tree 70497e8213f420ff6777bc4371bf9784d9c36e87
parent c52be0d21755e23d67128fbcaec17e5952f4af9e
author Matt Spinler <spinler@us.ibm.com> 1588013114 -0500
committer Matt Spinler <spinler@us.ibm.com> 1588013568 -0500

libmapper: Retry on wait timeouts

At times, early in the boot process, the async GetObject call used in
'mapper wait' can fail with an ETIMEDOUT, probably because the mapper
daemon is so busy introspecting new processes.  If this happens, retry
after a 5s wait. Also get a new connection to the bus and event loop
since when this wasn't done the retried method call would fail
immediately.

I first tried to put the ETIMEDOUT check in
async_wait_getobject_callback, but that caused hundreds of more retries
across all the mapper waits in use.

Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I7175e958fb633b2ecb7479a3307d6918ab29b9f7