3 ms·

I agree with the general notion, but not with the specifics. That is: It's nice to code on a system that "shares the exact same stack as what I host with", but

by batista 15y ago

I agree with the general notion, but not with the specifics. That is: It's nice to code on a system that "shares the exact same stack as what I host with", but just running on Linux is not doing it. Except if you only ever host with one single stack. Normally, you'd have different stacks for different projects, different technologies (Ruby, PHP, etc), others would have Apache, others NGINX, MySQL or Mongo, different major or minor versions, etc. So, I find the best way is to setup a VM with each hosting environment and work on that. You get easily reproducible stack, snapshots, isolation, and a cleaner main OS X or Linux install. With one click on the next image, you can completely change your environment.

regularfry 15y ago
We try quite hard to keep our stack homogeneous. We're Debian Squeeze from dev through to deployment, except for a very few Lenny services which haven't been upgraded yet (and won't be, until the service itself is replaced). This means that getting a new service up and running should, by and large, be trivial. There's no discussion over what the best platform for a given service is, it's just a question of how to get it working on Squeeze. If we have to package something up to make it deployable, that's what we do. Anyone coming to an arbitrary system knows what the tools are, and where to look. It's very, very simple.