gstreamer1 - add replacement for sysconf(_SC_NPROCESSORS_ONLN) on Tiger

[ken-cunningham-webuse]

I've made this function change three times so far, so I thought I'd finally submit it for consideration.

sysconf(_SC_NPROCESSORS_ONLN) is 10.5+ only.

this patch adds a suitable BSD replacement for Tiger.

[ken-cunningham-webuse]

and evidence

tigerg5:~ cunningh$ port -v installed gstreamer1
The following ports are currently installed:
  gstreamer1 @1.8.2_0 platform='darwin 8' archs='ppc'
  gstreamer1 @1.10.0_0 (active) platform='darwin 8' archs='ppc'

[dbevans (David B. Evans)]

Will look at this later today.

[ken-cunningham-webuse]

to make life easier:

$ cat testnumproc.c

#include <unistd.h>
#include <stdio.h>
#include <sys/resource.h>
#include <sys/types.h>
#include <sys/sysctl.h>


int main() {
        int nm[2];
        size_t len = 4;
        uint32_t count;
        uint32_t num_cpus;
          
        nm[0] = CTL_HW; nm[1] = HW_AVAILCPU;
        sysctl(nm, 2, &count, &len, NULL, 0);
          
        if (count < 1) {
                nm[1] = HW_NCPU;
                sysctl(nm, 2, &count, &len, NULL, 0);
                if (count < 1) { count = 1; }
         }
         
         num_cpus = count;
         
         printf("number of cpus %d \n\n", num_cpus);
 } 


$ which gcc
/opt/local/bin/gcc

$ gcc --version
powerpc-apple-darwin8-gcc-apple-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5666) (dot 3) (MacPorts apple-gcc42 5666.3_15)
Copyright (C) 2007 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

$ gcc -o testnumproc testnumproc.c

$ ./testnumproc
number of cpus 2 

[dbevans (David B. Evans)]

Revised patch that can be applied unconditionally

[dbevans (David B. Evans)]

I've just attached a revised version of your patch that can be applied unconditionally. This is more like something that the upstream developers might accept as a modification and that's where it really should be addressed.

I've successfully tested this on El Capitan and Tiger. Let me know if it works for you and, if so, I will commit.

[ken-cunningham-webuse]

beautiful. I see how you rounded it out to a more complete fix. I'll check it on a few more systems and get back to you, but I can't see what problems it might raise with the guards you've put on it. thanks, K

[ken-cunningham-webuse]

Looks good, Dave. Clean builds also on Leopard PPC, Snow Leopard with LibcxxOnOlderSystems, and Sierra.

Thanks for your attentions. I'll leave it to you to decide if this goes upstream, if that's OK. You have more cred up there than I do :>

K

[dbevans (David B. Evans)]

It needs more work to go upstream. The patch as it is works fine on Macs but the case where

__APPLE__

is not true is not handled correctly -- should go back to the original code again. I'll fix that and then commit the patch. Thanks for your contribution.

[dbevans (David B. Evans)]

In 4678a5ef/macports-ports:

gstreamer1: patch to fix build on 10.5 and earlier.

Fixes #52822