Run prelock and postlock synchronously to avoid race condition

As reported in issue #285, running postlock in a subshell and immediately
returning from lockselect() has a high chance of hitting `exit 0` before
xset and dunstctl get a chance to execute.

Run both prelock and postlock synchronously within lockselect().

betterlockscreen

@@ -278,7 +278,7 @@ postlock() {
 lockselect() {
 
     echof act "Running prelock..."
-    prelock &
+    prelock
 
     case "$1" in
         dim) if [ -f "$CUR_L_DIM" ]; then lock "$CUR_L_DIM"; else failsafe; fi ;;
@@ -291,7 +291,7 @@ lockselect() {
     esac
 
     echof act "Running postlock..."
-    postlock &
+    postlock
 }
 
 # calculate adjustments for hidpi displays