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
{{ message }}
Repository navigation
[BUG]: MultiThreadedExecutor in the driver where a single-threaded one would do #31
main() in waterlinked_dvl_driver.cpp uses MultiThreadedExecutor, but as written it does not seem to do anything over a single-threaded one. More of a simplification than a bug?
All the node callbacks are created without a callback group (L232-L284), so they share the node's default mutually-exclusive group. A MultiThreadedExecutor will not run them concurrently anyway.
By default all LifecycleNode callbacks are registered under the default callback group too.
So I believe a SingleThreadedExecutor would behave the same here and possibly provide some performance gains over a MultiThreadedExecutor.
Steps to Reproduce
N/A
Expected Behavior
Either use SingleThreadedExecutor since it is equivalent here, or, if concurrent in-flight commands were intended, set up reentrant or separate callback groups so the multithreading actually does something.
Activity