We’re using SG launcher, great central way to control software, artist friendly etc.
What is a difficult is debugging. When we launch houdini for example, launcher starts an xterm, runs the environment, starts houdini. If we want to inspect the environment that’s been created, we can’t; the xterm is read only.
Using in-house launchers at other studios I’ve been able to select the terminal type, and choose if its non-blocking or not. In this way if houdini crashes we get dropped back to a read/write prompt to see what happened, or we can ctrl-t (from gnome terminal say) to open a new tab which inherits the same environment as the original tab.
Artists love it too, as they don’t have to wait for the (often necessary) slow launch process while shotgun is looking up variables, rez goes and configure things etc… if they don’t need a refreshed environment, software can crash, artists swear, shrug, up-arrow and hit enter in the terminal, and they’re back up and running in seconds rather than minutes.
Have asked around, our td’s have done a few tests, they say its not easy to swap terminals, or if they tried it broke stuff.
Anyone got clues here? Would be super handy to solve!
This is SG Launcher under centos, using rez to do a lot of the heavy lifting.