regard to pixels and units
- we remove entry points we get to maintain for the whole 1.0
version of the API
- we make things simpler to bind for both manual language bindings
and automatic (gobject-introspection based) ones
- we have the simplest API possible while still exposing the
capabilities of the underlying GL implementation
When calling clutter_model_iter_next () / clutter_model_iter_prev () we need
to update the row for the iterator. In order to improve the peformance of
iterating this change adds a private row mutator and switches ClutterListModel
to use it.
In order to carry out various comparisons for e.g. identifying the first
iterator in the model we need a ClutterModelIter. This change switches from
creating a new iterator each time to reusing an existing iterator.
The flags field of ClutterActor should have accessor methods for,
language bindings.
Also, the set_flags() and unset_flags() methods should actively
emit notifications for the changed properties.
This is simply a wrapper around cogl_color_set_from_4f and
cogl_material_set_color. We already had a prototype for this, it was
an oversight that it wasn't already implemented.
There were several functions I believe no one is currently using that were
only implemented in the GL backend (cogl_offscreen_blit_region and
cogl_offscreen_blit) that have simply been removed so we have a chance to
think about design later with a real use case.
There was one nonsense function (cogl_offscreen_new_multisample) that
sounded exciting but in all cases it just returned COGL_INVALID_HANDLE
(though at least for GL it checked for multisampling support first!?)
it has also been removed.
The MASK draw buffer type has been removed. If we want to expose color
masking later then I think it at least would be nicer to have the mask be a
property that can be set on any draw buffer.
The cogl_draw_buffer and cogl_{push,pop}_draw_buffer function prototypes
have been moved up into cogl.h since they are for managing global Cogl state
and not for modifying or creating the actual offscreen buffers.
This also documents the API so for example desiphering the semantics of
cogl_offscreen_new_to_texture() should be a bit easier now.
These are necessary if nesting redirections to an fbo,
otherwise there's no way to know how to restore
previous state.
glPushAttrib(GL_COLOR_BUFFER_BIT) would save draw buffer
state, but also saves a lot of other stuff, and
cogl_draw_buffer() relies on knowing about all draw
buffer state changes. So we have to implement a
draw buffer stack ourselves.
Signed-off-by: Robert Bragg <robert@linux.intel.com>
It is valid in some situations to have a material layer with an invalid texture
handle (e.g. if you setup a texture combine mode before setting the texture)
and so _cogl_material_layer_free needs to check for a valid handle before
attempting to unref it.
Adds missing notices, and ensures all the notices are consistent. The Cogl
blurb also now reads:
* Cogl
*
* An object oriented GL/GLES Abstraction/Utility Layer
Redundant clearing of depth and stencil buffers every render can be very
expensive, so cogl now gives control over which auxiliary buffers are
cleared.
Note: For now clutter continues to clear the color, depth and stencil buffer
each paint.
The clutter frame source tries to average out the frame deltas so that
if one frame takes 1 interval plus a little bit of extra time then the
next frame will be 1 interval minus that little bit of extra
time. Therefore the deltas can sometimes be less than the frame
interval. ClutterTimeline should accumulate these small differences
otherwise it will end up missing out frames so the total duration of
the timeline will be a lot longer.
For example this was causing test-actors to appear to run very slow.
The method of ClutterTimeline that advances the timeline by a
delta (in millisecond) is going to be useful for testing the
timeline's behaviour -- and unbreak the timeline test suite that
was broken by the MasterClock merge.
With the introduction of the map/unmap flags and the split of the
visible state from the mapped state we require that every part of
a scene graph branch is mapped in order to be painted. This breaks
the ability of a ClutterClone to paint an hidden source actor.
In order to fix this we need to introduce an override flag, similar
in spirit to the current modelview and paint opacity overrides that
Clone is already using.
The override flag, when set, will force a temporary map on a
Clone source (and its children).
ClutterContainer provides a foreach_with_internals() vfunc for
iterating over all of a container's children, be them added using
the Container API or be them internal to the container itself.
We should be using the foreach_with_internals() function instead
of the plain foreach().
When replacing the fbo_handle with a new handle it first unrefs the
old handle. This was previously a call to cogl_offscreen_unref which
silently ignored attempts to unref COGL_INVALID_HANDLE. However the
new cogl_handle_unref does check for this so we should make sure the
handle is valid to avoid the warning.
Bug 1565 - test-fbo case failed in many platform
clutter_texture_new_from_actor was broken because it created the FBO
texture but then never attached it to the material so it was never
used for rendering. The old behaviour in Clutter 0.8 was to assign the
texture directly to priv->tex. In 0.9 priv->tex is replaced with
priv->material which has a reference to the tex in layer 0. So putting
the FBO texture directly in layer 0 more closely matches the original
behaviour.
When rendering a pango layout CoglPangoRenderer now records the
operations into a list called a CoglPangoDisplayList. The entries in
the list are either glyph renderings from a texture, rectangles or
trapezoids. Multiple consecutive glyph renderings from the same glyph
cache texture are combined into a single entry. Note the
CoglPangoDisplayList has nothing to do with a GL display list.
After the display list is built it is attached to the PangoLayout with
g_object_set_qdata so that next time the layout is rendered it can
bypass rebuilding it.
The glyph rendering entries are drawn via a VBO. The VBO is attached
to the display list so it can be used multiple times. This makes the
common case of rendering a PangoLayout contained in a single texture
subsequent times usually boil down to a single call to glDrawArrays
with an already uploaded VBO.
The VBOs are accessed via the cogl_vertex_buffer API so if VBOs are
not available in GL it will resort to a fallback.
Note this will fall apart if the pango layout is altered after the
first render. I can't find a way to detect when the layout is
altered. However this won't affect ClutterText because it creates a
new PangoLayout whenever any properties are changed.
Currently ClutterModel::get_iter_at_row() ignores whether we have
a filter in place. This also extends to the get_n_rows() method.
The more consistent, more intuitive and surely more correct way to
handle a Model with a filter in place is to take into account the
presence of the filter itself -- that is:
- get_n_rows() should take into account the filter and return the
number of *filtered* rows
- get_iter_at_row() should also take the filter into account and
get the first non-filtered row
These two changes make the ClutterModel with a filter function
behave like a subset of the original Model without a filter in
place.
For instance, given a model with three rows:
- [row 0] <does not match filter>
- [row 1] <matches filter>
- [row 2] <matches filter>
- [row 3] <does not match filter>
The get_n_rows() method will return "2", since only two rows will
match the filter; the get_first_iter() method will ask for the
zero-eth row, which will return an iterator pointing to the contents
of row 1 (but the :row property of the iterator will be set to 0);
the get_last_iter() method will ask for the last row, which will
return an iterator pointing to the contents of row 2 (but the :row
property of the iterator will be set to 1).
This changes will hopefully make the Model API more consistent
in its usage whether there is a filter in place or not.
Currently, there is no way for implementations of the ClutterModel
abstract class to know whether there is a filter in place. Since
subclasses might implement some optimization in case there is no
filter present, we need a simple (and public) API to ask the model
itself.
Setting the wrap mode on the PangoLayout seems to have disappeared
during the text-actor-layout-height branch merge so this brings it
back. The test for this in test-text-cache no longer needs to be
disabled.
We also shouldn't set the width on the layout if there is no wrapping
or ellipsizing because otherwise it implicitly enables wrapping. This
only matters if the actor gets allocated smaller than its natural
size.
Bug 1484 - Redraw ClutterClone when the source changes, even for
!visible sources
Connect to ::queue-redraw on the clone source and queue a redraw.
This allows redrawing the Clone when the source changes, even in
case of a non visible source actor.
http://bugzilla.openedhand.com/show_bug.cgi?id=1484
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Currently, all timelines install a timeout inside the TimeoutPool
they share. Every time the main loop spins, all the timeouts are
updated. This, in turn, will usually lead to redraws being queued
on the stages.
This behaviour leads to the potential starvation of timelines and
to excessive redraws.
One lesson learned from the games developers is that the scenegraph
should be prepared in its entirety before the GL paint sequence is
initiated. This means making sure that every ::new-frame signal
handler is called before clutter_redraw() is invoked.
In order to do so a TimeoutPool is not enough: we need a master
clock. The clock will be responsible for advancing all the active
timelines created inside a scene, but only when the stage is
being redrawn.
The sequence is:
+ queue_redraw() is invoked on an actor and bubbles up
to the stage
+ if no redraw() has already been scheduled, install an
idle handler with a known priority
+ inside the idle handler:
- advance the master clock, which will in turn advance
every playing timeline by the amount of milliseconds
elapsed since the last redraw; this will make every
playing timeline emit the ::new-frame signal
- queue a relayout
- call the redraw() method of the backend
This way we trade multiple timeouts with a single frame source
that only runs if a timeline is playing and queues redraws on
the various stages.
Bug 1547 - when an actor is unmapped while owning the focus, it should
release it
When an actor is unmapped while owning the focus, the should release it.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Bug 1138 - No trackable "mapped" state
* Add a VISIBLE flag tracking application programmer's
expected showing-state for the actor, allowing us to
always ensure we keep what the app wants while tracking
internal implementation state separately.
* Make MAPPED reflect whether the actor will be painted;
add notification on a ClutterActor::mapped property.
Keep MAPPED state updated as the actor is shown,
ancestors are shown, actor is reparented, etc.
* Require a stage and realized parents to realize; this means
at realization time the correct window system and GL resources
are known. But unparented actors can no longer be realized.
* Allow children to be unrealized even if parent is realized.
Otherwise in effect either all actors or no actors are realized,
i.e. it becomes a stage-global flag.
* Allow clutter_actor_realize() to "fail" if not inside a toplevel
* Rework clutter_actor_unrealize() so internally we have
a flavor that does not mess with visibility flag
* Add _clutter_actor_rerealize() to encapsulate a somewhat
tricky operation we were doing in a couple of places
* Do not realize/unrealize children in ClutterGroup,
ClutterActor already does it
* Do not realize impl by hand in clutter_stage_show(),
since showing impl already does that
* Do not unrealize in various dispose() methods, since
ClutterActor dispose implementation already does it
and chaining up is mandatory
* ClutterTexture uses COGL while unrealizable (before it's
added to a stage). Previously this breakage was affecting
ClutterActor because we had to allow realize outside
a stage. Move the breakage to ClutterTexture, by making
ClutterTexture just use COGL while not realized.
* Unrealize before we set parent to NULL in clutter_actor_unparent().
This means unrealize() implementations can get to the stage.
Because actors need the stage in order to detach from stage.
* Update clutter-actor-invariants.txt to reflect latest changes
* Remove explicit hide/unrealize from ClutterActor::dispose since
unparent already forces those
Instead just assert that unparent() occurred and did the right thing.
* Check whether parent implements unrealize before chaining up
Needed because ClutterGroup no longer has to implement unrealize.
* Perform unrealize in the default handler for the signal.
This allows non-containers that have children to work properly,
and allows containers to override how it's done.
* Add map/unmap virtual methods and set MAPPED flag on self and
children in there. This allows subclasses to hook map/unmap.
These are not signals, because notify::mapped is better for
anything it's legitimate for a non-subclass to do.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Bug 1228 - Unnecessary glColorMask on alpha drops performance
With DRI2, alpha is allowed in the window's framebuffer
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Bug 1513 - Allow passing in ClutterPickMode to
clutter_stage_get_actor_at_pos()
At the moment, clutter_stage_get_actor_at_pos() uses CLUTTER_PICK_ALL
internally to find an actor. It would be useful to allow passing in
ClutterPickMode to clutter_stage_get_actor_at_pos(), so that the caller
can specify CLUTTER_PICK_REACTIVE as a criteria.
Bug 1516 - Does not withdraw toplevels correctly
clutter_stage_x11_hide() needs to use XWithdrawWindow(), not
XUnmapWindow().
As it stands now, if the window is already unmapped (say the WM has
minimized it), then the WM will not know that Clutter has closed the
window and will keep the window managed, showing it in the task list
and so forth.
Bug 1517 - clutter_container_foreach_with_internals()
This allows us to iterate over all children (for things like maintaining
map/realize invariants) or only children that apps added and care about.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Bug 1561 - Bad code in clutter-alpha.c
The implementation of the easing modes equations followed closely
the JavaScript and ActionScript counterparts. Obviously, JS and AS
are not C-compatible, so later versions of gcc (4.4.0 for instance)
would complain about uninitialized variables and such. The code is
also obfuscated and hard to debug/understand.
For these reasons, the implementation should be unobfuscated and
sanitized.
A lockup was reported by fargoilas on #clutter and removing the server grab in
clutter_x11_texture_pixmap_sync_window fixed the problem.
We were doing an x server grab to guarantee that if the XGetWindowAttributes
request reported that our redirected window was viewable then we would also
be able to get a valid pixmap name. (the composite protocol says it's an
error to request a name for a window if it's not viewable) Without the grab
there would be a race condition.
Instead we now handle error conditions gracefully.
The documentation for the ::cursor-event says that the signal
is emitted only when the cursor position changes. Right now it
is being emitted every time we ensure the cursor position during
the Text paint sequence.
The clutter_text_ensure_cursor_position() function should check
whether the stored cursor position has changed since the last
paint, and emit the ::cursor-event signal only when there has
been a change.
We only want to eschew the pango_layout_set_width() on editable,
single-line Text actors because they can "scroll" the PangoLayout.
This fixes the ellipsization on entries.
In unifying the {gl,gles}/cogl.c code recently, moving most of the code into
common/cogl.c the gmodule.h include was also mistakenly moved.
Thanks to Felix Rabe for reporting this issue.
Note: I haven't tested this fix myself, as I'm not set up to be able to
build for OS X
Buffer objects aren't currently available for glx indirect contexts, so we
now have a fallback that simply allocates fake client side vbos to store the
attributes.
This makes the #if 0'd debug code that was in _cogl_journal_flush_quad_batch
- which we have repeatedly found usefull for debugging various geometry
issues in Clutter apps - a runtime debug option.
The outline colors rotate in order from red to green to blue which can also
help confirm the order that your geometry really drawn.
The outlines are not affected by the current material state, so if you e.g.
have a blending bug where geometry mysteriously disappears this can confirm
if the underlying rectangles are actually being emitted but blending is
causing them to be invisible.
Editable ClutterText will scroll when allocated less width than is
necessary to lay out the entire layout on a single line, so return 1 for
the minimum width in this case.
Approved by Emmanuele Bassi, fixes bug #1555.
Since we have to do (z_far - z_near) and use it in a division we
should check that the user is not passing a value that would
cause a division by zero.
If we need to check that the layout sequence is correct in
terms of order of execution and with respect to caching, then
having a CLUTTER_DEBUG_LAYOUT debug flag would make things
easier.
Ellipsizing was effectively broken for two reasons. There was a typo
in the code to set the width so it always ended up being some massive
value. If no height should be set on the layout it was being set to
G_MAXINT. Setting a height greater than 0 enables wrapping which so
ellipsizing is not performed. It should be left at the default of -1
instead.
Bug 1476 - JSON Parser memory leak
Static analysis of the code showed that the in-tree copy of
the JsonParser object leaks objects and arrays on parse errors.
Thanks to Gordon Williams <gordon.williams@collabora.co.uk>
If the stage is unrealized (such as will be the case if the stage was
created with clutter_stage_new) then it would set the size of the
stage display but it was not setting the fullscreen_on_map flag so it
never got the _NET_WM_STATE_FULLSCREEN property. Now it always sets
the flag regardless of whether the window is created yet.
There was also problems with dual-headed displays because in that case
DisplayWidth/Height will return the size of the combined display but
Metacity (and presumably other WMs) will sensibly try fit the window
to only one of the monitors. However we were setting the size hints so
that the minimum size is that of the combined display. Metacity tries
to honour this by setting the minimum size but then it no longer
positions the window at the top left of the screen.
