C# Discussion Boards - CodeProject

14-Nov-07 4:35

Hi,

how can I call static method on a Type variable, when following is known:

- there is a Class 'A', which implements a static method - public static int GetValue().

- there may by several subclasses 'B, C ,D ,E' where all of them derive from A , so they all derive GetValue() method.

- some of subclasses may want to override GetValue() ( declare public new static int GetValue()) and return different value than parent class ('A') returns.

now, I have a variable of type Type :

Type myType;

and I know that myType contains one of subclasses of A class.

how can I call static GetValue() on myType (withount instantiating it!!!)

I hope it's not confusing Smile | :)

)

thanx

zilo

Mr. Candyman14-Nov-07 4:53

Mr. Candyman

14-Nov-07 4:53

MethodInfo inf = myType.GetMethod("GetValue");
inf.Invoke(null, new object[] parameters );

if it's static it will ignore the first parameter (null);

Stevo Z14-Nov-07 4:56

14-Nov-07 4:56

thanx

that's what I used, but I (am) was curious if there is a "nicer" way how to do that.

zilo

Mr. Candyman14-Nov-07 5:38

Mr. Candyman

14-Nov-07 5:38

Well, guess you could use generics too, if reflection isn't nice enough for you (although that's not necessarily better way to do) Laugh | :laugh:

Stevo Z14-Nov-07 5:51

14-Nov-07 5:51

ok, it works, but the thing is, that when GetValue is not overridden in subclass, reflection doesn't return parent method but null. And when I think about it, it's probably better that way. I can call parent method anyway...

zilo

Mr. Candyman14-Nov-07 6:00

Mr. Candyman

14-Nov-07 6:00

Guess it's better with generics then, cuz you don't have to worry about wether it overrides or just inherits.

Stevo Z14-Nov-07 6:10

Bit Testing and operator unchecked()

14-Nov-07 6:10

Can you be more specific how to do that with generics?

Anyway , I can't use generics, cause that method is a part of a control (Form) and there is a bug regarding Designer and generics. It would not display child classes (Forms) in Designer at all. It's bug by design...

zilo

The cleaner method

Ennis Ray Lynch, Jr.14-Nov-07 9:26

Ennis Ray Lynch, Jr.

14-Nov-07 9:26

Involved not changing return types. Another option is to create a generic method.

Need a C# Consultant? I'm available.

Happiness in intelligent people is the rarest thing I know. -- Ernest Hemingway

Jeffrey Walton14-Nov-07 4:35

14-Nov-07 4:35

Hi All,

I'm attempting to test if a high bit is set on a octet. C# is stating:

An unhandled exception of type 'System.OverflowException' occurred in mscorlib.dll
Additional information: Value was either too large or too small for an unsigned byte.

Any ideas how to work with bit values in C# without losing semantics (readability)? At this point, I'd also take something that works as expected (or works as I expect).

Jeff

// vector< byte > 
ArrayList Octets = new ArrayList(); 
... 
Int32 i=0; 
while( i < Octets.Count ) 
{ 
    Byte b = Byte.Parse(Octets[i].ToString()); 
    b &= unchecked(Convert.ToByte(~0x80));        // Problem Here 

    if( 0x00 == b ) { break; } 

    .... 

    i++; 
}

Luc Pattyn14-Nov-07 5:08

Luc Pattyn

14-Nov-07 5:08

Hi Jeffrey,

I am not sure what type it is you originally put into the ArrayList, I will assume the
item type is byte.

Here are some ideas:

- I see no point in using Parse(...ToString()), that is turning a number to a string to
get a number again. why not treat the ArrayList item directly? Of course ArrayList only
returns type object, so you would have to cast to the original type, as in
(byte)Octets[i]

- in general, bit manipulations are easiest on unsigned numbers, so byte is OK

- your problem statement is not doing what you intend: ANDing with the complement of 0x80
is ANDing with 0x7F, that is stripping off the highest bit; a subsequent zero-test will
return true iff the original value was either 0x00 or 0x80.

