Lightnews — Scholar-powered news

pngwn @pngwn.at · 10h

I just spent 4 hours having my mind expanded with an ATProto OAuth shaped crowbar thanks to the inimitable @ngerakines.me

I have a very tiny brain but after I have gone over the content some more I am excited to play!

3 10

pngwn @pngwn.at · 10h

The information isn’t restricted. It is here and free to access, doesn’t require an account to view and for those wishing to engage, bluesky accounts are free.

5

pngwn @pngwn.at · 1d

Well, I’d be surprised if the original graph was correct in general tbh.

1

pngwn @pngwn.at · 1d

That is testament to the culture that the rust folks have built over the years and stands as a proof of the hard work that has been put in.

3

pngwn @pngwn.at · 1d

The results are legit in the context of the survey, but I’d argue the original graph title is a touch misleading. Surveys usually represent only a tiny minority of users.

What it does show is equally impressive though. That the most engaged part of the base identify this way.

2 4

pngwn @pngwn.at · 13d

There should be an option to use one of your new stamps.

pngwn @pngwn.at · 19d

i did a thing: bsky.app/profile/pngw...

pngwn @pngwn.at · 19d

We are switching to trusted publishing on the gradio repo but with 68 npm packages and no API/ scope-level connections, i decide to write a playwright script to take care of it:

gist.github.com/pngwn/e7b844...

automating npm trusted publishing connection for many packages

automating npm trusted publishing connection for many packages - login.ts

gist.github.com

1 4

pngwn @pngwn.at · 19d

The process is a two parter:

- login (user/pass + 2fa) and save the session state to disk
- run the script using the saved sessions state

pngwn @pngwn.at · 19d

Had an issue with cloudflare's bot detection a little but after I clicked through a few times it seems to let me through (although it got suspicious every time).

1

pngwn @pngwn.at · 19d

We are switching to trusted publishing on the gradio repo but with 68 npm packages and no API/ scope-level connections, i decide to write a playwright script to take care of it:

gist.github.com/pngwn/e7b844...

automating npm trusted publishing connection for many packages

automating npm trusted publishing connection for many packages - login.ts

gist.github.com

1 1

pngwn @pngwn.at · 19d

Looks really useful! What is the accessibility story for this feature?

pngwn @pngwn.at · 29d

You should go an read @43081j.com new article on why everything in the JS world is so messy. Written by the very person who is trying to tidy the mess.

James @43081j.com · 29d

some thoughts about the bloat introduced by edge-case first libraries

The bloat of edge-case first libraries

How building edge-case first led to bloated, overly-granular libraries and what we can do about it

43081j.com

2 6

pngwn @pngwn.at · 29d

Coming soon!

pngwn @pngwn.at · 29d

Thank you!n It's even cooler than i first thought! bsky.app/profile/pngw...

pngwn @pngwn.at · 29d

What's really cool is the CSS Highlight API is based around Range objects. These are really powerful primitives because they can capture a series of DOM elements or even parts of a text node. And you can access positional information of those ranges, enabling stuff like this.

pngwn @pngwn.at · 29d

What's really cool is the CSS Highlight API is based around Range objects. These are really powerful primitives because they can capture a series of DOM elements or even parts of a text node. And you can access positional information of those ranges, enabling stuff like this.

1

pngwn @pngwn.at · 29d

I would expect the dick in question to defend their poor behaviour. Anything else would be a disappointment.

1

pngwn @pngwn.at · 29d

Only typescript / js based ones, JSON schema is more portable so when working across languages is often preferred. So I’m not sure Standard Schema is more widely understood in reality, most users are silent.

1

pngwn @pngwn.at · 29d

It is still more surface area. Asymmetric marchers are agnostic, even if I don’t think people should use them. And why zod, why not another validation library? I do not and cannot use zod at work because it is JS only. Why not a JSON schema focused lib instead?

1 1

pngwn @pngwn.at · 29d

Yeah that’s fair I guess. I don’t generally agree with asymmetric marchers either in general.

pngwn @pngwn.at · 29d

Could be useful for property based testing but still don’t think it enables anything that isn’t already possible.

1

pngwn @pngwn.at · 29d

The reason it is important to consider best practice is because inclusion of a feature in core is tacit endorsement that it is something people should be reaching for in general where schema validation is useful for property-based testing but can cause issues with regular tests.

1

pngwn @pngwn.at · 29d

If this does get considered it should be an ‘expect’ extension, preferably in another library and not baked into core and ‘toEqual’. ‘ToMatchSchema’ would be better.

1

pngwn @pngwn.at · 29d

What is more, adding an additional feature for limited benefit bloats the surface area (and code) when test libraries are already difficult to reason about.

You can already do this kind of thing with a custom assertion.

1

pngwn @pngwn.at · 29d

I don’t think this makes much sense, while there are some interesting use cases it is a better practice for assertions to be concrete as they are easier to reason about than a validation schema and easier to ensure correctness. One tiny bug or misunderstanding and your tests could wrongly pass.

1 1

pngwn @pngwn.at · Sep 9

How to be a dick when people are doing their best to figure out how best to protect their users from a serious vulnerability.

2 7