这两种结构有什么区别

Question

There are 2 ways I can define this structure which I wish to pass as an argument to a pinvoke function. I wanted to know what was the difference between the 2

[StructLayout(LayoutKind.Sequential)]
public struct Rect {
   public int left;
   public int top;
   public int right;
   public int bottom;
}   

[StructLayout(LayoutKind.Explicit)]
public struct Rect {
   [FieldOffset(0)] public int left;
   [FieldOffset(4)] public int top;
   [FieldOffset(8)] public int right;
   [FieldOffset(12)] public int bottom;
}

From the definitions of layouts I found here , shouldnt both look the same in memory? Any advantages of one over the other?

Answer 1

From the definitions of layouts I found here, shouldn't both look the same in memory?

Yes, they will look the same in memory.

Any advantages of one over the other?

One's faster to type out, and is easier to read.

---

The use of FieldOffset is of course a useful tool; it's not like it's always useless, but if you just so happen to use it to explicitly lay out the fields in the manor in which they would be laid out by default, then it's useless. If you use it to lay out fields in a manor other than their default value (for example, having the overlap, adding padding space, having the underlying representation be in a different order than the declaration order, etc.), then it's not useless.

Answer 2

In theory these two are exactly the same. Typically you would use the explicit layout when you are interacting with non-managed code. This can be important because in the future "int" might not have 32 bits in it - meaning that the sequential layout might no longer be applicable. Hopefully this helps a bit!

Edit

One more thought is if you were mapping a .NET structure to an unmanaged union or using different types in .NET than in unmanaged code this would be really useful.

Second Edit

Others have noted that MS will "never" change the size of an int. I can agree this would be a breaking change and it is therefore highly unlikely. That said, being explicit about structure mapping when marshaling values between .NET and unmanaged code can still be a good idea. This is especially true if there is a possibility of unmanaged structure/layout changing in the future.

Answer 3

I don't think that there is a real better layout version for this problem. The second version with the Explicit LayoutKind though allows you to change the order of the structure's members as you indicate their position in the memory with the fieldoffest. The second version is easier to expand in my opinion.