- testing any bit can be achieved by ANDing and comparing to zero; testing the highest bit
would require something like: if ((b & bMSB)!=0) MSB_is_set; else MSB_is_clear;
where bMSB could be a byte variable initialized to 0x80u (u suffix for unsigned).

- you can use foreach instead of for to iterate over a collection. It does not need
an index variable, you can and must specify both the type and the name of the object itself
you are getting from the collection; as a bonus you don't need to cast the ArrayList
element anymore.

So this is what I would suggest:

byte bMSB=0x80u;
ArrayList Octets = new ArrayList(); 
... 
foreach (byte b in Octets) {
    if((b & bMSB)!=0) break;     // not sure what you want: == or !=
    .... 
}

Furthermore, you could use a generic collection (available since .NET 2.0), in this case
a List<byte> instead of an ArrayList, although the difference would be minor here.

Smile | :)

Luc Pattyn [Forum Guidelines] [My Articles]

this months tips:
- use PRE tags to preserve formatting when showing multi-line code snippets
- before you ask a question here, search CodeProject, then Google

Jeffrey Walton14-Nov-07 5:41

14-Nov-07 5:41

Hi Luc,

Thank you very much for the comprehensive response. Here's the next issue (I did not like the ToString() calls either). Changing:

Byte.Parse(Octets[0].ToString())

(Byte)Octets[0]

reults in Compiler Error:

Error CS0201: Only assignment, call, increment, decrement, and new object expressions can be used as a statement

More awkwardness: changing from Parse() and ToString() to

(Int32)Octets[0] / 40

results in Runtime Error:

An unhandled exception of type 'System.InvalidCastException' occurred in OIDFetch.exe
Additional information: Specified cast is not valid.

This is frustrating. I can understand abstracting away bit operations due to code portability. But when did operations on numbers stop being number-centric?

The situation reminds me of Java. Trying to read a file results in so many casts to so many readers, you lose the readability (semantics) of what you are trying to accomplish: FileReaders to ByteReaders to StreamReaders... I never went back to Java after college.

Jeff

Luc Pattyn14-Nov-07 5:52

Luc Pattyn

14-Nov-07 5:52

Hi Jeff,

I agree one should avoid casts most of the time.
In what I explained, there basically was only one, and that one is due to the fact
that ArrayList stores objects, not known types, so when retrieving an ArrayList element,
you must cast it again to its real type. (The new List<byte> alternative solves that!)

I cannot figure out what is causing your compile error CS0201; I would need a larger
code snippet (yours is less than one statement, there is no semicolon, no curly bracket).

Smile | :)

Luc Pattyn [Forum Guidelines] [My Articles]

this months tips:
- use PRE tags to preserve formatting when showing multi-line code snippets
- before you ask a question here, search CodeProject, then Google

Judah Gabriel Himango14-Nov-07 5:54

Judah Gabriel Himango

14-Nov-07 5:54

Jeff, some casting is going on here because you're using old stuff, ArrayList. Use List<byte> for the octets, and you won't need any casting.

Tech, life, family, faith: Give me a visit.
I'm currently blogging about: The Lord Is So Good
The apostle Paul, modernly speaking: Epistles of Paul

Judah Himango

Jeffrey Walton14-Nov-07 6:09

14-Nov-07 6:09

Hi Judah,

Thanks. I found it in System.Collections.Generic.List<byte>. Time for another rewrite (copy and paste is my friend).

Jeff

Guffa14-Nov-07 9:24

Guffa

14-Nov-07 9:24

Jeffrey Walton wrote:
Changing:
Byte.Parse(Octets[0].ToString())
to
(Byte)Octets[0]
reults in Compiler Error:
Error CS0201: Only assignment, call, increment, decrement, and new object expressions can be used as a statement

Then you obviously changed something more in that statement. If you only changed that, it would compile just fine.

