Comment 6 for bug 1255806

Still, could you please fix sbuild to not use options which do nothing Dmitrijs?

As Michael said, --fix-broken did, does and will not do anything for build-dep, so calling apt with it is just wrong, even if it was accepted in the past. The parsing of the commandline became stricter now as we get frequent reports from people who use this or that flag and assume that it has some effect because the apt-tools accepted them even if they are ignored.
I could nearly bet that the -f in sbuild is there because someone thought it would stand for --force and so helps in spreading the legend that -f would trigger some more reckless behavior, but it will not…

-o flags are btw always accepted. Regardless if the option in question even exists or is used. Were are wishes to change this as well (cupt e.g. does), but I think the -o flag should remain mostly unchecked to help with backward compatibility.

I btw don't agree with waiting for this to be done in a major release. There isn't any transition required. Either the failing scripts fails because they used apt wrongly, so breaking them now is good rather than breaking them at a point at which something really changes (like -f gets some sort of meaning for build-dep) or it breaks because it uses an option we forgot about. The later is a bug in APT we can fix easily and nothing special per se. Every change can include bugs after all. And as a by-product we get to know what users assume to be in working condition which we never thought about. If we never thought about it, we might break it in the future with real changes, so writing a testcase for it now might as well protect you from real fallout in the future (again).
[I e.g. would have never thought to --purge on build-dep. It might make sense for sbuild, but not for myself so I can perfectly imagine breaking it in some silly way – okay, I can't imagine how at the moment, but if its me creating bugs, everything is possible ;) ]