Asio 1.25.0 introduced ranged_parallel_group along with the overload of make_parallel_group that takes a dynamic range of async ops. For now, the ranged_parallel_group collects results of parallel ops into two/three std::vectors, which may not be the best choice under any circunstances. It would be better to allow users using other containers for collecting results. One may use absl::InlinedVector to avoid dynamic memory allocation if possible, or may write a special container which discards any value put into it to explicitly discard async ops results.
Asio 1.25.0 introduced
ranged_parallel_groupalong with the overload ofmake_parallel_groupthat takes a dynamic range of async ops. For now, theranged_parallel_groupcollects results of parallel ops into two/threestd::vectors, which may not be the best choice under any circunstances. It would be better to allow users using other containers for collecting results. One may useabsl::InlinedVectorto avoid dynamic memory allocation if possible, or may write a special container which discards any value put into it to explicitly discard async ops results.