apt cron.daily script doesn't check value-existence

I have the same problem. I have applied your patch manually. Why do you have inserted the ";" at the end of the first two new lines?

I'm Perl afflicted. That's probably why i tend to type ; after each line. It doesn't hurt though, in the shellscript.

also happens on one of my systems. The applied patch fixes the problem.

Same here. I gave up after numerous attempts to set a (empty) value for these keys but to no avail. Redirecting to stderr fixes.

+1

I use this workaround myself.

It is standard behaviour on Ubuntu Server/JeOS installs so it needs to be fixed IMHO

Please don't assign bugs to ubuntu-core-dev. It's sufficient to just leave it unassigned.

This bug was fixed in the package apt - 0.7.20.2ubuntu5

---------------
apt (0.7.20.2ubuntu5) jaunty; urgency=low

  [ Colin Watson ]
  * cmdline/acqprogress.cc:
    - Call pkgAcquireStatus::Pulse even if quiet, so that we still get
      dlstatus messages on the status-fd (LP: #290234).

  [ Michael Vogt ]
  * debian/apt.cron.daily:
    - do not clutter cron mail with bogus gconftool messages
      (LP: #223502)
    - merge fix for cache locking from debian (closes: #459344)
    - run update-apt-xapian-index (with ionice) to ensure that
      the index is up-to-date when synaptic is run (LP: #288797)

 -- Michael Vogt <email address hidden> Mon, 30 Mar 2009 13:22:28 +0200

As of April 24, 2009 this bug is back in Hardy, most likely a regression issue.

/etc/cron.daily/apt:
No value set for `/system/http_proxy/use_http_proxy'
No value set for `/system/http_proxy/host'
No value set for `/system/http_proxy/port'

The bug status makes this look like it's fixed, but it is still broken in Intrepid (and I can't figure out how to indicate that). As Intrepid is still a supported release, this seems like relevant information.

It seems the values for these variables are user based.

Then if you run the command "gconftool --get /system/http_proxy/use_http_proxy" with a user that is not the proper one, you get that error.

In /etc/cron.daily/apt you have:
admin_user=$(getent group admin|cut -d: -f4|cut -d, -f1)
and:
use=$(sudo -u "$admin_user" gconftool --get /system/http_proxy/use_http_proxy)

This mean the script uses the first user in the group admin, that in this case is my login user.

I added the root user in the group admin, look like it's fixed now. Had to check also if variables were set for user root though, they usually are.
Don't know why a cron job (running as root), would sudo quering the conf table of the the first admin user for the info on the proxy. Looks like a workaround or something unclean, even more for a server install.

This procedure is used also in other cron scripts.

Confirmed this is still present on Hardy.

Confirmed this is still present on Hardy.

This is present on Lucid (server)