You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I want to start the discussion on how to handle TF subscriptions and reconnects. We currently try to reconnect the ROSClient if we detect connection problems. We use the the service call (groovyCompatability=false) to establish a TF listener. Which again creates a topic for our client to listen to. However, the tf2_web_republisher will remove the publisher if there is no listeners for a given time period, determined by the timout option when setting up the subscriber. If the timeout is too low, the publisher will be removed before reconnection is done and the client would just listen to a topic noone is publishing to. We could of course use a really high timeout, but that does not seem like a good permanent solution.
Could we use that fact that tf2_web_republisher is using a latch topic to assume that we should receive a message fairly quick after a reconnect, and if not, we simply ask resends the request and subscribes to the new topic instead. Is this something we should pursue?
The text was updated successfully, but these errors were encountered:
I want to start the discussion on how to handle TF subscriptions and reconnects. We currently try to reconnect the ROSClient if we detect connection problems. We use the the service call (groovyCompatability=false) to establish a TF listener. Which again creates a topic for our client to listen to. However, the
tf2_web_republisher
will remove the publisher if there is no listeners for a given time period, determined by thetimout
option when setting up the subscriber. If the timeout is too low, the publisher will be removed before reconnection is done and the client would just listen to a topic noone is publishing to. We could of course use a really high timeout, but that does not seem like a good permanent solution.Could we use that fact that
tf2_web_republisher
is using a latch topic to assume that we should receive a message fairly quick after a reconnect, and if not, we simply ask resends the request and subscribes to the new topic instead. Is this something we should pursue?The text was updated successfully, but these errors were encountered: