> Since all modes are supported, you can unconditionally return 1 in the > code above, as is the case with fpu-387.h.
Indeed. I’ve committed the attached patch, which performs this trivial cleanup, adds some comments about unreachable cases (I don’t think we want to error out at runtime, but a comment in the code is the least we can do). It also cleans up the value we return for get_fpu_rounding_mode() when we don’t recognize the rounding mode: it used to be GFC_FPE_INVALID, which is wrong (it’s an “invalid” exception flag, not a marker for an invalid rounding mode). Now, it will return 0, i.e. a nonexisting rounding mode flag, which is more appropriate (though it should never happen in practice). Committed as rev. 212423 after testing on x86_64-apple-darwin13. Thanks, FX
z
Description: Binary data
