LFCS 2012: X and Wayland

Posted Apr 11, 2012 22:40 UTC (Wed) by mezcalero (subscriber, #45103)
Parent article: LFCS 2012: X and Wayland

The thing with socket activation for X is the weird allocation of sockets X needs. You match up one abstract namespace AF_UNIX socket with one AF_UNIX socket in /tmp with one TCP socket and then deduce the display number from that, and the simple thing is that I don't want to see code like that in systemd.

(Log in to post comments)

Posted Apr 11, 2012 23:24 UTC (Wed) by jake (editor, #205) [Link]

Couldn't systemd hand off all that information to the X server when starting it and allow it to do the "offensive" calculation?

Or perhaps work with the X folks to find some alternative way? We do own the whole stack, as Keith points out, and systemd *seems* like the right place to do this kind of thing ...

jake

Sep	MAR	Aug
	31
2014	2016	2023