'done_info' not generated and incorrect reward counting on maximum steps completion

[xyzeroth]

It seems that when an actor successfully completes the maximum number of steps (2048) without failure,

any 'done_info' is not generated at Line 214,

implementation-matters/src/policy_gradients/agent.py

Line 214 in 5ee6ecb

if is_done:

and no 'done_info' is appended to 'completed_episode_info' at Line 296.

implementation-matters/src/policy_gradients/agent.py

Line 296 in 5ee6ecb

# If some of the actors finished AND this is not the last step

Consequently, the reward is counted as -1, as observed in the code snippet at Line 324.

implementation-matters/src/policy_gradients/agent.py

Line 324 in 5ee6ecb

else:

If an actor reaches the maximum step without any failures, it should be considered 'done,' and the reward should be counted to the total reward as it is, rather than being set to -1.
Could you please examine how the reward is tallied when an actor successfully completes the maximum number of steps without any failures?