make bash stricter

[groob]

Closes #15

[zwass]

This change appears to cause the script to exit when we wait for MySQL to accept connections. Repros consistently for me if I do ./demo.sh reset && ./demo.sh up

[groob]

Trying to figure it out since -e will fail with non-zero exist status, but that's what we use to check for mysql connectivity.

I'll give it some thought and see if I can come up with a strategy. One way we could achieve this would be to move the for loop into a script which runs in the container, instead of initializing a new one. That way it will exit 0 when it finally connects, or exit 1 if 50 attempts failed.

[zwass]

Is the current strategy of manually handling the exit not sufficient? I am also concerned that -e will exit with no error printed.

[groob]

The reason the error is not printed is because we pipe the docker output to > /dev/null

IMO it's important to close #15, I've run into a few cases already where the port is being used(mostly because I test this set up over and over in a new folder) and the experience is pretty poor. The loop doesn't handle SIGINT well, so hitting CTRL+C doesn't help.

As far as how -e works, It will abort the rest of the script if the last command in a pipe exits non-zero. If there's a different straegy for handling it, we can try.

[directionless]

FWIW I usually handle this by setting -e globally, and toggling it off for specific places I need to check something. (set +e)