I'd split this into several issues:
(1) Does demonstrating breadth actively hurt your application, by creating the impression of a lack of focus? For example, might you do better on the job market with only a proper subset of your papers listed on your CV?
In general, I think the answer is no. Of course, it depends on the circumstances; for example, adding low-quality papers to your CV could actually hurt your chances, so it's not simply monotonic. However, I don't think I've seen any cases in which I thought a candidate was hurt by having good papers in several research areas.
(2) If two people have written the same number of papers of comparable quality, but one works in a single area and the other in several areas, which person will do better on the job market?
This is a trickier question. There are definitely cases in which each would have an advantage. Sometimes having many papers in one area can excite a research group in that area, and sometimes having great breadth can build broad support within a department for making a hire. Overall, you need to show enough depth in at least one area that you don't come across as a dilettante, and that you attract the interest of someone in the department, but otherwise it's not clear which approach is best. [I should say that for most beginners, it is better to focus. If you have only a few papers, then you very likely will look like a dilettante if they are all in different areas, and they would have to be awfully good to make up for that. However, it sounds like that is not an issue here.]
I wouldn't worry about this too much. I think it's largely a matter of personality: some people love to focus and get distracted if they are trying to juggle too many projects, while others feel terribly constrained if they are doing just one thing. If one approach lets you do your best work, and the other doesn't, then you should go with the one that fits you best.
(3) Is it better to write one great paper, or two very good papers in different areas?
This one really depends on who is evaluating the application, and on their judgement of the quality difference. However, the central issue here is quality vs. quantity, and breadth is a secondary concern. (In many cases, the answer won't depend on whether the two papers are in the same area.)
As for theoretical computer science, you're right that many mathematicians have little idea how to evaluate the quality of CS conferences, and this can be a problem. However, I don't think it's a big problem in practice, because people who are interested in the topic certainly know about this, and they are generally able to explain it to the rest of the department or the administration. Letters of recommendation also play a big role here: if well-known recommenders praise your work, then that carries more weight than whether the conferences stand out as impressive on a CV.