Subject: Re: Kernel include files proposal
To: der Mouse <mouse@Rodents.Montreal.QC.CA>
From: gabriel rosenkoetter <gr@eclipsed.net>
List: tech-kern
Date: 04/11/2001 10:03:58
On Wed, Apr 11, 2001 at 09:51:07AM -0400, der Mouse wrote:
> I find it astonishing that a NetBSD person could (apparently seriously)
> advance an argument that amounts to "we don't want you to build your
> own software, trust us, if you think you want something our software
> doesn't already do, you're wrong".  I'd expect that from a
> commercialware vendor.  Not from a free-source project.
> 
> I find it even more astonishing that this was sent some six hours ago
> and nobody else has said anything about it.  (I realize those six hours
> were mostly nighttime in North America, but still.)
> 
> Is this *really* the attitude you want to present?

I think you're reading a bit too much into this.

The point, as I grasp it, is that we've got a behemoth number of
includes that no userland software needs to or, in fact, DOES
reference, which makes their untar'ing time and disk usage (whether
in bytes or inodes) a waste for users of NetBSD.

The "we" that would be the "only" ones who need to know about stuff
in sys/dev is anyone interested in working NetBSD source... at which
point "we" have downloaded NetBSD source and have (looky there!)
another copy of all of these include files. And one that might be
out of sync with those installed in /usr/include/sys/dev, to boot!

While 2.5 MBs doesn't seem like much (though the argument that
that's less than 1% of most people's drives is ridiculous
considering the drives standard in many of our ports), it really is
useless waste. At least, in my humble opinion.

       ~ g r @ eclipsed.net