D6186: Implement software cursor in OpenGL backend

[David Edmundson]

davidedmundson added a comment.


  > Is this the correct way of doing such stuff?
  
  Pretty much. 
  Seems a sensible patch, screen recording aside we could be using a GL on a different backend that can't do planes.
  
  The obvious optimisation is that we don't need to make the new cursor texture every time, but it can be a member var reset on cursorChanged()
  
  > Should cursor update be included in damage regions?
  
  I would have thought so. Though I can't see where scene_qpainter does it. (could be a bug there?)
  
  > How can I write an autotest for it?
  
  A lot of the GL stuff isn't autotested.  and even if you did write a test, our current CI infrastructure couldn't actually run it.

INLINE COMMENTS

> scene_opengl.cpp:705
> +
> +    GLVertexBuffer* vbo = GLVertexBuffer::streamingBuffer();
> +    vbo->reset();

i don't think you need these 3 lines, you're  not doing anything with this vbo.

You do need vertex info from somewhere, but GLTexture has its own vbo.

REPOSITORY
  R108 KWin

REVISION DETAIL
  https://phabricator.kde.org/D6186

To: Kanedias, graesslin, davidedmundson
Cc: plasma-devel, kwin, #kwin, ZrenBot, spstarr, progwolff, lesliezhai, ali-mohamed, hardening, jensreuterberg, abetts, eliasp, sebas, apol, mart, hein, lukas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/plasma-devel/attachments/20170612/16a61b5e/attachment.html>