ffplay: eliminate possible resize of the window for first frame

[ami_stuff]

    if (is_full_screen && fs_screen_width) {
        w = fs_screen_width;
        h = fs_screen_height;
        fprintf(stderr,"1\n");
    } else if vp->width {
        calculate_display_rect(&rect, 0, 0, INT_MAX, vp->height, vp);
        w = rect.w;
        h = rect.h;
        fprintf(stderr,"2\n");
    } else if (!is_full_screen && screen_width && screen_height) {
        w = screen_width;
        h = screen_height;
        fprintf(stderr,"3\n");
    } else {
        w = 640;
        h = 480;
        fprintf(stderr,"4\n");
    }
    if (screen && is->width == screen->w && screen->w == w
       && is->height== screen->h && screen->h == h && !force_set_video_mode)
        return 0;

sometimes when I play a video at the beginning ffplay resizes window from 640x480 to actual
resolution of the video because code enters for first frame to "else" statement
(the rest of the 4 queued frames are displayed with correct size under (vp->width) statement),
so if possible I would like to see this unnecessary resize (flickering) for first frame removed.

this seems to remove (or reduce) the problem, but the first frame is not displayed at all
(so it fails for images)

display:
            /* display picture */
-            if (!display_disable && is->show_mode == SHOW_MODE_VIDEO)
-                video_display(is);

            pictq_next_picture(is);


+            if (!display_disable && is->show_mode == SHOW_MODE_VIDEO)
+                video_display(is);
            

            if (is->step && !is->paused)
                stream_toggle_pause(is);

No FFplay's output since it happens randomly with various video files
(sometimes easier to reproduce when -lowres option is enabled)

[ami_stuff]

It also happens with image as an input when you press "f" to enter fullscreen mode and go back again pressing "f" to windowed mode.

[cus]

Does the attached patch help for the original problem?

[ami_stuff]

Replying to cus:

Does the attached patch help for the original problem?

yes your patch fixes the problem, thanks

[cehoyos]

Is this problem still reproducible?

[ami_stuff]

Replying to cehoyos:

Is this problem still reproducible?

I belive it's fixed by 4fd6e5af1e334875eca4f803bbcfac9219b0524a.