Large flags enumerations in C#
Asked Answered
G

4

20

Hey everyone, got a quick question that I can't seem to find anything about...

I'm working on a project that requires flag enumerations with a large number of flags (up to 40-ish), and I don't really feel like typing in the exact mask for each enumeration value:

public enum MyEnumeration : ulong
{
    Flag1 = 1,
    Flag2 = 2,
    Flag3 = 4,
    Flag4 = 8,
    Flag5 = 16,
    // ...
    Flag16 = 65536,
    Flag17 = 65536 * 2,
    Flag18 = 65536 * 4,
    Flag19 = 65536 * 8,
    // ...
    Flag32 = 65536 * 65536,
    Flag33 = 65536 * 65536 * 2
    // right about here I start to get really pissed off
}

Moreover, I'm also hoping that there is an easy(ier) way for me to control the actual arrangement of bits on different endian machines, since these values will eventually be serialized over a network:

public enum MyEnumeration : uint
{
    Flag1 = 1,     // BIG: 0x00000001, LITTLE:0x01000000
    Flag2 = 2,     // BIG: 0x00000002, LITTLE:0x02000000
    Flag3 = 4,     // BIG: 0x00000004, LITTLE:0x03000000
    // ...
    Flag9 = 256,   // BIG: 0x00000010, LITTLE:0x10000000
    Flag10 = 512,  // BIG: 0x00000011, LITTLE:0x11000000
    Flag11 = 1024  // BIG: 0x00000012, LITTLE:0x12000000
}

So, I'm kind of wondering if there is some cool way I can set my enumerations up like:

public enum MyEnumeration : uint
{
     Flag1 = flag(1), // BOTH: 0x80000000
     Flag2 = flag(2), // BOTH: 0x40000000
     Flag3 = flag(3), // BOTH: 0x20000000
     // ...
     Flag9 = flag(9), // BOTH: 0x00800000
}

What I've Tried:

// this won't work because Math.Pow returns double
// and because C# requires constants for enum values
public enum MyEnumeration : uint
{
    Flag1 = Math.Pow(2, 0),
    Flag2 = Math.Pow(2, 1)
}

// this won't work because C# requires constants for enum values
public enum MyEnumeration : uint
{
    Flag1 = Masks.MyCustomerBitmaskGeneratingFunction(0)
}

// this is my best solution so far, but is definitely
// quite clunkie
public struct EnumWrapper<TEnum> where TEnum
{
    private BitVector32 vector;
    public bool this[TEnum index]
    {
         // returns whether the index-th bit is set in vector
    }
    // all sorts of overriding using TEnum as args
}

Just wondering if anyone has any cool ideas, thanks!

Gallantry answered 7/5, 2010 at 22:34 Comment(3)
You typed a few hundred lines of text here. Why didn't you just bite the bullet and type the original 40 lines? (You might have used 1<<1, 1<<2, ... instead of the multiplies but anyway...)Healey
Umm... I did want to present the solutions I had tried, "<<" and ">>" would have worked, but I don't exactly see the difference when they are only evaluated once. Its not like it takes me half an hour to type this out, a few hundred lines of text isn't much... I'd rather provide too much information than too little...Gallantry
you can use the full range of the ulong with the shift operator, you will just need to indicate to the compiler that the 1 you are shifting is a ulong. 1ul << 63, notice the 'ul' afer the 1.Forelady
M
13

You could write a T4 template to generate the enum :

Template (MyEnumeration.tt)

<#@ template language="C#" #>
<#@ output extension=".cs" #>
using System;

namespace MyNamespace
{
    [Flags]
    public enum MyEnumeration : ulong
    {
<#
    ulong value = 1;
    for(int i = 1; i <= 64; i++)
    {
#>
        Flag<#= i #> = <#= string.Format("0x{0:X8}", value) #>,
<#
        value = value << 1;
    }
#>
    }
}

