Thrown by gcc 4.9.2 on ppc64.

- %lu and casting to unsigned long will truncate values > 2TB
- Format of %.4g for (unsigned long) totalram is incorrect. And casting to
  (unsigned long) is incorrect anyways.
- %lu for memory_swappiness was incorrect, this is (uint64_t)
  not (unsigned long)
- don't print memory_swappiness as hex, instead indicate if set or not, and
  avoid printing NO_VAL64 if not set
- fix spacing and spread out the ternary operators

Bug 5408.

Bug 3851.

Update to commit 859f6c82

SIGUSR2 in slurmd would re-open the logging files on signal-driven
  basis, possibly resulting in deadlock due to interrupting a running
  thread.
bug 4281

Use %m instead, and print the raw number for posterity.

CID 178117

CID 178118, 178119

Coverity CID 144120

Coverity CID 48915

E.g., gpu:tesla:2 would have parsed as quantity "telsa" of gpu.
Change the index value to the end of the array (-1 index value).

Bug 4250.

GCC is complaining when incrementing a boolean variable, it should
be assigned a value of true or false, and not incremented with ++.

Bug 4212.

Bug 4171

Rather than end up with "%.-1s" printed out in the output as snprintf
refused to parse the format specifier.

Bug 4164.

Now that the x11_forwarding code is running under seteuid/setegid,
and not using _become_user in the stepd, the environment is not set
automatically so xauth was trying to alter /root/.Xauthority instead
of the user's. (That restriction is in place so hostkey auth can work
safely.)

While here, move the xauth cleanup into the x11 forwarding shutdown
thread and out of the stepd's mgr.c.

Caught by -Werror=unused-result.

x11 flags come in the cred here, not in the msg.

For reasons that aren't quite clear, this causes quite a few failures
in the regression suite.

As the X11 race has been solved elsewhere, reverting this is the safest
option for now.

This reverts commit 34b67180332cfec6f4a72cc79c670bd73395c3b7.