feat(routing): Support session based routing
The session id field in a message buffer is now used
directly for routing.
Change-Id: I3634c97588b11172db964b2d06c96c317d8b8ae3
Signed-off-by: E. Scott Daniels <daniels@research.att.com>
Routing table entry changes to pick up subid
Change-Id: If08dc21aae4acaab350ba75a8854ad2f24007b03
Signed-off-by: E. Scott Daniels <daniels@research.att.com>
Fix unit test for rmr_call
It was not properly setting message type and now that RMr
ensures that invalid message type is set by default on a newly
created message this was causing unit test to fail.
Change-Id: I50f08d1038ea7fca2a070cdd949657bfbc25f3fd
Signed-off-by: E. Scott Daniels <daniels@research.att.com>
Add application level tests
Added round robin and multi group application level
test scripts.
Change-Id: Ic6aebaf3bc1edb763decc7fd0aebb09df116f20c
Signed-off-by: E. Scott Daniels <daniels@research.att.com>
NNG based sub-id support added
Change-Id: I0d36b55bb90a315ba94c9476df88e2c7eac6c383
Signed-off-by: E. Scott Daniels <daniels@research.att.com>
Correct bug in app test script
Change-Id: I5b4a9f32aa1bc2907f320b8ad4628e0948062904
Signed-off-by: E. Scott Daniels <daniels@research.att.com>
Nano sub-id changes and unit test updates
Change-Id: Ia69f2fb33de3bbee2f33f9a4c5def779c872e52c
Signed-off-by: E. Scott Daniels <daniels@research.att.com>
Change nil-sub_id key to high order key
If there is no sub-id, then the key is based only on the
message type, but to allow for a sub-id == 0 the key
when there is no subscription id must be set to
0xffffffff00000000 + msg type.
New version for deb is 1.0.19
Change-Id: I55f89d368466a0137fdea99410c76ba72e1923ab
Signed-off-by: E. Scott Daniels <daniels@research.att.com>
diff --git a/test/unit_test.ksh b/test/unit_test.ksh
index 21848af..74bc048 100755
--- a/test/unit_test.ksh
+++ b/test/unit_test.ksh
@@ -267,6 +267,7 @@
show_all=1 # show all things -F sets to show failures only
strict=0 # -s (strict) will set; when off, coverage state ignored in final pass/fail
show_output=0 # show output from each test execution (-S)
+quiet=0
while [[ $1 == "-"* ]]
do
@@ -285,6 +286,7 @@
-s) strict=1;; # coverage counts toward pass/fail state
-S) show_output=1;; # test output shown even on success
-v) (( verbose++ ));;
+ -q) quiet=1;; # less chatty when spilling error log files
-h) usage; exit 0;;
--help) usage; exit 0;;
@@ -353,7 +355,12 @@
if ! ./${tfile%.c} >/tmp/PID$$.log 2>&1
then
echo "[FAIL] unit test failed for: $tfile"
- cat /tmp/PID$$.log
+ if (( quiet ))
+ then
+ grep "^<" /tmp/PID$$.log # in quiet mode just dump <...> messages which are assumed from the test programme not appl
+ else
+ cat /tmp/PID$$.log
+ fi
(( ut_errors++ )) # cause failure even if not in strict mode
continue # skip coverage tests for this
else