There's a fair amount of IO involved in taskgraph generation (reading config files, reading moz.build files, etc), but yeah.. I think we should continue using a separate subprocess for each generation here. I'd love to investigate using `asyncio` *within* a single invocation of `taskgraph` such that we can process kinds concurrently (as long as their kind dependencies are already processed), but that's a whole separate project.
Bug 1720715 Comment 10 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
Good call. There's a fair amount of IO involved in taskgraph generation (reading config files, reading moz.build files, etc), but yeah.. I think we should continue using a separate subprocess for each generation here. Haven't looked at what the "hack" is, but switching to a `concurrent.futures.ProcessPoolExecutor` might still be an improvement. I'd love to investigate using `asyncio` *within* a single invocation of `taskgraph` such that we can process kinds concurrently (as long as their kind dependencies are already processed), but that's a whole separate project.