The idea is that modern windows (10,11,12) boot fast enough through various tricks that booting is faster - or as fast as - awaking from hibernation.
TBH It's probably not wrong, I wouldn't know because for some unknown reason my motherboard sits doing nothing for over a minute before starting the POST. Once it's past that, I can be at a linux prompt in 20 seconds, or at the windows login within 40, depending on whether I interrupt rEFInd
Yeah, but only for the hardware you have - the upgrade path seems to force you to a 10 OEM license, I went 7 pro (non-oem) to 10 pro, then changed motherboard, so had to buy a new copy of 10 pro oem, then that CPU stopped working and had to buy a new CPU & motherboard, and so finally I'm on running my own KMS server - fuck paying for windows a 3rd time.
Competition doesn't necessarily imply that it bypasses sunk cost fallacy, or even that the competition won't do the same nonsense in a belief that feature X must be wanted.
Well, there probably needs to be a way to override the heuristic at least, sort of a 'this process is important, don't auto-kill it if trying to find memory'.
As for ssh specifically, I rarely ssh into my desktop machine, but I keep sshd running for just this kind of situation where I might need to try and rescue a swamped machine. So in most cases low-volume sshd use is exactly what is called for.
If you're running into the memory purge of doom on a server that's probably a whole different nightmare scenario.
malloc returning NULL has been a broken assumption for a long time though, and that isn't going to change afaik.
Running sshd as an on-demand (Type=socket) service would probably work better, since then the sshd process would be new and thus have a better heuristic score - also not be tying up memory sitting unused in the meantime.
systemd still seems to run it (Type=notify) with the -D option all the time though, at least on the systems I can check.
Dropbear is configured by default as a Type=socket service though.
I'm surprised you didn't implement a simple 8k bank switching scheme to utilise the rest of the 128K chip, it's really just a handful of 74244 buffers and a 74138 decoder stuck on an IO port.