[arandr-users] Rerandr3 and better save/restore mechanism for Arandr

[Vladimir Kudrya]

Hello everyone!
Some months ago I mailed chrysn about improving save/restore mechanisms 
in Arandr and my xrandr wrapper/daemon script called rerandr3. 
Unfortunately, I didn't get any response.
Anyway, some of you might be interested in this script and ideas behind it.

The project is hosted here: https://github.com/Vladimir-csp/rerandr3
Rerandr3 features:
- On-demand save of current outputs configuration in relation to 
combination of outputs and display device models. (complete with 
geometry, properties, primary screen).
- Restore outputs configuration for familiar combinations.
- Auto-enable and stack outputs on new unfamiliar combinations.
- Running as a daemon it can restore or auto-stack outputs on hot-plug 
in fully automatic mode, triggered by udev rule.
- Runs different sets of hooks depending on virtual screen size.
- Works around poor xrandr CRTCs handling.
- Temporarily preserves framebuffer size while switching outputs to 
prevent windows positioning mess
- Tries to avoid unnecessary switching, aplying only delta operations.

Throughout the past half a year I've squashed some bugs and it served me 
reliably and unobtrusively while moving between home and work and 
plugging in some new displays.

What it has to do with Arandr? It could provide a backend for smart 
save/restore and hotplug actions handling which Arandr is lacking. Or it 
could serve as a model for implementing such features in Arandr itself.
As a test, I once dirty-hacked (I'm not so comfy outside bash) Arandr 
purging everything except layout editor and "Apply" button, and adding 
trigger to run "rerandr3 hooks" and "rerandr3 save" on "Apply". Proved 
to be a very practical combination on the outside.

Cheers!