X-Git-Url: https://git.exim.org/exim.git/blobdiff_plain/8852dff2f3aa2b5a31505febb195d1c6fed620f1..c26e27d5b81ed5640c00ee87f1d4287fb066dc12:/test/scripts/0000-Basic/0548 diff --git a/test/scripts/0000-Basic/0548 b/test/scripts/0000-Basic/0548 index 7158ea584..a0a7fa048 100644 --- a/test/scripts/0000-Basic/0548 +++ b/test/scripts/0000-Basic/0548 @@ -2,6 +2,12 @@ # In this test, the primary host (normal IP address) defers the recipient, and # the secondary host (loopback address) defers the connection. # +#XXX the test is relying on one-second accuracy, which is too tight. +# The measured time is only one-second precision, so certain to sometime +# straddle a boundary, even ignoring scheduling and I/O-wait issues. +# Consider a factor of (?) two, in the retry section of the conf/ file +# and +1 /2 after the diff operation (runtest:463). +# need_ipv4 # exim -DSERVER=server -bd -oX PORT_D @@ -9,20 +15,27 @@ exim -DSERVER=server -bd -oX PORT_D exim -odi userx@test.ex Test message **** +# defers from both dump retry -sleep 2 -# At this point, the secondary host should have timed out +millisleep 4200 +# At this point, the primary host retry rule ehould have timed out +# and the secondary is ready for a retry exim -q **** dump retry -sleep 1 +millisleep 4200 +# This should be similar exim -q **** dump retry +# with luck this is tried within a second after the previous +# so the primary gets tried (it's a new RCPT so the existing retry record is irrelevant) +# but the secondary is not yet ready for a retry (host record) exim -odi usery@test.ex Test message 2. **** -sleep 1 +dump retry +millisleep 4000 exim -q **** killdaemon