Prior to this change, if the application within the enclave exited
shortly after sending a particularly slow (e.g. large) response, the
ingress proxy tasks would be aborted before the transfer could
complete, causing the response to be prematurely terminated. Instead
of aborting the tasks, send a shutdown signal and wait for their
completion.
Prior to this change, if the application within the enclave exited shortly after sending a particularly slow (e.g. large) response, the ingress proxy tasks would be aborted before the transfer could complete, causing the response to be prematurely terminated. Instead of aborting the tasks, send a shutdown signal and wait for their completion.
This fixes https://github.com/edgebitio/enclaver/issues/167.