uname output ?

[Paul Iadonisi]

On Fri, 2002-08-16 at 14:55, bscott at ntisys.com wrote:

[snip]

> 
>   Here's a radical idea: Get the source for the package and see where the
> hell it gets all these identification strings in the first place.  :)

  Very radical.  Which is what I've just done.  The -o option causes the
(preprocessor) value of HOST_OPERATING_SYSTEM to be printed out.  This
value is defined as $utils_cv_host_operating_system in the configure
script (it gets appended to confdefs.h, I think unconditionally). 
*This* value gets set in m4/host-os.m4 and appears to be based on
host_os which I think is set according to the guessed host os (or the
--host option given to configure).  There is a case statement that has a
list of about 35 or 40 possibilities.  If the host_os is 'linux*' the
value chosen is "GNU/Linux".  There is a heuristic at the bottom that
defaults to just uppercasing the first character of the host_os.
  There's not much available in the way of customizing it (short of
applying a patch to hard code something in place of "GNU/Linux".  Add to
that the fact that Red Hat's latest beta, Limbo2 (which does "ship" with
sh-utlis 2.0.12), returns "GNU/Linux" from the -o option and add to that
the fact that it's probably only Linux based OSes that ship with GNU
sh-utils, and I'd have to agree that this is a pretty useless feature.
  
-- 
-Paul Iadonisi
 Senior System Administrator
 Red Hat Certified Engineer / Local Linux Lobbyist
 Ever see a penguin fly?  --  Try Linux.
 GPL all the way: Sell services, don't lease secrets