|
PIEBALDconsult wrote: DataAdapters, DataSets, and DataGrid(View)s weren't created for actual real-world use -- their only purpose is allow MS demonstrators to quickly create barely-useful applications to make easily-impressed audiences go "ooohhh".
10+
I bug
|
|
|
|
|
Ok how do you recommend getting the data from sql server database into the application if you do not recommend the use of datasets?
|
|
|
|
|
that question has been answered here already.
|
|
|
|
|
Motion Seconded, on all points.
ragnaroknrol The Internet is For Porn[^]
Pete o'Hanlon: If it wasn't insulting tools, I'd say you were dumber than a bag of spanners.
|
|
|
|
|
Cynical, definately cynical.
Never underestimate the power of human stupidity
RAH
|
|
|
|
|
|
So, it's been a while since I did any threading and so I thought to review some BackgroundWorkers.
I made a test program and get this exception I cannot explain.
It occurs during the ProgressChanged event when reading a double value.
I'm using an ArrayList as argument where the first value is a double (for accurate progress report) and the second value is a string array.
The exception gets thrown when trying to assign the value of the ArrayList to a local double variable.
Can anyone tell me what went wrong and how I can fix it?
private void BW_Text_ProgressChanged(object sender, ProgressChangedEventArgs e)
{
ArrayList arg = (ArrayList)e.UserState;
double progress = (double)arg[0];
string[] content = (string[])arg[1];
textprogress += progress;
TSL_Text.Text = "Status: Reading..." + String.Format("{0:0.##}", textprogress) + "%";
Fill_ListView(true, content);
}Thanks!
|
|
|
|
|
Rather than just casting the UserState to an ArrayList, check it: use is or as and make sure it is what you think it is - a UserState could in theory be any Object, so it is well worth checking. I would use as as I like to see explicit null checks.
Do the same with the content of the ArrayList: check is is a double.
Since you don't say what the exception is, it could be caused by either; but check both - it doesn't add a lot of overhead and it makes your program a lot more robust.
Real men don't use instructions. They are only the manufacturers opinion on how to put the thing together.
|
|
|
|
|
I told you what the exception is, VS doesn't give any further explanation (which is why I'm so confused).
The code that throws it (according to VS) is the Application.Run line in the program class.
Though seemingly I was able to get rid of it by properly checking the userstate and always sending new ArrayList.
On the other hand, ReportProgress seems sorta broken. It happens once every blue moon compared to the DoWork event.
In DoWork I'm running a loop and part of it is the ReportProgress method, yet the first time the event is actually raised happens between randomly the 350th and 400th time of the loop.
How come?
When debugging I set break points in the DoWork and ReportProgress events and confirmed that DoWork is only actually raised after DoWork looped several times.
|
|
|
|
|
Megidolaon wrote: In DoWork I'm running a loop and part of it is the ReportProgress method, yet the first time the event is actually raised happens between randomly the 350th and 400th time of the loop.
That is perfectly feasible as the ReportProgress method raises the ProgressChanged event asynchronously. Presumably by the time the UI thread gets around to processing the first event the BackgroundWorker has called ReportProgress 350-400 times and quite a backlog has built up. If each ProgressChangedEventArgs references the same ArrayList then you will always see the most recent values and not those that were in force at the time the event was raised.
As a guess your code should not raise the event more frequently than once every 1000 loop iterations. Even that will probably be far too frequent and you should aim for no more than 2 per second if a display element is being updated.
Alan.
|
|
|
|
|
Thanks.
I guess I should better collect the data within the DoWork event and then report a while bunch at once.
But now the TargetInvocationException is back and I can#t find it, no matter where I set a breakpoint.
modified on Monday, August 16, 2010 9:47 PM
|
|
|
|
|
In the debugger, check to be sure the actual values are what you expect.
But, personally, I wouldn't use the ArrayList, I'd make separate fields in the ProgressChangedEventArgs.
|
|
|
|
|
Three remarks FYI:
1.
there is ProgressChangedEventArgs.ProgressPercentage which is of type int; its name suggests the value should be in the range [0,100] however the value is irrelevant to .NET and you can pass any value you choose, from int.MinValue to int.MaxValue, so you have a full 32-bit resolution at your disposal.
2.
I don't expect people be interested in a very accurate progress value; if you were to display a progress bar on a display with say 1280 pixels, then you would be able to show no more than 1281 different values anyway; yes you could just show a number, but that does seem a bit awkward.