The patch makes it avoid setting the minimum size when the stage is
fullscreen by checking the fullscreen_on_map flag. This also means we
can remove the static was_resizable flag which would presumably have
caused problems for multi-stage.
Adds a new property so that the selection color can be different from
the cursor color. If no selection color is specified it will use the
cursor color as before. If no cursor color is specified either it will
use the text color.
The debug macros for tracking reference counting of CoglHandles had
some typos introduced in c3d9f0 which meant it failed to compile when
COGL_DEBUG is 1.
Since the Cogl material branch merge when changing the color of a part
using pango attributes (such as using <span color="red" /> markup)
then it wouldn't return to the default color for the rest of the
layout. pango_renderer_get_color returns NULL if there is no color
override in which case it needs to revert to the color specified as
the argument to cogl_pango_render_layout. The 'color' member of
CoglPangoRenderer has been reinstated to store that default color and
now cogl_pango_render_set_color_for_part is the only place that sets
the material color.
The clutter_actor_animate*() family of functions should only connect
to the Animation::completed signal once, during the construction of
the Animation object attached to the Actor. Otherwise, the completed
signal handler will be run multiple times, and will try to unref()
the Animation for each call -- leading to a segmentation fault.
The Animation class is missing a ::started signal matching the
::completed one. A ::started signal is useful for debugging,
initial state set up, and checks.
Bug 1535 - Complete animation always unrefs ClutterAnimation (even
after g_object_ref_sink)
Animations created through clutter_animation_new() should not
automagically unref themselves by default on ::complete. We
only want that behaviour for Animations created by the
clutter_actor_animate* family of functions, since those provide
the automagic memory management.
ClutterGroup still ships with API deprecated since 0.4. We did
promise to keep it around for a minor release cycle -- not for 3.
Since we plan on shipping 1.0 without the extra baggage of the
deprecated entry points, here's the chance to remove the accumulated
cruft.
All the removed methods and signals have a ClutterContainer
counterpart.
Since we're planning to release 1.0 without any of the deprecated
API baggage, we can simply remove the set_uniform_1f() method from
ClutterShader public API and add it to the deprecated header.
The cogl_is_* functions were showing up quite high on profiles due to
iterating through arrays of cogl handles.
This does away with all the handle arrays and implements a simple struct
inheritance scheme. All cogl objects now add a CoglHandleObject _parent;
member to their main structures. The base object includes 2 members a.t.m; a
ref_count, and a klass pointer. The klass in turn gives you a type and
virtual function for freeing objects of that type.
Each handle type has a _cogl_##handle_type##_get_type () function
automatically defined which returns a GQuark of the handle type, so now
implementing the cogl_is_* funcs is just a case of comparing with
obj->klass->type.
Another outcome of the re-work is that cogl_handle_{ref,unref} are also much
more efficient, and no longer need extending for each handle type added to
cogl. The cogl_##handle_type##_{ref,unref} functions are now deprecated and
are no longer used internally to Clutter or Cogl. Potentially we can remove
them completely before 1.0.
A layer object may be instantiated when setting a combine mode, but before a
texture is associated. (e.g. this is done by the pango renderer) if this is the
case we shouldn't call cogl_texture_get_format() with an invalid cogl handle.
This patch skips over layers without a texture handle when determining if any
textures have an alpha channel.
Bug 1518 - [Patch] Widget derivied from ClutterText will crash on
key_press_event
In clutter_text_key_press() we are using G_OBJECT_TYPE_NAME to find
out the actor's type name. However, if some widget is derived from
ClutterText, when the key press handler is called, G_OBJECT_TYPE_NAME
will return the name of the derived widget.
The default implementation should get the binding pool for the base
class.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
ClutterText should offer multiple selection modes depending on the
number of pointer clicks.
Following the GTK+ conventions:
- double click selects the current word
- triple click selects the current line
Added support for registering a handler for the completed signal
directly amongst the varargs making it easier to attach code
to be executed when animations complete.
Bug 1529 - Selection bound out of sync with an empty Text actor
When the user clicks on a Text actor the cursor position and the
selection bound are set using bytes_to_offset(); if the Text is
empty, this means placing them both to 0. Setting the selection
bound to 0 means that when the user inserts a character by typing
it into the Text, the selection will be [0,1]; this will select
the first character, which will then be overwritten when typing
a new character.
The Text actor should, instead, check if there are no contents
and set the cursor position and the selection bound to -1.
The clutter_text_set_selection_bound() method should also validate
the value passed, in case it's bigger than the text lenght, or
smaller than -1.
The symbol name for cogl_vertex_buffer_draw_elements was wrong so it
ended up with no documentation. The name for the vertex attribute was
incorrect in cogl_vertex_buffer_add.
None of this code directly related to implementing CoglTextures, and the
code was needlessly duplicated between the GL and GLES backends. This moves
the cogl_rectangle* and cogl_polygon* code into common/cogl-primitives.c
makes which makes lot of sense since the two copies keep needlessly
diverging introducing or fixing bugs in one but not the other. For instance
I came accross one such bug regarding the enabling of texture units when
unifying the code.
The gtk-doc had a copy and paste error, saying the float array should be 8
elements per rectangle instead of 4. There was also no newline in the gles
code before the new function.
It's often nice to be able to draw a batch of vertices, even if these
have no texture coordinates. This add a cogl_rectangles, similar to
cogl_rectangles_with_texture_coords, only without.
There's no need to enable the texture target unless it is going to be
used for rendering. Enabling it directly with glEnable calls confuses
Cogl's state caching.
This is a replacement for the patch in bug 1483 which was reverted.
This reverts commit f9d996a460.
The change from calling glBindTexture to using the material API with
cogl_material_flush_gl_state does not always work because it doesn't
necessarily leave the active texture unit as GL_TEXTURE0. For example,
if the previously rendered texture was multi-layered then the last
thing cogl_material_flush_gl_state will do is select GL_TEXTURE1 just
to disable it.
Clutter was complaining about netural width smaller than minimum widths
(differences around 0.0005) by using an epsilon value of 1e-4 for these
floating point comparisons, these warnings have now been silenced.
Using test-cogl-vertex-buffer as a test case which is CPU bound due to
hls -> rgb conversions this alternative algorithm looked to be ~10%
faster when tested on an X61s Lenovo.
Queuing an animation on an actor cannot be done from within the
::completed signal handler, because we guarantee that the Animation
instance is valid and attached to the actor it animates for the
whole duration of the signal emission chain.
In order to queue animations you have to install an idle handler
on the main loop, and call clutter_actor_animate() inside it.
The documentation should be more clear about this caveat in the
memory management of ClutterAnimations created by the animate()
family of functions.
There are various constraints for when we can support multi-texturing and
when they can't be met we try and print a clear warning explaining why the
operation isn't supported, but we shouldn't endlessly repeat the warning for
every primitive of every frame. This patch fixes that.
This function was renamed a while ago in the .c file from
cogl_vertex_buffer_draw_range_elements but the corresponding .h and
doc/reference/cogl changes weren't made.
For convenience it is now valid to avoid a seperate call to
cogl_vertex_buffer_submit() and assume that the _draw() calls will do this
for you (though of course if you do this you still need to ensure the
attribute pointers remain valid until your draw call.)
Bug 1495 - Timelines run 4% short
Previously the timelines were timed by calculating the interval
between each frame stored as an integer number of milliseconds so some
precision is lost. For example, requesting 60 frames per second gets
converted to 16 ms per frame which is actually 62.5 frames per
second. This makes the timeline shorter by 4%.
This patch merges the common code for timing from the timeout pools
and frame sources into an internal clutter-timeout-interval file. This
stores the interval directly as the FPS and counts the number of
frames that have been reached instead of the elapsed time.
The Animation API should follow this pattern:
- functions with an Interval as part of the arguments should have
"interval" inside their name, e.g.:
clutter_animation_bind_interval
clutter_animation_update_interval
- functions dealing with property names should have "property"
inside their name, e.g.:
clutter_animation_has_property
clutter_animation_unbind_property
- unless -
- functions dealing with a property and its value should not
have any modifier, e.g.:
clutter_animation_bind
The change from update_property() to update_interval() frees up
clutter_animation_update(), to be added at a later date.
Bug 1454 - move queue_redraw virtualization to ClutterActor
The ClutterActor::queue-redraw signal allows parent containers to
track whether their children need a redraw.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
A common use of setting clip is to keep an actor inside its allocation;
right now to do this you have to set up a callback on notify::allocation.
There's no overhead added by sticking another bit in ClutterActor
clip-to-allocation that will clip painting to the allocation if set.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Bug 1419 - Add clutter_animation_bind, rename clutter_animation_bind_interval
This is a different approach from the clutter_actor_animatev
vector variant. The single call should be even easier on
automatic bindings, since calls can be chained like:
new Clutter.Animation({object: myactor}).bind("x", 42).bind("y", 43);
Note clutter_animation_bind_property which took a ClutterInterval
is renamed to clutter_animation_bind_interval for clarity, and to
discourage use since there are friendlier APIs about.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
If we are animating an actor using the clutter_actor_animate*() family
of functions we might want to expose a getter for the Animation instance
we are using, to avoid excessing bookkeeping.
The ::completed signal requires an emitter in case we want to stop
the animation and safely dispose it inside application code without
breaking invariants.
The memory management of the ClutterAnimation instances should be
heavily documented, given its "automagical" nature, so that other
people might understand it and avoid breaking it in the future.
It's reasonably normal for a relayout of the stage to cause the stage
to be queued for drawing; for this reason we should do the relayout before
we clear stage->update_idle. (But want to clear update_idle() before actually
doing the redraw to handle the corner case where the draw queues another
redraw.)
Bug 1499 - clutter_actor_notify_if_geometry_changed causes sync
layout cycles
Whenever clutter_actor_set_{width,height,x,y,size...} is used, extra
synchronous size requests are triggered in
clutter_actor_notify_if_geometry_changed.
If the get_preferred_width() and get_preferred_height() implementations
are particularly costly (e.g. ClutterText) this will result in a performance
impact.
To avoid excessive allocation or requisition cycles we use the
cached values and flags. If we don't have an allocation, we assume
that we need to notify all the properties; if we don't have a size
requisition we notify only width and height; finally, if we do have
a valid allocation we notify only on the changed values.
_cogl_add_path_to_stencil_buffer and _cogl_add_stencil_clip were leaving
the projection matrix current when calling cogl_rectangle which was
upsetting _cogl_current_matrix_state_flush.
Adds glFrustum wrappers (GLES only accepts floats not doubles, and GLES2
needs to use our internal cogl_wrap_glFrustumf)
Adds GL_TEXTURE_MATRIX getter code in cogl_wrap_glGetFloatv
Adds a GL_TEXTURE_MATRIX define for GLES2
Its not intended that users should use these with any other matrix mode, and
internally we now have the _cogl_current_matrix API if we need to play with
other modes.
If we later add internal flags to CoglMatrix then this code wouldn't
initialize those flags. The ways it's now done adds a redundant copy, but
if that turns out to be something worth optimizing we can look again at
using a cast but adding another way for initializing internal flags.
This is useful because sometimes we need to get the current matrix, which
is too expensive when indirect rendering.
In addition, this virtualization makes it easier to clean up the API in
the future.
Only have load-data-async and load-async properties, both are construct
only and the latter adds the former load-size-async behavior on top of
load-data-async.
This commit yesterday:
89e3e3a4cc
[animation] Add vector variants for ::animate()
broke the "fixed::" attribute on properties in clutter_actor_animate(),
because the fixed:: part is still on the string when it checks to see if
it's a valid property the class knows about.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
As a convenience, if NULL is passed for the text argument of
clutter_text_set_text() (and for consistency,
clutter_text_set_markup()), treat that the same as "".
Bug 1480 - Clutter groups don't output their name when
CLUTTER_DEBUG=paint
This is a really useful thing to do debugging-wise - as you can see
istantly when (and if) clutter objects are being drawn - as often
there's quite a hierarchy of objects in groups.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Add missing (out) annotations to the doc comments for ClutterActor methods
with multiple return values.
Fix the definition of clutter_actor_get_allocation_vertices() to be
consistent with the declaration and have verts[4] rather than verts[].x
http://bugzilla.openedhand.com/show_bug.cgi?id=1481
Bug 1501 - clutter_text_insert_text not working right with non-onebyte
character
In clutter_text_insert_text(), the position is expressed in characters, not
in bytes.
Actually, it turns out to be working on bytes, so when there are already
multi-byte character in the text buffer, insert text at the position after
the multi-byte character will not work right.
Also, the position is not updated after the insert work is done.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Bug 1500 - [Patch] clutter_text crash with non one-byte utf8 text exceed
max_length
In clutter_text_set_text_internal(), when text length in character is greater
than max_length, and there are multi-byte character in it, then the new text
string buffer is not malloc()'ed with right length. This will cause the app to
crash with segmention fault.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The fixed-point values are converted to integers, which obviously
messes up the conversion. Instead, they should be converted to
floating point values before normalizing to bytes for the assignment
to ClutterColor components.
$(builddir) isn't always defined in earlier versions of Automake
(although I can't track down exactly which version it appeared
in). According to the autoconf docs it is "rigorously equal to `.'" so
we might as well not use it.
The private CoglContext is created using g_malloc() and not
zeroed; this means we have to initialize the values we are
going to check, to avoid hitting garbage.
Thanks to Tommi Komulainen.
Commit a383929 added the $(srcdir) prefix to all of the source files
but some files are generated by the configure script and other make
rules so they actually live in $(builddir). Out-of-tree builds
therefore broke.
Bug 1438 - Implicit Animation API could use animatev variants
The clutter_actor_animate* family of functions use va_lists to
handle the property/value pairs for the final state of the
animation.
Language bindings have problems with variadic arguments functions,
and usually prefer vector-based API which allow a greater level
of control and conversion from native data types.
For each variadic arguments function in the clutter_actor_animate*
family there should be a vector-based version that takes:
- the number of property/value pairs
- a constant array of constant strings
- an array of GValues
Most of the internal implementation can be refactored from the
current one, thus both the var_args and the vector entry points
share a common implementation of the code; then, both versions
of the API are just loops over a list of arguments.
Based on a patch by: Robert Carr <carrr@rpi.edu>
Since the switch from fixed point to floating point, and the introduction
of CoglFixed, ClutterFixed has been typedef'd into a float. This makes
ClutterFixed the worst fixed point API ever.
Now that Clutter has been migrated to CoglFixed and gfloat whenever needed,
ClutterFixed can be safely removed.
The only thing that Clutter should still provide is ClutterParamSpecFixed,
for installing fixed point properties into GObject classes.
The ClutterFixed symbols have been entirely removed from the API.
A GValue containing a ClutterUnit should be transformable into a
GValue holding an integer, a floating point value or a fixed point
value.
This means adding more transformation functions when registering
the ClutterUnit GType.
The fog and perspective API is currently split in two parts:
- the floating point version, using values
- the fixed point version, using structures
The relative properties are using the structure types, since they
are meant to set multiple values at the same time. Instead of
using bare values, the whole API should be coalesced into two
simple calls using structures to match the GObject properties.
Thus:
clutter_stage_set_fog (ClutterStage*, const ClutterFog*)
clutter_stage_get_fog (ClutterStage*, ClutterFog*)
clutter_stage_set_perspective (ClutterStage*, const ClutterPerspective*)
clutter_stage_get_perspective (ClutterStage*, ClutterPerspective*)
Which supercedes the fixed point and floating point variants.
More importantly, both ClutterFog and ClutterPerspective should
using floating point values, since that's what get passed to
COGL anyway.
ClutterFog should also drop the "density" member, since ClutterStage
only allows linear fog; non-linear fog distribution can be achieved
using a signal handler and calling cogl_set_fog() directly; this keeps
the API compact yet extensible.
