Egregoros

Signal feed

Haelwenn /элвэн/ :triskell: 🔜fosdem

@lanodan@queer.hacktivis.me

🦊🦄⚧🂡ⓥ :anarchy: 👿🐧 :gentoo: :sun:
Pleroma maintainer (mostly backend); BadWolf developer; Gentoo contributor; Eternal upstreamer

Opinions are your employer

Arch users needs to pacman -R works-for-me arch-btw

Make the changes you want to see.

Just because computer bad: X5O!P%@AP[4\PZX54(P^)7CC)7}$EICAR-STANDARD-ANTIVIRUS-TEST-FILE!$H+H*

banner from: https://soc.flyingcube.tech/objects/56f79be2-9013-4559-9826-f7dc392417db
Federation-bots: #nobot

Posts

Latest notes

@lizzy And it feels so weird as someone which used to daily drive OpenSolaris/OpenIndiana, the destruction Oracle caused back then is probably by far the biggest reason I care about portability and am against implementation monoculture.

(Other reason against monoculture being security/resiliency, different baskets are unlikely to have the same flaws, something I apply for storage as well, never all the same brand)
@SuperDicq @volpeon The reasons why I'm taking it as unreliable are basically architectural, it'll work for most but the failure modes are the worst for a bootloader to have.

grub-install can literally fail to fully install and not report an error. Combine this with how modules have ABI breaks without backward compatibility, so you get the MBR or EFI executable updated to a new version but without their modules required for bootup.

Similarly grub-install often fails to detect the hardware, which is stuff that ought to be into a configuration (allowing a verification step) instead of magically done every time it's executed.

grub-mkconfig similarly can create config files that it then cannot parse, that one even predates grub 2.x