Closed psigen closed 8 years ago
This is a fantastic change - our ill-fated attempt at using Trollius introduced a lot of needless complexity into the planning pipeline. I am grateful that you removed it and left some specific comments above.
@jeking04 was attempting to use defer=True
earlier this week. I'd like to check with her before merging this pull request to see which features, if any, she is currently using.
I am good with merging this. I couldn't ever quite figure it out and ended up just using threading around the execute trajectory call rather than the defer flag.
On Saturday, March 26, 2016, Michael Koval notifications@github.com wrote:
This is a fantastic change - our ill-fated attempt at using Trollius introduced a lot of needless complexity into the planning pipeline. I am grateful that you removed it and left some specific comments above.
@jeking04 https://github.com/jeking04 was attempting to use defer=True earlier this week. I'd like to check with her before merging this pull request to see which features, if any, she is currently using.
— You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub https://github.com/personalrobotics/prpy/pull/278#issuecomment-201707494
Jen King
Thanks for the input, @jeking04! It sounds like you agree that this was an ill-fated endeavor. :sweat:
I am good with merging this after #282.
Trollius support never actually worked very well (or at all) for our planners, and with the addition of
prpy.futures
we have a much better interface for doing parallel operations.This removes the
defer
options on all prpy planning methods, and replaced theRanked
implementation with one that usesprpy.futures
and plain threading instead.