Finally, there is no ClutterStage:fog so it should be added.
The type machinery for CoglFixed should be implemented by COGL
itself, now that COGL exports the GType of its types.
This allows moving most of what ClutterFixed did directly to
CoglFixed where it belongs.
ClutterFixed as a type is going away, superceded by CoglFixed. The
fixed point entry points in the API should be ported to the
CoglFixed type so that they are useful again.
Since the conversion of a floating point value to a fixed point
value is already done in double precision we can safely expose
a macro that converts a double precision floating point value to
a CoglFixed one.
Grabs are an entirely evil way to override the whole event delivery
machinery that Clutter has in place.
A pointer grab can be effectively replaced by a much more reliable
::captured-event signal handler, for instance.
Sometimes, grabs are a necessary evil -- and that is why Clutter
exposes them in the API; that should not fool anyone into thinking
that they should be used unless strictly necessary.
When calling clutter_actor_animate() on an actor that is being
currently animated the default behaviour is to update the duration
of the animation; the easing mode; update all the common properties;
and finally add the new properties.
This:
clutter_actor_animate (actor, 500, CLUTTER_LINEAR,
"width", 100,
"height", 100,
NULL);
clutter_actor_animate (actor, 250, CLUTTER_EASE_IN_CUBIC,
"x", 100,
"y", 100,
"height", 200,
NULL);
Is logically equivalent to:
clutter_actor_animate (actor, 250, CLUTTER_EASE_IN_CUBIC,
"x", 100,
"y", 100,
"width", 100,
"height", 200,
NULL);
The documentation of the function should be slightly more verbose
in describing the default behaviour.
The generation of the GObject introspection data has broken
the distcheck phase.
The location of the header and source files should always be
depending on the $(top_srcdir) and $(srcdir) variables,
respectively; the special handling of the COGL API inside the
GIR generation should also take those two variables into
account.
COGL should ship its own pkg-config file, obviously still pointing
to Clutter's compiler flags and linking options, for COGL-specific
variables that might be queried at configure time.
For instance, it's easier (and less verbose) to do:
PKG_CHECK_EXISTS([cogl-gl-1.0],
[has_gl_backend=yes],
[has_gl_backend=no])
Than doing:
AC_MSG_CHECKING([for GL support in COGL])
cogl_backend=`$PKG_CONFIG --variable=cogl clutter-0.9`
if test x$cogl_backend = xgl; then
has_gl_backend=yes
AC_MSG_RESULT([found])
else
has_gl_backend=no
AC_MSG_RESULT([not found])
fi
ClutterShader contains a lot of duplication, as the vertex and fragment
shader code paths are mostly the same.
The code should be simplified by adding new internal functions that can
be called with a value from the already present ClutterShaderType
enumeration.
In the future it'll also be possible to deprecate the current split API
and expose the generic accessors instead.
The ClutterColor API has some inconsistencies:
- the string deserialization function does not match the rest of
the conversion function naming policy; the naming should be:
clutter_color_parse() -> clutter_color_from_string()
and the first parameter should be the ClutterColor that will
be set from the string, not the string itself (a GDK-ism).
- the fixed point API should not be exposed, especially in the
form of ClutterFixed values
- the non-fixed point HLS conversion functions do not make any
sense. The values returned should be:
hue := range [ 0, 360 ]
luminance := range [ 0, 1 ]
saturation := range [ 0, 1 ]
like the current fixed point API does. Returning a value in
the [ 0, 255 ] range is completely useless
- the clutter_color_equal() should be converted for its use inside
a GHashTable; a clutter_color_hash() should be added as well
- the second parameter of the clutter_color_shade() function should
be the shading factor, not the result (another GDK-ism). this way
the function call can be translated from this:
color.shade(out result, factor)
to the more natural:
color.shade(factor, out result)
This somewhat large commit fixes all these issues and updates the
internal users of the API.
The CoglPango code falls under the COGL "jurisdiction"; this means
that it cannot include Clutter headers unless strictly necessary.
The CoglPangoRenderer code was using the CLUTTER_NOTE() macro. Now
that COGL has it's own COGL_NOTE() similar macro, CoglPango should
use that and avoid including clutter-debug.h (which pulls in
clutter-private.h which in turn pulls in clutter-actor.h).
A new flag, COGL_DEBUG_PANGO, has been added to the COGL debug
flags.
Bug 1493 - GL ES does not work since Jan 9 in PowerVR SGX 535, Intel
The mapped flag needs to be set on the stage otherwise
clutter_actor_queue_redraw will never queue a redraw and never draw
anything.
In these two backends there is not really a way to hide the stage so
they both set the mapped flag immediatly when clutter_actor_show is
called.
Using glEnable() directly confuses COGL and can result in problems
like subsequent pick operations not working correctly. Get the
material for the ClutterTexture and call cogl_material_flush_gl_state()
instead.
http://bugzilla.openedhand.com/show_bug.cgi?id=1483
When event delivery is invoked by synthetic events through
clutter_do_event from inside an event handler clutter was silently
ignoring it, this warning will hopefully help resolving some issues.
In the future if we want to annotate matrices with internal flags, and add
caching of the inverse matrix then we need to ensure that all matrix
modifications are done by cogl_matrix API so we'd know when to dirty the
cache or update the flags.
This just adds documentation to that effect, and assuming the most likley
case where someone would try and directly write to matrix members would
probably be to load a constant matrix other than the identity matrix; I
renamed cogl_matrix_init_from_gl_matrix to cogl_matrix_init_from_array to
make it seem more general purpose.
The font options accessors in ClutterBackend only deal with const
cairo_font_options_t values, since:
- set_font_options() will copy the font options
- get_font_options() will return a pointer to the internal
font options
Not using const in these cases makes the API confusing and might lead
to erroneous calls to cairo_font_options_destroy().
gcc warns about casting a pointer to a guint because it is a different
size on 64-bit machines. However the pointer is only used as a hash so
it doesn't matter if we lose the most significant bits. The patch
makes it use GPOINTER_TO_UINT instead which first casts it to a gulong
and avoids the warning.
Instead of creating a separate thread for each texture load, the loads
are now queued in a thread pool which will use at most 3 textures at a
time. This avoids the potentially large overhead of creating and
destroying a thread for every image while still avoiding throttling
the CPU if a large number of textures are loaded at the same time.
If the load is cancelled the old code needed to join the thread which
effectively meant it had to wait for the load to finish. With this
model the threads are owned by the thread pool so there is no need to
wait for them to finish. Instead, each thread is given its own thread
data which we can use to mark the thread as aborted. Once the load is
finished the thread will check whether it is aborted before setting
the texture data. That way the ClutterTexture can just disown threads
that it needs to cancel.
The duration in ClutterMedia is currently expressed in integer multiples
of a second. This does not offer enough granularity for media playback
that has fractionary durations -- 2.3 seconds; 1 minute, 23.4 seconds;
1 hour, 23 minutes, 4.5 seconds; etc.
The duration value should be expressed in seconds with a sub-second
granularity; for this reason, a double should be used insted of an
unsigned integer.
The semantics haven't changed: the :duration property still exposes
seconds.
Bug 1474 - ClutterMedia interface
The GET_INTERFACE() macro is still using the old type name of
the ClutterMedia interface structure in the cast.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Bug 1473 - CoglPixelFormat enum data must be declared static
When registering an enumeration GType, the GEnumValue or GFlagsValue
arrays must be declared static; otherwise, you get a segmentation
fault when calling the function again.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Instead of using a fixed size array for storing the scenegraph sub-node
during event delivery we should use a GPtrArray. The benefits are:
- a smaller allocation
- no undocumented yet binding constraint on the scenegraph size
The environment variable to disable mipmapping should also be
a command line switch, and be handled like the rest of Clutter's
environment variables/command line switches.
The ClutterFixed type and symbols are now equivalent to the native
float type. In order to remove ambiguity and clean up the code, any
usage of ClutterFixed and relative macros has been removed.
Clutter is able to show debug messages written using the CLUTTER_NOTE()
macro at runtime, either by using an environment variable:
CLUTTER_DEBUG=...
or by using a command line switch:
--clutter-debug=...
--clutter-no-debug=...
Both are parsed during the initialization process by using the
GOption API.
COGL would benefit from having the same support.
In order to do this, we need a cogl_get_option_group() function in
COGL that sets up a GOptionGroup for COGL and adds a pre-parse hook
that will check the COGL_DEBUG environment variable. The OptionGroup
will also install two command line switches:
--cogl-debug
--cogl-no-debug
With the same semantics of the Clutter ones.
During Clutter initialization, the COGL option group will be attached
to the GOptionContext used to parse the command line options passed
to a Clutter application.
Every debug message written using:
COGL_NOTE (SECTION, "message format", arguments);
Will then be printed only if SECTION was enabled at runtime.
This whole machinery, like the equivalent one in Clutter, depends on
a compile time switch, COGL_ENABLE_DEBUG, which is enabled at the same
time as CLUTTER_ENABLE_DEBUG. Having two different symbols allows
greater granularity.
Mipmapped text is enabled by default in Clutter but it can cause
problems on some drivers so it is convenient to have an environment
variable to disable it for all apps.
An assert to verify there was no error when generating a buffer object
for the vertex buffer API was being hit when running the GLES1 conformance
tests.
Bug #1457 - Creating a new texture messes up the cogl material state
cache; reported by Neil Roberts
We still don't have caching of bound texture state so we always have to
re-bind the texture when flushing the GL state of any material layers.
Bug #1460 - Handling of flags in cogl_material_set_color
Cogl automatically enables/disables blending based on whether the source color
has an alhpa < 1.0, or if any textures with an alpha component are in use, but
it wasn't doing it quite right.
At the same time I removed some of the dirty flags which on second thought
are nothing more than micro-optimsations that only helped clutter the code.
thanks to Owen Taylor for reporting the bug
Since the CoglMatrix type was added for supporting texture matrices recently
it made sense to be consistent accross the Cogl API and use the Cogl type
over the GL style GLfloat m[16] arrays.
The INCLUDES directive should only contain pre-processor flags, since
we're passing it also to the introspection scanner.
Using AM_CFLAGS for compiler flags, like debug flags and maintainer
flags, is more indicated.
Bug 1442 - multistage, same-window resize events invalidate stage
When ensuring that the GL context is attached to the correct
ClutterStage we need to set the SYNC_MATRICES flag on the stage
itself. This is needed in case the size of the new stage does
not match the size of the old -- thus requiring a call to
glViewport() when the paint cycle starts.
cogl_wrap_glActiveTexture needs to call the GL version of
glActiveTexture otherwise the subsequent calls to glBindTexture will
all be using texture unit 0. This fixes test-cogl-multitexture.
Previously the texture unit settings were stored in growable GArrays
and every time a new texture unit was encountered it would expand the
arrays. However the array wasn't copied when stored in a
CoglGles2WrapperSettings struct so all settings had the same
array. This meant that it wouldn't detect that a different program is
needed if a texture unit is disabled or enabled.
The texture unit settings arrays are all now a fixed size and the
enabledness of each unit is stored in a bit mask. Therefore the
settings can just be copied around by assignment as before.
This puts a limit on the number of texture units accessible by Cogl
but I think it is worth it to make the code simpler and more
efficient. The material API already poses a limit on the number of
texture units it can use.
configure.ac: Check for gobject-introspection
build/introspection.m4: Include the file that defines the
GOBJECT_CHECK_INTROSPECTION m4 macro in case we want to disable
the introspection data generation.
clutter/json/Makefile.am: Build the .gir for "ClutterJson"
(json-glib as part of the Clutter library)
clutter/Makefile.am: Build the .gir for clutter, compile the
Clutter and ClutterJson girs into typelibs, and install them.
Also move GCC_FLAGS from $(INCLUDES) to $(AM_CFLAGS) since it includes
non-preprocessor flag like -Wall.
See also:
http://bugzilla.openedhand.com/show_bug.cgi?id=1450
Based on a patch by: Owen W. Taylor <otaylor@fishsoup.net>
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Add annotations such as (transfer-none) (out) (element-type ClutterActor),
and so forth to the doc comments as appropriate.
The annotations added here are a combination of the annotations previously
in gir-repository for Clutter and annotations found in a review of all
return values with that were being parsed with a transfer of "full".
http://bugzilla.openedhand.com/show_bug.cgi?id=1452
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Bug 1464 - clutter_timeline_list_markers should set n_markers=0 when
there are no markers
If no markers are registered and we bail out we should also set the
n_markers out parameter to 0 so that code checking the number of
markers will still work.
Bug 1465 - clutter_path_parse_description should check p==NULL
When accepting a stringified path description, ClutterPath methods
and the parser should not accept NULL or empty strings.
Bug 1463 - clutter_color_lighten(darken) unexpected because of
ClutterFixed=>float
These two functions contained ClutterFixed constants represented as
integers which are no longer valid since the cogl-float branch
merge. They are now converted to ClutterFixed from float constants
using CLUTTER_FLOAT_TO_FIXED (which is now a no-op).
Thanks to Zhang Wei for reporting.
This causes clutter to skip all the GL work of state changes and
texturing if the opacity was 0. This is done in ClutterTexture and not
ClutterActor to ensure that pre and post paint signals work correctly.
Other expensive actors should be doing the same thing.
The call to "cmp" to compare a built file with its current version
should use the -s (silent) command line switch. This avoids a ugly
message on the console when building Clutter the first time.
When setting the COGL texture handle for a ClutterTexture the
texture will be set as the first layer of the material used
by the ClutterTexture. The documentation should clarify this
point.
The :texture and :material properties of ClutterTexture use a
Clutter-provided GType shielding from CoglHandle. Since CoglHandle
now has a GType we can use COGL_TYPE_HANDLE instead.
This commit also removes the conditional compilation of the
:material property, as it makes little sense now that the
Materials API has landed.
COGL types should be registered inside the GType system, for
bindings and type checking inside properties and signals.
CoglHandle is a boxed type with a ref+unref semantics; slightly evil
from a bindings perspective (we cannot associate custom data to it),
but better than nothing.
The rest of the exposed types are enumerations or bitmasks.
The COGL_DEFINE_HANDLE macro generates a cogl_is_<type> function
as well, to check whether a CoglHandle opaque pointer is of type
<type>.
The handle for CoglMaterial does not export cogl_is_material() in
its installed header.
The rotation angle calculated in clutter_behaviour_rotate_alpha_notify
gets applied to each actor using clutter_behaviour_actors_foreach. The
angle is a ClutterFixed value. Before the cogl float branch merge it
was stuffed into a gpointer using GPOINTER_TO_UINT. The pointer was
then converted back to a uint and cast to a ClutterFixed which worked
out fine even for negative numbers.
After the cogl-float merge the angle is effectively a gfloat. This
gets cast to a uint and stored in a pointer and converted back to a
float via a uint at the other end. However this fails for negative
numbers because a uint -> float conversion can't take advantage of
overflow to preserve the sign so you end up with a large number.
It also had the side effect that it only rotated in whole degrees.
This commit fixes the problem by just passing the ClutterFixed value
inside a closure struct instead of trying to stuff it into a pointer.
Bug 1167 - clutter_effect_rotate improperly clamps negative angles
The rotation angle properties had a minimum value of 0.0 but the
rotation works fine with a negative value so the limitation is
unnecessary. This makes rotation using the ClutterAnimation API more
flexible because it was previously not possible to rotate
counter-clockwise from the 0.0 position.
When rendering a clone before this commit the clone's opacity was
combined with the opacity of the source but this is not usually the
desired effect. Instead the clone's opacity (combined with its
parents) should completely override the opacity of the source.
cogl_paint_init was a bit too miscellaneous; it mainly cleared the color, depth
and stencil buffers but arbitrarily it also disabled fogging and lighting.
