@MattHatton @NewtonMark setting aside any free speech issues, it wouldn't be the worst thing in the world if the AI web scrapers were only allowed to hammer right-wing websites...

@MattHatton @NewtonMark Presumably the issue would be with the first amendment rights of the person who published the program.

It turned out pretty well. Toppings are mushrooms, pepperoni, olives, and garlic.

It was cold, so I took the pizza dough out of the fridge in the morning. It got a bit more active than I expected.

@ebassi @blandford Sure, but it ended up splitting the type information into "things we can unregister" and "things we couldn't work out how to unregister in time for the GLib 2.0 release", which is not a great way to design an API.

@ebassi @blandford it's kind of amazing how much pain was caused by trying to support calling dlclose on a library that had registered a type.

@mjg59 Sounds kind of like the old Windows CardSpace API they built, except I don't think it ever saw support outside of Internet Explorer.

I wonder if they reused any concepts from that older system?

@zbrown @KekunPlazas I suspect part of it is to make it obvious when you're making a transition from the simple markup world to the full web.

If you've got a link in your simple HTML subset page, how do you know whether the target also uses this simple subset?

I suspect there is also an appeal of feeling like you're part of an exclusive club.

@decryption Poisoning the AI training data for your chosen topic is the new SEO?

@mjg59 For talking to an old server from a new client, I've used "HostKeyAlgorithms +ssh-rsa" and "PubkeyAcceptedAlgorithms +ssh-rsa". Presumably you can do something similar on the server side to let an old client communicate.

Might be easier to just use netcat though, if you trust the network.

@Gorfram @glent That's a fairly accurate description of the federal government. There is a similar arrangement for each state government, with the leader being the premier.

The problem with the headline is that all the context clues point to it referring to the state government, when it is instead referring to the federal government.

@jwz I'm not sure whether the Xwayland thing would have worked, even if you had been receiving messages.

In general, one client won't be able to reference surfaces created by another client. So that protocol would only be useful if the X client receiving the messages happened to be the Wayland server.

If the session-lock protocol expects you to provide a surface, then it would need to be a surface that you control.

@benno Maybe he should try the glasses trick like Dutton did?

I went to see Friendship at the cinema this morning. It looked like there were only two people who were attending together: everyone else spread out among the other seats.

@camwilson would the UAP still have that protection after having voluntarily deregistered?

@decryption Christopher Nolan would probably ridicule you for using a digital queue rather than standing in line for tickets at a traditional box office.

@sil Looking at https://developer.mozilla.org/en-US/docs/Web/API/Cache/addAll maybe you can?

It says the array can contain URLs or Request objects. If you pass Request object, you can set the cache option.

@sil From the way it reads, it should be enough to get the browser to try and revalidate it.

With everything going over TLS these days, the only other caches involved would be ones on your backend. And if you've designed the frontend so every request is a cache miss, why would you even bother with a CDN then?

@sil Would the Request.cache property documented here do the trick too?

https://developer.mozilla.org/en-US/docs/Web/API/Request/cache

It sounds like fetch(url, {cache: "no-cache"}) would force a cache revalidation. You could use "reload" if you really want to download the data again.