[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#980217: ipp-usb: should this be Multi-Arch: foreign?

Hi Didier,

I don't know how to correctly explain it in terms of Debian packaging rules, but conceptually, yes, ipp-usb:i386 should be satisfied by ipp-usb:amd64 and visa versa.

I'll Cc: this conversation to Zdenek, because the same idea may be applied to Fedora packaging too.

And probably, the same is true for the ARM32/ARM64 combination.

On 2/19/21 8:39 PM, Didier 'OdyX' Raboud wrote:
Hello Alexander,

As ipp-usb author; do you have an opinion/advice about this?

Many thanks in advance, cheers,


Le samedi, 16 janvier 2021, 12.19:15 h CET Simon McVittie a écrit :
Package: ipp-usb
Version: 0.9.16-1
Severity: minor

A recent update to wine32:i386 on my amd64 system pulled in libsane1:i386,
which Recommends ipp-usb. Because ipp-usb is Multi-Arch: no, apt
interprets this as wanting to install ipp-usb:i386, which conflicts
with ipp-usb:amd64.

ipp-usb seems to be a daemon (systemd service) that is contacted via
IPC, conceptually similar to avahi-daemon, dbus-daemon or cups - which
hopefully means that it doesn't matter which architecture's ipp-usb
you have, because libsane1:amd64, libsane1:i386 and libsane1:mipsel
all speak the same architecture-independent IPC protocol to communicate
with ipp-usb?

If that's the case, then ipp-usb should be marked Multi-Arch: foreign,
which means that libsane1:i386 Recommends: ipp-usb would be satisfied
by ipp-usb:amd64.


	Wishes, Alexander Pevzner (pzz@apevzner.com)

Reply to: