Open wjcoleman opened 1 year ago
Additionally, the --exec_time
(short option -e
) publisher option is not interpreted as seconds as the helptext and readme state, but rather microseconds. This should be either clarified or corrected in the application code.
System information
What is the current behavior?
For _flat_datalatency c++11 example, the publisher incorrectly exits the publishing while loop before sending the final
timestamp=0
sample to indicate to the subscriber app that the latency test is finished. This is applicable if using the either the--sample_count
or--exec_time
options for the publisher.Steps to reproduce the issue
-m 3
-m 3 -s 10
*Note this is applicable for each
--mode
option.Expected behavior
Publisher app should send a final sample with
timestamp=0
, which the subscriber waits for to indicate it should exit.Suggested solutions
For each occurrence of
count < options.sample_count
in _CameraImagepublisher.cxx, it should be replaced with:count <= options.sample_count
. (for all 4 modes)**unrelated: the help message for the subscriber says the short option for
--display_sample
is-d
. This is incorrect,-d
is for--domain
.-ds
the short option for--display_sample
.