excellent -I'm really glad java -jar -S gets "finished", would be great if it still worked without Bundler as well ...

I plan to test it without bundler too. 

@kares as far as I know this is working as expected. There aren't any automated tests that I can find around this functionality but the general expected outcomes work for me and the existing tests pass (though Travis seems to be consistently running this project, so I can't speak for the full build matrix).

I'd like to do a refactoring of the "runnable" responsibilities entirely (as well as make it available in JarMain as well as WarMain) but they are probably best served as different pull requests.

Hey Bo, yeah it seems all fine. There sure are no tests (besides the original code being a "draft" :)) I only did testing locally with a Rails app as well ... maybe @BanzaiMan can merge this than (refactoring to be done on another PR) ?

@kares ok. Make sure you tested after running rake jar. I didn't commit a regenerated jar (partially because I don't like committing build artifacts and partially because I only wanted to commit the working version).

@kares, have you seen those Travis failures before? I am not able to reproduce them locally or reason about what part of my change could have caused them to fail.

This is littered near the failures in the output on Travis:

fatal: attempt to fetch/clone from a shallow repository
fatal: Could not read from remote repository.

Please make sure you have the correct access rights
and the repository exists.

A cursory glance at the specs make it seem like it is bundling the local git repo as a gem in a test. My (still unsubstantiated) theory is that Travis has started only doing shallow clones of tested repositories to lessen network load.

bump

This looks extremely useful, and not having this is one of our largest pain points in our jruby deploy right now (not having a simple solution for utility servers). So I 👍 this.

I've fixed the ci failures. working on getting this merged now.

👏

I'd really like to see some specs for this. It would also be nice to refactor the runnable responsibilities as you describe (pretty much everyone who uses regular non-exectuable wars has this same need). I'll look into these things myself. let us know if you have any ideas for how to go about it.

I completely agree. There wasn't a real precedent for this stuff, but am happy to help out. I actually really want to do the refactoring but held back both because I didn't want this PR to be rejected due to too many changes and because of a lack of test infrastructure. If we can agree on a structure for that aspect, I'd be more than happy to help give these aspects a revamp.

Oh also @jkutner I didn't commit the compiled results (simply so as not to add repo weight until the changes were accepted). Be sure to do a compilation before any new gem builds!