Currently, we show the terminal output from a language runtime that fails to start, like this.
However, every single time I have tried to debug a startup failure, I have wanted to know what the environment variables were set to when the kernel was launched (as well as the exact launch command), and have wound up hacking print statements into kernel-wrapper.sh so I can see them. We should log these somewhere so that they can be brought up for debugging.
They don't need to be front-and-center, since showing all the environment variables at once creates a wall of text that makes the actual startup error hard to find, but they should be accessible somewhere.
Currently, we show the terminal output from a language runtime that fails to start, like this.
However, every single time I have tried to debug a startup failure, I have wanted to know what the environment variables were set to when the kernel was launched (as well as the exact launch command), and have wound up hacking print statements into
kernel-wrapper.sh
so I can see them. We should log these somewhere so that they can be brought up for debugging.They don't need to be front-and-center, since showing all the environment variables at once creates a wall of text that makes the actual startup error hard to find, but they should be accessible somewhere.