Lightnews — Scholar-powered news

James

@43081j.com

1.4K followers 220 following 1.9K posts

@e18e.dev - chaijs.com - tinylibs - parse5 - unjs - @lit.dev core contrib - github.com/43081j / Focusing on ecosystem perf & tooling 🥳

Posts Media Videos Starter Packs

James @43081j.com · 15h

i was curious what ESLint would look like if I was given free reign:

Before:
- 13,188KB
- 85 dependencies

After:
- 9,708KB
- 65 dependencies

one can dream 😅

2 13

James @43081j.com · 20h

great work! 🎉

James @43081j.com · 1d

and a big thanks to @vitest.dev who sponsored us today!

this will help a lot

James @43081j.com · 1d

you can sponsor @e18e.dev through GitHub Sponsors or OpenCollective

these funds will primarily be used for two things:

- distributed to projects/individuals in the wider JS community
- development of community projects/tools

thanks so much to those of you who already sponsor us 💙

e18e - Ecosystem Performance - Open Collective

The e18e community has the aim of improving performance of tooling and packages in the JavaScript ecosystem.

opencollective.com

1 9 24

James @43081j.com · 1d

Is there a post somewhere to catch up on this? 👀 I missed the announcement I guess

James @43081j.com · 1d

Best way to skip time waiting for a delayed flight: write more yaml parsing logic 😅

James @43081j.com · 1d

Next time for sure!

James @43081j.com · 2d

I got up super early to make it but the flight is delayed anyway 😅

2 3

James @43081j.com · 2d

See you later Amsterdam! 🖖 Has been good fun catching up with everyone

3 25

James @43081j.com · 3d

Also the first time we had most of the @e18e.dev early members in the same room 😅

James @43081j.com · 3d

What an excellent group 🙌

Best part of meetups and conferences

@viteconf.org

5 6 61

James @43081j.com · 5d

Was good fun! Ran out of time but hopefully got some useful stuff across 😀

James @43081j.com · 5d

Day 1 of viteconf has been good fun! Done my first talk in a very long time too 😬

2 25

James @43081j.com · 6d

It is difficult to describe 😅 it will mostly be done already for sure, somewhere. Just a way to render an "image" as ASCII. But the image is any "canvas" (just data of pixels or something)

James @43081j.com · 6d

Heading to Amsterdam for @viteconf.org 🥳

Come find me if you're going and say hi!

James @43081j.com · 6d

That's how it was in my old place too. An API for the UI, and various APIs that connects to

James @43081j.com · 6d

I was playing with rendering things to ASCII 😅 so basically wanted an engine that does all the math but leaves the output to me

1 1

James @43081j.com · 6d

I wish JS had a concept of symbol servers

Source maps are usually not loaded but are often the largest part of a package. It would be better if dev tools lazy loaded them

I did try to build this but node didn't support remote source maps at the time 😕

James @43081j.com · 6d

With this popular pattern of mixing your front-end and backend logic into one file - how many real world projects are architected that way?

Isn't it still true that most companies have a separate backend, usually not the same repo, and often not the same language?

Or has that changed? 👀

3 1 3

James @43081j.com · 6d

Congratulations 🙌

James @43081j.com · 7d

working on a zero-dependency lock file parser if anyone's interested!

this will likely be used in the e18e CLI and GitHub action.

useful where you don't want to pull a YAML parser in but do want to delve into a lockfile generically

GitHub - 43081j/lockparse: A tiny zero-dependency lockfile parser.

A tiny zero-dependency lockfile parser. Contribute to 43081j/lockparse development by creating an account on GitHub.

github.com

1 6

James @43081j.com · 7d

no worries! you are correct. i'll update when i can 👍

James @43081j.com · 7d

parsing lock files today - lot of repetition in them to say the least 😅

James @43081j.com · 7d

good feedback! i'll open a PR for it when i can 👍

1 1

James @43081j.com · 8d

Just in case you missed this

Still, that was this year some time. Change log doesn't seem to mention it

James @43081j.com · 8d

It turns out we were running a somewhat old Firefox (131). At some point between then and 141, the stack traces changed

But afaict chrome is the odd one out in that the at-format is some kind of standard

1 1