… failing one test each now
…at the test suite actually runs (but several tests fail)
…s (stcks, queues) we can inherit most of the tests from a hierarchy of parent test classes. Almost all tests now are in common.nqp and inherited by the individual files.
- FPQ_ITEMS() raised divide-by-zero before size was set (get rid of modulus operations) Changed Integer coercion to return .elements() Added .capacity() to return (apparent) allocated size More tests, with better descriptions More unification of the 3 different source files
…ill fails a does test that I can't quite explain. FPQ2 and FPQ3 are still aborting early with a floating point exception.
…to setup.pir so they build, and benchmarks.sh so the benchmarks run. Both types exhibit Floating point exceptions and fail tests/benchmarks
…hmark. Fix that and fix some output formatting
…sed benchmarks with a factored library file
…ing (but currently never shrinking flat memory buffer. Benchmark times are better-but-similar (33% faster) to RPA for push, ~200x better than RPA for pop.
…hmark pass. RPS shows marginally-better push performance than RPA due to less frequent memory reallocations. However pop performace is significantly less than RPA (0.17s for RPS compared to 0.05s for RPA). This demonstrates that the mechanism behind this type is flawed and a new algorithm needs to be determined.
…PQ has similar push/shift speeds, and is two orders of magnitude faster in shift operations than RPA. Is modestly slower than RPA for push.