On Mon, 7 Jul 2025, Martin-Éric Racine wrote:
>I'm not sure whether changing this behavior is remotely desirable.
>However, I do agree that if this makes ifupdown presume that an
>external IP has succesfully been acquired and the boot process continue
>to steps that launch network-dependant services, it is problematic.
I’m just interested in having the ability to remove this regression
against isc-dhcp-client. If there are use(r)s of dhcpcd-base for
which this backgrounding and dynamic switching is desired, then by
all means let them use it, but provide options.
I had to look up what IPv4LL/IP4LL even is (RFC 3927), and I can
say with confidence that, when I configure an interface to use DHCP,
I *never* want IPv4LL.
And as things are, both the backgounding and the addition of IPv4LL
are regressions that occur when using just the line…
iface eth0 inet dhcp
… relative to bookworm, where this worked as desired.
bye,
//mirabilos
--
15:41⎜<Lo-lan-do:#fusionforge> Somebody write a testsuite for helloworld :-)
