A few minor cleanups

[rdeits]

Thanks for making this available! I'm playing around with connecting PATH to https://github.com/rdeits/ConditionalJuMP.jl and noticed some opportunities to clean up the code a bit. This PR does two things:

1. get rid of the redundant solveLCP methods without var_name and con_name arguments and replace them with default values. Default argument values in Julia do exactly what you were doing by hand here, so there's no reason to write out the other methods.
2. get rid of the global user_f and user_j functions. You can call cfunction on a closure with no issues, so we can create a custom cfunction for the user's particular callback.

Currently I haven't observed (2) making any performance difference, but I'm planning to build further on that change to solve LCPs (without calling ForwardDiff.jacobian at every iteration).

[coveralls]

Coverage remained the same at 100.0% when pulling e29f4da on rdeits:cleanup into 7e4928a on chkwon:master.

[rdeits]

Oops, point (2) was not entirely true. I'll find a more elegant solution.

[chkwon]

Thanks again for your PR. Your project ConditionalJuMP.jl looks quite interesting. For complementarity constraints, I tried to do some work in https://github.com/chkwon/Complementarity.jl

See https://github.com/chkwon/Complementarity.jl/blob/master/MPEC.md