Correct: fluxes work by forming soluble salts of the metal oxides on the surface, thus allowing the metals to flow smoothly as they like. (This would all be so much easier in inert gas or vacuum -- indeed it is, I've done induction brazing this way -- such a beautiful sight, no flux needed at all, just start with clean parts!)
I strongly suspect the best cleaner is either a mild acid or base: base to clean up the residual rosin (which is an organic acid), or acid to break up the soaps (organic salts of metals). A solvent and detergent fraction keeps things suspended and dissolved. A complexing agent (throw in some EDTA??) might be effective, too.
My opinion of white residue, is that it's just not fully cleaned. Any solvent left on the board, has some rosin dissolved in it, which crystallizes out when it dries. The only way to prevent this is with a constant flow of pure distilled solvent: a vapor degreaser. The best you can approximate this without an explosion-waiting-to-happen, is with several washes of clean solvent, blow-drying the board inbetween (not to evaporate the solvent, but to blow it away while still liquid!)
And yeah, having the right solvent mix I'm sure accounts for a lot. 70% IPA evaporates to leave water, which will bead up and leave a speckled residue pattern. Not necessarily a reason to use 100% of something, because the polarity of water probably helps a lot too.
Ah well, fun stuff, chemistry. Unfortunately, doing more than armchair "research" (like, you know, anything actually
useful to discuss..
) is rather difficult. One would suppose -- hope? -- that industrial flux cleaners have done this research, so that the best suggestion would be simply to buy one of those, and enjoy the ease of use, at a modest cost hike over generic solvents.
FYI: toluene is not particularly harmful. Worse than acetone, say -- certainly, but it's not benzene (which you most definitely won't see at the hardware store).
Personally, I find ethanol, IPA, acetone and MEK quite suitable for the fluxes I use, without being too much trouble for the components I use. I would not recommend toluene, petroleum ether, mineral spirits or the like: nonpolar and aromatic solvents tend to dissolve many more plastics than these light polar solvents do. Even so, there are plenty of components affected, most often the ink labels on capacitors (electrolytics and film caps for example).
Tim