Develop Reference

Android new SyncRequest is ignored when another SyncRequest is being handled

I am making an app in which I have a local database and I'm using a SyncAdapter to sync this local database with the server, I don't have much experience with a SyncAdapter and I cannot seem to figure something out. So far I've implemented the "Run the sync adapter when content provider data changes" section from the Android documentation (https://developer.android.com/training/sync-adapters/running-sync-adapter), and initially it worked great but I started to notice something. When calling requestSync from inside the ContentObserver a new SyncRequest is queued by Android (I guess?) and then a little later executed but when my onPerformSync method from my SyncAdapter is executing, I find that if I make a new SyncRequest, this SyncRequest is completely ignored and not even executed later on. This is kinda annoying because when I update my database while my database is being synced, then it could happen that my updates do not reach my server (because the updates occured while the old data was being synced already). I cannot find much information about this behaviour, is this normal behaviour and if so how could I avoid this (without needing to write an entire queing system by myself)?
Here is the code from my ContentObserver (AppLogger is some custom logging system I made):
ContentObserver observer = new ContentObserver(null) {
#Override
public void onChange(boolean selfChange) {
super.onChange(selfChange, null);
}
#Override
public void onChange(boolean selfChange, #Nullable Uri uri) {
AppLogger.log(context, "AppBroadCastReceiver", "Requesting a sync for the Datamanager from the observer...");
ContentResolver.requestSync(mAccount, ApplicationProvider.AUTHORITY, new Bundle());
}
};
The onPerformSync method I used to test this behaviour:
#Override
public void onPerformSync(Account account, Bundle extras, String authority, ContentProviderClient provider, SyncResult syncResult) {
AppLogger.log(context, "DataManagerSyncAdapter", "Starting a sync attempt");
try{
Timer timer = new Timer();
CountDownLatch latch = new CountDownLatch(1);
timer.schedule(new TimerTask() {
#Override
public void run() {
latch.countDown();
}
}, 10000);
latch.await();
} catch(InterruptedException e){}
AppLogger.log(context, "DataManagerSyncAdapter", "Finishing a sync attempt");
}
And then the syncadapter xml:
<?xml version="1.0" encoding="utf-8"?>
<sync-adapter
xmlns:android="http://schemas.android.com/apk/res/android"
android:contentAuthority="com.example.getrequest.Providers.application_provider"
android:accountType="example.com"
android:userVisible="true"
android:supportsUploading="false"
android:allowParallelSyncs="false"
android:isAlwaysSyncable="true"/>
And I got as result in the log (when trying to test this behaviour):
AppBroadCastReceiver: Requesting a sync for the Datamanager from the observer...
DataManagerSyncAdapter: Starting a sync attempt
AppBroadCastReceiver: Requesting a sync for the Datamanager from the observer...
DataManagerSyncAdapter: Finishing a sync attempt
And then onPerformSync is never run again (atleast not in the next 20 minutes, after that I lost my patience). I also noticed setting android:supportsUploading="true" kinda solved my problem but then a ton of useless SyncRequests are handled which I don't even ask for (like one every minute almost).
I've also thought about maybe blocking access to the database until my SyncRequest is completely done but is this common practice? If I want to update multiple tables in the server based on multiple tables in my local database then isn't it better to only lock the database per table instead of locking everything until the SyncRequest is completed and added to this, does this really solve anything? At which point in the onPerformSync should I then unlock my database again? It looks to me that unlocking the database in the onPerformSync could always result in a database call being executed while the onPerformSync is still busy (even if the possibility is very small)? Any help or information about this would be greatly appreciated!
Edit:
When digging through the source code of the SyncManager (https://android.googlesource.com/platform/frameworks/base/+/master/services/core/java/com/android/server/content/SyncManager.java) I came across this:
// Check currently running syncs
for (ActiveSyncContext asc: mActiveSyncContexts) {
if (asc.mSyncOperation.key.equals(syncOperation.key)) {
if (isLoggable) {
Log.v(TAG, "Duplicate sync is already running. Not scheduling "
+ syncOperation);
}
return;
}
}
So I guess this is expected behaviour, which from my point of view does not make any sense at all but maybe I don't have enough experience with these sort of things. So how should I ensure that data updated during a onPerformSync is still getting updated (without writing tons of code myself) or how can I ensure that data is not being updated while onPerformSync is busy?

