My bad, i was a bit confused by the %Z choice and felt odd about it
and this explanation makes sense.

I've attached an updated patch and sha1.c that doesn't revert the
fixes in %Z discrepancies.

As for the dd, I'm not sure about the best way to test this, I've done:

tenshi% dd -quiet 1 -count 2 </dev/zero >> /tmp/venticonf.$pid || exit dd
tenshi% cat /tmp/venticonf.819083
tenshi% ll /tmp/venticonf.819083
--rw-rw-rw- M 8668 seh seh 1024 Sep 18 17:48 venticonf.819083

Which looks right, I think?

Cheers,
Sean
On Mon, Sep 17, 2018 at 4:49 PM <cinap_lenrek@felloff.net> wrote:
>
> this undoes a bunch of bugfixes.
>
> - %z format was changed to %Z because %z is used for pointer diff (our 6c yiels vlong there)
>
> -       dd -quiet 1 -bs 1024 -count 8 </dev/zero >> /tmp/venticonf.$pid || exit dd
> +       dd -quiet 1 -count 2 </dev/zero >> /tmp/venticonf.$pid || exit dd
>
> this is suspect, there was a bug there with devfs which doesnt allow unaligned
> writes. please check if thats not broken.
>
> +       sha1.$O\
>
> missing the file?
>
> --
> cinap