What to do with a paper with references but no citations?

Reject.

No work exists in isolation. The authors need to position their work in relation to existing work, and this requires more than just putting a bunch of papers in the bibliography section: it requires a detailed comparison. The authors failed to do this.

In any case, there will always be another conference for them to resubmit to.

My initial reaction is that they merely have encountered a set of issues for which there are no relevant papers they could cite. For example, where would you expect citations in there? Citations for the sake of citations are a Bad Thing. Citations should advance the position of the paper in a meaningful, shoulders-of-giants kind of way.

And initial reactions are often the right one^{[citation needed]}.

But upon further reflection, it occurs to me that with the vast multitude of papers out there, the likelihood of not having a single paper that could help advance this paper's position in a meaningful way is probably very, very small. What is much more likely is that the authors either did not do their due diligence in looking for work that could have advanced their position and/or saved them time, or worse, specifically excluded other research because glaring similarities. Both are common, and either is bad for the authors and the scientific community as a whole.

The main thing that concerns me is... where did they get the idea to do whatever is in the paper? Was it not at least partially based on some published work that they're either challenging or advancing? That is the most troubling thing to me.

It's not clear from your question, whether they simply don't regard other papers in the field, or they do have related work etc. but they omit the in-body-citations.

If it is the first case, I'd second Dave's answer (unless it's a brand new question with a brand new technique that solves it, and they clearly say that no related work can be found to the best of their knowledge).

However, if it is the second case, it seems quite technical issue that can be easily fixed, and in this case you can just mention in your review that references are missing and this should be fixed (also note this to the PC chair; s/he can condition the acceptance on fixing this issue)

The fact that it's a conference, in my eyes, makes it more flexible -- papers should be considered mainly by merit and not by technicalities. I can think of the opposite case, where the paper has all the citation, gets accepted, but in the camera-ready version all references are removed.