shell-recorder: Use clutter-computed sizes and scaling for the capture

Get from clutter the capture sizes and scale and don't mind
about doing any rounding here, as it might be different
from the one done at clutter level (causing mismatch and
not-working videos). Delegate this to clutter, and forget
about the internal details.

These values are then used to composte the image and set the video caps.

https://bugzilla.gnome.org/show_bug.cgi?id=7650111

src/shell-util.h

@@ -53,8 +53,9 @@ cairo_surface_t * shell_util_composite_capture_images (ClutterCapture  *captures
                                                        int              n_captures,
                                                        int              x,
                                                        int              y,
-                                                       int              width,
-                                                       int              height);
+                                                       int              target_width,
+                                                       int              target_height,
+                                                       float            target_scale);
 
 void shell_util_check_cloexec_fds (void);