[Dmitry Borodaenko] > This upgrade has cost me Friday evening and better part of Saturday, > and both were very dear to me. Please, pretty-pretty please: next > time you make changes like that to a package as important as > sysvinit, make absolutely sure it is tested.
I did test it, but not with cryptsetup. > If you can't test it yourself, find someone who can (e.g. get in > touch with me if you want to try it on a system that has root > partition encrypted with cryptsetup), or at least release it to > experimental first. If you want to get help out with new versions, please join us on the pkg-sysvinit-devel@lists.alioth.debian.org mailing list. I asked for testers and reviewers there, and the new features are discussed there first. > And, as is quite obvious from the patch, the damage is caused by a > really stupid mistake that could have been caught by a casual code > review. The patch is not correct. The intent of the change was to create a device named 'rootdev' in the freshly mounted tmpfs in /tmp/, and use it for fsck. Your patch do not do this. > Unstable is supposed to only hold packages intended for stable, last > two releases of sysvinit definitely belong to experimental. I agree. I should have sent it to experimental. I am sorry for all the problems I have caused. But I didn't, and now I have to handle the consequences. :( What is the problem you are seing? Your patch is undoing the change I did to avoid poking in /dev/shm/, so we need to come up with a different fix for your problem. Friendly, -- Petter Reinholdtsen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]