Richard and Mark - Thanks for the info!
- Ray Heath http://www.rayheath.com On Mon, 6 May 2002 10:44:54 -0700, Mark Hall (VC++) <[EMAIL PROTECTED]> wrote: >Just to follow up on the C++ sized enum: the bug with the underlying >type has been fixed and will show up in the next release. > >-----Original Message----- >From: Richard Grimes [mailto:[EMAIL PROTECTED]] >Sent: Sunday, May 05, 2002 4:00 AM >To: [EMAIL PROTECTED] >Subject: Re: [DOTNET] Schrodininger's Enum > >Hmmm, > >Its been a while since I did any quantum mechanics[1], semiconductor >research was just too much hard work, so I moved to software development >to >take a rest ;-). I think you'll find the answer to your question lies in >the >class heirarchy. An enum derives from System::Enum, which derives from >System::ValueType. So this makes an enum a value type. The ECMA spec >says >this: > >"An enum, short for enumeration, defines a set of symbols that all have >the >same type. A type shall be an enum if and only if it has an immediate >base >type of System.Enum. Since System.Enum itself has an immediate base >type of >System.ValueType (see Partition IV), enums are value types (see Chapter >12). >"[2] > >Also note that value types can be created on the stack and this is the >case >with enums - you do not use the managed new operator. (Actually I >dislike >the way that C# blurs the use of new, I think that you should be able to >call a ctor of a value type directly, just as you do in C++. The fact >that >C# makes you use new when calling the ctor of a value type confuses the >issue IMO.) Indeed, C++ is explicit in this respect, you declare a .NET >enum >with the __value keyword and if you try to create an instance with __gc >new >the compiler will generate an error. You could say that an enum is just >a >collection of named integral values, > >When you declare an enum it looks like you are deriving from an integral >type, but this is just a trick that compilers play on you, you are >merely >telling the compiler the type on which the enum is based. >Incidentally, the type of an enum highlights a bug in the C++ compiler. >If >you do this in C#: > >enum X : long {ONE}; > >you get a class derived from System::Enum with a member called value__ >that >is int64, the C# compiler will make the value__ member the type you >specify >in the declaration of the enum. If you do this in C++: > >__value enum X : __int64 {ONE}; > >you get a class derived from System::Enum with a member called value__, >but >regardless of the type you specify, value__ is always int32. Which is >not >what you want. > >So in summary, an enum is essentially a collection of named values, and >so >it is a value type. > >Richard > >[1] "FIR Studies of III-V Semiconductor Structures", 1990 >www.grimes.demon.co.uk/thesis/Firsumm.htm >[2] Common Language Infrastructure, Partition II Metadata > >----- Original Message ----- >From: "Ray Heath" <[EMAIL PROTECTED]> >To: <[EMAIL PROTECTED]> >Sent: Saturday, May 04, 2002 11:50 PM >Subject: [DOTNET] Schrodininger's Enum > > >Visual Studio MSDN Help Wrote: > >"An Enum is a named constant whose underlying type is any integral type >except Char. If no underlying type is explicitly declared, Int32 is >used. >Enum DERIVES FROM VALUETYPE, BUT IS NOT A VALUE TYPE. Programming >languages >typically provide syntax to declare an enumeration that consists of a >set >of named constants and their values." [1] > >Although a 'String Theory' kind of guy, I am not completely familiar >with >Quantum Object-Oriented Value/Non-Value Types. Is the upper-case type in >the above paragraph accurate? If so, could someone point me to a clear >explaination? > >Much of the literature seems to claim that Enum is indeed a value type >(though not a primitive value type). [2] > > >Thanks - Ray >www.rayheath.com > > >[1] ms-help://MS.VSCC/MS.MSDNVS/cpref/html/frlrfsystemenumclasstopic.htm >[2] e.g. Pattison, Ted. "Basic Instincts: Objects and Values Part 1." >(See >Figure 2) MSDN Magazine. June 2002: 115 > >You can read messages from the DOTNET archive, unsubscribe from DOTNET, >or >subscribe to other DevelopMentor lists at http://discuss.develop.com. > >You can read messages from the DOTNET archive, unsubscribe from DOTNET, >or >subscribe to other DevelopMentor lists at http://discuss.develop.com. > >You can read messages from the DOTNET archive, unsubscribe from DOTNET, or >subscribe to other DevelopMentor lists at http://discuss.develop.com. You can read messages from the DOTNET archive, unsubscribe from DOTNET, or subscribe to other DevelopMentor lists at http://discuss.develop.com.
