Wouldn't it be wild if you had to have a notification active (or hidden by the user) to run a process in the background reliably? Or if an app spawned more than 5-10 child processes (max of 32 children across all applications, systemwide), it and its children were extremely likely to get killed off when backgrounded?

That's the state of android 12, and with that in mind I can easily see how a complex application like Amazon store might get completely wrecked.

Two points that mostly agree with you. Abuse of authority on a large scale with greater power disparity *is* different than abuse of authority on a small scale (especially with the conspiracy and coverup). And second, most people aren't empowered to do anything to stop this behavior in construction (being a rich-people service), but they are able to stop buying and playing Activision Blizzard games.

All of the packages are signed and attributed to somebody. If there's a reproducibility problem it should affect all users semi-equally (depending on how the application is deployed). A small group of technically competent people (or an individual) can validate a Google Play or Windows Store deployed application for everyone and any malice attributed to the signer. But someone needs to do the legwork and the application needs to be set up for reproducible builds, neither of which is guaranteed.

I think breaking ublock origin is actively harmful to users, so we'll see if manifest v2 ever goes EOL without addressing the problems. Google's track record is spotty with what's best for users.

Just make the tax post-paid. Automobile registration renewals in most states are annual and at least in my state, that's where the BEV/hybrid road tax is applied (550 USD per vehicle this year). The BMV/DMV can plug an obdII dongle every couple of years to sync up and charge you the difference. Title transfer form has number of miles on it at transfer time signed by both old and new owner, which is used for registration. The BMV can just mail the old owner a bill. The solution doesn't need to be high tech and there are plenty of low tech checks that make evasion more trouble than the (at most) few hundred dollars a year is worth.

Sounds like the government version of an app asking for more permissions than it needs. Typical overreach.

What they don't know is if those miles were put on state or private roads or by farm use, etc. Granted, a statistically insignificant number of miles are driven off public roadways and it's better to take the technologically simpler approach like you said: tax road use at registration renewal.

So the init script managing mount/automount crashes. / is still mounted so the system is probably still functional-ish. init doesn't crash. Getty/login/sh shouldn't crash unless somehow they ended up in that directory. Updating boot0 might lose its shit. sysvinit, busybox init, and so forth... none of them *care* about mounts; they are a *user* problem. They don't track them, nor try to monitor which process is in which namespace: they give zero craps about anything except if there is a child process it needs to wait() on.

Simple, stupid, basically foolproof (for its *one* task), and hard AF use for slices/namespace/service isolation.

It could be designed differently, but such a design would end up being a maintenance burden based on the objectives of systemd: simplify management of processes in cgroups and namespaces (including mount namespaces).

Those are called missiles, not drones. To my knowledge no drone has ever shot down a manned fighter even though they carry air-to-air missiles. RPV pilots have substantially less situational awareness than their manned counterparts. Most of the current ones are designed to loiter for as long as possible which is an opposed design goal to fast and maneuverable.

In their current incarnation, they're only used to attack/patrol over ground targets in areas without anti-aircraft capability as they are exceptionally vulnerable to it.

Someday they'll be better, but that's a decade or more out until we actually see it on the battlefield.

The publication and abbreviation PNAS dates back to 1915, so this outcome was set in motion well before Paul Baran and Donald Davies, inventors of packet switching, were even a twinkle in their respective parents' eyes.

It's an attack on the chip they used. The fault lies with the IC manufacturer, NXP. Blaming electronics engineers at google for ICs failing to deliver on design promises is like blaming Dell and HP for spectre-related flaws. NXP is one of the biggest vendors in the smartcard IC space, so it makes sense to trust their word that the chips are secure, and for a long time they were secure to a reasonable degree.

So that's NDIS down. Only filesystem minidriver, winusb, bluetooth, and the rest of KMDF to go. And a scary amount of syscalls to port and support, some of which don't have kernel-equivalent subsystems. Call me when the ABI for easyanticheat/battleye/etc is finalized so I know what I need to start compiling out.

It's okay, dude, I have to carry two due to work. If you walk in with some business-looking chap, I bet you'll pass muster at least half the time.

I bet a majority of Android devices sold do. Besides, the important part isn't that it's Facebook, it's that Facebook is proving they can do it reliably at scale. It just needs some policy maker to latch onto the idea and make it mandatory.

That'd be nice. Can you even get aftermarket head units for cars anymore? Everything seems too tightly integrated and at the same time too distributed these days. Like in a modern VAG vehicle, you've got an optical bus separating the MMI controller from the audio amplifiers and another drop for the buttons. Seems like all the bus devices would be software drivers and that diversity would make it really difficult to come up with a generic unit.

Are you not familiar with xmpp/jabber? It does all of that.