Re: timeout function not called

[James Scott Jr <skoona verizon net>]

On Thu, 2007-10-04 at 13:27 -0700, v_wishful sandiego com wrote:

> Thanks James for giving me my first experience with Thread programming!
> I tried the code you sent me, but unfortunately, I still can't get the  
> progress_update
> function to run, looks like it never gets called. Any idea why this  
> might be happening?
> growPart still runs.
> 
> Thanks,
> 
> Vicki
> 


The code i offered is really high level - almost sudo code.  I prefer to
evaluate a fuller example, if you care to post the whole program - or
email it to me and I will email it back.

Things I would be interested in would be as follows;
1. did you do away with the fork() and wait() calls?
2. did you compile it with something like this - with gthread included?
    "# gcc -Wall `pkg-config --libs --cflags gtk+-2.0 glib-2.0
gthread-2.0` program.c "
3. Did you call g_thread_init() in main(). ?
    /*
     * Initialize GLib thread support and GTK
     */
    g_type_init ();
    g_thread_init (NULL);
    gdk_threads_init ();
    gtk_init (&argc, &argv);
    /* create app window */
    /* create progress bar dialog */

    /*
     * enter the GTK main loop
     */
    gdk_threads_enter ();
    gtk_main ();
    gdk_flush ();
    gdk_threads_leave ();
    return (0);

4. How long does growPart() run? longer than 100ms
    the g_timeout_add() is only for 50ms, and it waits 50ms before
making the first call kinda short, try 250 or 400

5. Assuming your created a dialog box for the progress bar and thats is
is visible it should have ran.  Stick a g_debug("update"); in the
progress_update() routine after the call to gtk_progress_bar_pulse()

6. I'm thinking the most likely cause is how the program was compiled.
send me some code to look at!

James,

> 
> [Hide Quoted Text]
> GTK program before.  The ret=wait(null) causes execution to stop at that
> 
> point and wait for growPart() to finish.  A way to get around the  
> logistics of fork() is to use threads directly.  Consider the following;
> 
> /* globals */
> gboolean  b_pulse_control = FALSE
> GThread  *grow_thread_id = NULL;
> gint   global_ret = 0;
> 
> 
> b_pulse_control = TRUE;
> timer = g_timeout_add(50, (GSourceFunc) progress_update,
> (gpointer)progressbar);
> 
> grow_thread_id = g_thread_create ( growPart_thread, &b_pulse_control,
> TRUE, NULL);
> 
> 
> static gboolean progress_update (gpointer progressbar)
> {
>       if (b_pulse_control)
>      {
>           gtk_progress_bar_pulse(GTK_PROGRESS_BAR(progressbar));
>           return true;
>      } else {
>          global_ret = GPOINTER_TO_INT( g_thread_join (grow_thread)  );
>          return FALSE;
>      }
> 
> }
> 
> static gpointer growPart_thread(gpointer b_pulse)
> {
> 
>       /* do work */
>       growPart();
> 
>       *b_pulse = FALSE;
>       g_thread_exit( GINT_TO_POINTER( 1 ) );
> 
> }
> 
> I would do the passing of control values via a structure to avoid  the  
> use of globals,
> but this should work for you.
> 
> 
> ------------------------------------------
> Invent your own San Diego at sandiego.com!
> 
> 
> _______________________________________________
> gtk-app-devel-list mailing list
> gtk-app-devel-list gnome org
> http://mail.gnome.org/mailman/listinfo/gtk-app-devel-list