Avoid enqueueing multiple jobs to fetch updates

Due to the way we rely on AlarmManager as a cron system to schedule new jobs to
fetch updates from event loop, when offline, it might happen that
multiple instances of the same job (FetchUpdateJob) are enqueued one
after the other, resulting in duplicated API calls when back online.

Here, we cancel any pre-existing "FetchUpdateJob" job before adding a
new one to the queue.

MAILAND-3074

app/src/main/java/ch/protonmail/android/api/segments/event/EventUpdaterService.java

@@ -25,6 +25,7 @@ import androidx.annotation.NonNull;
 import androidx.core.app.ProtonJobIntentService;
 
 import com.birbit.android.jobqueue.JobManager;
+import com.birbit.android.jobqueue.TagConstraint;
 
 import javax.inject.Inject;
 
@@ -70,6 +71,7 @@ public class EventUpdaterService extends ProtonJobIntentService {
 
 
     private void startService() {
+        mJobManager.cancelJobs(TagConstraint.ALL, FetchUpdatesJob.FETCH_UPDATE_JOB_TAG);
         mJobManager.addJob(new FetchUpdatesJob(eventManager));
         alarmReceiver.setAlarm(getApplicationContext());
     }

app/src/main/java/ch/protonmail/android/api/segments/event/FetchUpdatesJob.kt

@@ -26,11 +26,10 @@ import ch.protonmail.android.jobs.ProtonMailBaseJob
 import com.birbit.android.jobqueue.Params
 import timber.log.Timber
 import java.io.IOException
-import java.net.ConnectException
 import java.util.concurrent.TimeUnit
 
 class FetchUpdatesJob internal constructor(private val eventManager: EventManager) : ProtonMailBaseJob(
-    Params(Priority.HIGH).requireNetwork()
+    Params(Priority.HIGH).addTags(FETCH_UPDATE_JOB_TAG).requireNetwork()
 ) {
 
     constructor() : this(ProtonMailApplication.getApplication().eventManager)
@@ -58,4 +57,9 @@ class FetchUpdatesJob internal constructor(private val eventManager: EventManage
     }
 
     override fun onProtonCancel(cancelReason: Int, throwable: Throwable?) {}
+
+    companion object {
+
+        internal const val FETCH_UPDATE_JOB_TAG = "FetchUpdateJobTag"
+    }
 }