|
|
Hi everyone!
I have recently created a program when you click a button then a running (scrolling) string is appeared on the TextBox component....this line is "running" from right to left (like in Winamp) ...I used a timer object to alter string by replacing the leftmost character to the end of a string ...so it looked like it was moving ....but I don't understand:
every time i was using this code in Timer.Elapsed handler:
str = str.Substring(1, str.Length - 1) + str.Substring(1, 0);
txtMessage.Text = str;
so everytime the timer fires an event, a new string object is created (cause string are immutable in C#)
but in the System Process window it ate no memory at all!!!!
Does GarbageCollector work so fast?
|
|
|
|
|
Just how fast was this timer? Or just how slow was your PC? 
Seriously though, .NET applications get an allocation of memory when they start, unless you exceed that there would be no need to allocate more. The System Performanec window won't help you at all! For more details, see here[^] and here[^].You should never use standby on an elephant. It always crashes when you lift the ears. - Mark Wallace
C/C++ (I dont see a huge difference between them, and the 'benefits' of C++ are questionable, who needs inheritance when you have copy and paste) - fat_boy
|
|
|
|
|
When a new object is to be created and there is insufficient free memory inside your .NET app, the GC will try and free some, and when that were to fail, it would ask Windows for more memory. Every time the GC runs, it blocks your app, so it isn't really correct it is fast, all it does is steal the CPU to do its job. Which in most situations is quite all right.
What I tend to do in larger applications is have a Windows.Forms.Timer periodically send some statistical information to a Label, one useful item is Environment.WorkingSet (which I reduce by 1024*1024 to get it in megabytes). That often tells me how well the GC is doing, and gives me some clue when I am keeping on to too many objects (by accident).
|
|
|
|
|
Don't use Task Manager to see how much memory your app is using. What you actually seeing is how much memory the .NET CLR has reserved for your application, even if your app isn't using it all. Use Performance Monitor instead and the .NET Memory counters for a more accurate representation of your applications memory usage.
BTW: It's completely unneccessary to be creating and deleting so many strings. A better implementation would be to use an image of the string, then just move the image over 1 or pixels at a time for the scrolling effect.
|
|
|
|
|
Nick Reshetinsky wrote: Does GarbageCollector work so fast?
Yes, the garbage collector can work very fast for some types of memory (I am guessing for strings). Though some objects don't get garbage collected very fast (large objects, those with finalizers, and so on).
Strings are a special case, as they sometimes get interned (or, stored in a special way so it can be referenced by multiple objects). However, that is not typically done for runtime generated strings, such as those generated in the code you've posted above. So, they are not getting interned and the garbage collector is free to collect them.
Also keep in mind that you may only be generating a few unique strings (Maybe a few hundred), as I your scroller rotates (I imagine). So, even if your strings were interned, you still would not be eating up gobs of memory. And I'm sure your scroller isn't running too fast, which means that you'll not be consuming very much memory... even if the garbage collector ran every minute, you might not see a noticeable jump in memory in that timeframe.
|
|
|
|
|
strings get interned only if:
- they are literals, so the compiler interns them;
- the app explicitly calls string.Intern()
when they don't get interned, they don't get compared against each other either, so "you may only be generating a few unique strings" isn't really relevant; his code is generating millions of strings. And they get collected.
This[^] describes a string interning experiment showing it pays off when lots of compares are needed in a set with high multiplicity.
|
|
|
|
|
I know. You'll note that I said they don't typically get interned for runtime generated strings and that, if they (hypothetically) were (such as by using string.Intern), there were only a few unique strings so they would (hypothetically) not cause much of a memory footprint. By "unique", I meant the data would be unique, not the instances of that data ("A" and "A" are not unique, but they are two different instances). Was just pointing out some things the OP might want to think about (since the OP seems more interested in learning rather than solving a specific problem, as there isn't really a problem presented by the OP).
|
|
|
|
|
OK, we agree on the absence of a problem then. Time to move on.
|
|
|
|
|
I think, first, we should agree that we both agree there is no problem. And then we should agree that we both agree that we both agree there is no problem. 
|
|
|
|
|
I can't argue with that.
|
|
|
|
|
Hello, i'm trying to write a quick mock up system so my mother knows the colours of pictures when she cross stitches. I would like to import a picture overlay a grid i think it's 1mm boxes then be able to pick up the pixel or rgb colour. I've written the code to convert the colours to pattern chart colours but i was wondering if it was possible to overlay a grid then tell the colour of each square so i can easily convert it. I heard c# has a grid function as standard however i'm struggle to find it or use it
|
|
|
|
|
I'm unaware of any grid function, other than some Grid Controls which show tabular data.
showing a grid on top of an image can be accomplished easily with some code along these lines, using a Panel and assuming the image fits within the screen bounds:
public void myPanel_Paint(object sender, PaintEventArgs e) {
Graphics g=e.Graphics;
g.DrawImage(myImage, 0, 0);
int w=myImage.Width;
int h=myImage.Height;
int GRIDSTEP=20;
for (int x=0; x<w; x+=GRIDSTEP) g.DrawLine(Pens.Yellow, x, 0, x, h);
for (int y=0; y<h; y+=GRIDSTEP) g.DrawLine(Pens.Yellow, 0, y, w, y);
}
Alternatively, you could use an overload of DrawImage to get the image scaled if necessary;
note that I didn't use a PictureBox as I don't like them much.
|
|
|
|
|
hey thats the kinda thing i was looking for, quick question been a long time since i used forms, but when inserting the image is there a way of saying that the image = my image. I was under the impression it was something like backgroundimage = myimage(x,y,z). Also is there a way of picking the colour out to rgb?
|
|
|
|
|
I assume an image exists, maybe as a result of
Image myImage=Image.FromFile(filepath);
I avoid background images.