3.
Once you have settled on the exact range and resolution you want to pass around, you should make sure not to call ReportProgress() with the same progress value over and over, as each call causes a thread switch, and achieves nothing as no new information is passed anyway. So create a little method and compare with the previous value, as in:
public void MyReportProgress(int val) {
if (val!=previousVal) {
previousVal=val;
ReportProgress(val);
}
}
You may be surprised how much faster your actual operation runs once you stop passing unnecessary progress resolution back and forth (and stop repainting that progress bar abundantly).
|
|
|
|
|
Hi
I'm working on an client/server application which uses the UdpClient class to communicate over a network. The server end waits for a datagram to be received, after which it then processes the datagram and sends back a response. To stop the code from blocking, I am using Udpclient.BeginReceive and doing it all asynchronously. The BeginReceive method takes a delegate as its argument, which points to the method which processes the data (the callback method). What I need to know is whether the callback method runs in a separate thread? As usual the documentation is useless.
Thanks
modified on Sunday, August 15, 2010 11:35 AM
|
|
|
|
|
The callback method runs in a separate thread obtained from the thread pool.
|
|
|
|
|
|
Don't know about the OP but I found it useful.
I bug
|
|
|
|
|
Thanks.
|
|
|
|
|
How much memory can be consumed for a process that is running in
a) 32 bit environment
b) 64 bit environment
Is it 2G( less than ) per process? If so, is there any reason for such a limitation?
Suppose process is running with 2GB RAM. And if I am upgrading the RAM to 4GB, will there be any change in max memory?
|
|
|
|
|
Under 32 bit versions of Windows, most of the time processes are limited to 2GB of memory. You can specify the /3GB[^] switch in the BOOT.INI file for Windows XP and Server 2003 to increase that limit to 3GB.
For a comparison of 32 and 64 bits, try reading here[^].
|
|
|
|
|
There may be two parts to your question:
1. the amount of logical memory (the virtual memory address space) you can get, that is 2GB (sometimes 3GB) on Win32 (i.e. 50 to 75% of the theoretical 2^32). and a lot more on Win64 (not sure, could be up to 4000 times as much). That is individual to each process.
2. the amount of physical memory (how much real memory your app is allowed to use) your process is actually getting; this depends on other processes are running and doing. The sum over all processes cannot exceeed the amount of physical memory; all virtual excess will be swapped to disk (and may cause significant delays).
So if you have two or more processes aiming for lots of memory, and your system is Win32 with 3GB of RAM, each process is likely to get no more than 1GB of physical memory. By minimizing one's main form, the other may get some more.
ADDED
Here[^] are some more facts for different Windows editions.
/ADDED
|
|
|
|
|
I am making a desktop application that is supposed to read data sent via Http and i have no idea how to begin it. I have made a distributed application that read data from TCP by using TcpClient classes but this is a different thing all together. Where do i even begin?
Wamuti: Any man can be an island, but islands to need water around them!
Edmund Burke: No one could make a greater mistake than he who did nothing because he could do only a little.
|
|
|
|
|
|
Have a look at the WebRequest[^] class or you could use WCF
I know the language. I've read a book. - _Madmatt
|
|
|
|
|
I am developing a control that displays a grid of pictures. I want to be able to select a picture by clicking on it, and change its appearance (eg. with a border) to show that it is selected. I have created:
1. a class derived from UserControl that contains a PictureBox for the picture
2. a class derived from FlowLayoutPanel to contain the pictures. This overrides CreateControlsInstance() to return an instance of...
3. a class derived from Control.ControlCollection to contain the picture objects. This overrides the Add(Control value) method in order to add a handler to the Click event of the control being added to the collection. It also contains the click event handler, and this is where my problem lies:-
It seems to me that the logical next step is for the click event handler to identify the control that has been clicked (no problem) and to call its Select() method, which will then alter the appearance of the control as required. The problem is that Control.Select() cannot be overridden. This raises two questions:
Q1: what does Control.Select() actually do?
Q2: how do I implement the change in appearance of the control. Obviously I could just set the UserControl.BorderStyle property from within the click handler in collection class, but surely an object should make its own decision as to how it looks when it is selected.
I can only assume that I've misunderstood something, or am doing it all the wrong way. Can anyone advise?
Dave
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
