Fix error handling in incremental folder sync job
The comment about error handling that was originally included was a blind copy&paste from EwsFetchFoldersJob, where the initial sync can fail due to overload, which is handled by falling back to an identifier-only sync. In case of EwsFetchFoldersIncrJob there is no such fallback, so the error from the request must be propagated to the parent request. Failure to do so can cause the parent request to end without returning a result, which in turn can cause the resource state machine to be stuck on a pending sync job. In the past such problem almost never occurred, but nowadays with OAuth2 authentication every request can randomly fail due to auth error, so the problem became visible.
parent
3d8e4d17
Please register or sign in to comment