Update 3D Euler normal Riemann solver to more modern Fortran.

[ketch]

This is a minor modification of #71. It is a rewrite by @weslowrie of the 3D Euler normal Riemann solver, using more modern Fortran. I modified it as follows:

• use wave rather than qwave, following convention of all other Clawpack solvers
• Squash several separate commits into one, to keep history cleaner
• use 4-space indents, consistent with what all Clawpack code is supposed to use (some other Riemann solvers need to be fixed in this respect)
• Remove some comments that simply restate code
• Remove comment styling

[ahmadia]

+1. @weslowrie - You will probably want to review this.

[ketch]

I would like to put a 3D Euler test into PyClaw before merging this, so that we can check that it gives the same output as the current routine.

[weslowrie]

Looks good to me. Thanks for your edits.

The sample problems I have done are a 1D shock tube in a 3D domain, and a Sedov blast problem in 3D. Do you want me to add those python scripts to the repo?

[ketch]

@weslowrie Yeah, go ahead and submit a PR with the scripts. I can add a test or two that uses them.

[weslowrie]

@ketch I submitted a PR with the scripts. Hopefully they are useful for setting up tests.
For the Sedov blast problem, it could use some more resolution, but the full 3D problem might get to big for a simple test?

[ketch]

Closed in favor of #75.