Subject: Re: checkflist error codes
To: None <>
From: Derek God3 <>
List: current-users
Date: 11/05/2003 15:03:59 wrote:
> Thus spake Alan Barrett ("AB> ") sometime Today...
> AB> From: Alan Barrett <>
> AB> To: NetBSD current list <>
> AB> Subject: Re: checkflist error codes
> AB> Date: Wed, 5 Nov 2003 18:54:06 +0200
> AB> X-Spam-Level:
> AB>
> AB> On Wed, 05 Nov 2003, William Allen Simpson wrote:
> AB> > The most annoying problem with checkflist is that it aborts the
> AB> > process when there are extra files in the destdir.  Is there any reason
> AB> > why?
> AB>
> AB> I imagine it's so that developers will (eventually) notice the
> AB> problem and fix it.
> AB>
> AB> As a user, I just modify my copy of checkflist to treat certain types of
> AB> inconsistencies as non-fatal.
> [Not specifically to Alan!]
> This still does not address the fact that the flist is static and not
> conditional as to what you build for a distribution.  That needs
> desperately to be addressed.  I'm unfortunately up to my ass in alligators
> at the moment, so I have not the time but I have the motivation,
> believe me.
> Example:  I want to build a distribution with no /lib, /libexec or
> /rescue, and a statically-linked root. Or I might want to build a pure
> static distribution.  Or I might want to build a distribution that uses
> compressed manual pages.  Or I might want to build a distribution which
> will install nicely onto an embedded installation.
> But checkflist bites my knee when I do that.  I find myself adding '-V
> MAKEFLAGS=-i' to the build process, something which, as even a
> lightweight-by-comparison developer/programmer, I *really* don't like
> doing.
>                 --*greywolf;
> --
> NetBSD: Feed The Computer.

For my embeded NetBSD, I extended the "obsolete"
machanism to allow me to subract from both the
sets and flist, however, this is file by file and
is a pain to maintain - would be nice to do by "package"