Save event type logs

We want to add a reporting feature to our existing application.
For this purpose we are sending Events in JSON via HTTPS to a server application.
We need to remember Event-objects that could not be send to the server (No internet, server not reachable...). We are considering to store the events in a SQLite database and discard all Events that are older than 24 hours to prevent flooding our storage.
Another option would be to write the JSON-objects to a file and concat each new event when it could not be send to the server. The problem with this solution is, that it would be hard for us to discard logs older than 24 hours.
We store the event sin a table with the columns:
| id | json | created_at |
Can anyone recommend best practices for this use case?
Currently we tend to use the sqlite solution but we are wondering if there are any caveats that we are not aware of.

If you don't mind using third-party lib i can recommend android-priority-jobqueue. You can easily achieve what you are trying to do. You can always create job and it will handle itself. You can set if it needs network, if it is persistent (saved into DB when no network) and even you can customize your own retry logic.
Here's little example.
public class PostTweetJob extends Job {
public static final int PRIORITY = 1;
private String text;
public PostTweetJob(String text) {
// This job requires network connectivity,
// and should be persisted in case the application exits before job is completed.
super(new Params(PRIORITY).requireNetwork().persist());
}
#Override
public void onAdded() {
// Job has been saved to disk.
// This is a good place to dispatch a UI event to indicate the job will eventually run.
}
#Override
public void onRun() throws Throwable {
// yours code here
}
#Override
protected RetryConstraint shouldReRunOnThrowable(Throwable throwable, int runCount,
int maxRunCount) {
// An error occurred in onRun.
return RetryConstraint.createExponentialBackoff(runCount, 1000);
}
}
And you call it like this.
jobManager.addJobInBackground(new PostTweetJob("It works"));

use JobService(Android 5+ - lollipop and above) and AlarmManager (for android sdk<21 - pre lollipop) with this solution u can schedule any task and it would be performed. JobService was developed rxactely for tjis purposes(schedule and perform different tasks) maybe you can try JobIntentService it is would work on kitkat(android 4+) devices
P.S.
In that case you didnt need any third party libs and other dependrncies like firebase/google play services(like for FirebaseDispatcher)

Is using Executors.newSingleThreadExecutor recommended in app widget

Currently, in my main app, I am sending multiple texts to status bar object.
My status bar object, is going to display multiple texts sequentially, with sleep time of N seconds for each display interval.
Here's my implementation in my main app.
public synchronized void setNextText(final CharSequence text) {
if (executor == null) {
executor = Executors.newSingleThreadExecutor();
}
executor.execute(new Runnable() {
#Override
public void run() {
Fragment fragment = getTargetFragment();
if (fragment instanceof OnStatusBarUpdaterListener) {
((OnStatusBarUpdaterListener)fragment).setNextText(text);
try {
// Allow 1 seconds for every text.
Thread.sleep(Constants.STATUS_BAR_UPDATER_TIME);
} catch (InterruptedException ex) {
Log.e(TAG, "", ex);
}
}
}
});
}
Now, I would like to have the same behavior in app widget. I was wondering, is using Executor being recommended in app widget environment? If not, what class I should use to achieve the similar objective?
I do have experience in using HandlerThread + AlarmManager in app widget. It works good so far. However, the operation done by the runnable is one time. It doesn't sleep and wait.
The following is the code which I use to update stock price in fixed interval.
// This code is trigger by AlarmManager periodically.
if (holder.updateStockPriceHandlerThread == null) {
holder.updateStockPriceHandlerThread = new HandlerThread("updateStockPriceHandlerThread" + appWidgetId);
holder.updateStockPriceHandlerThread.start();
holder.updateStockPriceWorkerQueue = new Handler(holder.updateStockPriceHandlerThread.getLooper());
holder.updateStockPriceWorkerQueue.post(getUpdateStockPriceRunnable(...
}
However, I have a feeling that, for use case "display multiple texts sequentially, with sleep time of N seconds for each display interval", AlarmManager might not be a good solution. Imagine I have 100 texts. Having to set 100 alarms for 100 texts doesn't sound good...

An AppWidgetProvider is a subclass of BroadcastReceiver. Once your callback (e.g., onUpdate()) returns, your process can be terminated at any point.
If that is not a problem — if you fail to finish the semi-animation that you are doing, that's OK — using an Executor from onUpdate() could work.
If you want to make sure that the text changes go to completion, delegate the app widget updating to a Service, where you use your Executor. Call stopSelf() on the Service when you are done, so it can go away and not artificially keep your process around.

Well the singleThread instance work creates an Executor that uses a single worker thread. meaning only thread to process your operation. But in you case use at least two. Your operations sounds expensive.
To conclude your question stick with the executor service as it thread safe.

