Hi, I don't want to start a discussion. So no need to reply. I just wanted to clarify some things.
Quoting Sam Hartman (2021-02-24 23:12:11)
> I'm not at all convinced this is a good idea. We're replacing a great,
> well-tested facility--namely running maintainer scripts in root directories
> with a mechanism that requires support to be spread through each package.
This is not meant to replace running maintainer scripts in root directories
ever. This is an additional feature and is probably only going to be
implemented for the packages in the pseudo essential set. Since maintainer
scripts will retain their old behaviour if the DPKG_ROOT variable is empty,
this should not add bugs for normal operation where maintainer scripts are run
is the root directory. To reduce the perceived impact on "normal" maintainer
script usage even more, the patch could be amended to say:
if [ "$DPKG_ROOT" = "" ]; then
pam-auth-update --package $force
else
pam-auth-update --root "$DPKG_ROOT" --package $force
fi
That way it is clear that the "non standard" (and less tested) codepath is only
taken if dpkg was called with --root set.
> It's certainly too late for bullseye to consider something like this.
Yes, absolutely. Sorry for not having been clear on that.
> I'd be happy to sit down with one of the proponents over a voice call in May
> or June to understand the current state of this effort, what level of
> consensus has been achieved and consider next steps. I think that initial
> discussion would be too high bandwidth for email.
Thank you! :)
cheers, josch
signature.asc
Description: signature
