-
As part of turborepo adoption, what is the general guidance on tasks that were previously run as a root level task because its more efficient to do so (vs individual packages). For example, currently we run tests (jest) at the root of the repo - this is much faster because it avoids jest having to duplicate file traversal across the same files which is extremely expensive. I am assuming that turborepo is expecting all tasks to be run as package operations - which is potentially a tradeoff (leverage turborepo task running at the cost of inefficiency of running jest processes multiple times). It is not clear to me what approach comes out on top. I am assuming others have dealt with the jest use case in a mono repo - and would love to hear experiences moving from a root level to per package operation. Thank you. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 2 replies
-
After actually adopting turbo, I think the answer is to switch from root level execution to per package. The net benefit is greater then running under a single process. |
Beta Was this translation helpful? Give feedback.
After actually adopting turbo, I think the answer is to switch from root level execution to per package. The net benefit is greater then running under a single process.