This is the profile that I'm now using across a number of machines that I administer. Most of the tricks I use are fairly general and could be reused by anyone.

curl -s https://raw.github.com/matschaffer/profile/master/install | bash

From here you can now put any *.conf in ~/.profile.d and it'll get loaded in alphabetical order. You can also make directories that end in .conf which profile will then load all the .conf files in that directory in order.

Additionally, if you need to have OS-specific configuration just put a folder in .profile.d or .profile.d/core with the name returned by uname, for example .profile.d/Darwin. The profile checks uname when bash starts and loads any .conf files or directories in this OS-specific folder.

This profile loads init first, followed by all the .conf files in .profile.d/core, then any .conf files in .profile.d. This way you can put any general use configs in core and anything personal in .profile.d. If you want to keep your personal profile information in git, just set up core as a submodule to your personal repository.

This list is not guaranteed to be comprehensive, but here are some of the major features included as of the last time this document was updated.

• Reactive prompt - includes date and exit code of last command and git branch.
• Tab Completion for Git and Subversion
• Tab completion for Rake, Maven 2 and Ant
• Tab completion for ssh hosts on OS X.
• Git aliases
• 'safeedit' function that makes a timestamped backup copy of a file before editing
• Basic java set up on solaris and Darwin
• Maven memory settings
• 'profile_push' function for pushing these files out to other servers
• Auto setup of .foward file on Linux and Solaris
• Integration with CDArgs
• Integration with Autojump
• Integration with RVM

• 'link_dotfiles' command that will create symlinks for all the files listed in the dotfiles directory
• Tab completion for ssh hosts in Solaris/Ubuntu
• Single-command for SSH key distribution and profile distribution

As noted in the LICENSE file. This is all MIT licensed code, so do what you will. But to make things easy for contributions please keep all stuff in core as generic as possible so others can merge it into their own forks. If you need any personal stuff (e.g., API Keys), keep that in a separate git repository that loads this repo as a submodule.