Hello, hope this email comes in good faith. I was thinking of any benefits that could arise from moving the X11 socket from /tmp/.X11-unix to $XDG_RUNTIME_DIR/.X11-unix and I thought of a few. This thought was inspired by Wayland and their approach for sockets, and from how I see it, it may benefit XWayland.
- Moving the X11 socket to XDG_RUNTIME_DIR makes Xorg seem as if it is willing to adopt the XDG specification (which it already does with Xorg logs going to ~/.local/share/Xorg), but I feel as if this change would cement that further. - I think that the X11 socket being so easy to access (as it is 0777, and in /tmp) is a security flaw which shouldn't exist even in terms of backwards compatibility. I think this can be resolved by symlinking, but if most programs work fine without needing to symlink then that's amazing. - You wouldn't run into unpredictability issues regarding where the X11 socket is on many systems, as most systems now use XDG_RUNTIME_DIR. As for Windows, they can keep %TEMP% or %TMP% or whatever it is called. I also believe some of the Windows-specific functions could benefit from a rewrite, so moving the socket may bring that closer. I thought of these for a while and was thinking of where I could ask about it to but I want to ask about it here before taking any bigger steps. I admittedly am not the greatest fan of Xorg but I need to wait for the Wayland situation to improve on FreeBSD before I make the switch. It is okay now but not amazing in it's current state, while Xorg is just good enough on here for me to use it. I look forward to your (plural) views. Cheers, -- Artur Manuel (amadaluzia)
