|
I've run into a problem with .NET remoting. I need the client to instantiate an object on the server, however, intantiating that object requires I pass parameters to the constructor.
public class MyObj
{
public MyObj(string s, long l, SR someReferenceType)
{
}
}
RemotingConfiguration.RegisterWellKnownServiceType(typeof(MyObj), "MyURI", WellKnownObjectMode.Singleton);
RemotingServices.Connect(typeof(MyObj), "http://localhost/MyURI");
The RemotingServices.Connect call throws an exception because there is no default constructor. I need to somehow pass several value types and a remote object reference type to the object being constructed (it'd even be ok if I passed this data to the object after construction, but that doesn't seem possible either).
Any ideas?
---------------------------
He who knows that enough is enough will always have enough.
-Lao Tsu
|
|
|
|
|
Create your own RealProxy derivative that can incercept the construction calls and instantiate your class with the parameters provided. Should work since you're now the one controlling how objects are created now.
Microsoft MVP, Visual C#
My Articles
|
|
|
|
|
Thanks for the reply. I have now created a RealProxy derivative for my type. I've designed the RealProxy about as closesly as I could to some MSDN examples...is there anything else I need to do to make this work? I assume I've gotta hook up that proxy when I call .Connect to get the proxy on the client or similar, is that right?
---------------------------
He who knows that enough is enough will always have enough.
-Lao Tsu
|
|
|
|
|
You actually don't have to handle as much as you might think when creating your own proxy. You can choose to handle which methods you want. I recommend - if you haven't already - picking up a good book on .NET Remoting like "Microsoft .NET Remoting" from MS Press[^] or the one from Ingo Rammer[^] (supposed to be very good).
Microsoft MVP, Visual C#
My Articles
|
|
|
|
|
Look at the code: the are two similar methods. One declares all variables used in the method at the start of the method, the other one declares them when it needs then. May there be any lower CPU or memory cost in the second version of the method (the one at the bottom) ?
<br />
public void DoSomethingComplex(...)<br />
{<br />
int[] intarray = new ...<br />
bool somecondition = ...<br />
if (somecondition == true)<br />
intarray[1].Add ...
}<br />
<br />
public void DoSomethingComplex(...)<br />
{<br />
bool somecondition = ...<br />
if (somecondition == true)<br />
{<br />
int[] intarray = new ...<br />
intarray[1].Add ...
}<br />
}
Regards, Desmond
|
|
|
|
|
<br />
using System;<br />
<br />
class Test {<br />
<br />
<br />
public static void DoSomethingComplex(bool _someCondition) {<br />
int[] intArray;<br />
<br />
if (_someCondition) {<br />
intArray = new int[100];<br />
}<br />
<br />
}<br />
<br />
static void Main(string[] args) {<br />
DateTime startTime, endTime;<br />
int millisecondCount;<br />
int loopCount = 100000;<br />
<br />
startTime = DateTime.Now;<br />
for(int x = 0; x < loopCount; x++) {<br />
DoSomethingComplex(true);<br />
DoSomethingComplex(true);<br />
DoSomethingComplex(true);<br />
DoSomethingComplex(true);<br />
DoSomethingComplex(true);<br />
DoSomethingComplex(true);<br />
DoSomethingComplex(true);<br />
DoSomethingComplex(true);<br />
DoSomethingComplex(true);<br />
DoSomethingComplex(true);<br />
}<br />
endTime = DateTime.Now;<br />
millisecondCount = ((int)(endTime.Ticks - startTime.Ticks)) / 10000;<br />
Console.WriteLine(millisecondCount + " milliseconds" );<br />
<br />
startTime = DateTime.Now;<br />
for(int x = 0; x < loopCount; x++) {<br />
DoSomethingComplex(false);<br />
DoSomethingComplex(false);<br />
DoSomethingComplex(false);<br />
DoSomethingComplex(false);<br />
DoSomethingComplex(false);<br />
DoSomethingComplex(false);<br />
DoSomethingComplex(false);<br />
DoSomethingComplex(false);<br />
DoSomethingComplex(false);<br />
DoSomethingComplex(false);<br />
}<br />
endTime = DateTime.Now;<br />
millisecondCount = ((int)(endTime.Ticks - startTime.Ticks)) / 10000;<br />
Console.WriteLine(millisecondCount + " milliseconds" );<br />
<br />
}<br />
}<br />
|
|
|
|
|
The second version is better since you allocate the memory only if
if (somecondition == true)
Howver you can do the following:
<br />
public void DoSomethingComplex(...)<br />
{<br />
int[] intarray;
bool somecondition = ...<br />
if (somecondition == true)<br />
intarray = new ...
intarray[1].Add ...<br />
}
|
|
|
|
|
Unless you are into really performance critical applications (and if you are, why aren't you using C++?) I think good coding style and design is generally more important than worrying about how questions like this.
