From d80a17590a1aa6c65de0ef0f2e6e0bd273f4753e Mon Sep 17 00:00:00 2001 From: John Stebbins Date: Tue, 7 Aug 2018 13:34:36 -0700 Subject: fix race in getting sequence_id of completed job The sequence_id was only available for the WORKING state and not the WORKDONE state. But frontends poll for status periodically and can miss all status updates for the WORKING state if the file is very short or an error occurs early during transcoding. When WORKING status is missed, there was no way to know the sequence_id associated with the WORKDONE status. --- macosx/HBCore.m | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'macosx/HBCore.m') diff --git a/macosx/HBCore.m b/macosx/HBCore.m index 6c259e25a..b2252183d 100644 --- a/macosx/HBCore.m +++ b/macosx/HBCore.m @@ -456,7 +456,7 @@ typedef void (^HBCoreCleanupHandler)(void); } HBCoreResult result = HBCoreResultDone; - switch (_hb_state->param.workdone.error) + switch (_hb_state->param.working.error) { case HB_ERROR_NONE: result = HBCoreResultDone; -- cgit v1.2.3