Qt-contributors-summit-2012-QtOpenGL: Difference between revisions
Jump to navigation
Jump to search
No edit summary |
No edit summary |
||
Line 1: | Line 1: | ||
<div class="cpp-qt geshi"> | |||
# <div class="de1">Future plans</div> | |||
# <div class="de1"><span class="sy0">-</span> Desktop OpenGL <span class="nu0">3</span><span class="sy0">+</span> support<span class="sy0">,</span> ES <span class="nu0">3</span> support</div> | |||
# <div class="de1"><span class="sy0">-</span> QOpenGLWindow</div> | |||
# <div class="de1"><span class="sy0">-</span> Shared cross process graphics buffer APIs</div> | |||
# <div class="de2"> <span class="sy0">-</span> copy to texture</div> | |||
# <div class="de1"> <span class="sy0">-</span> bind to texture</div> | |||
# <div class="de1"> <span class="sy0">-</span> bind as framebuffer</div> | |||
# <div class="de1"> <span class="sy0">-</span> locking <span class="sy0">/</span> unlocking subrectangles <span class="kw1">for</span> SW <span class="sy0">/</span> HW access</div> | |||
# <div class="de1"> <span class="sy0">-</span> synchronization</div> | |||
# <div class="de2"> <span class="sy0">-</span> QPA <span class="sy0">/</span> preliminary ?</div> | |||
# <div class="de1"> <span class="sy0">-</span> IPC <span class="sy0">/</span> way of exposing buffer across processes</div> | |||
# <div class="de1"> <span class="sy0">-</span> file descriptor</div> | |||
# <div class="de1"> <span class="sy0">-</span> hardware address<span class="sy0">,</span> handle<span class="sy0">,</span> etc</div> | |||
# <div class="de1"> <span class="sy0">-</span> capabilities</div> | |||
# <div class="de2"> <span class="sy0">-</span> page flipping to the screen</div> | |||
# <div class="de1"> <span class="sy0">-</span> linear formats <span class="sy0">/</span> non<span class="sy0">-</span>linear formats</div> | |||
# <div class="de1"> <span class="sy0">-</span> orthogonal to the QPA APIs? multiple buffer abstractions</div> | |||
# <div class="de1"><span class="sy0">-</span> QOffscreenSurface</div> | |||
# <div class="de1"><span class="sy0">-</span> texture atlas API <span class="sy0">/</span> area allocator</div> | |||
# <div class="de2"><span class="sy0">-</span> simple imperative GLES2 painter suited <span class="kw1">for</span> Canvas 2D etc<span class="sy0">,</span> lose all the poorly <span class="sy0">/</span> unnecessary performing features of [http://doc.qt.io/QPainter.html <span class="kw5">QPainter</span>]</div> | |||
# <div class="de1"><span class="sy0">-</span> QPainterV2 API<span class="sy0">,</span> reduced [http://doc.qt.io/QPaintEngine.html <span class="kw5">QPaintEngine</span>] API</div> | |||
# <div class="de1"> <span class="sy0">-</span> learn from Skia</div> | |||
# <div class="de1"> <span class="sy0">-</span> improved API that discourages slow use cases</div> | |||
# <div class="de1"><span class="sy0">-</span> the problem of RGBA <span class="sy0"><-></span> <span class="me3">BGRA</span></div> | |||
# <div class="de2"> <span class="sy0">-</span> <span class="kw1">new</span> image formats?</div> | |||
# <div class="de1"><span class="sy0">-</span> image upload parts of scene graph adaptation layer</div> | |||
# <div class="de1"> <span class="sy0">-</span> better abstraction?</div> | |||
# <div class="de1"> <span class="sy0">-</span> move from declarative to QPA?</div> | |||
# <div class="de1"><span class="sy0">-</span> fast texture uploads</div> | |||
# <div class="de2"><span class="sy0">-</span> subpixel layout</div> | |||
# <div class="de1"> <span class="sy0">-</span> QScreen API <span class="kw1">for</span> accessing subpixel layout information</div> | |||
# <div class="de1"> <span class="sy0">-</span> subpixel layout information in [http://doc.qt.io/QImage.html <span class="kw5">QImage</span>] <span class="sy0">/</span> [http://doc.qt.io/QPixmap.html <span class="kw5">QPixmap</span>] ?</div> | |||
# <div class="de1"><span class="sy0">-</span> [http://doc.qt.io/QImage.html <span class="kw5">QImage</span>] vs [http://doc.qt.io/QPixmap.html <span class="kw5">QPixmap</span>]</div> | |||
# <div class="de1"><span class="sy0">-</span> [http://doc.qt.io/QPixmap.html <span class="kw5">QPixmap</span>]<span class="sy0">::</span><span class="me2">internalImage</span><span class="br0">(</span><span class="br0">)</span> accessor with no deep copy?</div> | |||
# <div class="de2"><span class="sy0">-</span> wrapping non<span class="sy0">-</span>[http://doc.qt.io/Qt.html <span class="kw5">Qt</span>] contexts in QOpenGLContext</div> | |||
</div> |
Revision as of 14:04, 25 February 2015
- Future plans
- - Desktop OpenGL 3+ support, ES 3 support
- - QOpenGLWindow
- - Shared cross process graphics buffer APIs
- - copy to texture
- - bind to texture
- - bind as framebuffer
- - locking / unlocking subrectangles for SW / HW access
- - synchronization
- - QPA / preliminary ?
- - IPC / way of exposing buffer across processes
- - file descriptor
- - hardware address, handle, etc
- - capabilities
- - page flipping to the screen
- - linear formats / non-linear formats
- - orthogonal to the QPA APIs? multiple buffer abstractions
- - QOffscreenSurface
- - texture atlas API / area allocator
- - simple imperative GLES2 painter suited for Canvas 2D etc, lose all the poorly / unnecessary performing features of QPainter
- - learn from Skia
- - improved API that discourages slow use cases
- - the problem of RGBA <-> BGRA
- - new image formats?
- - image upload parts of scene graph adaptation layer
- - better abstraction?
- - move from declarative to QPA?
- - fast texture uploads
- - subpixel layout
- - QScreen API for accessing subpixel layout information