[bitc-dev] Implementing closure conversion

[Jonathan S. Shapiro]

>I don't think it's unreasonable to implement (2) for IA-32, and (1) for
all other architectures...

Ironically, the unused arg probably won't hurt IA32 performance at all. In any case, we have worked very hard to ensure that the emitted C code will compile on any platform, so I would rather have one scheme, even at the expense of some performance.

Production compiler, of course, is another matter.

shap