From 122cbcc9f7a072c2a828334d3bce61f40de3244c Mon Sep 17 00:00:00 2001 From: "Isaev, Ilya" Date: Mon, 9 Dec 2024 14:46:35 +0100 Subject: [PATCH] Add RAII vs RAII + functions to open questions Signed-off-by: Isaev, Ilya --- rfcs/proposed/parallel_block_for_task_arena/README.md | 3 +++ 1 file changed, 3 insertions(+) diff --git a/rfcs/proposed/parallel_block_for_task_arena/README.md b/rfcs/proposed/parallel_block_for_task_arena/README.md index 240125e3ac..e57a6f7c4e 100644 --- a/rfcs/proposed/parallel_block_for_task_arena/README.md +++ b/rfcs/proposed/parallel_block_for_task_arena/README.md @@ -265,6 +265,9 @@ Some open questions that remain: only for already existing `task_arena`. Should it be possible for `this_task_arena` as well? * What should be expected from "Parallel Phase" API for `this_task_arena` when a calling thread doesn't yet have any associated arena? + * Should parallel phase API be limited only to RAII-only style? + * Are there any scenarios where incovinience of handling `scoped_parallel_phase` object is + not acceptable? * Are there additional use cases that should be considered that we missed in our analysis? * Do we see any value if arena potentially can transition from one to another state? * What if different types of workloads are mixed in one application?