🚀 go-pugleaf

Hi, Michael.  I apologize for arguing, I'm really not wanting to start
a debate.

	 Michael> [...] Configure should correctly figure out the
	 Michael> options to use, or it has a bug.  [...]

      Franc,ois> These statements look quite strong to me [...]

   Michael> If a program cannot be completely configure by autoconf,
   Michael> then other things need to be done.  GCC does not use
   Michael> autoconf, but it still needs no editing to be installed;
   Michael> similarly for gdb and the other cygnus-maintained
   Michael> utilities.

If I understand things correctly, Cygnus configuration process relies
on a database expressing the capabilities or features of a wide
variety of systems.  Editing *is* indeed required somewhere for any
system unknown to the database, preferably by completing the database,
of course, or else, the lazy way, by adjusting the Makefile.

   Michael> The coding standards make this quite clear.  Users should
   Michael> never need to edit machine generated files.

I read the GNU standards as discouraging the edition of generated
files, not prohibiting it.  I would perfectly agree with you in a
world in which all possible machines/systems would be tabulated for
Cygnus configure or properly checked by Autoconf configure's.  What
triggered me to reply is that I just felt incomfortable to call a
*bug* the fact that one machine is unknown to the configure process.

Passing by, one of my pleasures with the GNU people is the ease of
exchanges about bugs.  I do not feel that perennial reluctance to
which non-free software providers accustomed us.  I find it bizarre to
see myself in a position in which I think we are going too far :-).

					Keep happy!

-- 
Franc,ois Pinard        ``Vivement GNU!''        pinard@iro.umontreal.ca
(514) 588-4656                                    ...!uunet!iros1!pinard
Consider joining the League for Programming Freedom.  Email for details!