Description of problem:
I was dragging a window from an external 4K monitor to the laptop's 4K display and gnome-shell just froze up.

Looking at the back trace, the code is failing to verify that the monitor lookup function actually found a monitor.
This codebase obviously needs a static analyzer to verify that functions which can return NULL are actually checked
to see if they did return NULL.

The code also seems very weird in that it is freely mixing and matching ints and floats for pixels.
Why all the useless conversion? Pick one!

Version-Release number of selected component:
gnome-shell-3.24.2-1.fc26

Additional info:
reporter:       libreport-2.9.1
backtrace_rating: 4
cmdline:        /usr/bin/gnome-shell
crash_function: update_move
executable:     /usr/bin/gnome-shell
journald_cursor: s=f26a41dd0046472c99b00b7ae7f2971e;i=146688;b=b1ec753824bd400a9de5ac6ac484f740;m=afe466c63;t=552dff0f51011;x=eabb10a6fc5703d3
kernel:         4.11.6-301.fc26.x86_64
rootdir:        /
runlevel:       N 5
type:           CCpp
uid:            1001

Truncated backtrace:
Thread no. 1 (10 frames)
 #0 update_move at core/window.c:5660
 #1 meta_window_handle_mouse_grab_op_event at core/window.c:6091
 #2 meta_display_handle_event at core/events.c:304
 #3 event_callback at core/events.c:417
 #4 _clutter_event_process_filters at clutter-event.c:1913
 #5 emit_pointer_event at clutter-main.c:2011
 #6 _clutter_process_event_details at clutter-main.c:2327
 #7 _clutter_process_event at clutter-main.c:2548
 #8 _clutter_stage_process_queued_events at clutter-stage.c:1026
 #9 master_clock_process_events at clutter-master-clock-default.c:364