It no longer disables lighting, since we know Cogl never enables lighting and
disabling of fog is now handled with a seperate function.
Since I noticed cogl_set_fog was taking a density argument documented as
"Ignored" I've also added a mode argument to cogl_set_fog which exposes the
exponential fog modes which can make use of the density.
Bug #1453 - Asynchronous texture loading can starve cpu.
Add a mutex that is held in the loader threads during the image
decoding. We were spawning and starting a thread for each asynchronously
loaded texture. This can cause cpu / memory starvation when many pixbuf
loaders allocate their temporary memory at the same time.
Also added -fno-strict-aliasing to MAINTAINER_CFLAGS in configure.ac
to avoid incorrect warnings caused by the static mutex code.
The clutter_get_current_event_time() function will return the event
timestamp coming from a Clutter event. Clutter might synthesize or
throttle events, so the function cannot be used when dealing with
backend-specific use cases.
The X11 backend is the only backend supported by Clutter that makes
use of timestamps, so it's altogether fitting that it should come
with a specific function to deal with the timestamps of the X events.
In generate_enter_leave_events it passes the device pointer from the
event to set_motion_last_actor but it was reading it directly from
event->motion.device. However the function is also used to process
button events so it would read from the wrong location in this case.
The device location in the union happened to be in the same place as
the click_count field of ClutterButtonEvent so it only mattered if the
click_count is non-zero. The X11 backend doesn't set this but
Clutter-GTK does so it was causing a crash.
Bug 1178 - No enter / leave events on actors when pointer leaves the
stage window
This patch causes the Win32 backend to emit CLUTTER_LEAVE events when
a WM_MOUSELEAVE event is received in the same way that f505536 does
for the X11 backend.
Windows will only send WM_MOUSELEAVE events if they are previously
requested using TrackMouseEvent so this needs to be called whenever
the mouse enters the window. There is no WM_MOUSELEAVE event but we
can detect when the mouse enters because we get a WM_MOUSEMOVE event.
All GL functions that are defined in a version later than 1.1 need to
be called through cogl_get_proc_address because the Windows GL DLL
does not export them to directly link against.
Bug 1178 - No enter / leave events on actors when pointer leaves the
stage window
The patch is mostly thanks to Johan Bilien with small modifications
based on suggestions by Owen Taylor.
The X11 backend now listens for enter and leave notifications. Leave
notifications get translated directly to a CLUTTER_LEAVE
event. Clutter can detect these special events because the source
actor is NULL in which case it sets the source actor to the last known
actor and then sets the last known actor to NULL.
Enter notifications just get translated to CLUTTER_MOTION events which
will cause Clutter to generate an enter event through the usual code
path.
clutter-x11.h declares clutter_x11_set_display twice, which means code using
clutter doesn't compile cleanly with -Wredundant-decls.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Moved clutter_stage_set_title() before call to clutter_actor_realize()
to ensure _NET_WM_NAME is set early enough for use by interested WM's,
Signed-off-by: Shane Bryan <shane.bryan@linux.intel.com>
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The clutter_get_current_event_time() is a global function for
retrieving the timestamp of the current event being propagated
by Clutter. Such function avoids the need to propagate the
timestamp from within user code.
ClutterListModel has been added as a terminal class in case we
decided to change API or implementation.
Apparently, it's a lot more resilient than we expected -- or far
too few are using it and reporting bugs. Once common issue, though,
is that it cannot be subclassed, hence it's fairly limited in its
usage.
In the hope that more developers will start using it, here's a
patch that makes ListModel a fully subclassable object.
May whatever god you believe in have mercy on your eventual soul.
The main COGL header file is generated at configure time. If something
changes in the template, though, the file will not be regenerated.
Adding cogl.h to the BUILT_SOURCES list will allow the regeneration to
happen.
Key events can come from different input devices, so they need a
ClutterInputDevice member like the pointer-related events. This
field is reserved for future use.
To convert from a TextureQuality to a COGL filter enumeration
we need to perform two function calls. Since we need both filters
when calling cogl_texture_set_filters() we can safely consolidate
the two conversion functions into one. This also allows using a
shorter function name, thus reducing the lenght of the lines
involved and, more importantly preventing Emmanuele from crying.
This hides a number of internal structs and enums from the docs, and moves
some functions to more appropriate sections as well as misc description
updates (mostly for the vertex buffer api)
Fixes some blending issues when using color arrays since we were
conflicting with the cogl_enable state + fixes a texture layer
validation bug.
Adds a basic textured triangle to test-vertex-buffer-contiguous.
Otherwise if the dipose method is called twice the material will
already be destroyed when texture_free_gl_resources is called so it
will issue a warning about an invalid handle.
When the quad log contains multiple textures (such as when a sliced
texture is drawn) it dispatches the log with multiple calls to
flush_quad_batch and walks a pointer along the list of vertices.
However this pointer was being incremented by only one vertex so the
next quad would be drawn with three of the vertices from the last
quad.
The 'no_slice' property means the texture should never be sliced. We
want cogl to create a texture with any amount of waste so we pass
max_waste as -1. However this got broken in commit 168d55 so that the
meaning got negated (no_slice enabled slicing).
The current code that handles the invariant that the new parent
of an actor needing a layout should also be queued for relayout
is hitting the short-circuiting we do in the queue_relayout()
method.
In order to fix this we can forcibly set the actor to need a
width/height request and an allocation; then we queue a relayout
on the parent.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The :alignment property is prone to generate confusion: developers
will set it thinking that the contents of a ClutterText will
automagically align themselves.
Instead of using the generic term :alignment, and following the
GTK+ convention, we should use a more specific term, conveying the
actual effect of the property: alignment of the lines with respect
to each other, and not to the overall allocated area.
See bug 1428:
http://bugzilla.openedhand.com/show_bug.cgi?id=1428
When setting the cogl texture, the ClutterTexture takes a reference to
the Cogl handle but the material takes an additional reference when it
is set as the layer. Nothing was unrefing the temporary reference so
the textures were being leaked.
The opacity of the source actor when painted from the clone should be
combined with the opacity of the clone as well as the clone's parents,
instead of just the parents.
The quad drawing code keeps track of the number of texture units that
have the tex coord array enabled so that in the next call it can
disabled any that are no longer enabled. However it was using 'i+1' as
the count but 'i' is already set to 'n_layers' from the previous for
loop.
Therefore it was disabling an extra texture unit. This doesn't
normally matter but it was causing GLES 2 to pointlessly realize an
extra unit.
- In cogl-material.h it directly sets the values of the
CoglMaterialLayerCombineFunc to some GL_* constants. However these
aren't defined in GLES 2 beacuse it has no fixed function texture
combining. Instead the CGL_* versions are now used. cogl-defines.h
now sets these to either the GL_* version if it is available,
otherwise it directly uses the number.
- Under GLES 2 cogl-material.c needs to access the CoglTexture struct
so it needs to include cogl-texture-private.h
- There are now #define's in cogl-gles2-wrapper.h to remap the GL
function names to the wrapper names. These are disabled in
cogl-gles2-wrapper.c by defining COGL_GLES2_WRAPPER_NO_REMAP.
- Added missing wrappers for glLoadMatrixf and glMaterialfv.
- Renamed the TexEnvf wrapper to TexEnvi because the latter is used
instead from the material API.
Cogl previously tried to cache the currently bound texture when
drawing through the material API to avoid excessive GL calls. However,
a few other places in Cogl and Clutter rebind the texture as well so
this can cause problems.
This was causing shaped windows to fail in Mutter because
ClutterGLXTexturePixmap was binding a different texture to update it
while the second texture unit was still active which meant the mask
texture would not be selected when the shaped window was drawn
subsequent times.
Ideally we would fix this by providing a wrapper around glBindTexture
which would affect the cached value. The cache would also have to be
cleared if a selected texture was deleted.
If the source of a ClutterClone does not have a parent actor the
clone will be unable to paint it because it's missing an allocation.
A simple way to fix it is to make the ClutterClone act like a
"foster parent": when it is allocated it will check if the source
actor has a parent and if not it will allocate the source with
its preferred size.
Final bit of integration between ClutterActor and Pango: a simple
method for creating a PangoLayout, pre-filled with text and ready
to be rendered using cogl_pango_render_layout().
This should make writing new Actors rendering custom text in their
paint() implementation easy and reliable.
ClutterText should merge the PangoAttributes set by using
clutter_text_set_attributes() with the attributes generated by
parsing Pango markup.
For this to work we must parse the markup and merge the attributes
we get out of pango_parse_markup() with the attributes set by
the user.
Setting the markup or the attributes on an editable text should
not work for the time being.
When processing a motion event, we need to spin the event loop two extra
times to ensure that any enter/leave events that might have been synthesized
are pumped through (otherwise they end up being pushed down the queue and
never processed).
This tries to make a number of files more comparable with the intention of
moving some code into cogl/common/
Files normalized:
cogl.c
cogl-context.c
cogl-context.h
cogl-texture.c
Someone not sure which cogl_color_set_from_* version is "best" may use
set_from_4d because taking doubles implies higher precision. Currently
it doesn't have any advantage.
This makes it consistent with cogl_rectangle_with_{multi,}texture_coords.
Notably the reason cogl_rectangle_with_{multi,}texture_coords wasn't changed
instead is that the former approach lets you describe back facing rectangles.
(though technically you could pass negative width/height values to achieve
this; it doesn't seem as neat.)
The code is #if 0 guarded, but when uncommented it outlines all drawn
rectangles with an un-blended red, green or blue border. This may e.g. help
with debugging texture slicing issues or blending issues, plus it looks quite
cool.
Bug 1349 - Using the anchor point to set the scale center is messy
The branch adds an extra center point for scaling which can be used
for example to set a scale about the center without affecting the
position of the actor.
The scale center can be specified as a unit offset from the origin or
as a gravity. If specified as a gravity it will be stored as a
fraction of the actor's size so that the position will track when the
actor changes size.
The anchor point and rotation centers have been modified so they can
be set with a gravity in the same way. However, only the Z rotation
exposes a property to set using a gravity because the other two
require a Z coordinate which doesn't make sense to interpret as a
fraction of the actor's width or height.
Conflicts:
clutter/clutter-actor.c
When drawing a texture with waste in _cogl_multitexture_unsliced_quad
it scales the texture coordinates so that the waste is not
included. However the formula was the wrong way around so it was
calculating as if the texture coordinates are ordered x1,x2,y1,y2 but
it is actually x1,y1,x2,y2.
When the texture is sliced it drops back to a fallback function and
passes it the texture coordinates from the rectangle. However if no
tex coords are given it would crash. Now it passes the default
0.0->1.0 tex coords instead.
If no texture coordinates are given then texture_unsliced_quad tries
to generate its own coordinates. However it also tries to read the
texture coordinates to check if they are in [0.0,1.0] range so it will
crash before it reaches that.
Instead of having a separate set of font options that override the
backend options when clutter_set_font_flags is called, it now just
directly sets the backend font options. So now the font flags are just
a convenience wrapper around the backend font options.
This also makes the ClutterText labels automatically update when the
font flags are changed because they will respond to the 'font-changed'
signal from the backend.
Whenever a ClutterText is created it now connects to the font-changed
signal. When it is emitted the layout cache is dirtied and a relayout
is queued. That way changes to the font options or resolution will
cause an immediate update to the labels in the scene.
* generic-actor-clone:
Remove CloneTexture from the API
[tests] Clean up the Clone interactive test
Rename ActorClone to Clone/2
Rename ActorClone to Clone/1
Improves the unit test to verify more awkward scaling and some corresponding fixes
Implements a generic ClutterActorClone that doesn't need fbos.
Conflicts:
clutter/cogl/gl/cogl-texture.c
clutter/cogl/gles/cogl-primitives.c
* cogl-material:
clutter-{clone-,}texture weren't updating their material opacity.
Updates GLES1 support for CoglMaterial
Normalizes gl vs gles code in preperation for synching material changes
Removes cogl_blend_func and cogl_alpha_func
Fully integrates CoglMaterial throughout the rest of Cogl
[cogl-material] Restore the GL_TEXTURE_ENV_MODE after material_rectangle
[cogl-material] Make the user_tex_coords parameter of _rectangle const
[test-cogl-material] Remove return value from material_rectangle_paint
Add cogl-material.h and cogl-matrix.h to libclutterinclude_HEADERS
[cogl-material] improvements for cogl_material_rectangle
[cogl-material] Adds a cogl_material_set_color function
[cogl-material] Some improvements for how we sync CoglMaterial state with OpenGL
[cogl-material] Converts clutter-texture/clutter-clone-texture to the material API
[doc] Hooks up cogl-material reference documentation
Updates previous GLES multi-texturing code to use CoglMaterial
Adds a CoglMaterial abstraction, which includes support for multi-texturing
[doc] Hooks up cogl-matrix reference documentation
Adds CoglMatrix utility code
[tests] Adds an interactive unit test for multi-texturing
[multi-texturing] This adds a new cogl_multi_texture API for GL,GLES1 + GLES2
ClutterClone supercedes ClutterCloneTexture, since it can clone
every kind of actor -- including composite ones.
This is another "brain surgery with a shotgun" kind of commit: it
removes CloneTexture and updates every test case using CloneTexture
to ClutterClone. The API fallout is minimal, luckily for us.
The hope is that this function makes it easier to extend the font
settings with more flags without having to add a function for every
setting.
A new flag for enabling hinting has been added. If set, this changes
the font options on the global PangoContext and any newly created
PangoContexts. The options are only set if the flag is changed from
the default so it won't override any detailed setting chosen by the
backend.
cogl_set_source_color4ub was previously used to set a transparent
white source color according to the actors opacity, but since
cogl_set_source_color now always implies you want a solid fill we
we use cogl_material_set_color4ub to achieve the same result.
This updates cogl/gles in line with the integration of CoglMaterial throughout
Cogl that has been done for cogl/gl.
Note: This is still buggy, but at least it builds again and test-actors works.
Some GLES2 specific changes were made, but these haven't been tested yet.
This changes all GLES code to use the OpenGL function names instead of
the cogl_wrap_* names. For GLES2 we now define the OpenGL name to point
to the wrapper, as opposed to defining the wrapper to point to the
OpenGL name for GLES1.
I've also done a quick pass through gl/cogl.c and gles/cogl.c to make
them more easily comparable. (most of the code is now identical)
The GL blend function and alpha function are now controlled by the material
code, and even internally Cogl should now be using the material API when
it needs control of these.
This glues CoglMaterial in as the fundamental way that Cogl describes how to
fill in geometry.
It adds cogl_set_source (), which is used to set the material which will be
used by all subsequent drawing functions
It adds cogl_set_source_texture as a convenience for setting up a default
material with a single texture layer, and cogl_set_source_color is now also
a convenience for setting up a material with a solid fill.
"drawing functions" include, cogl_rectangle, cogl_texture_rectangle,
cogl_texture_multiple_rectangles, cogl_texture_polygon (though the
cogl_texture_* funcs have been renamed; see below for details),
cogl_path_fill/stroke and cogl_vertex_buffer_draw*.
cogl_texture_rectangle, cogl_texture_multiple_rectangles and
cogl_texture_polygon no longer take a texture handle; instead the current
source material is referenced. The functions have also been renamed to:
cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords
and cogl_polygon respectivly.
Most code that previously did:
cogl_texture_rectangle (tex_handle, x, y,...);
needs to be changed to now do:
cogl_set_source_texture (tex_handle);
cogl_rectangle_with_texture_coords (x, y,....);
In the less likely case where you were blending your source texture with a color
like:
cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */
cogl_texture_rectangle (tex_handle, x, y,...);
you will need your own material to do that:
mat = cogl_material_new ();
cogl_material_set_color4ub (r,g,b,a);
cogl_material_set_layer (mat, 0, tex_handle));
cogl_set_source_material (mat);
Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use
cog_rectangle_with_texure_coords since these are the coordinates that
cogl_rectangle will use.
