Closed wynnej1983 closed 11 months ago
Bun doesn't play nicely with turbo as far as my initial testing is going
Bun doesn't play nicely with turbo as far as my initial testing is going
Oh okay didn't know about that, if you find out anything more please share it 🤟
Oh okay didn't know about that, if you find out anything more please share it 🤟
Following up here - bun has been slowly fixing and improving it's workspace support! I have been using it to great success. However, there is at least one annoying thing left that's effecting me. If I change a package in any of my workspaces then bun i
tries to install my workspace dependencies from npm (which naturally fails).
To resolve, I have to remove the bun.lockb
file and re-install with bun i
.
Not ideal, but also not a back breaker
EDIT: Another issue, no way to easily manipulate workspace packages.
pnpm uninstall @auth/core --filter auth
would remove it from the auth
workspace. Alas...doesn't work with bun yet
Totally. I see what you're saying. I made a branch named bun
days ago, actually. It has the entire stack working fine, save for the database integration.
I'm thinking about removing turborepo entirely, as it adds extra complexity with little benefit with bun.
Give it a try!
These two issues track my work integrating bun into T4
https://github.com/cloudflare/workers-sdk/issues/3941 https://github.com/drizzle-team/drizzle-orm/issues/1224
I'm thinking about removing turborepo entirely, as it adds extra complexity with little benefit with bun.
Definitely removing turborepo (either temporarily or not) is not gonna hurt, but adding support for bun would be good for this project PR and developer experience (overall performance).
When turborepo (or even better, nx add support for bun, can consider adding it back later if necessary.
@timothymiller Can we close this issue since bun support is out?
I'm thinking about removing turborepo entirely, as it adds extra complexity with little benefit with bun.
Definitely removing turborepo (either temporarily or not) is not gonna hurt, but adding support for bun would be good for this project PR and developer experience (overall performance).
When turborepo (or even better, nx add support for bun, can consider adding it back later if necessary.
https://jordan-hall.github.io/nx-bun/ 😁 if you're interested
So I added Bun support last week. I am working on removing turborepo entirely today.
FWIW building the t4/ui package is much faster without turborepo caching, even on cached "turbo" runs.
Still working on how to execute bun scripts in parallel, like the top level "dev", which runs bun web and bun api. If you have any suggestions, let me know.
The Bun migration is complete, and all CI/CD is showing green.
34
second backend deploy times1:43
frontend deploy timesturbo
with run-p
bun install
timesbun install
times on first runJust a small consideration, @timothymiller you should remove turborepo from repo topics. Great job, I’m using this repo as a learning pathway and it’s very useful!
It should be pretty straightforward to implement, It also has a github action that we can use in the workflows