Define HAVE_KERN_PROC on FreeBSD to fix rust-lang/rust#54434 (#294)

[przygienda]

Define HAVE_KERN_PROC on FreeBSD to fix rust-lang/rust#54434 (#294)

that precisly broke my FreeBSD 11 build here

clang version 5.0.2 (based on LLVM 5.0.2)
Target: amd64-unknown-freebsd11.0
Thread model: posix

[backtrace-sys 0.1.35] cargo:warning=src/libbacktrace/fileline.c:113:45: error: use of undeclared identifier 'KERN_PROC'
[backtrace-sys 0.1.35] cargo:warning= return sysctl_exec_name (state, CTL_KERN, KERN_PROC, KERN_PROC_PATHNAME, -1,
[backtrace-sys 0.1.35] cargo:warning= ^
[backtrace-sys 0.1.35] cargo:warning=src/libbacktrace/fileline.c:113:56: error: use of undeclared identifier 'KERN_PROC_PATHNAME'
[backtrace-sys 0.1.35] cargo:warning= return sysctl_exec_name (state, CTL_KERN, KERN_PROC, KERN_PROC_PATHNAME, -1,
[backtrace-sys 0.1.35] cargo:warning= ^
[backtrace-sys 0.1.35] cargo:warning=2 errors generated.

[przygienda]

0.1.32 was compiling and working fine, 0.1.35 is broken per above

[alexcrichton]

cc @t6, can you help taking a look at this?

clang version 5.0.2 (based on LLVM 5.0.2)
Target: amd64-unknown-freebsd11.0

That's a very old Clang version. What FreeBSD version are you on?

FreeBSD < 11.3 is not supported anymore since last year.

[przygienda]

sigh, ok, yeah, I'm on old FreeBSD 11.0 ...

[przygienda]

FreeBSD 11.0 is 3 years old only. It's pretty frightening if such a base package is obsoleting that fast frankly. I'm relying on that in long chains of dependencies and things like GNU/C or C++ drag stuff for 20 years ...

[przygienda]

any way you gentlemen can add a feature for BSD 11? I'm not sure how to recursively force all packages to compile libbacktrace with such a feature but I can't move up from BSD 11 so it basically completely broke my code since stuff is being pulled in by sysctl which I don't want to factor out ...

[cemeyer]

In March 2020, 11.0 was three years and five months old, 11.1 was two years and eight months old, 11.2 was a year and eight months old, and 11.3 was eight months old.

Can we close this issue?

[alexcrichton]

Indeed!