For cogl_texture_polygon; as well as dropping the texture handle, the
n_vertices and vertices arguments were transposed for consistency. So
code previously written as:
cogl_texture_polygon (tex_handle, 3, verts, TRUE);
need to be written as:
cogl_set_source_texture (tex_handle);
cogl_polygon (verts, 3, TRUE);
All of the unit tests have been updated to now use the material API and
test-cogl-material has been renamed to test-cogl-multitexture since any
textured quad is now technically a test of CoglMaterial but this test
specifically creates a material with multiple texture layers.
Note: The GLES backend has not been updated yet; that will be done in a
following commit.
Step two: rename the object and its methods.
While we're at it, adhere more strictly to the coding style
practises; rename :clone-source to :source; add a setter method
for the :source property; take a reference on the source actor
to avoid it disappearing while we're still accessing it.
Like clutter_timeline_advance(), calling advance_to_marker() will
not emit ::new-frame for the frame where the marker is set on; it
will also not emit ::marker-reached for the marker we are advancing
the timeline to.
When calling clutter_timeline_advance(), a timeline will not emit
the ::new-frame signal for the frame we are advancing to, as this
would break the invariants of the timeline behaviour.
The documentation should make this clear.
The clutter_timeline_new() constructor is not checking the full
range of the passed :fps property. The ParamSpec of the property
and the setter method perform this check, so the ctor should as
well.
The "is-timeline-complete" condition is pretty long, spanning
four lines and four logical sub-conditions. It is possible to
neatly move it into an is_complete() function and make the
code more readable.
The long description of the ClutterTimeline class is very C
developer-oriented. Since many language bindings will refer to
the C API reference we should probably be more verbose and
language agnostic -- at least in the class description.
The methods documentation also requires a little pass to increase
the consistency of the terminology, the grammar and the syntax.
Finally, comments never killed anyone.
Since not every timeline will have markers it's unfair to make
all of them crete two empty hash tables (with a preallocated
fixed size).
This commit moves the responsibility of creating the hash tables
to the marker API itself, and adds the relative checks.
Since we are using milliseconds granularity to integrate timelines
with the GLib main loop, we cannot allow values of the :fps
property bigger than 1000. This means validating the fps value both
in the GParamSpec and the clutter_timeline_set_speed() accessor
function.
This should also fix floating point exceptions when trying to
perform "n_frames = milliseconds / (1000 / fps)".
See bug 1354:
http://bugzilla.openedhand.com/show_bug.cgi?id=1354
If a ClutterTexture does not sync size, it should be possible to
change the texture size without causing a relayout.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
There is some GL work and a repaint anytime the clip is set
or unset, so avoid that if it isn't really changed.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The stage will usually be painted before the first ConfigureNotify
arrives so we need to set the SYNC_MATRICES flag to ensure that the
viewport will be correct for that paint. Unfortunately this means that
the viewport will be set again once the ConfigureNotify is received
but compared to rendering an initial invalid scene I think it is the
lesser of two evils.
When calling clutter_behaviour_apply() the new actor is prepended
to the list of actors to which a behaviour is applied; this breaks
the rest of methods working on the actors list, e.g.:
# adding actors
apply(actor_0);
apply(actor_1);
apply(actor_2);
# expected: [ actor_0, actor_1, actor_2 ]
[ actor_2, actor_1, actor_0 ] = get_actors();
# expected: actor_2
actor_0 = get_nth_actor(2);
This commit fixes the inconsistency in the returned values.
The Cogl primitives broke for GLES 1.1 and 2 after the cogl-float
branch merge.
CoglPathNode was still being declared as GLfixed for the GLES backend
but it was being filled with float values so they were all ending up
as numbers < 1.
glDrawArrays was being called with GL_FIXED so this has been changed
to GL_FLOAT.
The scanline rasterizer had a leftover hardcoded ClutterFixed constant
to add a small amount to the height of each line.
struct _CoglFloatVec2 has been removed because it is no longer used
anywhere.
The special check to invert the progress when the timeline direction
is backwards is not necessary because the actual frame number will be
decreasing in that case. Inverting just makes it progress forwards
again.
This is more apparent since the float-alpha-value branch merge because
the clutter_linear function directly returns the value from
get_progress. For example in test-depth, the animations loop instead
of oscillating back and forth.
The documentation has been updated to reflect the fact that the anchor
point will move when the actor changes size if it was specified using
a gravity value. The new functions for setting the scale center and z
rotation gravity are also documented.
Currently only the Z axis rotation center can be set using a gravity
but the other rotations also store their center as an AnchorCoord for
consistency. Specifying the center as a gravity makes less sense for
the other axes because the actors have no size along the Z axis.
The rotation angles are now stored as gdoubles and the fixed point *x
entry points have been removed.
The Z rotation can now be set with a gravity center using the
following new function:
void clutter_actor_set_z_rotation_from_gravity (ClutterActor *self,
gdouble angle,
ClutterGravity gravity);
This sets the center point from which the scaling will occur. This can
be used insetad of the anchor point to avoid moving the actor. Like
the anchor point, it can be specified as either a coordinate in units
or a gravity enum.
To set the center you can use two new variants of set_scale:
clutter_actor_set_scale_full (ClutterActor *self,
gdouble scale_x,
gdouble scale_y,
int center_x,
int center_y);
or
clutter_actor_set_scale_with_gravity (ClutterActor *self,
gdouble scale_x,
gdouble scale_y,
ClutterGravity gravity);
The ClutterFixed variants of the set_scale functions have been removed
and the scale value is now always stored as a double.
This makes it so when the anchor point is set using a gravity enum
then the anchor point moves when the actor changes size. A new
property is added for the anchor point gravity. If the anchor point is
set from gravity then the position in units can also be retreived with
the regular API.
A new union type is used to store the anchor point with helper
accessor functions. The hope is these can be reused for the scale and
rotation center points.
The clutter_stage_get_resolution() and fixed-point API are just
shorthands for:
clutter_backend_get_resolution (default_backend);
And as such do not fit at all in the ClutterStage class. The only
reason for their existence was the ClutterUnit conversion macros,
which have now been fixed to use the default backend through a
function call instead.
Thus, we can safely remove the stage entry points.
The maintainer compiler flags we use trigger warnings and errors
in the autogenerated code that gtk-doc creates to scan the header
and source files. Since we cannot control that, and we must run
a distcheck with both --enable-gtk-doc and --enable-maintainer-flags
turned on, we need to use less-strict compiler flags when inside
the doc/reference subdirectories.
The way to do this is to split the maintainer compiler flags into
their own Makefile variable, called MAINTAINER_CFLAGS. The we
can use $(MAINTAINER_CFLAGS) in the INCLUDES or _CFLAGS sections
of each part of the source directories we wish to check with the
anal retentiveness suited for maintainers.
* float-alpha-value:
[script] Parse easing modes by name
[docs] Update the easing modes documentation
[animation] Implement new easing functions
[animation] Move the alpha value to floating point
Since we allow overriding the paint() implementation through the
::paint signal to change the way an actor is being painted, we
should also allow overriding the pick() implementation using a
::pick signal.
The script converted calls to COGL_FIXED_MUL(x,y) to (x*y). However
this fails for cases like this:
COGL_FIXED_MUL(a + b, c)
which become
(a + b * c)
The meaning of this is of course different because multiplication has
a higher precedence than addition.
This was causing breakages in cogl_texture_quad_sw when the vertex
coordinates are not in increasing order. This was the case in
test-backface-culling when NPOTs are not available.
Improve clutter_sinx() by replacing the low precision CFX_SIN_STEP
with a multiply/divide pair. This reduces the maximum error from
1.8e-04 to 2.4e-05.
http://bugzilla.openedhand.com/show_bug.cgi?id=1314
Based on a patch by Owen W. Taylor <otaylor@fishsoup.net>
Since a pick is really a paint operation, we can safely get
the allocation box, instead of using get_width() and get_height().
This should help cutting down the function calls. If we were
feeling adventurous, we could even use the allocation directly
from the private data structure.
Based on a patch by Gwenole Beauchesne <gbeauchesne@splitted-desktop.org>
Compute the value of the camera distance as exactly half the xx
component of the projection matrix. The heuristically derived
value for 60 degrees was off by about 0.016%, causing noticeable
blurring, and other field of view angles which didn't have the
heuristic adjustment off by much more.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
If an actor is not set as visible, or if it is in a section of
the scenegraph that it's set as not visible (e.g. one of the
parents is not visible) then we should not queue a redraw for
it.
Patch based on code from Michael Boccara <michael@graphtech.co.il>
The intention behind ::queue-redraw is to be able to block the
default handler by attaching a callback and calling one of the
g_signal_stop_emission variants.
However this doesn't work, because ::queue-redraw has the
G_SIGNAL_RUN_FIRST flag instead of G_SIGNAL_RUN_LAST.
The GValue and GParamSpec integration of ClutterUnit was still
using the old, fixed-point based logic.
Storing ClutterUnits in a GValue should use floating point values,
and ClutterParamSpecUnit should follow suit.
Instead of recomputing the number of units needed to fit in
an em each time clutter_units_em() is called, we can store this
value into the default Backend along with the resolution and
font name. The value should also be updated each time the
resolution and font are changed, to keep it up to date.
The coordinates of each ButtonEvent are relative to the stage that
received the event, so we should document this in the structure
annotation.
It should also be mentioned that the coordinates can be transformed
into actor-relative coordinates by using transform_stage_point().
An em is a unit of measurement in typography, equal to the point
size of the current font.
It should be possible to convert a value expressed in em to
ClutterUnits by using the current font and the current DPI as
stored by the default backend.
The stage-with/height-percentage converters had been broken by
the multiple-stages support of Clutter 0.8. They are also made
useless by the fact that Units are now floating point values.
The millimeters and typographic points converters also depended
on the default stage, but they can be reworked to use the default
DPI coming from the default Backend instead.
Boolean arguments for functions are pretty evil and usually
lead to combinatorial explosion of parameters in case multiple
settings are added.
In the case of the COGL texture constructors we have a boolean
argument for enabling the auto-mipmapping; it is conceivable that
we might want to add more settings for a COGL texture without
breaking API or ABI compatibility, so the boolean argument should
become a bitmask.
The internals have not been changed: instead of checking for
a non-zero value, we check for a bitmask being set.
ClutterMedia was a rough cut at a simple media API; it needs some
re-evaluation before 1.0 in order to keep it simple to use, and
simple to implement.
- ClutterMedia:position
The position property accessors collide with the corresponding
ClutterActor methods, which make it impossible to bind them in
high-level languages:
video_texture.set_position()
video_texture.get_position()
In order to resolve the collision, we have to go through the
GObject properties API:
video_texture.set('position', value)
value = video_texture.get('position')
A :position in seconds is also a GStreamer-ism, and should rather
be converted to a :progress property, with a normalized value
between 0 and 1. the current position in seconds would then simply
be progress*duration. For non-seekable streams, 0.0 would always
be returned. This makes it easier to use the progress inside
animations, Timelines or ClutterPath instances.
- ClutterMedia:volume should be renamed to :audio-volume and normalized
as well, instead of being a floating point value between 0 and 100.
- ClutterMedia:buffer-percent should just be :buffer-fill and normalized
between 0.0 and 1.0
This better reflects the fact that the api manages sets of vertex attributes,
and the attributes really have no implied form. It is only when you use the
attributes to draw that they become mesh like; when you specify how they should
be interpreted, e.g. as triangle lists or fans etc. This rename frees up the
term "mesh", which can later be applied to a concept slightly more fitting.
E.g. at some point it would be nice to have a higher level abstraction that
sits on top of cogl vertex buffers that adds the concept of faces. (Somthing
like Blender's mesh objects.) There have also been some discussions over
particle engines, and these can be defined in terms of emitter faces; so some
other kind of mesh abstraction might be usefull here.
Okey; to summarise the changes...
We have converted Clutter and Cogl over to using floating point internally
instead of 16.16 fixed, but we have maintained the cogl-fixed API as a
utility to applications in case they want to implement their own optimizations.
The Clutter API has not changed (though ClutterFixed and ClutterUnit are now
internally floats) but all Cogl entry points have been changed to accept floats
now instead of CoglFixed.
To summarise the rationale...
There have been a number of issues with using fixed point though out Clutter
and Cogl including: lack of precision, lack of range, excessive format
conversion (GPUs tend to work nativly with IEEE floats) and maintainability.
One of the main arguments for fixed point - performance - hasn't shown
itself to be serious in practice so far since we seem to be more limited
by GPU performance and making improvements regarding how we submit data to
OpenGL[ES]/the GPU has had a more significant impact.
Ref: The recent multiple rectangle queuing changes + the
cogl-texture-agressive-batching branch which show significant performance
gains, and that recent tests on the ipodtouch (ARM + MBX) also showed no
loss of performance running with floats.
So finally; please forgive the inevitable fallout, this is a far reaching
change. There are still a few known issues with the fixed to float
conversion but enough works for all our conformance tests to pass, and the
remaining issues hopefully wont be too tricky to solve. For reference two
tags will be available either side of this change: "cogl-fixed-end" and
"cogl-float-start"
The easing modes for a ClutterAlpha can either be parsed by using
the enumeration "nickname" (the shorthand form of the enumeration
value) or by using the common naming policy used in other
animation frameworks, like:
easeInCubic
easeOutElastic
easeInOutBounce
The ClutterAlpha API reference page should also list the
easing modes Clutter provides by default, by showing the
curves used by each entry in the AnimationMode enumeration.
We can also remove the incomplete graph showing the old
alpha functions.
Instead of using our own homegrown alpha functions, we should
use the easing functions also shared by other animation frameworks,
like jQuery and Tween, in the interests of code portability.
The easing functions have been defined by Robert Penner and
are divided into three categories:
In Out InOut
Each category has a particular curve:
Quadratic
Cubic
Quartic
Quintic
Sinusoidal
Exponential
Circular
In addition, there are "physical" curves:
Elastic
Back (overshooting cubic)
Bounce (exponentially decaying parabolic)
Finally, the Linear curve is also provided as a reference.
The functions are private, and are meant to be used only
through their logical id as provided by the AnimationMode
enumeration.
The tests should be updated as well to match the new
easing functions.
The current Alpha value is an unsigned integer that can be used
implicitly as a fixed point value. This makes writing an alpha
function overshooting below and above the current range basically
impossible without complicating an already complex code, and
creating weird corner cases.
For this reason, the Alpha value should be defined as a floating
point normalized value, spanning a range between 0.0 and 1.0; in
order to allow overshooting, the valid range is extended one unit
below and one unit above, thus making it -1.0 .. 2.0.
This commit updates the various users of the ClutterAlpha API
and the tests cases.
This commit also removes all the current alpha functions exposed
in the public API.
To avoid clashing with all the scripted changes, clutter-fixed.h and
clutter-units.h were manually converted to internally use floats instead of
16.16 fixed numbers.
Note: again no API changes were made in Clutter.
To deal with all the corner cases that couldn't be scripted a number of patches
were written for the remaining 10% of the effort.
Note: again no API changes were made in Clutter, only in Cogl.
This is the result of running a number of sed and perl scripts over the code to
do 90% of the work in converting from 16.16 fixed to single precision floating
point.
Note: A pristine cogl-fixed.c has been maintained as a standalone utility API
so that applications may still take advantage of fixed point if they
desire for certain optimisations where lower precision may be acceptable.
Note: no API changes were made in Clutter, only in Cogl.
Overview of changes:
- Within clutter/* all usage of the COGL_FIXED_ macros have been changed to use
the CLUTTER_FIXED_ macros.
- Within cogl/* all usage of the COGL_FIXED_ macros have been completly stripped
and expanded into code that works with single precision floats instead.
