© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
67d34a1
(Get the code!)
This is bizarre.
I did a fresh install of Lucid 10.04.1 onto a netbook, then plugged in my Option modem (0af0:d033).
I confirm that it caused my system to lock up.
I just spent the past 4 hours trying to bisect this issue, and I can't get it to fail reliably.
In my bisection adventure, I found a point where I could checkout and build/install a package, and no system freeze. I then checkout (build, install) the commit immediately before and lo! a system freeze! This makes me think that the package is broken up to the good commit. However, this hypothesis has two problems:
1) Yes, the "bad" commit immediately prior the "good" commit breaks as expected. The "bad" commit happens way after the commit we have in Lucid. So far, so good. Theoretically, all the commits between the "bad" one and the last Lucid one should all result in system freeze. Unfortunately, that is not the case. I picked several random commits between "bad" and "Lucid" and encountered success at each of them.
2) Simply downloading the exact modemmanager source from Lucid, rebuilding it where the only debdiff is in the changelog, and then installing it on my test system simply does not reproduce the failure. This makes no sense, since the code didn't change at all in this experiment. I made sure to reboot and purge the package before installing my test package as well. No luck.
My original intent was to try and find the one good commit to backport as an SRU into Lucid. However, with such unreliable bisection results, there is no way that I can safely do so in good conscience. I would essentially be backporting a random patch into Lucid and hoping that it fixed the problem, which is not allowed by SRU policy. Also, backporting the entire package from Maverick into Lucid is also disallowed by SRU policy.
My best recommendation at this point is to move your systems to Maverick.
If you are stuck on Lucid, please respond and we can try some more testing. Perhaps you can bisect more reliably on your machine than I can on mine which would help us to narrow down the issue.