Most matrices used by the UI toolkit are translation matrices, whose
inverses can be quickly computed by using the negated translation
vector.
Change-Id: I54a28a634a586085779bfc26f3a4160cd5ab2b22
Separate matrix passed to shadow system into two parts, one for
transforming the polygon XY points (using the actual draw matrix) and
a separate one which respects correct 4x4 3d rotations and
translations for determining Z values.
Change-Id: I7e30a84774a8709df6b2241e8f51fc5583648fe8
True 3d transformations are now supported by DisplayLists and the
renderer, initially with the translationZ property on view.
Renderer operations used directly by DisplayList (formerly,
clip/save/restore/saveLayer) are now more simply managed by allocating
them temporarily on the handler's allocator, which exists for a single
frame. This is much simpler than continuing to expand the pool of
pre-allocated DisplayListOps now that more operations are called
directly by DisplayList, especially with z ordered drawing.
Still TODO:
-APIs for camera positioning, shadows
-Make Z apis public, and expose through XML
-Make invalidation / input 3d aware
Change-Id: I95fe6fa03f9b6ddd34a7e0c6ec8dd9fe47c6c6eb
bug:10863163
This fixes two issues
The check for pure translation was incorrect. It was fixed and renamed
for clarity.
Certain matrix paths weren't setting kTypePositiveScale. For
simplicity (and because positive scale is simple to check) removed
flag in favor of dynamic checking.
Change-Id: Ic5ce235653ef49a68b8b242bd89fc2e95874ecc9
This change extracts the scale parameters of the current transform
to pass then to the font renderer. Rotation and perspective are
applied to the generated mesh inside the vertex shader. This limits
the number of glyphs we have to create in the font cache and thus
reduces memory churn.
Change-Id: Ic5b3bae2b2b0e0250a8ee723b071a1709725c749
The drawColorRects() method was clipping individual rectangles
using the wrong parameters left, top, right and bottom instead
of l, r, t and b. It also checked for count == 0 after the loop
when it should have checked for vertexCount == 0. The quickReject
is now not part of the loop since it's a bit overkill to perform
so many matrix multiplications. What we really care about is the
final quickReject performed on the max bounds of the entire set
of rectangles.
This change also replaces all instances of mSnapshot->transform
by currentTransform() to make the code slightly more readable.
Change-Id: I6485280414499716852f7dbfba186774eb6763d4
This change does not apply to drawPosText() and drawTextOnPath() yet.
Prior to this change, glyphs were always rasterized based on the
font size specified in the paint. All transforms were then applied
on the resulting texture. This creates rather ugly results when
text is scaled and/or rotated.
With this change, the font renderer will apply the current transform
matrix to the glyph before they are rasterized. This generates much
better looking results.
Change-Id: I0141b6ff18db35e1213e7a3ab9db1ecaf03d7a9c
Bug #7146141
When non-rectangular clipping occurs in a layer the render buffer
used as the stencil buffer is not cached. If this happens on a
View's hardware layer the render buffer will live for as long
as the layer is bound to the view. When a stencil buffer is
required because of a call to Canvas.saveLayer() it will be allocated
on every frame. A future change will address this problem.
If "show GPU overdraw" is enabled, non-rectangular clips are not
supported anymore and we fall back to rectangular clips instead.
This is a limitation imposed by OpenGL ES that cannot be worked
around at this time.
This change also improves the Matrix4 implementation to easily
detect when a rect remains a rect after transform.
Change-Id: I0e69fb901792d38bc0c4ca1bf9fdb02d7db415b9
Bug #7253839
1. Make sure we don't make GL calls while recording display lists
2. Disable an early and trivial clip optimization in font renderer
when a perspective transformation is used on the Canvas
Change-Id: I3f1052164239329346854f72d0a0d401fbfecf06
Some logic in the native matrix code would determine that a matrix was
'pureTranslate' based on the scale values of a matrix being close-enough to 1,
which was within a very small epsilon. This works in general, because screen space
coordinates make that epsilon value irrelevant, so close-enough really is close-enough.
However, TextView, when centering text, works in a coordinate system that is quite
huge, with left/right values about 500,000. These numbers multiplied times that small
epsilon value would give a result that was significant, and would cause a miscalculation
of up to 4-5 pixels, causing the snap that we'd see for a couple of frames as the
scale got "close enough" to 1.
The fix is to remove the optimization of "close enough". What we really need the matrix to
do is to identify itself as being translate-only when no scale as been set (which is the
default). For the purposes of that check, it is good enough to simply check the values against
1 directly. Similarly, the bounds-check logic needs to check against 0 and 1 directly.
Issue #6452687: Glitch when changing scale of a view containing text
Change-Id: I167fb45d02201fb879deea0e5a7ca95e38128e17
This optimization is currently disabled until Launcher is
modified to take advantage of it. The optimization can be
enabled by turning on RENDER_LAYERS_AS_REGIONS in the
OpenGLRenderer.h file.
Change-Id: I2fdf59d0f4dc690a3d7f712173ab8db3848b27b1
This change detects empty quads in 9patches and removes them from
the mesh to avoid unnecessary blending.
Change-Id: I4500566fb4cb6845d64dcb59b522c0be7a0ec704
This change also adds full support for local transformation matrices on
sweep and radial gradients.
Change-Id: Id8773bc0766575190e3f3d51984fc5e57b266c3f
The generator supports features that are not yet implement in the
renderer: color matrix, lighting, porterduff color blending and
composite shaders.
This change also adds support for repeated/mirrored non-power of 2
bitmap shaders.
Change-Id: I903a11a070c0eb9cc8850a60ef305751e5b47234
This change also modifies the way the clip is stored. The clip is now
always stored in screen-space coordinates.
Change-Id: I96375784d82dfe975bc6477a159e6866e7052487
This change also refactors the code base a bit by moving classes out of
OpenGLRenderer into separate headers/implementations. This makes the code
more manageable.
This change also adds documentation for implementation methods. The
undocumented methods are simply Skia's Canvas methods.
Change-Id: I54c68b443580a0129251dddc1a7ac95813d5289e
Even though there's an implementation for textured rects, drawBitmap() is not
hooked up yet as it will require a good texture cache.
This method is implemented using FBOs. There's currently an issue either in the
driver or in the Canvas renderer that forces the FBO to be fullscreen, which is
extremely expensive and yields terrible performance.
Change-Id: I148419195e12d45653c60186938aa78c23a68e2c
The OpenGL ES 2.0 renderer can now draw colored rectangles. At least there's
something on screen now.
Change-Id: I80a13ccc1dd56784edf74f2670a364f30700234a
Drawing two rectangles one after the other discards the second one because of
Z buffering issues. This will be fixed in another changelist.
Change-Id: Ida1b3cde8a78e60cacc07e477abc44def527ff67
This change adds partial support for the following transforms:
- scale()
- translate()
- rotate()
- setMatrix()
- getMatrix()
The transform is stored in a snapshot and saved/restored as needed.
The transform is currently not applied to the clip rect and is not
mapped to the vertex shader.
Change-Id: Id48993453311200804149917d0c126a4d0471226
The current implementation of clipRect() does not apply local transformations
before setting the new clip.
Change-Id: I5997871bb638dfcd1a8ef96354846af52427e445
