I need a bit of clarification about the definition of transcript of a simulator in Zero-Knowledge Proofs.
Basically, the way I understood it, the simulator is given access to the (potentially malicious) Verifier as a sub-routine, and the goal of the simulator is produce a transcript that's computationally indistinguishable from the transcript produced between the Prover and the (potentially malicious) Verifier.
My question is: Does the transcript of the simulator include the final Accept/Reject output of the verifier or not? In other words, if the verifier accepts in an interaction with the real Prover, should the simulator also be able to produce a transcript, that will be accepted by the verifier, or can the simulator just leave the verifier's output out.
If it leaves the verifier output, then the definition doesn't seem to make sense to me, because I can always generate a random transcript that looks like the interaction between prover and verifier. However, in most lectures, I don't see people requiring the result of verifier in the transcript, which is what's confusing me.