With Wayland 1.0, A Large TODO List Remains

[phoronix]

Phoronix: With Wayland 1.0, A Large TODO List Remains

Wayland 1.0 will be released as soon as today, but this doesn't mark the death of X11 and Wayland beginning to secure major traction on the Linux desktop...

http://www.phoronix.com/vr.php?view=MTIxMjM

[TheBlackCat]

For a project as central as this the TODO list doesn't strike me as being all that long. It only has 14 items if you exclude ports.

[renox]

A glyph cache makes sense for X because you can upload a glyph to the server and then tell the server to use it at several places, but for Wayland which has no drawing commands in the server (short of display this buffer), I don't get what is the point of a glyph cache?
Could someone explain?

[plonoma]

- Due to client side decorations, there needs to be a protocol for specifying title bar rectangle and the regions for the close button so that there can be a pop-up force-close dialog if applications don't respond to ping events.

Also being able to use escape button to force the application to minimize would be very practical.
(Do'nt use windowed mode since this could be abused and software can be made that you could select the program and set it in windowed mode.)

I hope the developers don't make the same mistake and add tons of things just for the sake of it.
Am a bit concerned about the glyph and settings protocol stuff.

Hopefully clone mode also allows pieces of screen to be cloned and (minimized) application windows.

Support for multiple screens can be done good this time.

[elanthis]

A glyph cache makes sense for X because you can upload a glyph to the server and then tell the server to use it at several places, but for Wayland which has no drawing commands in the server (short of display this buffer), I don't get what is the point of a glyph cache?
Could someone explain?

Indeed, not only is a glyph cache less useful, I'll argue it's outright harmful.

If you want high-speed 2D rendering of a GUI, you need to put the glyphs you're using into an atlas with the other textures and gradient source patterns you're using. If you stick with an atlas just for text, you're forced to split your draw calls between "image" draws and text draws, and then split those even further to deal with layering and control compositing.

Short version: each app should just create its own glyph cache in its own internal texture atlases to maximize drawing performance (and hence also battery life).