Assuming that one has begun a PhD program at a university which allows a student to choose an advisor after his/her first year in the program and the student has narrowed down on a couple of potential advisors, once a student has established contact with his potential advisor,
At what point do you discuss things like authorship, expectations and other such details? First meet, second meet or otherwise?
I assume the main issue for you is the question of authorship.
In academia, there are rules about authorship, and they differ between the disciplines slightly. You can expect that supervisor will request to be last author of every paper. I very much advise against trying to bargain around that, this is the only scientifically relevant reward your supervisor will get for supervising you. And since professors talk to each other, you may unknowingly ruin your reputation by trying to negotiate on this.
There are details about authorship of other people who might give you research ideas (like other grad students in the lab, outside collaborators) but I don't believe the general answer can be provided to you by the professor in a way that would protect you in the future if he will be trying to coax you into accepting another author that you don't see as having significant contribution to the paper. Usually can people agree on everything in general, but then it is always down to the question, what amount of work does qualify someone to demand co-authorship.
For example, a person who does some simple physical/electrical measurements is certainly not justified to be a co-author, but if someone performs experiment for you that requires trial of toxicity of the chemical substance including 50 live mice, and you depend on his knowledge and experiences to actually design the experiment properly, then such person will probably be expected to be co-author, especially if you only took the results and ran them through SPSS.