Ubuntu MID preseed uses experimental preseed value

[Emmet Hikory]

Binary package hint: debian-cd

/data/jaunty/preseed/ubuntu-mid/mid.seed uses the experimental preseed entry user-setup/allow-password-empty. As this is not going upstream, it is better to achieve the same result in an alternate way, and remove the associated diff from user-setup.

[Emmet Hikory (persia)]

Note that the debian-cd fix must be applied prior to the user-setup fix.

[Colin Watson (cjwatson)]

I'm not sure I agree that this should be removed. allow-password-empty is a useful facility for preseeding and e.g. much nicer than the fixed password hash that casper has to set. I wasn't convinced by the objections raised in Debian, particularly since this is preseeding-only, and I have no problem continuing to carry this change in Ubuntu.

[Emmet Hikory (persia)]

I'm not sure which is correct: the current implementation of user-setup/allow-password-empty is very clean, and possibly useful. I don't think it's the right long-term solution for MID, and I agree with the point of view that it's probably unsafe to really have no password on an installed system, and better to use an automatic-login facility (as even a well-known default password is less likely to get hit by an automated scan than a blank password).

That said, it seems to me that it would make sense to be consistent across our tools, and if we did want to preserve user-setup/allow-password-empty, use it also in other places where we wish to have this sort of password (e.g. casper), although I believe it to be late enough in the Jaunty cycle this is unwarranted.

[Mathieu Trudel-Lapierre (cyphermox)]

We're not shipping MID anymore, so I'll close these tasks as Invalid.