dev and prod environments

[shalunov]

Separate environments so the keys can be different in dev and in prod. Prod keys obviously should not be in the repo.

[shalunov]

Before writing any code, let's first discuss the proposed architecture. It's possible that this is a no-code ticket where we specify the private key file from the command-line in production or something to that effect.

[inetic]

I like the command line idea. If the command line argument isn't used, then hardcoded debug keys are be used and the user is notified on the std output.

[shalunov]

That's not the entire architecture. The client needs to know which public key to use. Also, we need to decide how to differentiate between production and development injectors and proxies if they're in the same swarm.

Feel free to propose something, or assign to me to specify.

[inetic]

Also, we need to decide how to differentiate between production and development injectors and proxies if they're in the same swarm.

I'd go with the approach that we don't mix production and development injectors in one swarm (same for proxies). That is, instead of injectors registering in DHT(SHA1("injectors")) they would register in DHT(SHA1("injectors" ++ seed)) where seed is specified on the command line (e.g. with -s) and by default it's just an empty string.

This has an advantage that concurrent tests won't interfere with each other.

When testing, the seed can be generated at random like this.

[inetic]

I have implemented what I described in my previous comment in this commit (as part of PR #59).

@shalunov, let me know if you want me to continue with setting up public and private keys using command line before @ghazel merges his ghazel branch (seems like it might introduce conflicts).