Resulting C# code (MyEnumeration.cs)

using System;

namespace MyNamespace
{
    [Flags]
    public enum MyEnumeration : ulong
    {
        Flag1 = 0x00000001,
        Flag2 = 0x00000002,
        Flag3 = 0x00000004,
        Flag4 = 0x00000008,
        Flag5 = 0x00000010,
        Flag6 = 0x00000020,
        Flag7 = 0x00000040,
        Flag8 = 0x00000080,
        Flag9 = 0x00000100,
        Flag10 = 0x00000200,
        Flag11 = 0x00000400,
        Flag12 = 0x00000800,
        Flag13 = 0x00001000,
        Flag14 = 0x00002000,
        Flag15 = 0x00004000,
        Flag16 = 0x00008000,
        Flag17 = 0x00010000,
        Flag18 = 0x00020000,
        Flag19 = 0x00040000,
        Flag20 = 0x00080000,
        Flag21 = 0x00100000,
        Flag22 = 0x00200000,
        Flag23 = 0x00400000,
        Flag24 = 0x00800000,
        Flag25 = 0x01000000,
        Flag26 = 0x02000000,
        Flag27 = 0x04000000,
        Flag28 = 0x08000000,
        Flag29 = 0x10000000,
        Flag30 = 0x20000000,
        Flag31 = 0x40000000,
        Flag32 = 0x80000000,
        Flag33 = 0x100000000,
        Flag34 = 0x200000000,
        Flag35 = 0x400000000,
        Flag36 = 0x800000000,
        Flag37 = 0x1000000000,
        Flag38 = 0x2000000000,
        Flag39 = 0x4000000000,
        Flag40 = 0x8000000000,
        Flag41 = 0x10000000000,
        Flag42 = 0x20000000000,
        Flag43 = 0x40000000000,
        Flag44 = 0x80000000000,
        Flag45 = 0x100000000000,
        Flag46 = 0x200000000000,
        Flag47 = 0x400000000000,
        Flag48 = 0x800000000000,
        Flag49 = 0x1000000000000,
        Flag50 = 0x2000000000000,
        Flag51 = 0x4000000000000,
        Flag52 = 0x8000000000000,
        Flag53 = 0x10000000000000,
        Flag54 = 0x20000000000000,
        Flag55 = 0x40000000000000,
        Flag56 = 0x80000000000000,
        Flag57 = 0x100000000000000,
        Flag58 = 0x200000000000000,
        Flag59 = 0x400000000000000,
        Flag60 = 0x800000000000000,
        Flag61 = 0x1000000000000000,
        Flag62 = 0x2000000000000000,
        Flag63 = 0x4000000000000000,
        Flag64 = 0x8000000000000000,
    }
}

In order to edit T4 templates, I recommend you use a T4 editor plugin like this one (this gives you syntax highlighting and Intellisense)

Medullated answered 7/5, 2010 at 23:14 Comment(4)
You hardly need a T4 system to do this. A simple loop in any scripting programming language can print the core of this.Healey
Yes, but T4 in built into Visual Studio, which makes it very convenient for that kind of things...Medullated
Nice. Is it possible to go above 64?Rialto
@arao6, no. For that you would need an integer type with more than 64bits... and there isn't one (well, there is BigInteger of course, but it can't be used in an enum)Medullated
D
16

Why not just do:

public enum MyEnumeration : ulong 
{ 
    Flag1 = 1, 
    Flag2 = 1 << 1, 
    Flag3 = 1 << 2, 
    Flag4 = 1 << 3, 
    .
    .
    .
    Flag30 = 1 << 29, 
    Flag31 = 1 << 30, 
    Flag32 = 1 << 31
}
Dressel answered 7/5, 2010 at 22:52 Comment(3)
hmm... bit shifting behind 8 bits seems weird to me... I'm not sure if the compiler handles this automatically, but technically, shouldn't (1 << 8) == 0 regardless of the size of the datatype on little endian systems? I may be completely out to lunch, I'm not sureGallantry
@LorenVS, you can use the full range of the ulong with the shift operator, you will just need to indicate to the compiler that the 1 you are shifting is a ulong. 1ul << 63, notice the 'ul' afer the 1.Forelady
@Gallantry - The compiler is fine with that. It is, however, wrapped to the data-size. So actually for int/Int32, 1 << 33 is the same as 1 << 1. Since (note Chris's comment, though) we're using ulong in this case it is % 64, so 1 << 65 is the same as 1 << 1Muscarine
M
13

You could write a T4 template to generate the enum :

Template (MyEnumeration.tt)

<#@ template language="C#" #>
<#@ output extension=".cs" #>
using System;

namespace MyNamespace
{
    [Flags]
    public enum MyEnumeration : ulong
    {
<#
    ulong value = 1;
    for(int i = 1; i <= 64; i++)
    {
#>
        Flag<#= i #> = <#= string.Format("0x{0:X8}", value) #>,
<#
        value = value << 1;
    }
#>
    }
}