If you want to get a pixel at a known coordinate (x,y), just do Color color=myImage.GetPixel(x,y);
If you want to pick a pixel with the mouse, and assuming the image is shown unscaled, you could use a MouseMove handler:
private void myPanel_MouseMove(object sender, MouseEventArgs e) {
color=myImage.GetPixel(e.X, e.Y);
myLabel.Text="Color at ("+e.X+", "+e.Y+") = "+color.ToString();
}
If you need more, I suggest you start reading the MSDN documentation on the relevant classes and methods.
|
|
|
|
|
namespace WindowsFormsApplication1
{
public partial class Form1 : Form
{
Bitmap DrawArea;
Bitmap myImage;
public Form1()
{
DrawArea = new Bitmap(pictureBox1.Size.Width, pictureBox1.Size.Height);
pictureBox1.Image = DrawArea;
InitializeComponent();
}
private void Form1_Load(object sender, EventArgs e)
{
}
private void button1_Click(object sender, EventArgs e)
{
try
{
OpenFileDialog open = new OpenFileDialog();
open.Filter = "Image Files(*.jpg; *.jpeg; *.gif; *.bmp)|*.jpg; *.jpeg; *.gif; *.bmp";
if (open.ShowDialog() == DialogResult.OK)
{
pictureBox1.Image = new Bitmap(open.FileName);
Image myImage = Image.FromFile(open.FileName);
}
}
catch (Exception)
{
throw new ApplicationException("Failed loading image");
}
}
private void button2_Click(object sender, EventArgs e)
{
Graphics g = Graphics e;
g.DrawImage(myImage, 0, 0);
int w = myImage.Width;
int h = myImage.Height;
int GRIDSTEP = 20;
for (int x = 0; x < w; x += GRIDSTEP) g.DrawLine(Pens.Yellow, x, 0, x, h);
for (int y = 0; y < h; y += GRIDSTEP) g.DrawLine(Pens.Yellow, 0, y, w, y);
}
|
|
|
|
|
I strongly recommend you do all painting in a Paint handler. Maybe you should read this[^], it shows the normal way for painting (and animating) things.
|
|
|
|
|
This is one way to skin the cat:
1. Start with WPF, not winforms it'll be easier
2. Use a unform grid and create set up the elemt size to be square (height and width the same).
3. Bind the grid up to the converted color chart object you have so one element = one square
4. Use a template in each square, you need two templates:
a. A color template that fills in the square color with a SolidColorBrush
b. A color chart template that renders the colour as a symbol.
5. Apply a TemplateChooser to swap between the two.
Sorry this is a little light on detail, it's been a few months since I've handled WPF....Dalek Dave: There are many words that some find offensive, Homosexuality, Alcoholism, Religion, Visual Basic, Manchester United, Butter.
Pete o'Hanlon: If it wasn't insulting tools, I'd say you were dumber than a bag of spanners.
|
|
|
|
|
Hi everyone.
Quick question: is reference comparison atomic? For example:
Object a, b;
....
if (a == b)
....
....
Or should I use the following?
Interlocked.ReferenceEquals(a, b);
I imagine if the method exists it's because a==b wouldn't be atomic, which I guess makes sense because we're reading two references, and changes could occur in between the reads. Am I right?
Sorry about the trivial question, but I want to make sure and my brain isn't working very well at the moment (too much coding...).
Thanks modified on Tuesday, March 16, 2010 11:00 AM
|
|
|
|
|
What your code is doing is:
- read a
- read b
- perform the comparison
Any other thread may write between those two reads. Being "atomic" doesn't apply to the comparison as it only deals with local values.
There is no Interlocked.ReferenceEquals, you're just calling the inherited Object.ReferenceEquals.
AFAIK there is no way to read two references atomically (so that no other thread can write between the two reads).
|
|
|
|
|
Daniel Grunwald wrote: What your code is doing is:
- read a
- read b
- perform the comparison
Any other thread may write between those two reads. Being "atomic" doesn't apply to the comparison as it only deals with local values.
Exactly.
Daniel Grunwald wrote: There is no Interlocked.ReferenceEquals, you're just calling the inherited Object.ReferenceEquals.
True! I hadn't realised Interlocked didn't override the method.
Daniel Grunwald wrote: AFAIK there is no way to read two references atomically (so that no other thread can write between the two reads).
Well, surely with a lock.
Thanks for your answer.
|
|
|
|
|
Yes, you are right.
For reference types, using a==b will comapare references and not the values that we generally mean to have.
a.Equals(b) is a correct way to compare two reference types. This will comapre values and not memory address.
|
|
|
|
|
That's not what I meant. The issue wasn't whether I was comparing references or values, but whether it was atomic or not. It's not atomic.
Thanks anyway.
|
|
|
|
|
even if it were atomic, what good would it be, as the result of the comparison is to be used for some purpose while it also is volatile. What you probably want is a lock over some code of which the comparison is only a fraction.
lock(myLock) {
if (a==b) doSomething();
}
|
|
|
|
|
Yes, I know, but thanks. I was just wondering about it, I'm not actually using it anywhere.
It's pretty obvious that it isn't atomic, though, it's just that my brain isn't working properly and I wanted to make sure I got it right.
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
