Type-Ia supernovae (SNe Ia) are thermonuclear bombs in which about one solar mass of carbon and oxygen are burned into iron-peak elements. The fuel is apparently a white dwarf stellar remnant. SNe Ia became popular about 15 years ago, when it became clear that they can serve as excellent cosmological distance indicators. In 1998, they provided the first evidence that the cosmic expansion is accelerating, apparently under the influence of an enigmatic "dark energy".   However, despite their confident use for cosmology, a major embarrassment remains: no one knows, based on direct evidence, what exactly is exploding, Two scenarios have been on the table for a long time for explaining how a white dwarf can ignite and explode as a SN Ia. In the "single-degenerate" picture, a white dwarf  accretes matter from a companion  "normal" star  (i.e. a star with a classical equation of state) , until approaching the Chandrasekhar limit and igniting. In the  "double-degenerate" picture, a close white-dwarf binary loses energy and angular momentum to gravitational waves, until the two white dwarfs merge, thus starting the ignition and the thermonuclear runaway. However, both scenarios have theoretical and observational problems, and little or no direct evidence. Measurement of SN Ia rates, as a function of  cosmic time and environment, can shed light on this problem. I will show how, recently, many different measurements are converging toward a single SN Ia "delay-time distribution". This is is the SN Ia rate, as a function of time, that would follow a hypothetical short burst of star formation, i.e., it is the Green's function of SNe Ia. The emerging function is remarkably similar to what one expects from white dwarf mergers, based directly on the fundamentals of gravitational wave emission.