Piskvor left the building

@GAIUS: this happened to me, in a brand new RAID1 array: new disk failed at ~1 week of operation, its sibling followed on the same day, leaving the array running on a single, older disk. That one managed to stay up until the NBD replacements arrived. (Could have been a submarine batch, it was some months after that)

@joojaa: "lame superpowers" would be just "powers", no?

Excellent. The script handled migration to a newer kernel flawlessly :) Thanks again!

That matches, I suppose, but not sure how to switch to "{old/new} kernel" code branch based on this.

That should handle both old, new, and newest RPis now. Still thinking how to do the initramfs-update-too-large thing, and how to handle update-rpi-initramfs if used with -b (not generating into /boot)

Ok, I made my changes into github.com/Piskvor/custom-rpi-initramfs , with the most relevant commit being github.com/Piskvor/custom-rpi-initramfs/commit/…

Sure; a repo would work, even better. As for the script: I think the currenttype="${currentversion: -4}" check breaks on old pis (w/o -v[78]\+). My current version is 4.14.34+ , therefore it would compare on ".34+" and fail to ever generate for any other kernel...

Ah, that would help, too. As for "runs on default" - I guess the easiest way would be to make this into an installable package. Still not entirely out-of-the-box, but could be installed in an unattended way.

MODULES=dep helps, but a) not a lot and b) two initramfs-s still wouldn't fit into my /boot

/boot can be resized (impractical) or initramfs generated elsewhere and then copied (the method I use when generating manually - more coding required).

Excellent, thank you; that looks like a good starting point.

...and it will generate a new initramfs, under a different name, which won't match the one in config.txt. Good catch, that was a huge blind spot.

Wouldn't update-initramfs be invoked on kernel upgrade?

@Schwern: That would be...unpopular. However, so would any choices that the UK could take at this point; IMNSHO the UK will now face a civil uprising no matter which scenario happens; it's just not clear who will take part in it (in every case, someone will feel that HM's government has betrayed them so much that they're no longer obliged to cooperate with it).

Aha, it's en.wikipedia.org/wiki/The_Zahir then - I knew it reminded me of something. I do concede that your points are valid and the answer is applicable :)

That's an interesting proposition, but one that's hard to think of - can an unthinkable question even make sense? (OMG, it has already begun! ;)) More specifically, is this enough to keep the readers' attention, or will they shed the load by reducing to "so there's this lethal virus, except this time it's a meme"?

@Legisey: Wouldn't the usual human reaction kick in, shedding the load with "meh, that's just bloody nonsense"?

In Europe, the Schengen agreement has made this not only easy, but even advisable: At some points, there is no simple route from Czech Republic to Germany except via Poland: openstreetmap.org/… In usual traffic, you'd spend about two minutes in Poland (the street is even called Three Countries' Avenue). It's almost unnoticeable - except for the gas stations that have seized the opportunity of lower gas taxation in Poland, compared to CZ and DE ;)

@David Murdoch: en-GB. (That was the 1-second course "A brief introduction to i18n and l10n" ;))

@Rory Alsop♦: I'd say a certain segment of the population will go for "unicorn" ;)

Some more recent old news: "Claimed DigiNotar hacker: domain administrator password of the DigiNotar network was Pr0d@dm1n." Source: theregister.co.uk/2011/09/06/comodohacker_claims_diginotar_hack (FWIW, most password strength checkers rate this password as "very strong".)