@Suiseiseki @nyanide @sysrq @enigmatico
> webp
Belch! 🤮
@condret
> debian maintainers have the hybris to patch packages, because they believer they know better than the devs
And that is a good thing, because software developers often assume that you want new features when you only want security updates. Updating is fine and dandy until things start breaking as a result. Backporting security fixes if good, I wish it was still more common, sadly it isn't anymore.
@Forestofenchantment @get @Suiseiseki @nyanide @sysrq @enigmatico
@condret
That's what you should always be doing, not for Debian, but in general, end users should never report problems upstream, that's what they have their distro's maintainers for — if they decide that it's indeed a problem with software and not their build of it, they report it further upstream 🤷
@Forestofenchantment @get @Suiseiseki @nyanide @sysrq @enigmatico
@condret
"Hello, I'd like to report a problem: I have this binary that someone else has built for me and it does not work",— WTF is this shit? They won't even be able to tell you how to reproduce the problem even if they tried really hard: they simply don't know what flags the software was built with.
@Forestofenchantment @get @Suiseiseki @nyanide @sysrq @enigmatico
@condret
> distros should pack our stuff without modifications
No, just no! I probably won't be able to use half of the stuff I use with such an approach: systemd dependency, musl incompatibilities and so on — all of this has to be patched to work well with the distro's base system.
And there is nothing wrong with back-porting security patches — because I don't want those coming bundled with 5 new bugs or incompatibilities.
@Forestofenchantment @get @Suiseiseki @nyanide @sysrq @enigmatico
@condret
Free software devs are not "licensing out" their stuff — they shouldn't have control over how others use it, nor they should provide support for binaries built with modifications.
End user reports the issue to the package maintainer, who in turn checks if it's a problem with their modifications/configuration or an upstream bug, and acts accordingly — that is how it used to be and I don't see what was wrong with that.
@Forestofenchantment @get @Suiseiseki @nyanide @sysrq @enigmatico
@condret
Entrusting the upstream developer with security fixes is a horrible idea: you update the package to get a fix for the new exploit only to find out that UI has changed, or config file format — because it was a good idea that happened to coincide with the security fix, and now you have to waste time on fixing your configs — no, thank you! There are exceptions, but in general… Fuck this! It's like building on quicksand.
@Forestofenchantment @get @Suiseiseki @nyanide @sysrq @enigmatico
@condret
Why bother typing them in? Ask AI assistant built-into your terminal emulator to help you with that 😜
@Forestofenchantment @get @Suiseiseki @nyanide @sysrq @enigmatico
@Forestofenchantment
Bloat isn't software that you might need or not and that you can easily uninstall like ffmpeg of libreoffice, but thing that build an infrastructure of its own, like systemd or pulseaudio, bringing in a myriad of dependencies, adding another layer of abstraction that you might not want on this machine — or at all! But somehow distro maintainers still assume that you have to have them just for the sake of uniformity.
@Forestofenchantment
I'm perfectly fine with these things when they are optional, an luckily, there are still distros that allows you to not have them, but at times it gets ridiculous — why would any application, even indirectly, depend on systemd. How can e.g. an instant messenger dictate you what init system to use.
Maintainers of distros that allow for such flexibility have to invest a great deal of effort to work this all around.