chore(atomic, headless): make bueno external

[fpbrault]

This PR makes bueno external in atomic and headless.

https://coveord.atlassian.net/browse/KIT-3551

[github-actions]

Pull Request Report

PR Title

✅ Title follows the conventional commit spec.

Live demo links

• Storybook

• Playwright report

Bundle Size

File	Old (kb)	New (kb)	Change (%)
case-assist	239.2	235.9	-1.4
commerce	338.1	339.1	0.3
search	407.7	409.5	0.4
insight	394.2	395.4	0.3
recommendation	250.1	248.5	-0.6
ssr	401.7	403.5	0.5
ssr-commerce	350.3	351.4	0.3

SSR Progress

Use case	SSR (#)	CSR (#)	Progress (%)
search	39	44	89
recommendation	0	4	0
case-assist	0	6	0
insight	0	27	0
commerce	0	15	0

Detailed logs

search : buildInteractiveResult
search : buildInteractiveInstantResult
search : buildInteractiveRecentResult
search : buildInteractiveCitation
search : buildGeneratedAnswer
recommendation : missing SSR support
case-assist : missing SSR support
insight : missing SSR support
commerce : missing SSR support

[alexprudhomme]

why is this needed again ? I don't see the point of externalizing bueno. If I were to make a library that uses zod for schema validation, I feel like I would bundle those validation functions with my library. This is not a framework headless and atomic depend on to work. This is a library for utility functions.

[fpbrault]

why is this needed again ? I don't see the point of externalizing bueno. If I were to make a library that uses zod for schema validation, I feel like I would bundle those validation functions with my library. This is not a framework headless and atomic depend on to work. This is a library for utility functions.

The main reason to externalize Bueno was to avoid code duplication when a end-user bundles Atomic or Headless in their own project: https://coveord.atlassian.net/browse/KIT-3181 I dont know if there was other reasons @louis-bompart ?

[alexprudhomme]

why is this needed again ? I don't see the point of externalizing bueno. If I were to make a library that uses zod for schema validation, I feel like I would bundle those validation functions with my library. This is not a framework headless and atomic depend on to work. This is a library for utility functions.

The main reason to externalize Bueno was to avoid code duplication when a end-user bundles Atomic or Headless in their own project: https://coveord.atlassian.net/browse/KIT-3181 I dont know if there was other reasons @louis-bompart ?

But do users really install bueno ? If they don't, they should not have duplication right? I feel like bueno is more an internal package for us that we made before there were popular schema validation libraries like zod.

[fpbrault]

why is this needed again ? I don't see the point of externalizing bueno. If I were to make a library that uses zod for schema validation, I feel like I would bundle those validation functions with my library. This is not a framework headless and atomic depend on to work. This is a library for utility functions.

The main reason to externalize Bueno was to avoid code duplication when a end-user bundles Atomic or Headless in their own project: https://coveord.atlassian.net/browse/KIT-3181 I dont know if there was other reasons @louis-bompart ?

But do users really install bueno ? If they don't, they should not have duplication right? I feel like bueno is more an internal package for us that we made before there were popular schema validation libraries like zod.

Thats a good point, I see ~20k weekly downloads on npmjs.org, but I would not be surprised if these are all from us/our CI...

[louis-bompart]

why is this needed again ? I don't see the point of externalizing bueno. If I were to make a library that uses zod for schema validation, I feel like I would bundle those validation functions with my library. This is not a framework headless and atomic depend on to work. This is a library for utility functions.

The main reason to externalize Bueno was to avoid code duplication when a end-user bundles Atomic or Headless in their own project: https://coveord.atlassian.net/browse/KIT-3181 I dont know if there was other reasons @louis-bompart ?

But do users really install bueno ? If they don't, they should not have duplication right? I feel like bueno is more an internal package for us that we made before there were popular schema validation libraries like zod.

Thats a good point, I see ~20k weekly downloads on npmjs.org, but I would not be surprised if these are all from us/our CI...

It is very unlikely that we have that many users for Bueno (, one of these days, we might want to consider putting the 🪓 into Bueno in favour of another mainstream library like Zod.

However, the concept/philosophy here still applies:

• We want to externalize our dependencies for CDN users to facilitate chunking wherever possible and allow loading dependencies twice. In the present case, we see that Bueno is used by both Headless & Atomic, which means that if we don't externalize Bueno, CDN users will load it twice (is that a big deal, not much 'cause Bueno is very, very, very light)
• We want to externalize our dependencies for NPM users to avoid them bundling the dependencies twice or more. Granted, in the case of Bueno, the chances are minimal, and the gains are abysmal.

[louis-bompart]

I took the liberty of fixing those poor tests (& removing a few cats from the code 😿). I peppered my changes with comments to explain the gist of it, can talk you thru tmrw if you want!
d2aa4e5