I am currently having problems with the sysSleep testGroup, virtually all Linux/Unix platforms fail the test with a too long time.
Currently all Linux/Unix are running in Virtualbox VMs on a macOS 10.15.7 (Catalina, Darwin Kernel Version 19.6.0) Host. It appear that the inherent time variations of the Host and the Guests are adding up. I did not have this problem in the past on a slower machine, I guess it is because I have more than one user active now, and the task switching causes the problem. I have not migrated the Windows and macOS VMs yet so I do not know if they are also affected, but to give you an idea here some timing snapshots: OS Call Time Min Max Delta+ Actual delay CentOS9 SysSleep(0,016) took 0,065303 but should take between 0,016 and 0,042 seconds 0,026 0,049303 Fedora35 SysSleep(4,99) took 5,032791 but should take between 4,99 and 5,016 seconds 0,026 0,042791 OpenIndiana SysSleep(0,333) took 0,360608 but should take between 0,333 and 0,359 seconds 0,026 0,027608 OpenSuse15 SysSleep(0,016) took 0,054691 but should take between 0,016 and 0,042 seconds 0,026 0,038691 Ubuntu22 SysSleep(2,4) took 2,528106 but should take between 2,4 and 2,426 seconds 0,026 0,128106 LinuxMint SysSleep(2,4) took 2,456762 but should take between 2,4 and 2,426 seconds 0,026 0,056762 LinuxMint SysSleep(0,0) took 0,041531 but should take between 0 and 0,026 seconds 0,026 0,041531 FreeBSD SysSleep(3) took 3,065961 but should take between 3 and 3,026 seconds 0,026 0,065961 FreeBSD SysSleep(0,333) took 0,365399 but should take between 0,333 and 0,359 seconds 0,026 0,032399 Would it be ok to loosen this requirement? Can something else be done? Each VM gets two cores and 2 GB memory, that should be sufficient I think? Hälsningar/Regards/Grüsse, ooRexx oor...@jonases.se
_______________________________________________ Oorexx-devel mailing list Oorexx-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/oorexx-devel