Android background threading

I'm making image processor app. I need to scan the phone for pictures and list them with their number of pixels. So that's gonna be a a large impact on performance and as I understood, I need to make it work on background thread.
So my question is, what is the best approach for this? I understand that IntentService may be the best solution, but I'm not sure how I will implement progress bar with it, and I need to return Picture objects and later update the UI on shuffle button. I'm doing update with Glide library so that's gonna go smooth.
Reading about Asynctasks, I stumbled about comments how it's bad and leads to leaks in memory and should avoid using it. rXJava is too complicated at the moment.
This is my code:
Main activity:
#OnClick(R.id.shuffle)
public void shuffleList() {
Collections.shuffle(listOfImageFiles);
recyclerViewAdapter = new PictureRecycleViewAdapter(listOfImageFiles, this);
recyclerView.swapAdapter(recyclerViewAdapter, false);
recyclerViewAdapter.notifyDataSetChanged();
}
#OnClick(R.id.scan)
public void processImages() {
//progress bar
listOfPictures = new ArrayList<>();
//Gets data from default camera roll directory. Note that some of the phone companies have different file paths. So instead of hardcoding string paths, I used this instead.
String path = Environment.getExternalStoragePublicDirectory(Environment.DIRECTORY_DCIM).getPath();
File filePath = new File(path);
listOfImageFiles = scanPhotos(filePath);
// async?
for (File file : listOfImageFiles
) {
Bitmap bitmap = BitmapFactory.decodeFile(file.getPath());
//int is sufficient for most today's pixels. long would be overkill - 4 vs 8 bytes
int pixels = bitmap.getHeight() * bitmap.getWidth();
listOfPictures.add(new Picture(file.getPath(), pixels));
}
}
public List<File> scanPhotos(File directory) {
List<File> listOfPictures = new ArrayList<>();
try {
File[] files = directory.listFiles();
for (File file : files
) {
if (file.isDirectory() && !file.isHidden()) {
listOfPictures.addAll(scanPhotos(file));
} else {
if (file.getName().endsWith(".jpg") || file.getName().endsWith(".jpeg") || file.getName().endsWith(".png")) {
listOfPictures.add(file);
}
}
}
} catch (Exception e) {
Log.e(e.getMessage(), e.getMessage());
}
return listOfPictures;
}

IntentService
IntentService is definitely a valid approach. You can use Broadcasts to return your result to another component of the app, be it Activity or another Service, for example:
Start the IntentService - if you need some parameters, place them in the Extras of the service intent.
Your IntentService runs on the background thread until the computation is finished.
Upon finishing, send a broadcast with computation result placed in intent extras.
In your activity, register a BroadcastReceiver that will listen for your computation result broadcast.
Upon getting the broadcast in your Activity, retrieve the computation result from intent extras.
You might also implement broadcasts received by your Service for things like cancellation of the computation or updating the parameters.
One of the advantages of IntentService is that you can easily integrate it with the JobScheduler API to defer execution until certain system conditions are met.
Alternatives
You can use a bus library, such as https://github.com/greenrobot/EventBus to communicate between Activity and Service - the only problem is, EventBus won't work with remote services (running in a separate process).
Like you've mentioned, using RxJava with IO and computation schedulers is also a good idea.
AsyncTask is fine as long as you not tie it with a hard reference to an activity - don't implement it as an inner class of Activity and if you want to communicate the result back, do it through a WeakReference<T>

AsyncTask is fine, you just need to be careful with its implementation.
However, for longer running tasks there are better options. IntentService is a good option.
When it comes to a responsive UI when using an IntentService you could add two things.
Notifications
Create an ongoing notification that indicates that your App is working on something. This lets users know that their CPU cycles are being eaten by something in the background and they are less likely(?) to be confused and cranky about their device running slower.
Additionally, it gives your App more of an allowance for staying alive when Android is looking for background Apps to kill to release memory.
EventBus
You can make UI reporting extremely simple by using an EventBus library. I am personally a fan of greenbot/EventBus, but there are others.
Example
In your Activity:
#Subscribe(threadMode = ThreadMode.MAIN)
public void onProgressEvent(ProgressEvent event) {
mProgressBar.setProgress(event.value);
}
In your IntentService:
EventBus.getDefault().post(new ProgressEvent(5000));

