Separated nrepl code, added CLI options, added failover

[chrisbetz]

Hi,

I'm not sure whether you're interested in this sort of changes or not, so feel free to just drop this PR. But for my use-case (combining GorillaREPL with ApacheSpark) it's really important, so it might be interesting for others. Main rational is I needed to avoid dependency conflicts.

Here's what I did:

• I separated nrepl code from gorillaREPL webserver to allow for both a standalone version (nrepl + GorillaREPL) and a GorillaREPL-version with the ability to connect to an existing nrepl server.
• I added CLI options to configure GorillaREPL
• I added the ability to re-establish a connection from GorillaREPL to the nrepl server. This is still work in progress, as the current REPL-state might get lost and the end-user will not be notified, but it's actually better than breaking the GorillaREPL (in my eyes).

Feel free to take whatever is necessary and wanted and be assured that I find GorillaREPL a really cool project!

Cheers,

Chris

---

This change is 

[JonyEpsilon]

Hi Chris,

thanks, this is great :-) That is a feature I'd really like to get into the next release (see #137). I should have time in a few weeks to sit down and put a release together, so will return to this PR then.

Thanks again,

Jony

[chrisbetz]

Oh, you're welcome. Didn't see that discussion before ;)

Just take your time!

On Tue, Jun 30, 2015 at 12:35 PM, Jony Hudson [email protected]
wrote:

Hi Chris,

thanks, this is great :-) That is a feature I'd really like to get into
the next release (see #137
#137). I should have
time in a few weeks to sit down and put a release together, so will return
to this PR then.

Thanks again,

Jony

—
Reply to this email directly or view it on GitHub
#220 (comment)
.

[JonyEpsilon]

There's a broader discussion that might be of some interest too, thinking about what the relationship should be between nREPL, Gorilla etc. It's at #184 in case you'd like to look :-)