- Uses of cogl_fixed_* have been replaced with single precision math.h
alternatives.
- Uses of COGL_ANGLE_* and cogl_angle_* have been replaced so we use a float for
angles and math.h replacements.
This simplifies the mucking about with the model-view matrix that was previously
done which improves its efficiency when scaling is necessary.
Notably: There should now be no performance advantage to using
ClutterCloneTexture as a special case clone actor since this method is just as
efficient.
The unit test was renamed to test-actor-clone.
Many use cases for clonning an actor don't require running a shader on the
resulting clone image and so requiring FBOs in these cases is overkill and
in-efficient as it requires kicking and synchronizing a render for each clone.
This approach basically just uses the paint function of another actor to
implement the painting for the clone actor with some fiddling of the model-
view matrix to scale according to the different allocation box sizes of
each of the actors.
A simple unit test called test-actors2 was added for testing.
It's more sensible to use 2^n-1 for a max tile-waste value rather
than 2^n, so change the value default from 64 to 63. Example:
191 and 192 will both be sliced to 128+64 rather than having
191=>128+64, 192=>256.
http://bugzilla.openedhand.com/show_bug.cgi?id=1402
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
While X11 Pixmap and Window types only have 32-bits of data, they
are actually 'unsigned long'. Change the "window" and "pixmap"
property of ClutterX11TexturePixmaps to be ulong.
This fixes 64-bit bugs where ClutterGLXTexturePixmap passed a
reference to Pixmap to g_object_get("pixmap", &pixmap, ...);
http://bugzilla.openedhand.com/show_bug.cgi?id=1405
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Since the stage in the EGL native backend only has one size, and it
is determined at realization, we can simply set the SYNC_MATRICES
private flag and let _clutter_stage_maybe_setup_viewport() set up
the GL viewport at the first redraw.
Both clutter_alpha_new_with_func() and clutter_alpha_set_func()
will not register a global alpha function, so we need to update
the documentation to explicitly say so.
The animation mode parameters and properties are now slightly
anonymous unsigned longs, so we need to clarify in the documentation
that the user should either pass a ClutterAnimationMode value or
the result of registering an alpha function.
The animation mode symbolic id might come from the AnimationMode
enumeration or from the clutter_alpha_register_*() family of
functions. For this reason, we should use a gulong instead of
ClutterAnimationMode whenever we have an "animation mode" parameter
or property.
In order to unify alpha functions and animation modes in ClutterAlpha
we should be able to register alpha functions and get a logical id
for them; the logical id will then be available to be used by
clutter_alpha_set_mode().
The registration requires API changes in ClutterAlpha constructors
and methods. It also provides the chance to shift ClutterAlpha
towards the use of animations modes only, and to alpha functions
as a convenience API for language bindings alone.
It looks like the changes to cogl-gles2-wrapper.h were accidentally
committed to the actual file instead of the patch in commit
de27da0e. This commit moves the changes back into the patch so
cogl-gles2-wrapper.h is reverted back to master.
The patches have been updated to apply cleanly.
The patches for the g_warnings in clutter-actor.c have been removed
because master now uses CLUTTER_UNITS_FORMAT so they aren't
necessary. The clutter-units.h patch now sets CLUTTER_UNITS_FORMAT to
'f'.
The changes from the GL version of cogl-texture.c have been mirrored
in the GLES version. This adds the cogl_texture_new_from_bitmap
function and fixes the build errors.
If you try to use the CLUTTER_ACTOR_IS_* macros defined in ClutterActor
like this:
typedef struct { unsigned int reactive : 1; } foo_t;
foo_t f; f.reactive = CLUTTER_ACTOR_IS_REACTIVE (actor);
It will blow up because while the macros evaluate to 0 they can also
evaluate to non-zero values. Since most of the boolean flags in
Clutter and Clutter-based code are going to be stored like in the
example above, we should change the macros and let them evaluate
stricly either to 0 or to 1.
The Effects API and all related symbols have been superceded by
the newly added Animation API and clutter_actor_animate().
This commit removes the Effects implementation, the documentation
and the interactive test/example code.
The ::load-finished signal is emitted only when loading a texture
using clutter_texture_set_from_file(). Since this breaks user
expectations and consistency, we should also emit ::load-finished
when loading a texture from image data.
* async-textures:
Whitespace fixes in ClutterTexture
[async-loading] Do not force the texture size on async load
[async-loading] Update asynchronous image loading
Add API for extracting image size from a file
Update/clean and apply the async-texture patch from bug #1144
Since Clutter changed to using a layout scheme the handling_configure
flag no longer works because the allocate method is not invoked
immediately during the call to set_size from the ConfigureNotify
handler. However it is also no longer neccessary because the resizes
are effectively batched up until a relayout is run so it won't cause
an infinite loop of resize and notify events anyway.
* animation-improvements:
[docs] Add ClutterAnimatable to the API reference
Add license notice to ClutterAnimation files
[docs] Update the ClutterAnimation section
[animation] Extend ClutterAnimation support to all objects
[animation] Use ClutterAnimatable inside Animation
[animation] Add ClutterAnimatable
[animation] Allow registering custom progress function
[animation] Interval::compute_value should return a boolean
Animate ClutterColor properties
ClutterText already has code to try to preserve the x position when
moving up or down. A target x-position is stored and the cursor is
positioned at the nearest point to that in the appropriate line when
up or down is pressed. However the target position was never cleared
so it would always target the x-position of the cursor from the first
time you pressed up or down.
To fix this the patch clears the target position in set_position and
then sets it after the call in real_move_up/down. That way pressing
up or down sets the target position and any other movement will clear
it.
To get an index for the pixel position in the line
pango_layout_line_x_to_index is used. However when x is greater than
the length of the line then the index before the last grapheme is
returned which was causing it to jump to the penultimate
character. The patch makes it add on the trailing value so that it
will jump to the last character.
The old function ended up returning the length of the string when pos
was zero. This caused it to insert characters at the end when the
cursor was at the beginning of the string.
If an unbound control key is pressed (such as Ctrl+R) it would insert
a rectangle into the text.
Also zero is considered a valid unicode character by
g_unichar_validate so pressing a key such as shift would cause the
current selection to be deleted. The character isn't actually inserted
because insert_unichar disallows zeroes.
The GLES 2 wrapper needs to set up some state before each
draw. Previously this was acheived by wrapping glDrawArrays. Since the
multiple-texture-rectangle branch merge, glDrawElements is used
instead so we also need a wrapper for that.
It was also directly calling glBindTexture. GLES 2 uses a wrapper for
this function so that it can cope with GL_ALPHA format textures. The
format of the current texture needs to be stored as well as the target
and object number for this to work.
The BindingPool constructor should only check for duplicate pools
and then set the :name constructor-only property. If a BindingPool
is created without a name we also make a fuss about it.
It is also possible to simply dispose of a binding pool using
g_object_unref(), as long as it has been created by using
clutter_binding_pool_new() or directly with g_object_new(). Only
BindingPools attached to a class are not owned by the user.
ClutterBindingPool is already "problematic" in terms of memory
management for language bindings and gobject-introspection. It
also lacks a GType.
Turning ClutterBindingPool into a GBoxed would not make much
sense, since it does not adhere to the copy/free semantics. It
could be referenced/unreferenced, but in that case we can just
as well use GObject as a base class instead of reimplemeting
a ref-counted object and then boxing it.
ClutterBindingPool is obviously a terminal class, so we just
hide the instance and class structures.
The size of the texture as retrieved by the filename should
be set as the image size, not as the actor size, in order to
respect the :sync-size property.
When the asynchronous loading process terminates, we queue
a relayout so that the scene is updated.
Provide a main loop-based fallback to the asynchronous loading in
case the GLib threading support hasn't been enabled. This also
allows us to clean up the asynchronous loading machinery and have
it behave consistently across different scenarios.
Emit the ::load-finished even if the asynchronous loading from
disk was not enabled.
Finally, block clutter_texture_set_from_file() until we have an
image width and height, so that querying the texture actor size
after set_from_file() will still yield the correct result even
when asynchronous loading is set.
For the asynchronous loading we need a function call that parses
a file, given its path, and retrieves the image width and height.
This commit adds cogl_bitmap_get_size_from_file() to the CoglBitmap
API.
Add a ClutterStage::queue-redraw signal.
The purpose of this signal is to allow combining the Clutter redraw
idle with another redraw idle such as gtk's (or any other one really;
this is desirable anytime Clutter is not the only thing drawing to
a toplevel window).
To override the default, you would connect to ::queue-redraw and then
stop the signal emission.
By creating an ARGB texture for 24bpp pixmaps we were exposing an undefined
alpha channel to the blending and texture combine stages which resulted in
nasty artefacts. (This issue was seen on i945 + DRI2)
Since we only update the GL viewport when we receive a ConfigureNotify
event on X11, we also need a function to allow other toolkits to tell
a stage that the viewport should be updated.
This commit adds clutter_stage_ensure_viewport(), a function that simply
sets the private SYNC_MATRICES flag on the stage and then queues a
redraw.
This function should be called by libraries integrating Clutter with
other toolkits, like clutter-gtk or clutter-qt.
Continuation of the fix in commit 00a3c69868.
Instead of using a separate flag for the resize process, just
delay the setting of the CLUTTER_ACTOR_SYNC_MATRICES flag on the
stage to the point when we receive a ConfigureNotify event from
X11.
This commit will break the stage embedding into other toolkits.
There is a race condition when we resize a stage before showing
it on X11.
The race goes like this:
- clutter_init() creates the default stage and realize it, which
will cause a 640x480 Window to be created
- call set_size(800, 600) on the stage will cause the Window to be
resized to 800x600
- call show() on the stage for the first time will cause COGL
to set up an 800 by 600 GL viewport
- the Window will be mapped, which will cause X to notify the
window manager that the Window should be resized to 800x600
- the window manager will approve the resize
- X resizes the drawable to 800x600
To fix the race, we need to defer COGL from setting up the viewport
until we receive a ConfigureNotify event and the X server has resized
the Drawable.
In order to defer the call to cogl_setup_viewport() we add a new
private flag, CLUTTER_STAGE_IN_RESIZE; the flag is checked whenever
we need to change the viewport size along with the SYNC_MATRICES
private flag. Thus, cogl_setup_viewport() will be called only if
SYNC_MATRICES is set and IN_RESIZE is not set.
Some of the read-write properties of ClutterText were missing
an implementation in clutter_text_get_property(), as well as
the :position and :selection-bound properties being wrongly
converted from fixed point to integer, passing through floating
point values.
ClutterUnits should not be used interchangeably as, or with
ClutterFixed values. ClutterUnits should also not be assumed
to be integers.
This commit fixes the last few improper usages of ClutterUnit
values, and adds a CLUTTER_UNITS_FORMAT macro for safely printing
ClutterUnit values with printf().
* animatable-iface:
[docs] Add ClutterAnimatable to the API reference
Add license notice to ClutterAnimation files
[animation] Use ClutterAnimatable inside Animation
[animation] Add ClutterAnimatable
Instead of limiting the use of ClutterAnimation to ClutterActor
instances, relax the constraint to include all GObject classes.
ClutterAnimation is not using actor-specific API, since it is
only using properties.
The only actor-based API is the clutter_actor_animate() family
of functions.
ClutterAnimation should check if the object is implementing the
Animatable interface, and if so delegate to it the computation
of the value along the interval initial and final value, depending
on the progress.
The ClutterAnimatable interface is meant to be used by GObject
classes to override the value computation for an animatable
property within the boundaries of an interval.
It is composed of a single virtual function, animate_property();
its implementation will receive the ClutterAnimation used to
animate the object; the property name; the initial and final
interval values; and the progress factor as retrieved by the
Alpha object bound to the Animation instance.
A ClutterInterval can change the way the progress is computed
by subclassing and overriding the ::compute_value() virtual function.
It should also be possible to register a custom progress function
in the same way it is possible to register a custom transformation
function between two GValues.
This commit adds an internal, global hash table that maintains a
GType <-> progress function association; each ClutterInterval
will check if there is a progress function registered for the
GType of the initial and final values of the interval and, if
it has been found, it will call it to compute the value of the
interval depending on the progress factor.
If the computation of the interval value depending on the progress
was not successful, ClutterInterval::compute_value() should return
this information to the caller.
Bug 1289 - Draw multiple glyphs at once
The multiple-texture-rectangle branch adds a new Cogl texture function
called cogl_texture_multiple_rectangles which is used to draw multiple
rectangles out of a texture using a single GL call. This is
significantly faster than drawing the rectangles with individual calls
on some platforms. The Pango renderer now uses this to speed up
rendering.
The conflicts are just due to the whitespace fixes in cb569a5.
Conflicts:
clutter/cogl/gl/cogl-context.c
clutter/cogl/gl/cogl-context.h
clutter/cogl/gl/cogl-texture.c
An editable ClutterText should not use pango_layout_set_markup(),
as the contents of the text actor will not match the text.
Only read-only text actors should parse the contents for Pango
markup.
If the Text actor is neither editable nor has its cursor set
to visible, then we should not be ensuring the cursor position.
This fixes a failure in the conformance test unit for the
layout cache.
Merge branch 'text-actor'
* text-actor: (108 commits)
Re-align ClutterText header file
[text] Fix cursor sizing
Comments and whitespace fixes to ClutterText
[docs] Add newly added :single-line-mode accessors
Update the ignore file
[tests] Add text field interactive test
[text] Add single-line-mode to ClutterText
[text] Fix the deletion actions
[text] Use cached length when possible
[tests] Add unit for the ClutterText:password-char property
[docs] Update the Text section
[text] Coalesce text visibility and password character
Allow localizations to change the text direction
Clean up the update_pango_context() function
Pass the PangoContext, not the MainContext
Revert the logic of the PangoContext check
Remove the binding pool entry from the list
Remove BindingPool::list_actions()
Add ClutterActor::create_pango_context()
Rename the PangoContext creation functions
...
The cursor should be slightly smaller than the height of the actor, to
allow for painting a border. Let's pad it by 1 pixel on the top and 1
on the bottom.
Also, we should use the cursor size everywhere and not use hardcoded
magic numbers.
Allow using ClutterText as a single line text field. This is useful for
text fields that accept just a single line of contents by default, and
respond to the Enter key press to execute some action.
The :single-line-mode property enables this behaviour inside ClutterText
by clipping and scrolling the contents of the PangoLayout if they do
not fit the allocated width of the Text actor.
When using the delete-prev action from the end of the text we end
up either missing the first glyph we have to delete or falling
through the last one in the text.
This commit fixes both issues.
Since clutter_text_set_text() measures the length of the text
each time, we should use the cached length instead of recomputing
the text length each time. This should save us some time when
dealing with long, multi-byte texts.
The rest of Cogl expects the texture mode to be GL_MODULATE so it
needs to be restored after calling cogl_material_rectangle. Otherwise
cogl_texture_rectangle will fail to blend with the Cogl color properly
and all of the labels will be black.
Using two properties to set a password entry can be construed as
both cumbersome and a gtk-ism. And rightly so on both counts.
The :text-visible property has also conflicting semantics with the
:cursor-visible one: while the latter hides the cursor, the former
changes the display of the contents of the Text actor. It is, thus,
not a matter of "visibility" but of "rendering".
Instead of setting the :text-visible and :invisible-char properties
to have a password text field, the Text actor should just have a
single :password-char property holding a Unicode character. If the
value of the :password-char is non-zero, the Text actor will use the
Unicode character to render the contents of the text entry.
