Drip sync

Tests are about usefulness, not testing for the sake of testing.

One thing that is useful, is that user reported issues get tests, if someone cared enough to make an issue, it should get a test. There have been 1k issues on BlockNote (a rich text editor library). That would be minutes of tests

I don't really buy that either though. Sure, you can do things concurrently, but CI machines aren't known to be fast and it'll likely be concurrent to the number of CPU cores. So, like 2 or 4.

I very much would question this. 2s is nothing to be proud of. Imagine thousands of tests.

Having a few e2e tests like this, sure. But each and every test being isolated doesn’t feel right

Yea I’ve noticed the same too. I’ve been a bit more aggressive with closing issues until they prove an actual problem. And not one has proven me wrong yet

3. I’m waiting for private data to implement lists. Sure I could do public lists but I don’t wanna store private ones.
Added start and end dates recently. I don’t know what a prompt is here.

4. Can easily be added, not totally clear why you’d need it though.

1. Almost intractable, you’d need copies of both data sources and it isn’t entirely clear what reading the same book means.

2. Data is the hardest part of any book project.

Oh, very nice!

It's unfortunately very hard to get book data.

Wish you best on your app!

Took a couple of hours, but way easier than editing the @bookhive.buzz logo took, and now I have an SVG that I can actually edit around with.

After a bit more tweaking, Pepper is now on the GitHub project page:

github.com/nperez0111/t...

It is all open source.

Editor (core & React) is MPL

There are additional packages which are GPL or commercial license. But they aren’t required and can be built on your own

Can definitely recommend Hono for a web standards based server!

It is like express, if it were designed in the last five years!

As someone who works on rich text editors for a living with blocknotejs.org

Yes, rich text editors are extremely difficult to program for, because most often UIs are 1 way so it is simpler to represent, whereas text editors need to both make sense UX wise and programmatically which is hard.

I understand that this will store the PLC ops efficiently, but I don’t quite understand the sort of data you’d get out of holding this.

Yea, I can’t believe that I even watched that video. What a waste of time

I think it uses a buffer polyfill: jimp-dev.github.io/jimp/guides/... Bottom of the page

Do you want the sharp NPM package to be in pure JS, or to be able to run it within the browser (e.g. WASM)?

I can see why someone wouldn't ship an image format conversion to a web app.

Yep that will force unmount and mount again

I haven’t run into another use case that requires the whole module to be top level awaited.

Maybe for slow initialization and then exposing a synchronous api on top?