refactor: switch globby to tinyglobby

[benmccann]

🔗 Linked issue

❓ Type of change

• 📖 Documentation (updates to the documentation, readme, or JSdoc annotations)
• 🐞 Bug fix (a non-breaking change that fixes an issue)
• 👌 Enhancement (improving an existing functionality like performance)
• ✨ New feature (a non-breaking change that adds functionality)
• 🧹 Chore (updates to the build process or auxiliary tools and libraries)
• ⚠️ Breaking change (fix or feature that would cause existing functionality to change)

📚 Description

https://npmgraph.js.org/?q=globby - 23 dependencies
https://npmgraph.js.org/?q=tinyglobby - 2 dependencies

📝 Checklist

• I have linked an issue or discussion.
• I have updated the documentation accordingly.

[pi0]

I appreciate the PR.

I'm aware of tinyglobyy and we are progressively migrating to it (or alternatives) across unjs, however it has a big difference from globby where it does not have integration with npm ignore to auto support gitignore and other repo ignore paths (gitignore patterns are not same as glob ignore patterns, and they are relative to git root)

This lack of feature had caused significant runtime performance regressions (with unbuild/mkdist for one instance) which was far beyond the small install size benefits, therefore I want to be careful about migration.

Do you think we can either followup this first in upstream or help making an internal util that uses tinyglobby+ignore?

[benmccann]

Oh, that's good to know! I wasn't aware that it was causing performance issues that you'd run into. I've filed an issue upstream and will check to see what the tinyglobby author thinks: SuperchupuDev/tinyglobby#92

[SuperchupuDev]

FYI many performance issues were solved with the release of 0.2.11 which happened this week

[benmccann]

I do know that the one nuxt performance issue I saw was reported as being fixed with 0.2.11: nuxt/nuxt#30137

@pi0 would you be able to check if it fixes your issue too or whether the .gitignore functionality is still needed to address the performance issue you saw?

[pi0]

(will test and confirm locally) I think more generally it would be also a behavior change if certain paths from public/ for example previously being ignored with respect to gitignore and now become scanned.

[pi0]

All is good in terms of gitignore support (globby seems was opt-in so not breaking)

The only remaining issue is negated support for ignore (see fixture and test) ~> pending SuperchupuDev/tinyglobby#70