In computer science, what should be the typical author list if it's postdoc's idea and professor's student?

In general, author lists in CS are usually in order of contribution. For practical reasons this often means that you end up with student/postdoc/professor, simply because that's also the order of effort that the various parties can typically put into the project. This would also be my "default" if there are no special circumstances that warrant a different author list.

Some possible special circumstances:

• You, the postdoc, don't only have the idea but also do most of the actual work in the project. In that case it would be warranted to put your name first. However, note that for a PhD student being first author is often more useful than for you, so take this into account when distributing tasks for the project. That doesn't mean that the postdoc should gift the PhD student with first authorship if not warranted, but often it's better for everybody if the project is laid out in a way that the PhD student gets enough tasks so that their first authorship is actually warranted.
• The professor is not interested in the project and does not want to get involved. In this case he may not even be part of the project at all and it becomes student/postdoc. However, if you want to go for that, make very, very sure that the professor is also ok with not being on the paper. There are many CS labs where the professor assumes to be on every paper that their students write. If you can't live with taking such a professor on, don't work with their students. If you decide to kick off the professor (even if warranted due to contributions), you may not only get into a fight yourself, but you may also make life much harder for the student.

Note that there is in my experience rarely a case where the ordering student/professor/postdoc would be warranted. Actual contributions rarely seem to work out like that, and it's also much less useful for the postdoc than some may think - the postdoc does not become the senior author in the collaboration just because they have been put last in the list.

In my part of computer science (theory), the student and postdoc should be listed in alphabetical order, and the professor should not be listed.