Subject: Re: patch for share/man/man4/Makefile to help small machines
To: None <current@NetBSD.ORG>
From: Julian H Stacey <Julian.H.Stacey@regent.e-technik.tu-muenchen.de>
List: current
Date: 10/09/1994 16:32:19
> From: "Charles M. Hannum" <mycroft@gnu.ai.mit.edu>
> It is not desirable to omit the other architectures from the standard
> share/man build, primarily for the obvious reason that `share' is
> supposed to be sharable between machines of possibly different
> architecture.

Agreed.

But neither is it nice when the make goes recursive if it can't cd to a non
existant directory, & I simply don't have the space for other architectures
on that box (PC532).  (I dont have a net on the PC532, I figure _loads_ of
other people don't want man's for other arch's forced on them either).

I _do_ maintain a full source tree in the spirit of completness, 
to give to friends, for dial in access on my large i486 box, but on my PC532
I _must_ strip the slack.

What would be ideal would be if someone could fix the NetBSD mk macros,
so if a SUBDIR element doesn't exist, the make doesnt go recursive
atemmpting the CD, (FreeBSD doesn't suffer this, perhaps NetBSD has been fixed
recently, but it certainly used to suffer this).

If mk macros were to be fixed, that'd be the ideal solution:
	- If I then strip the slack, & make -i,  it'll work,
	- & your big systems will be happy too, compiling manuals to
	  export shared to nets.

I spend most time working on FreeBSD & PC532 specific NetBSD things,
so don't personally want to fix the fragile NetBSD share/mk/*	;
anyway, I figure the fixes will be more popular & acceptable
if they come from someone like you :-)

I'd be happier to not need my patch man/man4/Makefile, if only someone would
fix share/mk/* SUBDIR handling so `su ; make -i' didnt recurse infinitely
on absent directories 	:-)

Best Regards

Julian Stacey <stacey@freefall.cdrom.com>