test: Remove java comparison tool

[laanwj]

There is a growing number of issues with it, and no one is really maintaining it anymore. From #4545 (comment):

Seems I'll just have to set a deadline: I'm going to remove use of the unmaintained java-based comparison framework after the 0.13 branch-off (projected 2016-07-06, see #7679). I hope the Python framework will be a full replacement by then, but as this is holding up progress in other areas, that will not affect the decision.

@MarcoFalke worded it quite well too:

If they don't and no one knows how to fix, we can remove them right now. Because the only thing they'd be doing then, is slowing down travis.

• The effort of porting the java comparison tool to Python is going very well, see e.g. Continuing port of java comparison tool #8141. If there is something left to do, I'm sure it can be resolved before the 0.14 release.
• The test is blocking changes such as Address mempool information leak and resource wasting attacks. #7093 (Address mempool information leak and resource wasting attacks).
• This test takes quite a while and is holding up enabling the tests for win32: tests: Re-enable RPC tests for Windows #8227.
• comparison tool support seems to have been dropped from the v0.14 release of bitcoinj (discovered @droark)

This removes all the support for the tool, and removes it from depends (and hopefully all other references to it).

Not to be backported to 0.13.

[maflcko]

Thank you!

ACK 3c5251d

[sipa]

Concept ACK

[laanwj]

@theuni can you have a look at the depends/build system changes here?

[droark]

ACK

Definitely not the person to sign off on the depends/build system changes but what I see strikes me as pretty thorough. I did a bit of a deep dive awhile back and seem to recall what's here as being everything that's related to the tool.

[theuni]

Woohoo, great to see this! Thanks to everyone involved.

Build-side looks good. The .travis.yaml can drop the hostname hack since we no longer require java, but that can be done as a follow-up.

utACK.

[laanwj]

@theuni Did I remove the right hack?

[maflcko]

Yes, eb0c52e looks right. (c.f. #8253)

[theuni]

@laanwj yup, thanks.