Resulting C# code (MyEnumeration.cs)

using System;

namespace MyNamespace
{
    [Flags]
    public enum MyEnumeration : ulong
    {
        Flag1 = 0x00000001,
        Flag2 = 0x00000002,
        Flag3 = 0x00000004,
        Flag4 = 0x00000008,
        Flag5 = 0x00000010,
        Flag6 = 0x00000020,
        Flag7 = 0x00000040,
        Flag8 = 0x00000080,
        Flag9 = 0x00000100,
        Flag10 = 0x00000200,
        Flag11 = 0x00000400,
        Flag12 = 0x00000800,
        Flag13 = 0x00001000,
        Flag14 = 0x00002000,
        Flag15 = 0x00004000,
        Flag16 = 0x00008000,
        Flag17 = 0x00010000,
        Flag18 = 0x00020000,
        Flag19 = 0x00040000,
        Flag20 = 0x00080000,
        Flag21 = 0x00100000,
        Flag22 = 0x00200000,
        Flag23 = 0x00400000,
        Flag24 = 0x00800000,
        Flag25 = 0x01000000,
        Flag26 = 0x02000000,
        Flag27 = 0x04000000,
        Flag28 = 0x08000000,
        Flag29 = 0x10000000,
        Flag30 = 0x20000000,
        Flag31 = 0x40000000,
        Flag32 = 0x80000000,
        Flag33 = 0x100000000,
        Flag34 = 0x200000000,
        Flag35 = 0x400000000,
        Flag36 = 0x800000000,
        Flag37 = 0x1000000000,
        Flag38 = 0x2000000000,
        Flag39 = 0x4000000000,
        Flag40 = 0x8000000000,
        Flag41 = 0x10000000000,
        Flag42 = 0x20000000000,
        Flag43 = 0x40000000000,
        Flag44 = 0x80000000000,
        Flag45 = 0x100000000000,
        Flag46 = 0x200000000000,
        Flag47 = 0x400000000000,
        Flag48 = 0x800000000000,
        Flag49 = 0x1000000000000,
        Flag50 = 0x2000000000000,
        Flag51 = 0x4000000000000,
        Flag52 = 0x8000000000000,
        Flag53 = 0x10000000000000,
        Flag54 = 0x20000000000000,
        Flag55 = 0x40000000000000,
        Flag56 = 0x80000000000000,
        Flag57 = 0x100000000000000,
        Flag58 = 0x200000000000000,
        Flag59 = 0x400000000000000,
        Flag60 = 0x800000000000000,
        Flag61 = 0x1000000000000000,
        Flag62 = 0x2000000000000000,
        Flag63 = 0x4000000000000000,
        Flag64 = 0x8000000000000000,
    }
}

In order to edit T4 templates, I recommend you use a T4 editor plugin like this one (this gives you syntax highlighting and Intellisense)

Medullated answered 7/5, 2010 at 23:14 Comment(4)
You hardly need a T4 system to do this. A simple loop in any scripting programming language can print the core of this.Healey
Yes, but T4 in built into Visual Studio, which makes it very convenient for that kind of things...Medullated
Nice. Is it possible to go above 64?Rialto
@arao6, no. For that you would need an integer type with more than 64bits... and there isn't one (well, there is BigInteger of course, but it can't be used in an enum)Medullated
B
8

Use 1UL like this

[Flags]    
    public enum EventType : ulong
    {
        f1= 1,
        f2= 1 << 1,
        f3= 1 << 2,
        ......

        f... = 1 << 30,
        f... = 1UL << 31,
        f...  = 1UL << 32        
    }
Boscage answered 27/2, 2019 at 8:57 Comment(1)
This is by far the easiest solution to implement and use. Not sure why it doesn't have more votes.Vasiliu
F
0

Well to address the endianes you have two options that I can think of off the top of my head

1- Handle the serialization your self and use System.Net.IPAddress.HostToNetworkOrder to ensure consistent byte ordering on the wire and of course do the reverse with System.Net.IPAddress.NetworkToHostOrder when you deserialize.

I have two old blog posts on the topic of binary serialization, they could do with an update but it is a starting point.

http://taylorza.blogspot.com/2010/04/archive-binary-data-from-structure.html
http://taylorza.blogspot.com/2010/04/archive-structure-from-binary-data.html

2- Serialize to XML, in which case endianes is not an issue but of course there are other downsides such as payload size and general performance.

Forelady answered 7/5, 2010 at 22:48 Comment(0)

© 2022 - 2024 — McMap. All rights reserved.