Add some XXX comments about when we should re-poll

This commit is contained in:
David Glasser
2013-12-12 15:55:06 -08:00
committed by Nick Martin
parent a6c8b5eef3
commit cce9d36dfa

View File

@@ -294,6 +294,15 @@ _.extend(OplogObserveDriver.prototype, {
// In various circumstances, we may just want to stop processing the oplog and
// re-run the initial query, just as if we were a PollingObserveDriver.
//
// XXX We should call this when we detect that we've been in FETCHING for "too
// long".
//
// XXX We should call this when we detect Mongo failover (since that might
// mean that some of the oplog entries we have processed have been rolled
// back). The Node Mongo driver is in the middle of a bunch of huge
// refactorings, including the way that it notifies you when primary
// changes. Will put off implementing this until driver 1.4 is out.
_pollQuery: function () {
var self = this;