This commit removes the following methods:
clutter_text_set_text_visible()
clutter_text_get_text_visible()
clutter_text_set_invisible_char()
clutter_text_get_invisible_char()
And the following properties:
ClutterText:text-visible
ClutterText:invisible-char
In favour of:
clutter_text_set_password_char()
clutter_text_get_password_char()
And:
ClutterText:password-char
Thus making obvious what use the property and accessor methods are
for and simplifying the process of creating a simple password text
field to:
text = clutter_text_new ();
clutter_text_set_password_char (CLUTTER_TEXT (text), '*');
The locale translators of Clutter are also the ones that should set
the default direction of the text in a Clutter user interface.
This commit adds a translatable string that defines the direction
of the text; the translation authors will change it to the correct
value and that will determine the default direction.
The default text direction can be overridden by using the
CLUTTER_TEXT_DIRECTION environment variable, or by using the
--clutter-text-direction command line switch. In any other case,
the locale will determine the text direction, as it should.
It was always reading one pixel lower than requested. If y was 0 then
it would try to read below the lowest line.
Thanks to Geoff Gustafson for spotting.
Clutter has a set of command line options that are added to every
application by means of clutter_init() or by obtaining the Clutter
GOptionGroup and using g_option_context_parse(). Thus, every Clutter
application will automatically have an --help command line switch
showing the list of options and their description.
At the moment, Clutter does not enable localization of the help,
thus making it less than useful on non-English locales.
This patch enables the machinery to create a localization file and
load it when initializing Clutter, by means of the GLib macros and
locale.h API we already use.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
When updating the PangoContext with the current options (font name,
options, resolution) pass the PangoContext instead of the Clutter
MainContext structure pointer.
When removing a binding entry from the binding pool we should not
only remove it from the hash table, but also from the linked list
we use to iterate over inside the block/unblock_action() pair.
The clutter_binding_pool_list_actions() was not implemented. The
utility of a call listing all the action names is also debatable:
all the functions related to the key bindings take the key symbol
and modifiers -- except the block_action() and unblock_action()
pair.
Sometimes an actor needs to set specific font rendering options on
the PangoContext without changing settings for every other text-rendering
actor.
In order to do this, we need a new public method to create a Pango
context object -- preset with all the default settings -- owned by the
developer and not shared with the rest of Clutter.
This new method is called clutter_actor_create_pango_context(); while
it does not strictly depend on a ClutterActor, it is a good idea to
have it inside the ClutterActor API to map the current get_pango_context()
method and in case we start storing screen-specific data to the Actor
itself during the 1.x API cycle.
The _clutter_context_create_pango_context() should create a new
context; the function returning the PangoContext stored inside the
MainContext structure should be named _get_pango_context() instead.
The following functions are fixed:
clutter_animation_set_actor
clutter_animation_set_timeline
clutter_animation_set_alpha
This is related to bug 1392 which discusses the problem for
behaviour_set_alpha.
Bug 1392 - behaviour_set_alpha set same alpha twice lead to warning
and destroy the input alpha
The following functions are fixed:
clutter_actor_set_shader
clutter_alpha_set_timeline
clutter_behaviour_set_alpha
clutter_clone_texture_set_parent_texture
They either now reference the new value before destroying the old
value, or just return immediately if the values are the same.
It previously attempted to set the mode on the alpha using
clutter_animation_set_mode_internal, but this was setting the mode on
priv->alpha. At that point in the code priv->alpha is always NULL.
clutter_animation_set_mode_internal now takes a parameter to specify
which alpha to modify.
As of now, a key binding installed into a BindingPool is always there
and cannot be changed.
This is problematic for sub-classes trying to override the callback
or the action for a given key binding.
This commit adds the ability to override the closure for an existing
key binding inside a binding pool -- assumed the caller knows the
key symbol and modifiers used to install the key binding in the first
place.
Otherwise the call to clutter_alpha_set_func sets the mode back to
CLUTTER_CUSTOM_MODE so clutter_alpha_get_mode won't get back the same
value that was set.
* units-rework:
[texture] Do not mix fixed point and units values
[tests] Fix the actor detection
[units] Do not use fixed point and units interchangeably
The API has been changed to take an explicit length for the number of
texture coordinates passed, and it's now documented that if there are
more layers to the current material than the number of texture coords
passed, then default coordinates will be generated for the other
layers.
cogl_material_rectangle should now handle the case where a single
sliced texture is supplied as a material layer by falling back to
cogl_texture_rectangle. We are nearly at the point that
cogl_texture_rectangle could be deprecated. A few issues remain
though, such as not considering waste in cogl_material_rectangle.
The other colors of a material; such as the ambient and diffuse color are
only relevent when we can enable lighting. This adds a basic unlit
color property.
Later cogl_set_source_color can be integrated to either modify the color
of the current source material, or maintain a special singlton CoglMaterial
that is modified by calls to cogl_set_source_color and implicitly made
current.
This flattens the three functions: cogl_material_flush_gl_material_state,
.._flush_gl_alpha_func and .._flush_gl_blend_func into one:
cogl_flush_material_gl_state which doesn't takes a material handle. (the handle
is instead taken from the context.)
This has allows us to avoid re-submitting some state to OpenGL when the
material has not been replaced.
Note: Avoiding redundant state changes for material layers isn't dealt with
in this patch.
This converts clutter-texture to use the new CoglMaterial API instead of
cogl_texture_rectangle.
This change does not add multi-texturing support to Clutter, it's more about
aiming to deprecate cogl_texture_rectangle, and moving everything over to the
cogl_set_source(material); cogl_draw_somthing(); coding pattern.
With this patch, the code can be built with/without CoglMaterial support so we
can do comparisons of the old/new code for a transient period. (look for the
USE_COGL_MATERIAL define)
Note: The FBO paths haven't currently been tested, so they may need some
tweaks.
Removed trailing white space from the following files:
- clutter-clone-texture.c
- clutter-texture.c
- clutter-texture.h
- cogl/cogl-texture.h
- cogl/gl/cogl-context.c
- cogl/gl/cogl-texture.c
- cogl/gl/cogl-context.h
The direction of the text depends on the locale, and it is the
basic setting needed to enable internationalization of user
interfaces.
This commit allows setting the direction of the PangoContext instance
used by Clutter by using the CLUTTER_TEXT_DIRECTION environment
variable, or by passing the --clutter-text-direction command line
argument. Valid values are:
ltr - for left-to-right locales
rtl - for right-to-left locales
The default is LTR.
Ideally, this should be a value set by the localization teams on the
PO file, but this step requires some build system surgery to allow
the translation of the Clutter strings.
When calling clutter_backend_get_font_name(), if no default font
name has previously been set, we just set the default and return
a pointer to it - like we do for the font options.
Instead of storing the default font name and size as a pre-processor
macro, use the newly added ClutterBackend API to retrieve the current
default font from the backend.
The default backend stores some of the global defaults, like the
font options, text resolution, double click settings. It should also
store the default font name, to allow various text-based actors to
share the same settings.
When the font name changes, the ::font-changed signal is emitted,
to allow actors to pick up the change.
When the ClutterBackend notifies of changes in the resolution or
font options, update the PangoContext stored by Clutter's main
context. This allows changing the backend font-related settings at
runtime.
The PangoContext should be stored once, and inside the main
Clutter context. Each actor for which clutter_actor_get_pango_context()
has been called will hold a reference on the Pango context as well.
This makes it possible to update the text rendering for Clutter
by using only public API.
Rendering text inside an actor is pretty much impossible without
using internal API to create the various pieces like the PangoContext
and the font map.
Each actor should have the ability to create a PangoContext, which
is the only object needed to generate layouts and change the various
Pango settings.
This commit adds a clutter_actor_get_pango_context() function that
creates a PangoContext inside the ClutterActor private data and allows
the creation of PangoLayouts when needed. If the actor already
has a PangoContext, the same instance is returned.
The PangoContext is created only on demand.
The ClutterBackend instance at the moment lacks the ability to
notify runtime changes of the font options and the resolution.
For this reason, this commit adds a ::resolution-changed and a
::font-changed signals to the Backend class.
The ::resolution-changed signal is emitted when set_resolution()
is called with a different DPI; ::font-changed is emitted when the
cairo_font_options_t* changes from the default.
test-cogl-material now runs on GLES 1 using the PVR GLES1 SDK (though since
only 2 texture units are supported the third rotating light map doesn't show)
Note: It currently doesn't build for GLES 2.0
My previous work to provide muti-texturing support has been extended into
a CoglMaterial abstraction that adds control over the texture combine
functions (controlling how multiple texture layers are blended together),
the gl blend function (used for blending the final primitive with the
framebuffer), the alpha function (used to discard fragments based on
their alpha channel), describing attributes such as a diffuse, ambient and
specular color (for use with the standard OpenGL lighting model), and
per layer rotations. (utilizing the new CoglMatrix utility API)
For now the only way this abstraction is exposed is via a new
cogl_material_rectangle function, that is similar to cogl_texture_rectangle
but doesn't take a texture handle (the source material is pulled from
the context), and the array of texture coordinates is extended to be able
to supply coordinates for each layer.
Note: this function doesn't support sliced textures; supporting sliced
textures is a non trivial problem, considering the ability to rotate layers.
Note: cogl_material_rectangle, has quite a few workarounds, for a number of
other limitations within Cogl a.t.m.
Note: The GLES1/2 multi-texturing support has yet to be updated to use
the material abstraction.
Multitexturing allows blending multiple layers of texture data when texturing
some geometry. A common use is for pre-baked light maps which can give nice
lighting effects relativly cheaply. Another is for dot-3 bump mapping, and
another is applying alpha channel masks.
The dot-3 bump mapping would be really nice one day, but currently cogl doesn't
support lighting so that's not dealt with in this patch.
notable limitations:
- It can only texture rectangles a.t.m - and like cogl_texture_rectangle there
is no support for rotated texturing.
- Sliced textures are not supported. I think I've figured out how to handle
layers with different slice sizes at least for rectangular geometry, but I'm
not sure how complex it becomes once rotations are possible and texturing
arbitrary cogl_polygons.
- Except for this new API, cogl still doesn't know about more than one texture
unit, and so has no way of caching any enables related to other units. So that
things don't break it's currently necessary to disable anything to do with
additional units as soon as we are done with them which isn't ideal.
- No clutter API yet.
This follows the convention of GtkLabel/GtkEntry in GTK+ and the old
ClutterEntry.
It makes it easier to use strlen/strcmp etc on the output, since we can
assume that it is always a string.
This commit also updates the test unit for ClutterText to verify that
the clutter_text_get_text() function also returns an empty string when
a ClutterText actor has been created.
The clutter-private.h header already includes cogl-pango.h with
the correct inclusion path, because the main context stores a
pointer to the font map.
There is no need for clutter-text.c to include cogl-pango.h
again since it already includes clutter-private.h.
Instead of returning CLUTTER_X11_FILTER_CONTINUE always from
clutter_x11_handle_event() return CLUTTER_X11_FILTER_REMOVE if
the event was on a stage and translated to a ClutterEvent.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
The cairo pkg-config is crafted such that #include <cairo.h> should be used.
If building with a private build of cairo installed to a custom prefix
<cairo/cairo.h> can cause an un-intended cairo header to be picked up during
compilation due to /usr/include being in the header search path.
Our beloved math.h exports, from bits/mathcalls.h, a bare "y1" symbol.
Apparently, it's unthinkable for code including <math.h> to also declare
arguments or variable named "y0", "y1" and "yn".
Anyway, the quick fix already used elsewhere in Clutter's codebase is
to rename the colliding variables "y_0", "y_1" and "y_n" - and
obviously everything similar to them as well, using the same pattern.
The name of the parameter in the header and the one in the gtk-doc
annotation on top of a function must match.
Unfortunately, there is an index() function declared inside strings.h
which makes gcc complain for the "index" argument as soon as we
enable the extra compiler flags we use when distchecking.
Hence, we need to rename "index" to "index_" in the header and in
the source files.
* cairo-texture:
[cairo-texture] Remove the construct only restriction on surface size
[cairo-texture] Silently discard 0x0 surfaces
Re-indent ClutterPath header
Add a test case for the new cairo path functions
Add clutter_path_to_cairo_path and clutter_path_add_cairo_path
Warn instead of returning in the IN_PAINT check
Small documentation fixes
Print a warning when creating a cairo_t while painting
Do not set the IN_PAINT flag inside the Stage paint
Set the IN_PAINT private flag
[docs] Add ClutterCairoTexture to the API reference
Add ClutterCairoTexture
Require Cairo as a Clutter dependency
Conflicts:
Fix merge conflict in clutter/clutter-path.h
It is possible to change the surface size after construction with
clutter_cairo_texture_set_surface_size so it doesn't seem right to
restrict changing the properties.
clutter_cairo_texture_resize_surface_internal is called in a handler
for the notify signal. It is called there rather than directly in the
set_property handler so that changing both properties in a single
g_object_set will only cause one resize. The constructed override is
no longer needed.
resize_surface_internal will now bail out if the size of the surface
is already the right size.
Like we did for ClutterActor in commit cdb78ec4, fix ClutterTexture
usage of CoglFixed and ClutterUnit values without conversion between
the two types.
Clutter units are, at the moment, implemented as a value in fixed point
notation using the same format as CoglFixed. This is, though, an
implementation detail. For this reason, units should not be treated as
CoglFixed values and should be converted to and from fixed point using
the provided macros.
This commit updates the usage of units and fixed point values in
ClutterActor and rationalises some of the transformation code that
heavily relied on the equivalency between them.
Some drivers (e.g. Nvidia) get upset if you try to create multiple glx pixmaps
for the same server side pixmap object, even though you might have unique
client side names, we now avoid hitting this problem by destroying the current
glx pixmap early within clutter_glx_texture_pixmap_create_glx_pixmap.
The calculation for cubic bezier curves had an extra multiplication by
3 which was causing the curve to go over 1.0 very quickly. This had
the affect of making test-animation appear to complete much before the
completed signal is emitted.
The patch makes it cast to double before subtracting the original
value from the target value. Otherwise if the target value is less
than the original value then the subtraction will overflow and the
factor will be multiplied by a very large number instead of the
desired interval.
The problem is demonstrable using the border-width property of
ClutterRectangle.
The current CairoTexture can be created with a surface size of 0
by 0 pixels, but a warning will be printed.
Worse, the surface can be resized to be 0 by 0 pixels without a
warning. The :surface-width and :surface-height properties accept
a minimum value of 0, and not check is performed on either the
constructor or set_surface_size() parameters to enforce the "greater
than zero" rule.
The correct and consistent behaviour is to allow a 0 by 0 pixels
surface size everywhere; inside surface_resize_internal(), the
current surface will be destroyed and if either :surface-width or
:surface-height are set to 0, the resizing terminates.
Attempting to create a Cairo context from a CairoTexture with
either :surface-width or :surface-height set to 0 will result in
a warning.
This allows:
- creating a CairoTexture with :surface-width or :surface-height
set to zero and delaying the surface resize at a later point;
- resizing the surface to 0 by 0 pixels to destroy the image
surface used internally;
- increase the consistency in the usage of CairoTexture.
After fixing the cursor position issues around the initial
glyph of the layout, the selection position needs fixing as
well.
The fix is similar: check if the position of the selection
is 0 and provide a fast path by setting the offset to 0.
The gdouble value represents an interval along the path from 0.0 to
1.0. This makes more sense than using an alpha value because paths are
not directly related to ClutterAlphas and the rest of the Clutter API
tends to expose gdouble arguments.
This is the port of the patch attached to bug 1266:
ClutterLabel support for pango_layout_set_height()
to the ClutterText actor. This patch adds support for the
pango_layout_set_height() function inside the layout creation
and caching, so that setting the height of a ClutterText will
make PangoLayout try to fill out the given height before
ellipsizing the contents of the ClutterText actor.
The original patch is from Havoc Pennington.
This commit ports the patch to ClutterText and applies a small
change to the height computation pointed out by Behdad Esfahbod
on GNOME's Bugzilla:
http://bugzilla.gnome.org/show_bug.cgi?id=560931
That the layout heights when doing a size request should be
set to be G_MAXINT instead of -1.
