A post about the distributed version control system "git" in two parts.
Part One: Git Puns
#ubuntu-one... No thanks! I'll stick with my home-brewed git + server + usb drive solution. My git repos breed like rabbits!
Which basically sums up my opinion on ubuntuone. But I thought that the "my git repos breed like rabbits" was both accurate (git repositories are designed to be replicated in their entirety), and a sort of funny way to put it. And being the kind of person that I am, I decided to see what other (potentially dirty) puns I could make about git. Here's what I came up with:
what did one git repo say to another git repo? pull my diff
what did mama git say when she found her remote in his room making new branches? octopus merge this instant!
what did one git remote say to entice another remote to branch? it's ok we can just tell them we were cherry picking later.
what did dr. git say when a repo complained of bloating? git gc
I should point out that these four puns all demonstrate a factual feature of git, though the "pull my diff" isn't exactly what happens.
"Octopus Merge" is the method that git uses when there are a lot of divergent branches (more than three) that need to be merged together. Similarly "cherry picking" is a way to manually select what changes get merged together if you're not ready to do full merges, and git gc is the cleanup script that goes through and re-compresses and prunes the database so that your repo works faster and with less disk space.
Anyway, I'm out of puns, you all are welcome to join in.
Part Two: Atypical uses of Git.
I'm sure I've written a bunch here about how I'm not really a programmer, and while this is true I do use git a lot. In part I think this is because git is really mostly an ad-hoc file system and also given how I write, the kind of writing I do isn't that different from programming.
So aside from storing my writing projects, and my orgmode, I do things like store all of my mail directories in git. Which you might think is kind of weird, but the truth is that it makes keeping lots of computers in sync a rather simple proposition, and its damn fast.
I also have a directory I call "garen" (but used to call "main") that is basically my home directory. It has all my emacs lisp files, most of my non-mail related scripts, various configuration files. and so forth. It started out as a backup and workspace for smaller projects, but it's since morphed into "that one thing I need to have of my computer in order to actually work." When I was setting up the server it took a thousand things that might have been huge headaches and made them non-issues. Here's what this repo looks like:
emacs/ This is where my emacs-lisp files all live. I have a 'init.el' file which is basically the standard .emacs file, and a 'gui-init.el' file for code that I only want to run if I'm running desktop where I'll be running non-console emacs frames. As a result on my machines my .emacs file looks like this:
(load "~/garen/emacs/gui-init.el") (load "~/garen/emacs/init.el")
With the first line commented out if needed. End result, emacs loads the same everywhere, no thinking.
scripts/ I add this to my path, so that any little bit of bash script that I want to be able to use is accessable and the same on all my machines.
configs/ Generally my format is to have config_file.machine_name, for example: bashrc.leibniz. In the case of the bashrc, I have a ".common" file that has everything that all my machines need, while the machine specific files have everything that's... well specific, and a source statement for the common file. So my "real" .bashrc looks like this:
And everything stays in sync between the machines. How cool is that.
That's sort of the most important thing. The great thing is that this makes setting up a new user account on a server, or a box itself a piece of cake.
Food for thought!