[Gtk-sharp-list] Gtk.Application.Invoke + anonymous delegates == lots of uints?
Rafael Teixeira
monoman at gmail.com
Sat Dec 9 08:22:47 EST 2006
Answered too fast, sorry!!!
Well looking at your code it shows the kind of problem I'm talking about:
int ctr = 0;
void test()
{
while(true)
{
Gtk.Application.Invoke( delegate {
ctr++;
Label1.Text = ctr.ToString();
});
Thread.Sleep(10);
}
}
You are autoboxing ctr, in every call to ToString(). It is better to
call some static method of Convert that accepts ints as a parameter to
be formatted, like
Label1.Text = Convert.ToString(ctr);
:)
:)
On 12/9/06, Rafael Teixeira <monoman at gmail.com> wrote:
> It is worse: If heap-buddy is showing them up and GC collects them, it
> means they are boxed uints (probaly autoboxed ones), what means a lot
> of overhead creating/managing objects around the value type, beyond
> the memory comsuption.
>
> Definitively we need to find the culprit, probably someone passing a
> uint to some method that wants an object. Can you help to better
> assess where it could be?
>
> Thanks,
>
> On 12/8/06, 32bitkid <jim at 32bitkid.com> wrote:
> > I seem to be getting a weird allocation of uints when using
> > Gtk.Application.Invoke If i run the following code through heap-buddy:
> > using System;
> > using Gtk;
> > using System.Threading;
> >
> > public class MainWindow : Gtk.Window
> > {
> > Gtk.Label Label1;
> >
> > public MainWindow() : base ("Test")
> > {
> > this.DeleteEvent += new DeleteEventHandler(OnDeleteEvent);
> >
> > Label1 = new Gtk.Label("");
> > this.Add(Label1);
> > this.ShowAll();
> >
> > Thread x = new Thread(new ThreadStart(test));
> > x.IsBackground = true;
> > x.Start();
> > }
> >
> > int ctr = 0;
> >
> > void test()
> > {
> > while(true)
> > {
> > Gtk.Application.Invoke( delegate {
> > ctr++;
> > Label1.Text = ctr.ToString();
> > });
> > Thread.Sleep(10);
> > }
> > }
> >
> > protected void OnDeleteEvent( object sender, DeleteEventArgs e)
> > {
> > Application.Quit();
> > e.RetVal = true;
> > }
> > }
> >
> > I'll get a lot of uints allocated on the heap. Most seem to be cleaned up
> > by the garbage collection, but still it doesnt seem quite right. Here is the
> > output from heap-buddy, after running for about 5 seconds: $ heap-buddy
> > types 10 sorted by count
> >
> > Type # Total AvSz AvAge BT#
> > uint 557439 6.4M 12.0 0.0 18337
> > string 18558 863k 47.6 0.6 18418
> > byte[] 18400 357k 19.9 0.0 18400
> > GLib.TimeoutHandler 18392 718k 40.0 0.6 4
> > GLib.Timeout/TimeoutProxy 9196 143k 16.0 0.8 2
> > System.EventHandler 9196 359k 40.0 0.6 2
> > Gtk.Application/InvokeCB 9195 179k 20.0 0.6 1
> > object[] 9186 717k 80.0 0.0 9186
> > System.Collections.ArrayList 9182 179k 20.0 0.0 9182
> > System.Text.StringBuilder 9179 215k 24.0 0.0 9179
> >
> > Is this a bug, or am I just doing something very wrong. That just seems
> > like a LOT of uints. This is basically a stub from another Gtk# program I'm
> > working on that is fairly thread intestive. I don't remember seeing this
> > kind of behavior before 1.2. I am running mono 1.2.2, on Fedora Core 5. Any
> > help would be greatly appriciated. Thanks
> > ________________________________
> > View this message in context: Gtk.Application.Invoke + anonymous delegates
> > == lots of uints?
> > Sent from the Mono - Gtk# mailing list archive at Nabble.com.
> >
> > _______________________________________________
> > Gtk-sharp-list maillist - Gtk-sharp-list at lists.ximian.com
> > http://lists.ximian.com/mailman/listinfo/gtk-sharp-list
> >
> >
> >
>
>
> --
> Rafael "Monoman" Teixeira
> ---------------------------------------
> "The reasonable man adapts himself to the world; the unreasonable one
> persists in trying to adapt the world to himself. Therefore all
> progress depends on the unreasonable man." George Bernard Shaw
>
--
Rafael "Monoman" Teixeira
---------------------------------------
"The reasonable man adapts himself to the world; the unreasonable one
persists in trying to adapt the world to himself. Therefore all
progress depends on the unreasonable man." George Bernard Shaw
More information about the Gtk-sharp-list
mailing list