The libgo implementation of math.Ldexp declared the libc "ldexp" as taking an 'int' exponent argument, which is not quite right for 64-bit platforms (exp arg is always int32); this could yield incorrect results for exponent values outside the range of Minint32/Maxint32. This patch by Than McIntosh fixes this by upating the type for the libc version of ldexp, and adding guards to screen for out-of-range exponents. This fixes https://golang.org/issue/21323. Bootstrapped and ran Go tests on x86_64-pc-linux-gnu. Committed to mainline.
Ian
Index: gcc/go/gofrontend/MERGE =================================================================== --- gcc/go/gofrontend/MERGE (revision 250873) +++ gcc/go/gofrontend/MERGE (working copy) @@ -1,4 +1,4 @@ -db685a1a9aa8b3b916dd6d1284895e01d73158e1 +5fd112e5c2968e94761c41519c451d789e23a92b The first line of this file holds the git revision number of the last merge done from the gofrontend repository. Index: libgo/go/math/ldexp.go =================================================================== --- libgo/go/math/ldexp.go (revision 250873) +++ libgo/go/math/ldexp.go (working copy) @@ -13,10 +13,15 @@ package math // Ldexp(NaN, exp) = NaN //extern ldexp -func libc_ldexp(float64, int) float64 +func libc_ldexp(float64, int32) float64 func Ldexp(frac float64, exp int) float64 { - r := libc_ldexp(frac, exp) + if exp > MaxInt32 { + exp = MaxInt32 + } else if exp < MinInt32 { + exp = MinInt32 + } + r := libc_ldexp(frac, int32(exp)) return r }