Fix various custom script processing issues

The following issues were addressed. Note that these are edge cases.

* If an empty `/cron-scripts` directory was mounted in the container, an
  error would occur.
* If files in the `/cron-scripts` directory had spaces in the name, this
  would cause issues.
* stdin was held open when running `docker exec` which would cause hangs
  and other issues. The `exec` command is now run non-interactively.
* Custom scripts that failed would interrupt processing of scripts after
  it, if any.

README.md

@@ -213,6 +213,8 @@ above. However, it is explicitly specified for example purposes.
 * All cron task shell scripts run at the same interval defined by `NEXTCLOUD_CRON_MINUTE_INTERVAL`.
 * Modification of your own shell scripts on the host do not require that you restart/recreate the
   container (only when volume mappings change in the YAML file).
+* If a custom script in the `/cron-scripts` directory fails, it will not impede the processing of
+  other scripts in the directory.
 
 ## Debugging
 

scripts/cron-tasks.sh

@@ -21,9 +21,10 @@ echo "> Nextcloud Container ID: ${containerId}"
 
 run_scripts_in_dir() {
     cd "$1"
-    for script in *.sh; do
-        echo "> Running Script: $script"
-        nextcloud_exec "$containerId" "$(< $script)"
+    find . -type f -name '*.sh' -print0 |
+    while IFS= read -r -d '' file; do
+        echo "> Running Script: $file"
+        nextcloud_exec "$containerId" "$(cat $file)" || continue
     done
 }
 

scripts/nextcloud-exec.sh

@@ -9,7 +9,7 @@ nextcloud_exec_no_shell() {
         exec_user="--user $NEXTCLOUD_EXEC_USER"
     fi
 
-    docker exec $exec_user -i "$containerId" "$@"
+    docker exec $exec_user "$containerId" "$@"
 }
 
 nextcloud_exec() {