Re: bin/57820: devel/rcs fails on 10.0_RC2 works on 9.3

[RVP <rvp%SDF.ORG@localhost>]

On Tue, 16 Jan 2024, George Georgalis wrote:

> If my goal is minimal OS, how might I determine the sets required
> for a given task? Do the sets have definitions for what they contain
> or should not contain? It is natural that pkgsrc is just checking if
> the OS is NetBSD if there is no definition of what sets provide.
I think going with the "Installation without X11" choice should work out
better than going with the "Minimal installation" selection.

> Not sure about the nb10rc2 base make on nfs observation,
> but I'll start a new thread if that returns as I resume using
> nfs.

Yeah, that NFS issue puzzles me too (server not supporting nanosec
mod times? uid/gid mixup? perm. issues?...) but, for shared NFS
pkgsrc build, make sure you do:

a) sync the time on all your machines (as Martin said)
b) set `WRKOBJDIR=' and `PACKAGES=' to some local dir.
c) one NetBSD, either use make(1) with the std. paths, or bmake(1) with
   if you want a different prefix. Don't mix these (the gurus have
   spoken!).

> I suspected it was a date GNUism and had no clue why it would come up.
> Obviously, there was an unchecked error in the rcs build, no rcsfreeze.1
> man page...

Due to a lack of time, I didn't go find out _why_ that rule is
being run when the pkgsrc Makefile has `post-configure' step to
stop just that:

```
# configure will be modified by pkgsrc framework with timestamp update.
# sync timestamp with man/REL to avoid regeneration of man.
post-configure:
	${TOUCH} -r ${WRKSRC}/man/REL ${WRKSRC}/configure
```

So, the whole man-page generation dance shouldn't happen at all
_even if_ soelim wasn't there and pkgsrc had to build it--and
whatever a groff build depends on--on a minimal-package system (which
seems to include: Perl, Python, Rust...I stopped this recursive
dep. build on Ubuntu right there :)).

-RVP