Re: SCO/Xinos OpenServer 7 pkgsr port after bootstrap bmake: no system rules (sys.mk).

[Roland Illig <roland.illig%gmx.de@localhost>]

Am 28.11.2019 um 19:40 schrieb Boyd Lynn Gerber:
> bmake -dA 2>&1 | grep 'sys\.mk'
> Expanding "sys.mk"... /usr/pkg/share/mk/sys.mk
> Global:MAKEFILE = /usr/pkg/share/mk/sys.mk
> Global:.MAKE.MAKEFILES = /usr/pkg/share/mk/sys.mk
> Parse_SetInput: file /usr/pkg/share/mk/sys.mk, line 0, fd -1, nextbuf
> 805b4c0, arg 807495c
> Global:.PARSEFILE = sys.mk
> MAKEFILE         = /usr/pkg/share/mk/sys.mk
> .MAKE.MAKEFILES  = /usr/pkg/share/mk/sys.mk
> MAKEFILE         = /usr/pkg/share/mk/sys.mk
> .MAKE.MAKEFILES  = /usr/pkg/share/mk/sys.mk

From this log I conclude that sys.mk can be found now. Otherwise it
should look like this:

$ bmake -dA 2>&1 | grep 'sys\.mk'
Expanding "sys.mk"...
bmake: no system rules (sys.mk).

Therefore I wonder whether your very first error message about the "no
system rules" is correct, or at least consistent to the behavior from
your latest mail. In my mind, they just don't fit together.

I also don't think that running "bmake" searches in a different
directory than running "bmake -dA".

Does SCO have some tool like strace, ktrace or ktruss, so that you could
record the syscalls that the failing bmake makes? That should reveal
where it fails to find sys.mk.

I'm clueless right now. Can you perhaps search for yourself, now that
you know how to get the debug output from bmake?

In your first mail you wrote:

> I then try to do a bmake as a user and as root

Could it be that you installed installed pkgsrc as root and later ran
bmake as user, and that user doesn't have permissions to access
/usr/pkg? It's just a wild guess.

Roland