First of all the Boolean type is said to have a default marshal type of a four-byte value. So the following code works:
struct A
{
public
The line commented with 'FAILS, WOW, WTF?' fails because of the way boolean comparison is performed. It is comparing 2 to 1:
IL_007e: ldc.i4.1
IL_007f: ldloca.s 3
IL_0081: ldflda valuetype Test/A Test/Broken::a
IL_0086: ldfld bool Test/A::bValue1
IL_008b: ceq
ceq ends up comparing 1 to the byte in bValue, which is 2.
The funny thing is that if (broken.a.bValue1) will test 'true' because it's non-zero.
As far as the other problem (broken.a.iValue2 == 4), it went away when I applied:
[MarshalAs (UnmanagedType.Bool)]
to both boolean fields in the structures. This makes sure the booleans are marshaled as an integer (4 bytes in .NET).
It would appear earlNameless is correct, as adding another structure of ints:
struct C
{
public int iValue1;
public int iValue2;
}
to the end of the union seems to correct at least part of the problem. However, this is still flawed since the boolean will only consider a single-byte value and as demonstrated is not dependable. Finally the best answer I've come up with is to use a custom type for the marshaling.
[Serializable]
[ComVisible(true)]
public struct BOOL : IComparable, IConvertible, IComparable<BOOL>, IEquatable<BOOL>, IComparable<bool>, IEquatable<bool>
{
private uint _data;
public BOOL(bool value) { _data = value ? 1u : 0u; }
public BOOL(int value) { _data = unchecked((uint)value); }
public BOOL(uint value) { _data = value; }
private bool Value { get { return _data != 0; } }
private IConvertible Convertible { get { return _data != 0; } }
#region IComparable Members
public int CompareTo(object obj) { return Value.CompareTo(obj); }
#endregion
#region IConvertible Members
public TypeCode GetTypeCode() { return Value.GetTypeCode(); }
public string ToString(IFormatProvider provider) { return Value.ToString(provider); }
bool IConvertible.ToBoolean(IFormatProvider provider) { return Convertible.ToBoolean(provider); }
byte IConvertible.ToByte(IFormatProvider provider) { return Convertible.ToByte(provider); }
char IConvertible.ToChar(IFormatProvider provider) { return Convertible.ToChar(provider); }
DateTime IConvertible.ToDateTime(IFormatProvider provider) { return Convertible.ToDateTime(provider); }
decimal IConvertible.ToDecimal(IFormatProvider provider) { return Convertible.ToDecimal(provider); }
double IConvertible.ToDouble(IFormatProvider provider) { return Convertible.ToDouble(provider); }
short IConvertible.ToInt16(IFormatProvider provider) { return Convertible.ToInt16(provider); }
int IConvertible.ToInt32(IFormatProvider provider) { return Convertible.ToInt32(provider); }
long IConvertible.ToInt64(IFormatProvider provider) { return Convertible.ToInt64(provider); }
sbyte IConvertible.ToSByte(IFormatProvider provider) { return Convertible.ToSByte(provider); }
float IConvertible.ToSingle(IFormatProvider provider) { return Convertible.ToSingle(provider); }
ushort IConvertible.ToUInt16(IFormatProvider provider) { return Convertible.ToUInt16(provider); }
uint IConvertible.ToUInt32(IFormatProvider provider) { return Convertible.ToUInt32(provider); }
ulong IConvertible.ToUInt64(IFormatProvider provider) { return Convertible.ToUInt64(provider); }
object IConvertible.ToType(Type conversionType, IFormatProvider provider) { return Convertible.ToType(conversionType, provider); }
#endregion
#region IComparable<bool> Members
public int CompareTo(BOOL other) { return Value.CompareTo(other.Value); }
public int CompareTo(bool other) { return Value.CompareTo(other); }
#endregion
#region IEquatable<bool> Members
public bool Equals(BOOL other) { return Value.Equals(other.Value); }
public bool Equals(bool other) { return Value.Equals(other); }
#endregion
#region Object Override
public override string ToString() { return Value.ToString(); }
public override int GetHashCode() { return Value.GetHashCode(); }
public override bool Equals(object obj) { return Value.Equals(obj); }
#endregion
#region implicit/explicit cast operators
public static implicit operator bool(BOOL value) { return value.Value; }
public static implicit operator BOOL(bool value) { return new BOOL(value); }
public static explicit operator int(BOOL value) { return unchecked((int)value._data); }
public static explicit operator BOOL(int value) { return new BOOL(value); }
public static explicit operator uint(BOOL value) { return value._data; }
public static explicit operator BOOL(uint value) { return new BOOL(value); }
#endregion
#region +, -, !, ~, ++, --, true, false unary operators overloaded.
public static BOOL operator !(BOOL b) { return new BOOL(!b.Value); }
public static bool operator true(BOOL b) { return b.Value; }
public static bool operator false(BOOL b) { return !b.Value; }
#endregion
#region +, -, *, /, %, &, |, ^, <<, >> binary operators overloaded.
public static BOOL operator &(BOOL b1, BOOL b2) { return new BOOL(b1.Value & b2.Value); }
public static BOOL operator |(BOOL b1, BOOL b2) { return new BOOL(b1.Value | b2.Value); }
#endregion
#region ==, !=, <, >, <=, >= comparison operators overloaded
public static bool operator ==(BOOL b1, BOOL b2) { return (b1.Value == b2.Value); }
public static bool operator !=(BOOL b1, BOOL b2) { return (b1.Value != b2.Value); }
#endregion
}
It is working as designed.
Here is what is happening:
Take the new int[] { 2, 4 } and lets marshal it into A, B, Broken, and Broken2. The last is the same as Broken, but with fields' order reversed (first b, then a).
If we marshal the ints into these structures we get the following values in memory:
So what is happening is the following:
So for A, the first int gets converted to 1, for B, the second int gets converted to 1, for Broken, since B is the last field, its rules apply, and hence the second int gets converted to 1. Similarly for Broken2.