Expand the profile screen to simulate bluetooth actions separately and handle them properly in the native code

[shankari]

No description provided.

[shankari]

Starting	Enter	Range	Exit

[shankari]

Note that 7cb3882 requires e-mission/e-mission-data-collection#221 and e-mission/cordova-usercache#51 to work

[shankari]

Start scan	Range	Range	Beacon found	Beacon lost when stopped scan

[shankari]

@JGreenlee @the-bay-kay @louisg1337 here are initial native code changes and templates to save entries and simulate transitions through the UI

@JGreenlee you should be able to use this to fill in the matching and conditional surveys for the draft trips
@the-bay-kay and @louisg1337 you should save similar entries into the usercache from your code so that the code that @JGreenlee writes can read them and populate the surveys accordingly

I still have some examples of reading and writing objects from the native code coming up, but need to go to meetings for a bit.

NOTE: The geofence exit simulator does not currently work; that is also upcoming

[JGreenlee]

@JGreenlee you should be able to use this to fill in the matching and conditional surveys for the draft trips

I have been working on this yesterday and today. I will get as far on it as I can today / this weekend. However, I don't have a lot of time to work on it this weekend.

[JGreenlee]

Draft PR is up at #1145

I think it might be done, but I don't have time today to test it with the "simulation" UI

[shankari]

Hack to test native read and write

Android:

+        if (actionString.equals(ctxt.getString(R.string.transition_ble_beacon_found))) {
+            BluetoothBLE enterWrapper = BluetoothBLE.initFakeWithEventType("REGION_ENTER");
+            UserCacheFactory.getUserCache(ctxt).putSensorData(R.string.key_usercache_bluetooth_ble, enterWrapper);
+
+            for (int i = 0; i < 5; i++) {
+                BluetoothBLE rangeWrapper = BluetoothBLE.initFakeWithEventType("RANGE_UPDATE");
+                UserCacheFactory.getUserCache(ctxt).putSensorData(R.string.key_usercache_bluetooth_ble, rangeWrapper);
+            }
+
+            BluetoothBLE[] currentEntries = UserCacheFactory.getUserCache(ctxt).getLastSensorData(R.string.key_usercache_bluetooth_ble, 5, BluetoothBLE.class);
+            System.out.println("[BLE native] Found "+ currentEntries.length+" entries");
+            System.out.println("[BLE native] First entry is "+ currentEntries[0]+" last entry is "+currentEntries[currentEntries.length -1]);
+            System.out.println("[BLE native] while handling transition " + actionString);
+        }
+

iOS:

+    // HACK to test read and write in the native code
+    // Should be removed when we read and write directly from the service
+    if ([transition isEqualToString:CFCTransitionBeaconFound]) {
+        BluetoothBLE* enterWrapper = [[BluetoothBLE new] initFakeWithEventType:@"REGION_ENTER"];
+        [[BuiltinUserCache database] putSensorData:@"key.usercache.bluetooth_ble" value:enterWrapper];
+
+        for (int i = 0; i < 5; i++) {
+            BluetoothBLE* rangeWrapper = [[BluetoothBLE new] initFakeWithEventType:@"RANGE_UPDATE"];
+            [[BuiltinUserCache database] putSensorData:@"key.usercache.bluetooth_ble" value:rangeWrapper];
+        }
+
+        NSArray* currentEntries = [[BuiltinUserCache database] getLastSensorData:@"key.usercache.bluetooth_ble" nEntries:5 wrapperClass:BluetoothBLE.class];
+        NSLog(@"[BLE native] Found %lu entries", currentEntries.count);
+        NSLog(@"[BLE native] First entry is %@, last entry is %@", currentEntries[0], currentEntries[currentEntries.count-1]);
+        NSLog(@"[BLE native] while handling transition %@", transition);
+    }
+

Full patch:
hack-to-test-native-read-write.patch

[JGreenlee]

Bundling JS fails at this line.
What is it supposed to do anyway? It doesn't appear to do anything

[shankari]

It removes monitorResult and in_range from sampleBLEDevices[uuid] and retains the rest in noResultDevice. We then save noresultDevice - it doesn't make sense to save the fake results along with it.
https://stackoverflow.com/a/34710102

what is the error that you get? Is this primarily in production mode? I am able to build in dev mode just fine.

[JGreenlee]

I dont' see anywhere that noResultDevice is used, though

SyntaxError: ... Identifier '_' has already been declared.
it's trying to destructure the object and assign both monitorResult and in_range to a variable called _.

Yes this happened with build-prod-android

[shankari]

Yes this happened with build-prod-android

Ah I haven't tried the prod builds yet. A simple workaround would be to use { monitorResult: p0, in_range: p1, ...} to avoid reusing the _ variable.

I dont' see anywhere that noResultDevice is used, though

Ah, before my most recent set of commits, when i used a newly added native mock function, I was saving the values directly using the usercache. See 7cb3882

When I changed to use mocks, I forgot to clean these up. I can do that now, or you can comment them out to unblock yourself.

[JGreenlee]

same here

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 17 lines in your changes are missing coverage. Please review.

Project coverage is 26.42%. Comparing base (0f6aabc) to head (ccab0ca).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1144      +/-   ##
==========================================
- Coverage   27.32%   26.42%   -0.90%     
==========================================
  Files         113      114       +1     
  Lines        4816     4980     +164     
  Branches     1032     1062      +30     
==========================================
  Hits         1316     1316              
- Misses       3498     3662     +164     
  Partials        2        2              

Flag	Coverage Δ
unit	26.42% <0.00%> (-0.90%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files	Coverage Δ
www/js/bluetooth/BluetoothScanPage.tsx	0.00% <ø> (ø)
www/js/usePermissionStatus.ts	0.00% <0.00%> (ø)

[shankari]

It's good that I first fixed #1148 before merging this because 0eebe4b did generate a regression

Ensure that we fail on error
Copied config.cordovabuild.xml -> config.xml and package.cordovabuild.json -> package.json
Copying fake FCM configurations for android and iOS
Setting up all npm packages
npm ERR! code 1
npm ERR! The git reference could not be found
npm ERR! command git --no-replace-objects checkout 1.8.5
npm ERR! error: pathspec '1.8.5' did not match any file(s) known to git

[shankari]

I am going to ignore the codecov regression since our current coverage is only 27%, and we can write the test cases for these as we work through the backlog