0

I've written a Muzei plugin that uses the Flickr API to check for and fetch a remote image once an hour. I've noticed that sometimes my app will get in to a bad state however and will stop updating. I've checked the shared prefs (where Muzei stores the next update time) and I've noticed that the scheduled_update_time_millis variable is no longer present in my art source xml file.

I've double checked my RemoteMuzeiArtSource.onTryUpdate() method call and I've verified that it should always either call

long CHECK_INTERVAL_MILLIS = 1 * 60 * 60 * 1000L; // 1 hr
scheduleUpdate(System.currentTimeMillis() + CHECK_INTERVAL_MILLIS);

or throw a RetryException

I'm unable to consistently repro getting into this bad state unfortunately, so I'm not sure how it happens.

Full source for my RemoteMuzeiArtSource class lives here: CactusArtSource

ianhanniballake
  • 191,609
  • 30
  • 470
  • 443
Andrew Flynn
  • 1,501
  • 12
  • 17

1 Answers1

0

Answering my own question: turns out that any RuntimeException thrown during execution of onTryUpdate or onUpdate will result unsurprisingly in not getting through the entire method and thus (in my case) not scheduling a new update. In my case, it was a RetrofitError being thrown during a bad network event. I threw a little try-catch handling at it and seems to have fixed the error

Andrew Flynn
  • 1,501
  • 12
  • 17