Fix Julia 4.0 Regressions

Negating Int32 via bitfield operations expands the type to Int64
Uint32 is now used with a reinterpret cast

Pretty printing of true/false typeclasses fixed

Fixes Issue #2
pull/4/head
fundamental 2014-09-29 13:37:57 -04:00
parent 4fa945b175
commit 67c26b7742
1 changed files with 11 additions and 5 deletions

View File

@ -248,7 +248,7 @@ function rtosc_argument(msg::OscMsg, idx::Int)
t |= (uint64(msg.data[@incfp(arg_pos)]) << 8);
t |= (uint64(msg.data[@incfp(arg_pos)]));
if(typeChar == 'h')
return int64(t)
return reinterpret(Int64, t)
elseif(typeChar == 'd')
return reinterpret(Float64, t);
else
@ -257,7 +257,7 @@ function rtosc_argument(msg::OscMsg, idx::Int)
elseif(typeChar in "f")
return reinterpret(Float32,msg.data[arg_pos+(3:-1:0)])[1]
elseif(typeChar in "rci")
i::Int32 = 0
i::Uint32 = 0
i |= (uint32(msg.data[@incfp(arg_pos)]) << 24);
i |= (uint32(msg.data[@incfp(arg_pos)]) << 16);
i |= (uint32(msg.data[@incfp(arg_pos)]) << 8);
@ -267,7 +267,7 @@ function rtosc_argument(msg::OscMsg, idx::Int)
elseif(typeChar == 'c')
return char(i)
else
return i
return reinterpret(Int32, i)
end
elseif(typeChar in "m")
m = Array(Uint8, 4)
@ -277,7 +277,7 @@ function rtosc_argument(msg::OscMsg, idx::Int)
m[4] = msg.data[@incfp(arg_pos)]
return m
elseif(typeChar in "b")
len::Int32 = 0
len::Uint32 = 0
len |= (msg.data[@incfp(arg_pos)] << 24);
len |= (msg.data[@incfp(arg_pos)] << 16);
len |= (msg.data[@incfp(arg_pos)] << 8);
@ -312,8 +312,14 @@ function showField(io::IO, msg::OscMsg, arg_id)
if(issubtype(typeof(value), Array))
value = value'
end
fieldClass = dict[typeChar]
if(fieldClass == true)
fieldClass = 'T'
elseif(fieldClass == false)
fieldClass = 'F'
end
@printf(io, " #%2d %c:", arg_id, typeChar);
println(dict[typeChar]," - ", value)
println(fieldClass," - '", value,"'")
end