0
$\begingroup$

Rosanswers logo

It may be useful


Originally posted by jihoonl on ROS Answers with karma: 634 on 2012-11-21

Post score: 1


Original comments

Comment by felix k on 2013-09-29:
I'd argue that only publishers should be latched on topics which use case is indifferent of knowing they are latched. Could you think of some example?

Comment by jihoonl on 2013-12-08:
One usecase is rosbridge node which is a subscriber of topic message but also publisher to the web clients. The problem was that one rosbridge client which came late couldn't receive the topic message if two rosbridge clients attempt to subscribe a latch topic.

Comment by jihoonl on 2013-12-08:
um.. maybe this issue describes better

https://github.com/RobotWebTools/rosbridge_suite/issues/1

$\endgroup$

1 Answer 1

0
$\begingroup$

Rosanswers logo

I'd think comparing the time of ones own (successfull) subscription to the timestamp in the message could serve as a workaround, if the mechanical details are checked. Like that the subscriber initialization probably returns before the subscription at the rosmaster is completed.


Originally posted by felix k with karma: 1650 on 2013-09-29

This answer was ACCEPTED on the original site

Post score: 0


Original comments

Comment by dornhege on 2013-09-29:
This assumes that a message's time stamp is (near) the current time, when it was sent, which isn't necessarily true.

Comment by jihoonl on 2013-12-08:
One workaround we found is using rospy api. http://wiki.ros.org/rospy/Overview/Publishers%20and%20Subscribers#Connection_Information.

Comment by jihoonl on 2013-12-08:
This question was to support rosbridge. and we got it workaround with this pull request. Thanks all

https://github.com/RobotWebTools/rosbridge_suite/pull/58

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.