I prefer to declare locals at the point where I going to use them, not detached at the head of the function. To me this localises the use of the variable, and is just better style. Anyway, we shouldn't try to second guess the compiler - for all we know, the compiled code is near identical in both cases.
|
|
|
|
|
Hi,
I would like to send data from one instantiation of class to a textbox in my Form.
public class Form1 : System.Windows.Forms.Form
{
private System.Windows.Forms.TextBox textBox1;
.
.
}
public class AnotherClass
{
public void SomeMethod(String msg)
{
Form1.textBox1.Text = msg;
}
}
I have tryed setting the textBox1 public. I have try creating an accessor method. But no luck.
Any comments or suggestions would be very welcome.
Thanks in advance.
Jack
|
|
|
|
|
That isn't exactly proper programming practice. It could be done, but really, ALL your code for manipulating the controls on the form should be done on the form, not in a class that doesn't (shouldn't) know what a form is.
If all you doing is what it's your example, that code really shouldn't be in a seperate class. In order for it to work, you'd have to also pass a reference to the textbox to the SomeMethod function.
public void SomeMethod(ref TextBox tbToModify, String strMessage)
{
tbToModify.Text = strMessage;
}
RageInTheMachine9532
|
|
|
|
|
I want to write a application it like Microsoft's PowerPoint User interface..
when user add a new file my application can new a file but the file is adding to other form and can scroller to view and edit others file just like powerpint
how can i write the application or where has similar code...
please help me....thanks
|
|
|
|
|
You can add Microsoft PowerPoint Object Library to your project and use its powerpoint functionality. You can do it by add refrence-->CoM tab.
Mazy
"A bank is a place that will lend you money if you can prove that you don't need it." - Bob Hope
|
|
|
|
|
sorry...i just want a interface like powerpoint which can satisfy my need.(can add a new file to a new form and can scroll up and down all forms just like powerpoint)
I do want't just use a powerpoint component....
everyone please help...thanks your ideas~~~
|
|
|
|
|
|
I am looking for a good book that explores .net diagnostics in depth such as topics about Tracing, Listeners, Event Logs etc. in .NET. Does anyone have any suggestions? I can't find any so far.
Mark Sanders
|
|
|
|
|
See the Microsoft Patterns and Practices[^] section. They have books available for download as PDFs (you can also buy them as books, but the PDFs are complete). One or two go into depth with instrumenting your application, like the Enterprise Instrumention block.
As far as trace listeners, the documentation for the TraceListener in the .NET Framework SDK should be sufficient. There isn't much you can do with these, though, which is where the EI comes into play.
Microsoft MVP, Visual C#
My Articles
|
|
|
|
|
I have my multi-threaded plugin app that I wrote. In the dispose method of a plugin if I issue a Application.Exit it exits both the plugin and the parent app.
Is this normal, seems kinda strange to me...
Matthew Hazlett
Windows 2000/2003 MCSE
Never got an MCSD, go figure...
|
|
|
|
|
Um, this is what it's supposed to do. That's why it's Application.Exit. Read the documentation.
Besides, you don't need to exit anything when your plugin is disposed. The reference object is freed. You really don't even need to implement IDisposable unless you have native resources or other resources you want cleaned up. This is how a garbage collection platform - like .NET and Java - work.
Microsoft MVP, Visual C#
My Articles
|
|
|
|
|
Hi folks
I've got an array that i am using as a refference to COM that will store the IntPtr's returned from the IEnumID.
<br />
[PreserveSig]<br />
int Next(<br />
int celt,
[MarshalAs(UnmanagedType.LPArray, SizeConst = 100)]<br />
IntPtr[] rgelt,
out int pceltFetched);
While i have it marshaling as an LPArray, the problem i face is that the pointer points to the first element (as far as I can see), and while, forwarding this array to the Next produces a new result (Assumably the next folder in the enumeration), it is only ever allocated to the first possition. The values for the celt increments properly, but the pCeltFetched is always equal to 1 after the first method call.
