On  8/06/10 07:59 PM, Steve McKinty wrote:
Why are these "not relevant"? In my experience they are mostly
used in scripts and Makefiles, to ensure that the right code path
is taken. Won't removing them break older configure-style scripts,
i.e. ones that test things like "if [ ! vax ]" etc.?
Is the gain of removal worth more than potential breakage of
older (especially FOSS) code?

...
/usr/bin/iAPX286
/usr/bin/i286
/usr/bin/i860
/usr/bin/pdp11
/usr/bin/u3b
/usr/bin/u3b2
/usr/bin/u3b5
/usr/bin/u3b15
/usr/bin/vax
/usr/bin/u370


Can you tell me, with a straight face - and proof- that
Solaris 2.6 or later will even run on *any* of the above
processors?


The above processor truth types are all links to machid,
which contains this comment in $SRC/cmd/machid/machid.c:

/*
 * This program replicates the function of the links from a machine name
 * (such as sun4c) through /usr/kvm to true or false as appropriate.  It
 * knows the correct special cases.
 *
 * IMPORTANT NOTE:
 *
 * Do not modify this program to know about additional special cases or
 * reflect new platforms or instruction set architectures.  This is a
 * deprecated interface and strictly for backwards compatibility.  This
 * is psarc/1992/171.  Note the following excerpt from the opinion:
 *
 *    It is most important to note that the manual page states in
 *    the NOTES section:  "The machid family of commands is
 *    obsolete.  Use uname -p and uname -m instead."
 *
 *    The intent of Kernel Architecture Project team is to provide
 *    only enough functionality to mimic the existing definitions
 *    on the SPARC and Intel x86 versions of Solaris 2.x.  No new
 *    identifiers will ever be added to the documented and
 *    undocumented identifiers listed above.
 */



I think it is long past time that we actually obsoleted
the above interfaces.

Pete: +1


James C. McPherson
--
Senior Software Engineer, Solaris
Oracle
http://www.jmcp.homeunix.com/blog
_______________________________________________
opensolaris-arc mailing list
opensolaris-arc@opensolaris.org

Reply via email to