store: Pass dont_block in first poll attempt after a failure

[chrisbobbe]

Fixes #979.

Discussion:
#979 (comment)

---

Tested manually, with and without this change:

1. Log out from CZO (too active; events coming in frequently); connect to a not-very-active org

2. Make an artificial network error in the getEvents binding:

diff --git lib/api/route/events.dart lib/api/route/events.dart
index bd14521c7..da7bf4fea 100644
--- lib/api/route/events.dart
+++ lib/api/route/events.dart
@@ -1,6 +1,7 @@
 import 'package:json_annotation/json_annotation.dart';
 
 import '../core.dart';
+import '../exception.dart';
 import '../model/events.dart';
 import '../model/initial_snapshot.dart';
 
@@ -27,6 +28,7 @@ Future<InitialSnapshot> registerQueue(ApiConnection connection) {
 Future<GetEventsResult> getEvents(ApiConnection connection, {
   required String queueId, int? lastEventId, bool? dontBlock,
 }) {
+  throw NetworkException(routeName: 'foo', message: 'foo', cause: 'foo');
   return connection.get('getEvents', GetEventsResult.fromJson, 'events', {
     'queue_id': RawParameter(queueId),
     if (lastEventId != null) 'last_event_id': lastEventId,

3. Watch the logs to see retries going by
4. Comment out the throw NetworkException line and do a hot-reload
5. Watch the logs for dont_block in the GET /events, and watch the app to see if the loading indicator goes away quickly
   • Without the change, I don't see dont_block, and the loading indicator doesn't go away until the next event comes in
   • With the change, I see dont_block, and the loading indicator goes away quickly

[chrisbobbe]

I found a comment that needed updating; revision pushed.

[gnprice]

I'm planning to include this PR in a v30.0.261 release tomorrow. It'd be great to have it merged first.

Just read through the code; comments below. I've also included this in a build I'm using now on my own phone.

[gnprice]

This is a bit confusing because of what I think is a double negative. This value takes effect precisely on the first attempt, right? Then it's set to a different value for the next attempt.

[gnprice]

I think probably just dropping this comment clears things up. The reader can see a few lines below how this gets mutated after the first attempt.

[chrisbobbe]

This value takes effect precisely on the first attempt, right? Then it's set to a different value for the next attempt.

Yeah, that's right.

[chrisbobbe]

Thanks!! Revision pushed.

[gnprice]

Thanks! Looks good; merging.