Android dynamic UI update from AsynchTask/Handler and Thread priorities

I am using an AsynchTask to host a simulator that runs indefinelly and posts the results after each simulation step.
Limiting the simulation loop in background at a maximum of 25Hz, and only calling a javascript function with the results, it works "fine".
Apart from updating a webgl model in a browser, what looks fast enough, I have two more things to update from the Android UI: the FPS indicator and the panel with TextViews representing some of the values. If we forget about the FPS:
The onProgressUpdate() function is already limited to be called at 25Hz, to refresh the model. Now I use another time variable to limit, inside this method, the call to another method that updates the UI panel textViews. It is limited to 1Hz, less than what I actually wanted but fast enough for the kind of information. The method is as clean as possible, all the views are previously loaded to a variable that I keep to not load them every time.
What is the effect: looks like updating 5 textViews takes like one second where all the UI freezes, the touch moves are very very laggy...
I decreased the priority of the background task with:
#Override
protected Boolean doInBackground(ModelSimulation... params) {
Thread.currentThread().setPriority(Thread.MIN_PRIORITY);
...
And used Thread.yield() at the end of the doInBackground method. This improves the behavior to what I explained, without these commands, the behavior is even worst.
My questions are:
-Can I reduce even more the priority if instead of using a background task I use a handler and my own Thread?
-Will a service improve the behavior of the UI?
-Why updating 5 textViews takes so long compared with calling a javascript function that finally will have to use the gpu to change the webgl model?
-Is Android not prepared in any sens to do dynamic applications? How applications like the ones to test sensors update so fast the UI? because there are not standar components like the textViews? (like browser going faster than a textView)
Note: even reducing the refreshing limitations, it produce a laggy effect every time the HUD is updated. In fact I talk about 5 textViews but only updating the FPS indicator produces the same pause. Looks like the only fact of having to switch to the UI thread already consumes this time.
Edit 1:
#Override
protected Boolean doInBackground(ModelSimulation... params) {
Thread.currentThread().setPriority(Thread.MIN_PRIORITY);
if(simulator.getSimulatorStatus().equals(SimulatorStatus.Connected)){
try {
while (true){
//TODO Propagate
long dur = (System.nanoTime()-time_tmp_data);
if(dur<Parameters.Simulator.min_hud_model_refreshing_interval_ns){
try {
long sleep_dur = (Parameters.Simulator.min_hud_model_refreshing_interval_ns-(System.nanoTime()-time_tmp_data))/1000000;
if(sleep_dur>0){
Thread.sleep(sleep_dur);
}
} catch (InterruptedException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
time_tmp_data = System.nanoTime();
SpacecraftState sstate = propagate();
int progress = (int)((extrapDate.durationFrom(finalDate)/mission.sim_duration)*100);
if(sstate!=null){
SimResults results = new SimResults(sstate, progress);
simulator.getSimulationResults().updateSimulation(results.spacecraftState, results.sim_progress);
publishProgress();
}
if(isCancelled())
break;
Thread.yield();
}
} catch (OrekitException e) {
// TODO Auto-generated catch block
e.printStackTrace();
simulator.showMessage(simulator.getContext().getString(R.string.sim_orekit_prop_error)+": "+e.getMessage());
}
}
return true;
}
#Override
protected void onProgressUpdate(Void... values) {
//Update model by push
simulator.getSimulationResults().pushSimulationModel();
//Update GUI HUD
if(time_tmp_gui==0 || (System.nanoTime()-time_tmp_gui)>Parameters.Simulator.min_hud_panel_refreshing_interval_ns){
time_tmp_gui = System.nanoTime();
simulator.getSimulationResults().updateHUD();
}
}
If I comment the line simulator.getSimulationResults().updateHUD(); or directly the contents of the method, it works "fine". And this method is only changing some textviews text:
public synchronized void updateHUD(){
//Log.d("Sim",System.currentTimeMillis()+": "+"pre update gui 1");
activity.runOnUiThread( new Runnable() {
#SuppressLint("ResourceAsColor")
public void run() {
if(view != null){
if(panel_time != null)
panel_time.setText(info.time.replace("T", " "));
if(panel_progress != null)
panel_progress.setProgress(info.progress);
if(panel_vel != null){
panel_vel.setText("Vel. "+String.format("%.2f", info.velocity)+" Km/s");
if(info.velocity>config.limit_velocity)
panel_vel.setTextColor(activity.getResources().getColor(R.color.panel_limit));
else
panel_vel.setTextColor(activity.getResources().getColor(R.color.panel_value));
}
if(panel_accel != null){
panel_accel.setText("Accel. "+String.format("%.2f", info.acceleration)+" Km/s2");
if(info.acceleration>config.limit_acceleration)
panel_accel.setTextColor(activity.getResources().getColor(R.color.panel_limit));
else
panel_accel.setTextColor(activity.getResources().getColor(R.color.panel_value));
}
if(panel_radium != null)
panel_radium.setText("Orbit radium: "+String.format("%.1f", info.orbit_radium)+" Km");
if(panel_mass != null)
panel_mass.setText("Mass: "+String.format("%.1f", info.mass)+" Kg");
if(panel_roll != null)
panel_roll.setText("Rol: "+String.format("%.1f", (180*info.roll/Math.PI))+"º");
if(panel_pitch != null)
panel_pitch.setText("Pitch: "+String.format("%.1f", (180*info.pitch/Math.PI))+"º");
if(panel_yaw != null)
panel_yaw.setText("Yaw: "+String.format("%.1f", (180*info.yaw/Math.PI))+"º");
}
}
});
//Log.d("Sim",System.currentTimeMillis()+": "+"post update gui 1");
}
Edit 2: I can actually remove the runOnUiThread since it is already at that thread, but the effect is the same, this is not the problem.
Edit 3: I tried to comment all the lines of the method updateHUD() and leave only these two:
if(panel_time != null)
panel_time.setText(info.time.replace("T", " "));
The effect is almost the same, if I touch any textView, the animation goes by steps like periodically freezing
Edit 4:
I noticed that the process inside the AsyncTask was taking longer than the available step time so it was never sleeping. I established a safe guard time of 10ms that is slept even if the simulation step is longer than the available time. So, I have minimum 10ms free of each 100ms. The efect stills the same. I am updating at 25Hz the browser and 1Hz a single textview text. If I disable the textview update, the webgl model animates smoothly. On the other hand, if I enable the textview update too, every time the text is updated, there are some miliseconds where the browser animation and its response to touches are blocked. This effect gets worst if I increase the task priority. I tried setting a huge guard of 500ms but the freezing effect stills appearing. I am using XWalkView, can it be something blocking the interaction of this view when UI Thread is acting?
I can't understand why a 4 core 2 RAMgb device needs way more time to compute the same simulation than in Linux or windows desktop PC. I have 25Hz-->40ms of available time and the steps take almost 70ms. In a PC I could keep the simulation at 25Hz in real time. Is there so much shit running in background in Android compared to other OS?

