There are a handful of devices that refuse to use our AesGcmProvider,
and as a result they would crash with AssertionErrors when downloading
avatars. Still haven't found why, but for now, probably best to stop
these devices from crashing, since it puts them in a crash loop, and the
app is still usable without avatars.
We have to make some changes, and it's gotten to the point where
maintaining it as a separate library is more hassle than it's worth,
especially with Google releasing WorkManager as the preferred job
scheduling library.
1) Move contact URI, contact photo URI, and custom label
into recipient database, so there are no longer any
contact DB queries during Recipient object loading.
2) Use a SoftHashMap so that any referenced Recipient objects
can't get kicked out of the cache.
3) Don't load Recipient objects through the provider during sync.
This was a super expensive thing to do, and blew up the cache.
4) Only apply changes to Recipient objects during sync if they
are in the cache. Otherwise, there should be no outstanding
references, and the changes are fine going exclusively to
the DB.
