I've started a branch called "autopilot" that tries to hive out the parts that require user input from those that don't. The point is to allow a full install with blank passwords to be run automatically (for instance, as part of a Vagrant machine startup, which is my current goal); for a local version, this is fine. The passwords are only necessary if running the thing on the open web; if you just want to test a local server, it doesn't really matter what they are.
It also means that git clone commands need be done using the http protocol instead of the git one. The federalist example now does that, and I'll start to migrate some slightly more complicated example bookworms to that as well.
To handle passwords, I think the best strategy would be first to setup a server with blank passwords, and then have a script that interactively alters the .my.cnf files and configuration files with the user-generated passwords. That way further changes won't require maintaining two separate forks.
I've started a branch called "autopilot" that tries to hive out the parts that require user input from those that don't. The point is to allow a full install with blank passwords to be run automatically (for instance, as part of a Vagrant machine startup, which is my current goal); for a local version, this is fine. The passwords are only necessary if running the thing on the open web; if you just want to test a local server, it doesn't really matter what they are.
It also means that git clone commands need be done using the http protocol instead of the git one. The federalist example now does that, and I'll start to migrate some slightly more complicated example bookworms to that as well.
To handle passwords, I think the best strategy would be first to setup a server with blank passwords, and then have a script that interactively alters the .my.cnf files and configuration files with the user-generated passwords. That way further changes won't require maintaining two separate forks.