Re: /usr/src/Makefile.inc1 typo?

On Mon, Dec 12, 2022, 2:21 PM Mark Millard <marklmi@yahoo.com> wrote:

> On Dec 12, 2022, at 12:39, bob prohaska <fbsd@www.zefox.net> wrote:
>
> > It looks like there might be a typo (or nomenclature change)
> > in armv7.
> >
> > Last attempt at installkernel stopped with
> > make[1]: "/usr/src/Makefile.inc1" line 163: Unknown target armv7:armv7
> > . . .
>
> You suffer from bad timing relative to main's [so: 14's]:
>
> Sun, 11 Dec 2022
>    . . .
>     • git: 83bf6ab56829 - main - uname: switch machine to HW_MACHINE_ARCH
> Piotr Kubaj
> . . .
> Mon, 12 Dec 2022
>    . . .
>     • git: 85dd85323614 - main - Revert "uname: switch machine to
> HW_MACHINE_ARCH" Piotr Kubaj
>     • Re: git: 83bf6ab56829 - main - uname: switch machine to
> HW_MACHINE_ARCH Piotr Kubaj
> . . .
>
> Having a system built from a source tree from between the 2
> commits one ends up with a broken default native build
> environment with complaints about "target NAME:NAME" when the
> correct naming for the context has 2 distinct names, such as
> arm and armv7 in your context. (That need not be the only
> complaint.)
>
> It may be that explicitly listing TARGET=arm TARGET_ARCH=armv7
> on either the make command line(s) or in the environment(s)
> might allow a valid build from the messed up context. (But
> such is untested.) I've no clue if a full rebuild-from-scratch
> might be needed. If it works, you would likely need to be
> explicit, even for installkernel and installworld, not just
> for building.
>

Libc is all that's needed. Use "make MACHINE=arm" though maybe it's a
general workaround.

Warner

The build(s) should be based on source that does not contain
> the 83bf6ab56829 change.
>
> ===
> Mark Millard
> marklmi at yahoo.com
>
>
>