Re: [SCM] Debian package checker branch, master, updated. 2.5.9-27-g9e61a25

[Jakub Wilk <jwilk@debian.org>]

* Niels Thykier <niels@thykier.net>, 2012-06-27, 17:12:
> +    if ( -r '/proc/cpuinfo') {
> +        my $jobs = 0;
> +        my $cpus = 0;
> +        open my $fd, '<', '/proc/cpuinfo'
> +            or fail ("failed to open /proc/cpuinfo: $!");
> +        while (<$fd>) {
> +            next unless m/^cpu cores\s*:\s*(\d+)/;
> +            $jobs += $1;
> +            $cpus++;
> +        }
> +        close $fd;
> +
> +        if ($cpus) {
> +            # Running up to twice the number of cores usually gets the most out
> +            # of the CPUs and disks but it might be too aggresive to be the
> +            # default for -j. Only use <cores>+1 then.
> +            debug_msg (1, "Number of cores: $jobs (CPUs: $cpus)");
> +
> +            return $jobs + 1;
> +        }
> +        debug_msg (1, "Did not find any core information for any cpus in /proc/cpuinfo?");
> +    }
> +
> +    # No decent number of jobs? Just use 2 as a default
> +    return 2;
> +}

I would advice against parsing /proc/cpuinfo. Format of this file varies 
with architecture, and it's very easy to get it wrong. For example, the 
subroutine returns 5 on my machine (even though it has only one CPU with 
2 cores) and 2 on gabrielli.d.o (16 cores).

How about something like this:

my $cpus = int(`nproc` or 1)
return $cpus + 1

?

--
Jakub Wilk