Notices by Empterdose (mrdos@hachyderm.io)

@thomasfuchs I had a bike headlamp once that charged with a Type-A-to-Type-A cable. I think I've also seen them used for power banks.

@thomasfuchs The M.C. Escher-designed ThinkPad in the bottom left is doing my head in. I wonder what model that was.

@dalias I wasn't going to say anything because my experience isn't directly relevant, but no one else has said much, so...

In the past, when I needed to prioritize reading speed, I bought a used Plextor drive. (Mine was a PX-716UF, but really anything made before Phillips licensed the name in 2010 should be decent.) At the time, I was ripping audio CDs, largely brand new and pressed (promo copies sent to a student radio station). But it seemed to do really well with the few pieces of older media that I tried it with, too. I don't specifically recall running into anything that it simply couldn't rip, but also, these memories are over a decade old now.

Anyway, my point is that you might want to consider prioritizing drive speed, because then the things it can handle will be dealt with quickly, and you'll know in a hurry if it won't be able to read with something.

@dalias Oh is _that_ why so many threads are broken? I hadn't realized but that makes sense.

@dalias Is `a` statically-provable to have fewer than 20 values, permitting the GCC optimizer to jump on the UB of an out-of-bounds dereference and compile out the `return`?

@dalias @Codeberg Yeah, agreed. Reputation systems or even basic social graph adjacency checks (“Have these users interacted before? Have both of these users interacted with the same repositories before? Do these users have overlapping organization memberships?”, etc.) have performance costs, and can be tricky to balance, but are well worth it. Would actually be a really neat case study to show how to add that sort of protection to an existing code base.

@ryanc I was under the impression that the general consensus was to use “A” unless you had a specific requirement for “B” compatibility, but I'm not sure where I got that from.

@thomasfuchs Could it be a failing battery rather than a software problem? I once knew an older (2014 model) MacBook Air that consistently died without warning once the battery was down to about 30%. One battery replacement later and it was fine again.

@thomasfuchs ...with a Blackberry-style keyboard. I know this was a real device, but I can't shake the feeling that it's a @NanoRaptor special.

@thomasfuchs curl -s 'https://api.launchpad.net/devel/ubuntu/+archive/primary?ws.op=getPublishedBinaries&binary_name=linux-image-generic&distro_arch_series=https://api.launchpad.net/devel/ubuntu/focal/amd64&status=Published' | jq -r .entries[0].binary_package_version