Port ioctl prototype check to CMake #3

This looks very good – like the right solution to replace pull request #2.

I am going to test your PR in a moment. BTW, just one detail; unsigned long and unsigned long int are the same thing (I don't know if you want to preserve both here)

@selk "unsigned long int" v.s. "unsigned long" is the questional and WIP point.
I don't know the historical reason why aRts introduced the check.

Far as I can see, it is safe to remove unsigned long int:

http://man7.org/linux/man-pages/man2/ioctl.2.html
https://www.freebsd.org/cgi/man.cgi?query=ioctl&sektion=2
https://man.openbsd.org/ioctl.2
https://docs.oracle.com/cd/E26505_01/html/816-5167/ioctl-2.html

I recently tested your (WIP) patch under Musl and it worked. :-)

https://mirror.git.trinitydesktop.org/gitea/TDE/arts/src/branch/master/artsc/artsdsp.c#L78

long and long int are the same data type. I suggest we drop long int and update this PR and arts code accordingly so that we have only HAVE_IOCTL_INT_INT_DOTS and HAVE_IOCTL_INT_ULONG_DOTS.

Any feedback is welcome.

Just only HAVE_IOCTL_INT_ULONG_DOTS is sufficient, others should be the standard compliance?

Tests 1 – HAVE_IOCTL_INT_INT_DOTS and 2 – HAVE_IOCTL_INT_ULONG_DOTS are necessary. Test 3 – HAVE_IOCTL_INT_ULONGINT_DOTS seems to be actually the same as 2, because long and long int determine the same thing.

Therefore, it seems that we can exclude test 3 as well as the condition in artsc/artsdsp.c.

changed as suggested

It looks good. 👍

perfect, thanks!