How much should minisat be cleaned up?

[horenmar]

There are various low hanging fruit to use in improvement of MiniSat, starting from bringing in work from niklasso/minisat, through modernizing it to use C++11 (which would also allow us to stop it from forcing __STDC_LIMIT_MACROS and __STDC_FORMAT_MACROS defines) all the way through fixing up things like Minisat::Vec implementation (ATM it leaves performance on the table for no reason what-so-ever).

Which of these are worth it, and which are not?

[horenmar]

I added some integration tests to ensure that potential future changes don't break things.

[dralley]

This fork is 74 commits ahead, 68 commits behind niklasso:master. Is there any reason why those latter 68 commits were disincluded? Since this fork isn't strictly ahead of the original one, it's hard for me to judge what the differences are.

[cernoch]

@dralley We started at version 2.2, which is the lastest release, widely used and at least “tested by experience”. Since there are no automated tests, I wasn't sure about the code quality of the other 68 commits.

Nevertheless, many of these changes look reasonable (e.g. niklasso@369fc3f). Please feel free to start backporting!