cleanup(core): migrate to nanospinner

[43081j]

Migrates from ora to nanospinner, a much lighter package.

this should roughly be a drop in replacement and will be faster/lighter

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Updated (UTC)
nx-dev	✅ Ready (Inspect)	Visit Preview	Apr 26, 2025 3:37pm

[nx-cloud]

View your CI Pipeline Execution ↗ for commit 356e47f.

Command	Status	Duration	Result
nx affected --targets=lint,test,build,e2e,e2e-ci	❌ Failed	42m 57s	View ↗
nx run-many -t check-imports check-commit check...	✅ Succeeded	15s	View ↗
nx-cloud record -- nx-cloud conformance:check	✅ Succeeded	2s	View ↗
nx-cloud record -- nx format:check	✅ Succeeded	2s	View ↗
nx-cloud record -- nx sync:check	✅ Succeeded	1s	View ↗
nx documentation	✅ Succeeded	1m 24s	View ↗

---

☁️ Nx Cloud last updated this comment at 2025-04-26 16:15:40 UTC

[JamesHenry]

@43081j I was thinking we would go for nanospinner for this, but found an issue that I linked to from the umbrella cleanup issue: e18e/ecosystem-issues#117 (comment)

Please could you confirm this library does not have this issue? (Should be able to modify the linked stackblitz) and how the two libraries compare?

[43081j]

I'll take a look when I can 👍

The two libraries are pretty well aligned, so it comes mostly down to personal preference. picospinner is a bit more customisable

Both maintainers are active in our community too, so you really could just flip a coin here 😅

[JamesHenry]

@43081j I added picospinner to my comparison with ora and nanospinner:
https://stackblitz.com/edit/stackblitz-starters-dmnx9n?file=package.json&view=editor

It seems like it doesn't actually allow for the error handling at all. Maybe it is calling process.exit itself?

• ora: works as expected but bloated
• nanospinner: has double printing issue on exit
• picospinner: does not allow for exceptions to be handled

We cannot proceed until one of nanospinner or picospinner matches the functionality/robustness of ora I'm afraid

[43081j]

All good, I'll look into this today or tomorrow 👍

I'm mobile at the minute so can't dig into it but I'll figure it out once I'm at a laptop

[JamesHenry]

Hi @43081j,

Following up on this one:

With the latest versions of picospinner and nanospinner, it's nanospinner that seems to best match ora for our usage:

It's because usmanyunusov/nanospinner#32 was explicitly addressed.

tinylibs/picospinner#5 remains in limbo somewhat. I think we could probably press ahead with nanospinner for now, but remain open to picospinner in future?

[43081j]

its been a while so i forget what the disagreement was, but as far as i remember, it was because none of these spinners should really allow output while they're enabled

when the spinner is enabled, you wouldn't expect output to occur (because we're outputting the spinner frames).

errors are no different and shouldn't output either.

picospinner seems to be doing the right thing here and we're wrong for not stopping the spinner before outputting errors.

however, that does make your try/catch a little more verbose, so i understand why ora and nanospinner would add special behaviour for it.

we should go with nanospinner for that reason

[43081j]

i've rebased and updated to use nanospinner 👍

[JamesHenry]

Awesome, thanks! Please can we add the usual lint import check to prevent regressions