v2.6.0

ParallelTestRunner v2.6.0

Diff since v2.5.1

Merged pull requests:

• Use Base.Semaphore to control test execution parallelism (#119) (@giordano)
• Fix init value of count in _count_child_pids (#126) (@giordano)
• Add a check that there are no running workers at end of the tests (#127) (@giordano)
• Fix styling in print_header function (#128) (@christiangnrd)
• Features for CUDA.jl conversion (#129) (@maleadt)