The problem with simple listen and repeat approach is that you end up going around in circles, switching quickly between different mental processes of working out what to play, deciding how to play it and committing to memory. But that constant going over causes confusion. You start to remember things that should be discarded, and forgetting things that should be remembered. It leaves a footprint of errors behind and the memory imprint is weak because the source was flaky. This is, IMHO and IME, a really ineffective way of learning.
The end goal is to make my own score for what I intend to play and learn from that. I try to separate the transcription from the learning/memorisation. I won't bore everyone with the details, but the general idea is to delay learning/memorisation until I have written out in full what is to be learned. Generally this involves creating my own detailed score which I confirm as good (error-free) before using it as the single learning reference.
I'm not saying it is a perfect system, but it is the one that has given me the most success in dramatically reduced learning time, deeper understanding of the music and order-of-magnitude improvement in long-term retention and recall.
YMMV
And my cheat sheet is riddled with typos ta boot
