Actually it is still a bug. If the driver requires that libdrm be at a certain level then it should check at load time for such a dependency and report back if it detects if you are running an old version. This usually gives the developer/user a clue that an old lib is installed in the LDPATH.
Given that all packages in this case are simply snapshots of a development trunk I don't think it's a real issue. Plus on a production system you have a package manager to take care of that.
Your accusations may be right, but still give some credit to him since he listened to the suggestions from the kernel regression related comments and I think he really tried his best this time.
Even if this time it was not a real/big bug I'm sure it will be able to find some nasty ones in the future.
Certainly, this is a step in the good direction.