Is this simplification of GTK+ code correct? - c

Question

I found the following GTK+3 code in Zetcode. It creates an animation using the cairo library while displaying an image:
#include <cairo.h>
#include <gtk/gtk.h>
/* compile with
*
* gcc spectrum.c -o spectrum `pkg-config --cflags --libs gtk+-3.0`
*
* */
static void do_drawing(cairo_t *);
struct {
gboolean timer;
cairo_surface_t *image;
cairo_surface_t *surface;
gint img_width;
gint img_height;
} glob;
static void init_vars()
{
glob.image = cairo_image_surface_create_from_png("beckov.png");
glob.img_width = cairo_image_surface_get_width(glob.image);
glob.img_height = cairo_image_surface_get_height(glob.image);
glob.surface = cairo_image_surface_create(CAIRO_FORMAT_ARGB32,
glob.img_width, glob.img_height);
glob.timer = TRUE;
}
static gboolean on_draw_event(GtkWidget *widget, cairo_t *cr,
gpointer user_data)
{
do_drawing(cr);
return FALSE;
}
static void do_drawing(cairo_t *cr)
{
cairo_t *ic;
static gint count = 0;
ic = cairo_create(glob.surface);
gint i, j;
for (i = 0; i <= glob.img_height; i+=7) {
for (j = 0 ; j < count; j++) {
cairo_move_to(ic, 0, i+j);
cairo_line_to(ic, glob.img_width, i+j);
}
}
count++;
if (count == 8) glob.timer = FALSE;
cairo_set_source_surface(cr, glob.image, 10, 10);
cairo_mask_surface(cr, glob.surface, 10, 10);
cairo_stroke(ic);
cairo_destroy(ic);
}
static gboolean time_handler(GtkWidget *widget)
{
if (!glob.timer) return FALSE;
gtk_widget_queue_draw(widget);
return TRUE;
}
int main(int argc, char *argv[])
{
GtkWidget *window;
GtkWidget *darea;
init_vars();
gtk_init(&argc, &argv);
window = gtk_window_new(GTK_WINDOW_TOPLEVEL);
darea = gtk_drawing_area_new();
gtk_container_add(GTK_CONTAINER (window), darea);
g_signal_connect(G_OBJECT(darea), "draw",
G_CALLBACK(on_draw_event), NULL);
g_signal_connect(G_OBJECT(window), "destroy",
G_CALLBACK(gtk_main_quit), NULL);
gtk_window_set_position(GTK_WINDOW(window), GTK_WIN_POS_CENTER);
gtk_window_set_default_size(GTK_WINDOW(window), 325, 250);
gtk_window_set_title(GTK_WINDOW(window), "Spectrum");
g_timeout_add(400, (GSourceFunc) time_handler, (gpointer) window);
gtk_widget_show_all(window);
gtk_main();
cairo_surface_destroy(glob.image);
cairo_surface_destroy(glob.surface);
return 0;
}
I achieve exatly the same result if I remove the do_drawing() function and move its code to the on_draw_event(), like:
#include <cairo.h>
#include <gtk/gtk.h>
/* compile with
*
* gcc spectrum.c -o spectrum `pkg-config --cflags --libs gtk+-3.0`
*
* */
struct {
gboolean timer;
cairo_surface_t *image;
cairo_surface_t *surface;
gint img_width;
gint img_height;
} glob;
static void init_vars()
{
glob.image = cairo_image_surface_create_from_png("beckov.png");
glob.img_width = cairo_image_surface_get_width(glob.image);
glob.img_height = cairo_image_surface_get_height(glob.image);
glob.surface = cairo_image_surface_create(CAIRO_FORMAT_ARGB32,
glob.img_width, glob.img_height);
glob.timer = TRUE;
}
static gboolean on_draw_event(GtkWidget *widget, cairo_t *cr,
gpointer user_data)
{
cairo_t *ic;
static gint count = 0;
ic = cairo_create(glob.surface);
gint i, j;
for (i = 0; i <= glob.img_height; i+=7) {
for (j = 0 ; j < count; j++) {
cairo_move_to(ic, 0, i+j);
cairo_line_to(ic, glob.img_width, i+j);
}
}
count++;
if (count == 8) glob.timer = FALSE;
cairo_set_source_surface(cr, glob.image, 10, 10);
cairo_mask_surface(cr, glob.surface, 10, 10);
cairo_stroke(ic);
cairo_destroy(ic);
return FALSE;
}
static gboolean time_handler(GtkWidget *widget)
{
if (!glob.timer) return FALSE;
gtk_widget_queue_draw(widget);
return TRUE;
}
int main(int argc, char *argv[])
{
GtkWidget *window;
GtkWidget *darea;
init_vars();
gtk_init(&argc, &argv);
window = gtk_window_new(GTK_WINDOW_TOPLEVEL);
darea = gtk_drawing_area_new();
gtk_container_add(GTK_CONTAINER (window), darea);
g_signal_connect(G_OBJECT(darea), "draw",
G_CALLBACK(on_draw_event), NULL);
g_signal_connect(G_OBJECT(window), "destroy",
G_CALLBACK(gtk_main_quit), NULL);
gtk_window_set_position(GTK_WINDOW(window), GTK_WIN_POS_CENTER);
gtk_window_set_default_size(GTK_WINDOW(window), 325, 250);
gtk_window_set_title(GTK_WINDOW(window), "Spectrum");
g_timeout_add(400, (GSourceFunc) time_handler, (gpointer) window);
gtk_widget_show_all(window);
gtk_main();
cairo_surface_destroy(glob.image);
cairo_surface_destroy(glob.surface);
return 0;
}
So... I wonder... Am I missing something here (loss of generality)?
Or was the call to do_drawing() in function on_draw_event() of the original code redundant?
Thanks

Answer 1

Yes. It's correct on compiler's side.
But it's a good practice for functions to solve exactly one task. do_drawing is busy with drawing lines and pixels, on_draw_event is busy with processing events. Probably in this code snippet there is no real reason to make a separate function, but usually on_draw_event would be much more complicated.