Jeffrey Walton wrote:
changing from Parse() and ToString() to
(Int32)Octets[0] / 40
results in Runtime Error:
An unhandled exception of type 'System.InvalidCastException' occurred in OIDFetch.exe
Additional information: Specified cast is not valid.

When you are unboxing a boxed value, you have to use the exact data type of the boxed value. Once you have unboxed it to get the byte value, you can cast it to an int if you like.

Experience is the sum of all the mistakes you have done.

Jeffrey Walton14-Nov-07 6:05

14-Nov-07 6:05

Hi Luc,

I hope it's not bad form to spawn a response for each question... I think it keeps dialogue exchange manageable.

I am not sure what type it is you originally put into the ArrayList, I will assume the item type is byte.

I used ArrayList due to a recomendation on the C# usenet group. I despise the concept of ArrayList of Objects for numbers (integers, shorts longs, etc)... I don't like the casting. I understand one may have an array of type Employee, where this makes sense.
For a 32 bit integer or 8 bit byte, this is offensive to me. What I want is a vector of bytes - no casting, perform math on the elements, etc. If C# wants to call it a collection, that is fine. But don't take away semantics of primitive operations such as v[2] = v[0] * v[1].

Luc Pattyn wrote:
Furthermore, you could use a generic collection (available since .NET 2.0), in this case
a List<byte> instead of an ArrayList, although the difference would be minor here.

I'd like to try use this... Where does one find it? I did not see it in System.Collection. I imagine it is elsewhere? I am using VS 2005. I have .Net 1.1 and 2.0 Configuration Wizards, so I imagine the runtimes are installed.

Jeff

Jeffrey Walton14-Nov-07 6:08

14-Nov-07 6:08

Hi Luc,

Found it in System.Collections.Generic.List<byte>.

Jeff

Luc Pattyn14-Nov-07 6:39

Luc Pattyn

14-Nov-07 6:39

Hi Jeff,

"collection" (lower-case!) is the general term for an object holding a lot of objects that
can be enumerated somehow; so arrays, ArrayLists, Lists, Dictionaries, all are collections.

Array, ArrayList containing numbers, and List<some_numeric_type> v all would support
v[2] = v[0] * v[1].

Jeffrey Walton wrote:
I used ArrayList due to a recomendation on the C# usenet group.

Either that recommendation dates back from the .NET 1.x times, or it is suboptimal.
Basically List<T> is an ArrayList that only allows you to add objects of type T, and
automatically returns a type T when you access its items.

Starting with VS2005 (=.NET 2.0) a new Win app automatically gets a
"using System.Collections.Generic" and no longer a "using System.Collections".

Smile | :)

Luc Pattyn [Forum Guidelines] [My Articles]

this months tips:
- use PRE tags to preserve formatting when showing multi-line code snippets
- before you ask a question here, search CodeProject, then Google

Guffa14-Nov-07 9:15

Guffa

14-Nov-07 9:15

Jeffrey Walton wrote:
b &= unchecked(Convert.ToByte(~0x80)); // Problem Here

The 0x80 literal is an int value. When you use the bitwise completent operator (~) on it, it will switch all bits in the int value, which results in the value -0x7FFFFF7F. This is clearly way too large to fit in a byte.

Experience is the sum of all the mistakes you have done.

Jeffrey Walton14-Nov-07 11:33

Bit Testing And operator~

14-Nov-07 11:33

Hi Guffa,

Guffa wrote:
The 0x80 literal is an int value.

It's a shame C# has a penchant for the Int conversion, even when I specify otherwise. For example, the following:

Byte b = 0x80;
b = ~b;

results in a Compiler Error:

Cannot implicitly convert type 'int' to 'byte'. An explicit conversion exists (are you missing a cast?)

Humorous... When I attempt to cast, it complains also. The problems would go away if the compiler would listen to the human operator and simply compliment the Byte value b.

Jeff

Jeffrey Walton14-Nov-07 11:49