🚀 Launching https://t.co/H4IjcyUgzN
Prepare for the German Einbürgerungstest 🇩🇪:
- Real practice questions
- English translations
- Progress tracking
- 100% FREE and ad-free
Start your citizenship journey today! 🌟
#GermanCitizenship#Einbürgerungstest#FreeResources
@ashh640@tailwindcss@PostCSS The problem with "apply" is that it will create a new class duplicating those rules and defeating the main advantage of having a logarithmic growth of the overall CSS bundle. I find the best of both worlds can be to use postcss-modules "composes" to create zero cost aliases
Get to know the elements composing our #DesignSystem. An engineering point of view over the journey from initial idea to full adoption. What we learnt and what challenges we faced handling the complexities of an evolving system across different domains. https://t.co/Hlk45y8JUY
@brian_d_vaughn May want to give a look at https://t.co/Lwq4kY9DaZ
The idea is using various postcss steps to generate the long class strings based on arbitrary numbers of compositions and export them as Js packages. The drawback is needing some kind of bundling or templating step to consume JS
@lmatteis The other big problem is that the more a language (or a DSL) is optimized for a certain problem (or project), the more having stable requirements and assumptions becomes important. That's why in the real world there is always a compromise between maintainability and flexibility
@lmatteis The core question is "what helps maintaining large projects?". If the answer is automatically prevent as many errors as possible and be concise and easy to read, then I'd probably say strongly typed languages with tools for DSL creation via different meta programming paradigms
@lmatteis I find the code of theia ide quite inspiring in this regard. They used inversify.js quite cleverly and achieved real plug in isolation both on backend and frontend, to the point that you can enable or disable whole UI features by only adding or removing one dep in package.json
@buildsghost I think that is expected that a medium to complex component would use many of the core building blocks od the DS. The main 2 problems that I recon are 1) bundling multiple presets and optional features in 1 component 2) double bundling of the shared core components across the app