This is the last set of fixes that I had waited to apply until the other pull requests resolved. I think that this should apply cleanly after applying pull request #80 (as I used that as the base for this branch).
Notably, this pull request includes:
Changes that uses OpenMP more efficiently (which can wildly speed up permutation testing when there are multiple cores available).
Using the "divsufsort" library to calculate the suffix array (SA) for the t-tuple and (various) lrs tests. This is much faster and more memory efficient than the prior approach, but this does impose an additional build dependency.
Some additional scientific computing work and bug fixes for the chi-square tests.
Establishes the cutoff for the restart testing using simulation, rather than making the assumption that the underlying distribution is binomial. Note: This particular set of functionality fixes a problem in the underlying special publication.
Make a few changes to silence warnings from clang's static analysis tool. Most of these were technically issues, but didn't cause problems in practice.
This is the last set of fixes that I had waited to apply until the other pull requests resolved. I think that this should apply cleanly after applying pull request #80 (as I used that as the base for this branch).
Notably, this pull request includes:
Closes #78, closes #79, closes #82, closes #61, closes #51, closes #82, closes #61, closes #86, closes #55, closes #88, closes #89, closes #90, closes #91, closes #93, closes #95, closes #56, closes #92, closes #52, closes #53.