Bug#673461: [Pkg-puppet-devel] Bug#673461: facter: processor fact is wrong on some arches

[Stig Sandbeck Mathisen]

Stephen Gran sg...@debian.org writes:

 I recently broke a few things (oops, mea culpa) by assuming rather
 naively that all machines have processors. Apparently, facter does not
 agree.

 A quick looks says:

[...]

 It seems as though this fact really only works the way it's intended
 to work on intel/amd chips?

That is rather strange. I'll investigate.

Is this a mix of physical and virtual servers?

-- 
Stig Sandbeck Mathisen


pgprCBFQmTcWt.pgp
Description: PGP signature

Bug#673461: [Pkg-puppet-devel] Bug#673461: facter: processor fact is wrong on some arches

[Stephen Gran]

Hi,

This one time, at band camp, Stig Sandbeck Mathisen said:
 Version: 1.6.2-1
 
 micah anderson mi...@riseup.net writes:
 
  Can you try the newest bpo version of facter to see if it still is
  the same?
 
 in 1.5.7, facter parses /proc/cpuinfo, which unfortunately has a
 different format for each architecture.
 
 In 1.6.2 (commit at http://deb.li/ii11z), the processor facts
 started using sysfs instead. I'll mark this bug as closed, with this
 version.

It's quite a bit better, in that it now gives the correct processorcount
on linux machines.  It still doesn't give a processorcount on kfreebsd
at all.  It seems this is trying to use sysfs as well - that won't work
on kfreebsd.

While processorcount is now correct, it seems that other processor
information is still missing on s390, s390x, ppc, mips and mipsel.  That
just seems to be missing switches in the case statement.

Cheers,
-- 
 -
|   ,''`.Stephen Gran |
|  : :' :sg...@debian.org |
|  `. `'Debian user, admin, and developer |
|`- http://www.debian.org |
 -


signature.asc
Description: Digital signature