Unfortunately, the patch seems to break the conformance suite
when setting the ellipsization on the layout.
The clutter_text_set_selection() function is a convenience
method for setting the cursor position and the selection
boundary to a given position in a single call, with sanity
checks for the positions.
The clutter_text_get_selection() function was not checking the
passed argument, and was still accessing the contents of the
Text actor using clutter_text_get_text().
This commit also adds the last few gtk-doc annotations missing
from ClutterText.
Instead of repeating the same code for the ::activate signal
emission, use the clutter_text_activate() function inside the
'activate' key binding handler.
The clutter_text_activate() function will emit the ::activate
signal if the :activatable property is set.
This function is useful for subclasses or application code, for
example if we are going to use ::captured-event or ::key-press-event
signal handlers to intercept the Return key and emit the ::activate
signal ourselves.
Instead of installing the line-start and line-end key bindings
using the bare ClutterBindingPool API, we can use the internal
clutter_text_add_move_binding(), which automatically installs
the same key binding with the Shift modifier mask.
This allows selecting when pressing Shift+Home or Shift+End.
The selection behaviour is still incorrect around the zeroeth
position, with all the text after the first line being selected.
We can control the width of the cursor when painting by using
a simple property.
The magic -1 number passed to the setter method will reset the
cursor size to the default one of 2px.
The getter method will return an unsigned integer with the
current size.
Moving the text by a "page" depends on being able to define a
"page size" in terms of lines of text. Since we don't define
something similar to an Adjustment that allows us to handle this
behaviour, we should defer the paging implementation to a higher
level class based on ClutterText.
Another fix for the key navigation behaviour around the zeroeth
glyph in the layout.
This commit adds a fast path for for the zero index when the
cursor position is set as zero, in case we are using the line-start
or line-end key bindings, similarly to what we did in commit
be64cbcdc2 for the move-up and
move-down bindings.
The behaviour of ClutterText around the initial position is still
a little bit erratic.
This commit fixes the key navigation with Up and Down arrows around
the first line of text.
We loose precision with a direct conversion for PangoUnits to
pixels, so we should do the conversion as needed, inside the
callers of clutter_text_position_to_coords().
We don't allow changing the cursor geometry inside the ::cursor-event
signal handlers; for starters, it would make binding the signal a
huge mess, and it would also potentially break the whole actor.
The ClutterParamSpecFixed constructor is declared in the header
as taking ClutterFixed values, but the implementation takes
ClutterUnit values instead. This obviously works because fixed
and units are exactly the same type.
This commit fixes the wrong parameters.
Fix the CairoTexture description, and some of the comments inside
the code, especially with regards to the alpha channel unpremultiplication
that we have to perform each time we upload the image surface to
GL.
If you create a Cairo context in the middle of a paint run and then
you destroy it, the CairoTexture will have to upload the contents of
the image surface to the GL pipeline. This usually leads to slow
downs and general performance degradation.
ClutterCairoTexture will warn to the console if Clutter has been
compiled with the debug messages and if create() or create_region()
are called while an actor is in the middle of a paint.
Since the CLUTTER_ACTOR_IN_PAINT private flag is set as part
of the paint process by clutter_actor_paint(), there is no
need to set it inside the ClutterStage paint function.
When calling clutter_actor_paint() we should be setting the
CLUTTER_ACTOR_IN_PAINT private flag. This allows signalling
to each Actor subclass that we are effectively in the middle
of a paint sequence. Actor subclasses can check for this
private flag and act based on its presence - for instance to
avoid recursion, or to detect performance degradation cases.
Move the ClutterCairo actor from a separate library to an in-tree
actor.
ClutterCairoTexture is a simple texture subclass that allows you
to retrieve a Cairo context for a private image surface. When the
Cairo context is destroyed it will cause the image surface
contents to be uploaded to a GL texture.
The image surface used is not hardware accelerated.
If we create the PangoContext for ClutterText inside the class
initialization we might not have a Clutter main context yet.
Ideally, we should store the Pango context inside the main context
and create it on clutter_init(), but for now we can lazily create
the PangoContext when we initialize a ClutterText instance for the
first time.
For the time being, just don't include them or compile them; the
files will be removed from the repository as soon as all the
documentation annotations have been ported over to ClutterText.
We allow KeyEvents with a key symbol of '0' to fall through only
if they are marked as synthetic. Otherwise we discard them without
mercy.
Synthetic events are useful to test ClutterText behaviour; in fact,
we do use them inside the test suite exactly for that reason.
I understand we are not Pascal developers, and we don't have to
use cute and cuddly names like "i_am_an_integer_counter", but
a ClutterButtonEvent should be stored inside an "event" variable.
Using "bev" instead? Mmmh, not so much.
ClutterText should use the paint opacity for both text and
cursor.
ClutterLabel had the wrong behaviour, as it set the actor's
opacity using the text color's alpha channel, and ClutterEntry
completely disregarded the actor's opacity when painting the
cursor.
This commit harmonizes the ClutterText behaviour to always
use a composition of the actor's paint opacity and the text
and cursor alpha channel values, thus behaving more
consistently with the rest of Clutter.
When inserting text on a key press event we should also truncate
the selection.
We should not truncate the selection when inserting any Unicode
character, since changing the selection also changes the cursor
position - and one of the invariants we inherited from ClutterEntry
is that inserting characters programmatically does not change the
cursor position.
If a selection has been truncated inside a key binding handler,
we should just return and let the usual key event handler continue.
This fixes the case where we deleted a selection using the Delete
or the Backspace keys.
ClutterText should use the newly added ClutterBindingPool API to
handle key events, instead of its homegrown code.
This commit removes the action/mapping code and defers the entire
key binding matching to a ClutterBindingPool created inside the
Text class initialization function.
The clutter_text_get_chars() function returns a section of the
contents of the Text actor, delimited by a start and an end position.
This commit adds the implementation for that function and a test
unit that guarantees the offset-to-bytes computations are correct.
The :cursor-color-set property is a read-only property that
reflects whether the ClutterText actor is going to use the
color set inside the :cursor-color property when painting
the cursor.
We should re-use the offset_to_bytes() implementation from ClutterEntry
as it guaranteed some behaviour and sanity checks that we want to keep
inside ClutterText.
A ClutterText can be put in "password mode" by setting the
text as "invisible": every character inside the Text actor's
contents will be replaced when building the Pango layout with
a specific Unicode character.
The Unicode character is set to '*' by default, but the user
can be changed using the provided API.
ClutterText should have the same properties as ClutterLabel.
While at it, we can finally fix the disconnect between the wrap
and wrap-mode properties and its accessors, that we perpetuated
from GtkLabel.
The ClutterText:line-wrap property and ClutterText:line-wrap-mode
are mapped to the set_line_wrap(), get_line_wrap() and
set_line_wrap_mode(), get_line_wrap_mode() accessor functions
respectively. This should simplify bindings the Vala ones that
map a property into a method.
Bug 1003 - Add clutter_actor_take_key_focus()
The grab_key_focus() method is just a simple wrapper around
clutter_stage_take_key_focus() that removes the need to get
the ClutterStage of an actor in order to set the key focus.
Based on a patch by Xan López.
Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
Signed-off-by: Øyvind Kolås <pippin@linux.intel.com>
ClutterText replaces ClutterLabel, so it should expose the same
kind of API - ideally with the minimal amount of changes, so that
the porting is trivial.
The TidyText actor is meant as a replacement for both ClutterLabel
and ClutterText.
Any text-displaying and editing actor should derive from ClutterText
and implement the various visual cues to differentiate the editable
from the non-editable state. Those visual cues usually belong to
a high-level toolkit, especially if themeing is involved.
When the apply_transform_to_point() and its relative variant
landed in Clutter 0.3, the initial approach was to modify the
passed vertex as an in-out parameter. This was later dropped
in favour of a more consistent out parameter.
Unfortunately, the implementation never changed: both methods
where modifying the passed vertex with the partial results of
the computations.
This commit copies the contents of the "point" ClutterVertex
argument inside a stack variable and, for good measure, constifies
the argument.
Thanks to Thomas Steinacher for catching this in Python.
* clutter/cogl/gl/cogl-defines.h.in:
* clutter/cogl/gl/cogl-context.h:
* clutter/cogl/common/cogl-mesh.c: Rename the glBufferDataSub
function to glBufferSubData. When calling glXGetProcAddress with
the former Mesa returns a stub dispatch function which will
segfault if you try to use it. With NVIDIA it returns NULL so
_cogl_features_init decides the card doesn't have VBO support.
* clutter/clutter-behaviour-ellipse.c (actor_apply_knot_foreach):
Don't set the depth if there is no x or y tilt. That way it can
still be used in conjunction with ClutterBehaviourDepth. Thanks to
Tonny Tzeng.
There's no point in clearing the index array because it is always the
same sequence of indices regardless of the vertices. Instead it is
just added to when there are more vertices than ever before.
* clutter/Makefile.am:
* clutter/clutter.h: Add ClutterBindingPool to the build.
* clutter/clutter-binding-pool.c:
* clutter/clutter-binding-pool.h: Add ClutterBindingPool, a data
structure meant to hold (key symbol, modifiers) pairs and associate
them to a closure. The ClutterBindingPool can be used to install
key bindings for actors and then execute closures inside the
key-press-event signal handlers, removing the need for big
switch() or if() blocks for each key.
* clutter/clutter-event.c: Consistently use "key symbol" instead
of "key value".
* clutter/clutter-event.h: Add more modifier masks.
* clutter/clutter-marshal.list:
* tests/conform/Makefile.am:
* tests/conform/test-binding-pool.c:
* tests/conform/test-conform-main.c: Add ClutterBindingPool
conformance test.
* tests/interactive/Makefile.am:
* tests/interactive/test-binding-pool.c: Add interactive test (and
example code) for the ClutterBindingPool usage.
Bug 1309 - clutter_timeline_new and clutter_timeline_set_speed
have two standard of the fps limitation
* clutter/clutter-timeline.c:
(clutter_timeline_class_init): Set the maximum value of the
:fps property to be G_MAXUINT. (Zhang Wei)
* clutter/clutter-path.h:
* clutter/clutter-path.c: Implementation of new ClutterPath object
to represent a path combining straight line and bezier curve
elements.
* clutter/clutter.h: Include clutter-path.h and remove
clutter-behaviour-bspline.h
* tests/interactive/test-threads.c (test_threads_main):
* tests/interactive/test-script.c:
* tests/interactive/test-behave.c (test_behave_main): Use new path
API
* clutter/clutter-effect.c: Use the new ClutterBehaviourPath API.
* clutter/clutter-bezier.h:
* clutter/clutter-bezier.c: Moved bezier curve handling code out
from clutter-behaviour-bspline.c to a separate file.
* clutter/clutter-behaviour-path.h:
* clutter/clutter-behaviour-path.c: Reimplemented to work with a
ClutterPath
* clutter/clutter-behaviour-bspline.h:
* clutter/clutter-behaviour-bspline.c: Removed
* clutter/Makefile.am: Add clutter-path and clutter-bezier, remove
clutter-behaviour-bspline.
* tests/conform/test-path.c: New automatic test for ClutterPath
consistency
* tests/conform/test-conform-main.c (main): Add test_path
* tests/conform/Makefile.am (test_conformance_SOURCES): Add
test-path.c
* clutter/clutter-sections.txt: Add ClutterPath docs
* clutter/clutter.types:
* clutter/clutter-docs.xml:
* doc/reference/clutter/clutter-animation-tutorial.xml: Remove
mention of ClutterBehaviourBspline
* clutter/clutter-marshal.list: Add VOID:UINT
* clutter/cogl/gl/cogl-texture.c (cogl_texture_new_from_foreign,
(_cogl_texture_quad_hw, cogl_texture_polygon),
(_cogl_texture_quad_sw): Support GL_ARB_texture_rectangle textures
* clutter/glx/clutter-glx-texture-pixmap.c: Use rectangle textures
when NPOTs are not available or it is forced by the
CLUTTER_PIXMAP_TEXTURE_RECTANGLE environment variable.
* clutter/cogl/gl/cogl.c (cogl_enable): Allow enabling
GL_TEXTURE_RECTANGLE_ARB.
* clutter/cogl/cogl-path.h:
* clutter/cogl/common/cogl-primitives.c:
* clutter/cogl/common/cogl-primitives.h:
* clutter/cogl/gl/cogl-primitives.c:
* clutter/cogl/gles/cogl-primitives.c: Changed the semantics of
cogl_path_move_to. Previously this always started a new path but
now it instead starts a new disjoint sub path. The path isn't
cleared until you call either cogl_path_stroke, cogl_path_fill or
cogl_path_new. There are also cogl_path_stroke_preserve and
cogl_path_fill_preserve functions.
* clutter/cogl/gl/cogl-context.c:
* clutter/cogl/gl/cogl-context.h:
* clutter/cogl/gles/cogl-context.c:
* clutter/cogl/gles/cogl-context.h: Convert the path nodes array
to a GArray.
* clutter/cogl/gl/cogl-texture.c:
* clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure
* clutter/cogl/common/cogl-clip-stack.c:
* clutter/cogl/common/cogl-clip-stack.h: Simplified the clip
stack code quite a bit to make it more maintainable. Previously
whenever you added a new clip it would go through a separate route
to immediately intersect with the current clip and when you
removed it again it would immediately rebuild the entire clip. Now
when you add or remove a clip it doesn't do anything immediately
but just sets a dirty flag instead.
* clutter/cogl/gl/cogl.c:
* clutter/cogl/gles/cogl.c: Taken away the code to intersect
stencil clips when there is exactly one stencil bit. It won't work
with path clips and I don't know of any platform that doesn't have
eight or zero stencil bits. It needs at least three bits to
intersect a path with an existing clip. cogl_features_init now
just decides you don't have a stencil buffer at all if you have
less than three bits.
* clutter/cogl/cogl.h.in: New functions and documentation.
* tests/interactive/test-clip.c: Replaced with a different test
that lets you add and remove clips. The three different mouse
buttons add clips in different shapes. This makes it easier to
test multiple levels of clipping.
* tests/interactive/test-cogl-primitives.c: Use
cogl_path_stroke_preserve when using the same path again.
* doc/reference/cogl/cogl-sections.txt: Document the new
functions.
fallbacks are being used
* glx/clutter-glx-texture-pixmap.c:
clutter_glx_texture_pixmap_using_extension now checks to see if
priv->use_fallback is TRUE not just that the tfp extension is
available.
border artifacts
* clutter/cogl/gl/cogl-texture.c: Set the wrap mode of a texture
on demand
Instead of setting the wrap mode once per texture at creation, it
is now changed whenever the texture is drawn. The previous value
is cached so that it isn't changed if the value is the same.
This is used in _cogl_texture_quad_hw to only enable GL_REPEAT
mode when the coordinates are not in the range [0,1]. Otherwise it
can pull in pixels from the other edge when the texture is
rendered off-pixel.
Most cards don't actually support GL_QUADS and they are deprecated in
GL 3.0 so there is a chance it will perform faster with GL_TRIANGLES
even though it has to submit two extra vertices.
The glyphs are queued into an array of rectangles instead of being
drawn directly. Whenever a different texture is used or the sequence
is complete the array is flushed.
This is based on a patch from Owen Taylor.
This takes an array of sets of 8 floats to describe the rectangles. It
tries to send the geometry with a single glDrawArrays as far as
possible. cogl_texture_rectangle is now just a wrapper around
cogl_texture_multiple_rectangles.
cogl_texture_quad_hw and _sw now just add vertices to the vertex
array. The last texture used is stored and if a different texture is
encountered then flushes the vertices. cogl_texture_rectangle always
flushes the vertices after calling either of the functions.
