Lightnews — Scholar-powered news

Wes @notwes.bsky.social · 2d

Ha, I doubt that’s what any of the npm folks would say I have done.

Wes @notwes.bsky.social · 2d

It is entirely reasonable to take this direction, but unfortunately it is clearly out of the realm of possibility with the current investment from GH/Msft. And frankly, we are all better off looking toward vendors like @socket.dev than trying to get npm to directly built out detection.

1 5

Wes @notwes.bsky.social · 5d

Well, I did the free trial. And TIL that Apple never figure out that "natural scrolling" on a trackpad is *ENTIRELY* different from it on a mouse. This is neigh on unusable.

1

Wes @notwes.bsky.social · 5d

FWIW, this is more nuanced in many use cases. We bundle cli’s for workflow tooling we use in CI to improve execution times (no full install) for example. I will back you up any time blaming TS people for bringing many problems on themselves, but it needs nuance.

1

Wes @notwes.bsky.social · 5d

Yeah the reason we don’t is because our e2e tests run on real infra and real projects. So that requires publishing. This is why I want platform features to support these workflows.

1 1

Wes @notwes.bsky.social · 5d

At work we have a pretty great set of integration test flows we use where we publish and use prereleases for this. All that said, avoiding that complexity is always preferable when possible.

1 1

Wes @notwes.bsky.social · 5d

There are a ton of good ways to do this, what I posted is more of a “make it part of the platform” instead of everyone having these unique solutions on their own.

1 1

Wes @notwes.bsky.social · 6d

lol, probably the first time I said it was in some stage of frustration induced burnout. Some combination of deadpan and acceptance. 🤣

3

Wes @notwes.bsky.social · 6d

I *tried to find a way to kill it instead*. Fixing it was a last resort when that fails.

3

Wes @notwes.bsky.social · 6d

I’m not sure folks who don’t know me understand my involvement in the project. I am not working on it because I think it is the best choice for most things.

I am working on it to stabilize the ecosystem and unblock node core for better future http APIs.

1 3

Wes @notwes.bsky.social · 6d

github.com/vltpkg/vltpk...

[FEATURE] VSR: Release Manager · Issue #1220 · vltpkg/vltpkg

Description Allow for two staged publishing by introducing a blessed top-level key explicitly for "staged packages" in packuments (ex. stagedVersions). Published versions to the staging version hav...

github.com

1

Wes @notwes.bsky.social · 6d

What is dead can never die. Express has been a dead project for years now.

2 4

Wes @notwes.bsky.social · 6d

Enough that most "simple" dynamic pages are better off with a static host and a slightly more complicated build process. And if that doesn't meet your needs there are plenty of saas offerings to do a tiny bit of server side dynamism that we didn't have 10 years ago when this was a good idea.

1 2

Wes @notwes.bsky.social · 6d

This is literally ejs. And there are reasons why we moved away from this for most web pages. Even with my *love* of simplicity I moved away from dynamic rendered ejs views.

1 1

Wes @notwes.bsky.social · 6d

So honestly, I double down on still believing in the goal. I double down on Mastadon and ActivityPub needing to learn from Bluesky (centralized UX on federated foundations). I just hope that this story can play out before the crappy market forces drive the outcomes Cory is concerned with.

1

Wes @notwes.bsky.social · 6d

Idk, maybe someone like Cory Doctorow could find a way, but sadly he also needs to see the goal reached before buying in. doctorow.medium.com/https-plural...

Bluesky and enshittification

No one is the enshittifier of their own story.

doctorow.medium.com

1 1

Wes @notwes.bsky.social · 6d

I see this both with my tech and non-tech friends. It is hard to fully articulate this without relying on some technobabel (I say this in the most loving way), and that almost always looses people.

1 1

Wes @notwes.bsky.social · 6d

This is a "goal", and one I have believed in from the early days. That said, I don't think that goal has been achieved *yet*. I am glad to see that the folks working at @bsky.app *believe in this goal*. I just wish there was a way to communicate this goal that was widely accepted.

hailey @hailey.at · 6d

No, you are not our product. Our product _is_ the code that we write and distribute for others to use. For others to take and build their communities. Making the users the product is exactly the model that never works and continues to never work.

1 2

Wes @notwes.bsky.social · 6d

Yep, them changing over time is one of the key design constraints I have been keeping in mind. My goal is to be data driven and have the tools *ready* and easy to use to measure and understand those changes over time.

1

Wes @notwes.bsky.social · 6d

Also, can you link the PR? I know the kinds of things we are thinking about on the Express side, would be interested in seeing if it is similar to what we see.

1

Wes @notwes.bsky.social · 6d

Have I told you about the stuff we are working on in the Express Perf WG? Maybe some future sharing opportunities for things like this, as we plan to document and give built in tooling (wrapping bench-node in this case) for folks looking to give this kind of empirical feedback in our PRs.

2 1

Wes @notwes.bsky.social · 6d

And the theory here is that it has an optimization for that which is only triggered in the "large array" case (presumably with some trigger point where it kicks in)? Because otherwise they would all be the same *very fast* noop?

Wes @notwes.bsky.social · 6d

Yeah I started looking for a link to the the right v8 code and force-ably closed the tab. Not today @james.sumners.info, I have important things to do 🤣.

2 1

Wes @notwes.bsky.social · 6d

Looking forward to a recording if it will be available! In the mean time, I have lots of interested folks I will share your slides with at Netflix.

It is based on your work that my team is planning to start the move toward native ESM only for all JS libraries across the company. Thank you so much!

5

Wes @notwes.bsky.social · 6d

Yeah I read this earlier and almost got nerd sniped into looking at this lol, but my hunch was something that might show from one of these flags. So many little array optimizations with seemingly random cutoffs (they are not random, they just seem random until you see the v8 code). So 🍿 for me.

1 1