Open y-nk opened 3 years ago
Thanks @adamfratino ! Honestly if it's not being maintained it would nice for all people who jumped on board to actually give it back to the community. I understand @jaredpalmer has intent to earn money with turborepo but it's an ass move to lock us up like this (leaving this OSS project dead in favor of a paid one). Of course, we could be doing forks, but there would be "the battle of the forks" and it'd be rather nice to pass ownership to someone (not me) who has intent to drive this project forward for all the people already in.
@y-nk +100!
My team and I chose TSDX for an internal project last December. We're currently having conversations about replatforming because we don't have faith in the future of TSDX.
The main issue I have with all this is the lack of transparency. I've also lost faith in recommending/using any future tools from @jaredpalmer based on this experience. It's simply too risky.
Any alternatives?
as alternatives:
I find this TS starter template useful: https://www.metachris.com/2021/04/starting-a-typescript-project-in-2021/
I would like to contribute. If its possible...
@y-nk Do let us know if manage to clinch that multi-year SLA for an open source project. While not a traditional approach, given the impressive size of your engineering pool I am sure corporate will be happy paying to retain key respository contributors.
@MarcelCutts Is that sarcasm?
As said I'm grateful for the work accomplished and I won't change my position on that, but that the same time you gotta understand that when you put something out there and it's that great, people using it are gonna have a minimum of expectations ; and it's okay to not want to carry that burden, no one wants to ever bully you into forced labor... but the least people want is simple/minimum communication.
"is the project dead in favor of a paid one?"
if so, simply put a message in the readme or archive this one would suffice to be clear about it so that newcomers won't think "oh great! i can start my codebase with it." ; again i'm not judging, if that's the destiny of this repo, then be it.
...and if you ever care on the TypeScript community as much as you did when you started this project... as proposed, maybe ask someone "who has time and dedication to continue what i started because i just don't have the capacity at the moment" and add them as contributor so that the project can outlive you and continue to help other people.
And now I need to replace tsdx with webpack back in my @react-google-maps/api library and dependencies.
A big problem is the typescript official website pointing to this tool as a starter https://www.typescriptlang.org/docs/bootstrap. I will try and raise an issue with them to remove and maybe suggest an alternative.
EDIT: lol, someone raised this issue 9 hours ago https://github.com/microsoft/TypeScript-Website/issues/2045 and it was subsequently merged too. Pretty much ninja'd in GitHub time.
In my @react-google-maps/api library we had replaced tsdx with rollup this month and it is doing great. If somebody need an example, look into my package source code.
For the past couple of weeks, I have started working on a new library which aims to bootstrap a typescript project quickly with some sane defaults. For now, I am only focusing on bootstrapping and not other things. See https://github.com/SayakMukhopadhyay/ts-boot.
The reason I am starting a new project is to eliminate tech debt from tsdx author. Contributions and inputs welcome.
Sad to see this situation, but we really appreciated tsdx
at @we-io & @hotelery.
Thanks for to good work @jaredpalmer!
Time to move on for us. Thank you!
I think this one is also good https://github.com/unjs/siroc
Also nowadays one can use https://nx.dev/ it works for single or monorepo
Today it was announced that Turborepo is joining Vercel and the Turborepo CLI is now open-source.
So, can turborepo be considered as a spiritual successor to TSDX?
Does anyone have a good way to eject?
@colinrobertbrooks is that a non-explicit way to confirm that tsdx should be considered as deprecated/not-maintained anymore?
I saw this tweet the other day by @jaredpalmer
We’ll likely be either merging tsdx or an equivalent into turbo. I’m not a huge fan of plugins, so it could end just being built in https://twitter.com/jaredpalmer/status/1469161070258393092
I wish we could have an official statement from the owners of the repo to take an action
@y-nk, no. It's just information as Turborepo has been discussed above.
as alternatives:
- a fork of tsdx: https://www.npmjs.com/package/dts-cli
- microbundle
- just use rollup/eslint/prettier/typescript/jest/...
+1 for microbundle. Took me days to get up and running with tsdx because of all the outdated dependencies. Dropped microbundle in and it worked straight away.
Another alternative suggestion, found at: https://github.com/hoeck/simple-runtypes/issues/63
https://github.com/egoist/tsup looks like a good candidate
I'm wondering if Jared is gonna give up TSDX, wouldn't it be nice to add some deprecation warning in README and point new comers to an alternative like tsup ? I don't think just leave the repo here a good idea.
+1 for a note explaining this isn't being maintained (if that is the case). Spent a frustrating day-and-a-half trying to get it to play with React 18 and github actions; now finding parcel won't hot reload which is really the whole point!
+1 for a note explaining this isn't being maintained (if that is the case). Spent a frustrating day-and-a-half trying to get it to play with React 18 and github actions; now finding parcel won't hot reload which is really the whole point!
In my job we need to move to plain rollup, typescript and everything. Because We initialize with this our Design System Implementation in React, and we need to keep update. This library has not been updated for more than a year.
I recommend the same alternatives that @aladdin-add.
as alternatives:
- a fork of tsdx: https://www.npmjs.com/package/dts-cli
- microbundle
- just use rollup/eslint/prettier/typescript/jest/...
But specially the last one because the full control of styles. (CSS/SCSS Modules) and injection with rollup-plugin-styles if you want to make a component library.
In case it helps anyone, I went a different way - reverted to trusty old create-react-app, refactored into lib/ and example/ and added back microbundle to publish it. The empty shell is available here if anyone wants to look at it:
https://github.com/sandtreader/rafiki/releases/tag/EMPTY-SHELL
The resulting (empty, useless, for now!) npm module is here:
https://www.npmjs.com/package/@sandtreader/rafiki
I can imagine at some point ejecting CRA and going more custom but this will be fine for now.
I tried to use tsdx recently but couldn't get it to work in the context of a component library for react due to the lack of maintenance.
So I created a solution, and I thought I would drop it here to see if it could help anyone:
I should have done this long ago (not sure why i didn't), but seeing new comments still submitted here to this day, it started to feel necessary.
There's now a PR to propose a simple message about this being a zombie repo. There's little to do now to make a change. A single review and a merge, and we'll close this issue for good.
I'm not sure who/how to push for this. I hope maintainers will receive a notification, care to read and won't take it the wrong way. We're just trying to help here.
It's not an offensive question, but a more a friendly business-related inquiry.
I'm really loving the project and the overall initiative, but on my side I can't afford to base my work on it if it's dying/going to die anytime soon. I'm not asking for an answer with some commitment of any sort, but just a general pulse and the overall mindset of the authors. I understand maintaining an open-source project on spare time is a huge work and I've been reading numerous stories about it, so please do consider I'm really understanding if you guys are busy/having a hard time/etc. I'm currently asserting basing 3 majors libraries of the company I work in, which will be under ~15~20 frontend guys of various background and so I need to confidence that this tool will be solid enough to support them for the years to come.
Thanks for the work accomplished so far anyway :)