An now for the stupid question of the day: md5sum differences for "/bin/cp"

[Will McDonald]

On 09/03/06, David Tonhofer, m-plify S.A. <d.tonhofer at m-plify.com> wrote:
> Does anyone know why e.g. /bin/cp should yield different
> checksum on different machines that have the same OS (RH4.0)?

That is weird. I've just checked across a bunch of CentOS4.2 systems
and they all display the same behaviour, i.e. different md5sums.

[wmcdonald at willspc tmp]$ md5sum local.cp webdev1.cp
d5a727482a4fa3c0a0057233ff4e166f  local.cp
faf79484269dd8c3ea6d06914f0bf479  webdev1.cp

I copied and "string"ed two copies of cp and the string differences
were minimal...

[wmcdonald at willspc tmp]$ diff localcp.strings webdev1cp.strings
2,4d1
< N OC
< @ OC,;/D
< @ OC

Yet the initial md5sum in rpm -q --dump appears identical...

[wmcdonald at willspc tmp]$ rpm -q --dump coreutils | grep /bin/cp
/bin/cp 55920 1124790615 4d5e0ec0bbe25886c9dcb45b879382e3 0100755 root
root 0 0 0 X
[wmcdonald at willspc tmp]$ ssh webdev1 "rpm -q --dump coreutils | grep /bin/cp"
/bin/cp 55920 1124790615 4d5e0ec0bbe25886c9dcb45b879382e3 0100755 root
root 0 0 0 X

And yet the coreutils package successfully passes rpm's verify.

Will.