Treat Inf/NaN specially for %d formatting

Author: quinnj

stdlib/Printf/src/Printf.jl

@@ -38,6 +38,8 @@ Base.string(f::Spec{T}; modifier::String="") where {T} =
         f.precision == 0 ? ".0" : f.precision > 0 ? ".$(f.precision)" : "", modifier, char(T))
 Base.show(io::IO, f::Spec) = print(io, string(f))
 
+nonfinitefmt(s::Spec{T}) where {T} =
+    Spec{Val{'g'}}(s.leftalign, s.plus, s.space, s.zero, s.hash, s.width, s.precision)
 ptrfmt(s::Spec{T}, x) where {T} =
     Spec{Val{'x'}}(s.leftalign, s.plus, s.space, s.zero, true, s.width, sizeof(x) == 8 ? 16 : 8)
 
@@ -264,7 +266,7 @@ end
 # integers
 toint(x) = x
 toint(x::Rational) = Integer(x)
-toint(x::AbstractFloat) = x > typemax(Int128) ?
+toint(x::AbstractFloat) = !isfinite(x) ? x : x > typemax(Int128) ?
                             BigInt(round(x)) : x > typemax(Int64) ?
                             Int128(round(x)) : Int64(round(x))
 
@@ -273,6 +275,9 @@ toint(x::AbstractFloat) = x > typemax(Int128) ?
         spec.leftalign, spec.plus, spec.space, spec.zero, spec.hash, spec.width, spec.precision
     bs = base(T)
     arg2 = toint(arg)
+    if !isfinite(arg2)
+        return fmt(buf, pos, arg, nonfinitefmt(spec))
+    end
     n = i = ndigits(arg2, base=bs, pad=1)
     x, neg = arg2 < 0 ? (-arg2, true) : (arg2, false)
     arglen = n + (neg || (plus | space)) +
@@ -682,6 +687,9 @@ end
 
 function plength(f::Spec{T}, x) where {T <: Ints}
     x2 = toint(x)
+    if !isfinite(x2)
+        return plength(nonfinitefmt(f), x)
+    end
     return max(f.width, f.precision + ndigits(x2, base=base(T), pad=1) + 5)
 end
 

stdlib/Printf/test/runtests.jl

@@ -433,6 +433,8 @@ end
     # 37552
     @test @sprintf("%d", 1.0e100) == "10000000000000000159028911097599180468360808563945281389781327557747838772170381060813469985856815104"
     @test @sprintf("%d", 3//1) == "3"
+    @test @sprintf("%d", Inf) == "Inf"
+    @test @sprintf(" %d", NaN) == " NaN"
 end
 
 @testset "integers" begin