Fix mem_max_rss measurement on macOS (#173)

On BSD/darwin, ru_maxrss is defined in "bytes", unlike Linux
where it is "kilobytes".  Therefore, this code was wrongly
multiplying the result by 1024.  Rather than just fixing it
here, it seems better to reuse the already correct `get_max_rss`
function (used for calculating `command_max_rss`) elsewhere.

pyperf/_collect_metadata.py

@@ -23,6 +23,7 @@
                                get_logical_cpu_count, format_cpu_infos,
                                set_cpu_affinity)
 from pyperf._formatter import format_timedelta, format_datetime
+from pyperf._process_time import get_max_rss
 from pyperf._utils import (MS_WINDOWS,
                            open_text, read_first_line, sysfs_path, proc_path)
 if MS_WINDOWS:
@@ -199,10 +200,7 @@ def collect_system_metadata(metadata):
 
 def collect_memory_metadata(metadata):
     if resource is not None:
-        usage = resource.getrusage(resource.RUSAGE_SELF)
-        max_rss = usage.ru_maxrss
-        if max_rss:
-            metadata['mem_max_rss'] = max_rss * 1024
+        metadata["mem_max_rss"] = get_max_rss()
 
     # Note: Don't collect VmPeak of /proc/self/status on Linux because it is
     # not accurate. See pyperf._linux_memory for more accurate memory metrics.