Sway: Sway 0.12

Created on 5 Jan 2017  路  6Comments  路  Source: swaywm/sway

List of things I want to ship in 0.12. Anyone's welcome to work on this stuff (or anything else). This list is unordered.

  • [x] Redshift support
  • [x] Implement improved IPC security flags

Most helpful comment

No, it's mostly happened on IRC. I'll give the basic rundown here:

Problems with wlc:

  • Performance (CPU, GPU, and power usage)
  • Design complexity
  • Maintenance (@Cloudef doesn't have time to work on it and few others contribute, which is compounded by the over-complex design)

Advantages to switching:

  • We can design our compositor with the hindsight wlc didn't have and avoid many problems
  • Easier to implement surface-layers if we have our own compositor, as well as other protocols we'll need in the future
  • Sway gets far more maintenance attention than wlc

Also, solving some of the problems in wlc that we need solved would necessitate design decisions that would be worse off for both projects.

All 6 comments

Executive decision: postponing the following tasks indefinitely:

  • [ ] Get surface-layers protocol upstreamed in wayland-protocols (discussion on wayland-devel)
  • [ ] Update swaylock, swaybar, swaybg to use surface-layers protocol
  • [ ] Implement swayslop based on surface-layers
  • [ ] Implement orbital/weston's screenshooter protocol and update swaygrab to use it
  • [ ] Discuss permission management protocols with wayland-devel
  • [ ] Implement KDE's DPMS protocol and write a DPMS daemon
  • [ ] Implement KDE's modesetting protocol and work with KDE to improve arandr-style client interfaces

And these permenantly:

0.12 will be a much smaller release, mostly to ship redshift and bugfixes on a reasonable timeframe. The primary goal of 0.13 will be to replace wlc with our own wayland compositor implementation.

What's the benefit of switching from wlc to a custom wayland compositor? Is there a discussion about this somewhere?

No, it's mostly happened on IRC. I'll give the basic rundown here:

Problems with wlc:

  • Performance (CPU, GPU, and power usage)
  • Design complexity
  • Maintenance (@Cloudef doesn't have time to work on it and few others contribute, which is compounded by the over-complex design)

Advantages to switching:

  • We can design our compositor with the hindsight wlc didn't have and avoid many problems
  • Easier to implement surface-layers if we have our own compositor, as well as other protocols we'll need in the future
  • Sway gets far more maintenance attention than wlc

Also, solving some of the problems in wlc that we need solved would necessitate design decisions that would be worse off for both projects.

0.12-rc1 has been released, please test it!

0.12-rc2 released, please test it! The following changes were included: 0.12-rc1...0.12-rc2

Was this page helpful?
0 / 5 - 0 ratings

Related issues

soymjolk picture soymjolk  路  3Comments

DpoBoceka picture DpoBoceka  路  4Comments

ghost picture ghost  路  4Comments

WhyNotHugo picture WhyNotHugo  路  3Comments

Olie440 picture Olie440  路  4Comments