Am I marshaling it incorectly? I've tried using SafeArray with an IntPtr type, but the compiler keeps spitting it out.
I know that they are marshaled as in only by default, but is there some special handling procedure for selecting the index that the array should begin at?
I have also tried converting the array to a unmanaged structure, so that i can simply use a pointer to handle it, instead of marshaling to and from managed code with each call. However, i get null pointer exceptions, and it's not liking it. I'm calling like this:
IntPtr[] EnumerationArray = new IntPtr[100];<br />
IntPtr ArrayHandle = IntPtr.Zero;<br />
Marshal.StructureToPtr(EnumerationArray, ArrayHandle,true);
Cheers
Cata
|
|
|
|
|
Hi, I don't know the answer to your question, but this 10 minute MSDN video[^] could very well provide the information you need to solve the problem yourself. It discusses the different ways that items are marshalled in platform invoke calls, though not specifically about COM.
BTW, assuming I knew the answer to your problem, it would be difficult to answer your question without also knowing what the COM interface actually expects. You've only listed the part that isn't using the interface correctly. [edit]I might have assumed incorrectly there, if you are referring to the common enumeration interfaces that COM defines.[/edit]
John
"You said a whole sentence with no words in it, and I understood you!" -- my wife as she cries about slowly becoming a geek.
|
|
|
|
|
This is how it is written in C:
HRESULT Next(
ULONG celt,
LPITEMIDLIST *rgelt,
ULONG *pceltFetched
);
where celt is the number of items in the array that is forwarded.
rgelt is a pointer to the array.
pceltFetched is a pointer to a value that holds the number of elements in the returned array.
I shall check this video out.
Cheers
Cata
|
|
|
|
|
That's the way it's supposed to work. Read the documentation in the Platform SDK for IEnumXXXX. It's just like an IEnumerator in .NET, only COM's IEnumXXXX implementations can fetch multiple results depending on what you pass into celt (the first param to Next). If you know the size of the object (like an ITEMIDLIST, which you can get with Marshal.SizeOf), you can get the offset the address and get the next object in the array using other methods in the Marshal class, or using an unsafe context.
Another way would be to create an array of ITEMIDLIST (or whatever struct you need, so long as it's defined and attributed properly) and pin the object to gets its address like so:
ITEMIDLIST[] items = new ITEMIDLIST[40];
GCHandle handle = GCHandle.Alloc(items);
int len;
iEnum.Next(items.Length, handle.AddrOfPinnedObject(), out len);
handle.Free(); Then you should be able to access the elements of the array with managed code.
Microsoft MVP, Visual C#
My Articles
|
|
|
|
|
Hmmmm.
Maybe i've misinterpreted what i read.
I was under the assumption that the array that i hand to the Next object, is something akin to a stack that new PIDL's are added too, and the celt is the number of items currently on the array, or the index at which the next PIDL is added.
Seems kind of dumb now i think about it.
The PIDL that i am getting back in the first part of the array (I only get one PIDL at a time), is the one i'm looking for?
Can i continuously increment next, and store the PIDL in a purely managed array?
Cheers
Cata
|
|
|
|
|
The Catalyst wrote:
Can i continuously increment next, and store the PIDL in a purely managed array?
Sure. Why not? You should use a list, though, since you'd have to resize the array yourself, or use any length property that the enum might have.
The Catalyst wrote:
I was under the assumption that the array that i hand to the Next object, is something akin to a stack that new PIDL's are added too, and the celt is the number of items currently on the array, or the index at which the next PIDL is added.
Nope. The first param is how large the array is. Since an array is just a pointer, you have to tell native APIs how many elements are in that array so it can use offsets to fill the addresses. This is true of any native API that uses arrays (even character arrays like LPTSTR).
Microsoft MVP, Visual C#
My Articles
|
|
|
|
|
Thanks Heath, most helpful. I was planning on using an arraylist. 
One more question.
I'm running my call through Next, like this:
IntPtr[] EnumerationArray = new IntPtr[10];<br />
int ArrayLengthIn = 10;<br />
int ArrayLengthOut = 0;<br />
<br />
ContentEnumeration.Next(ArrayLengthIn,EnumerationArray,out ArrayLengthOut);
But I only ever get a single value returned.
I'm not sure why this is, as i know that there is more than 1 folder on my C drive. I'll have a play with one of the pointers i get back from the array, but can you see any reason why it's only returning a single value?
Cata
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