There must be another issue with your code. Try posting your AsyncTask in here.
You could also try something very basic like:
Create a new Thread that loops every 25Hz and update your UI by using the post() method of your UI elements or the runInUiThread() of your Activity. See if there's any code still running inside the UI Thread, that could do heavy work, that can be done outside the UI Thread.

I tried literally everything except for the most logic thing, trying the application without the debugger connected.
The reason to have slower simulation than in a PC, to freese UI events... all because the debugger takes a lot of resources from the device. So, I guess that from this point and avobe I will have to test the application without debugger, what forces me to reboot the phone each time to avoid the "waiting for debugger to connect".
Thank to all who tried.

I could be wrong, but I think that yours problem in synchronization on simulator.getSimulationResults() object. I can't see the realization of the simulator class and realization of the object returned by getSimulationResults(), but I suppose that getSimulationResults() returns the same object every time? If so, then it can be looks like this:
In the AsyncTaks call simulator.getSimulationResults().updateSimulation(...). If this method is synchronized, then this call will be lock the SimulationResults object for AsyncTaks thread.
updateSimulation(...) returns, and publishProgress() is called, but publishProgress() is only schedule the onProgressUpdate(Void... values) in the UI thread.
The new iteration in the AsyncTaks thread can be started befor the UI thread gets the control and executes onProgressUpdate(Void... values). So, AsyncTaks thread goes to the first step.
The UI thread gets the control and executes the onProgressUpdate(Void... values) and synchronized void updateHUD() methods, but updateHUD() can't be executed, because SimulationResults object is locked by the AsyncTaks thread in the updateSimulation(...) method. So the UI thread returns the control to the OS. This may occur many times.
So, onProgressUpdate(Void... values) method and all events in the UI thread can be executed only if the UI thread gets the control in the right moment when updateSimulation(...) method is not called in the AsyncTask thread.
You can check this idea by replacing the public synchronized void update HUD() on the public void update HUD(), and write something randomly in the TextView.
In any case, the use of AsyncTask in this case is not the best idea. AsyncTask's are executed in the TheadPool, but in the Android system this pool can consist from only one thread. So, all AsyncTask's will be executed one by one in the one thread.