what are the tangible end-user advantages of compiling everything from source?
>>101561160use of system-libs over bundled libs whenever possiblereduction of attack surfaceno need to install additional "dev" packages*-bin installs still available if you wantOnly reason to use something other than Gentoo is if you actually NEED an immutable system.
>>101561160its fun to do if you got no one to talk to
>>1015611600.4% performance boostThat is all.
>>101561160Simultaneously having both stability and rolling-release.Easy flexibility. Want matplotlib to only support a Qt5 GUI and not the other ones? Just set a use flag.See a patch you want? No need to wait for it to trickle down to your distro's repos, just apply it yourself.Between debloating packages, running a custom kernel, and optimizing everything for your system, your desktop can often feel slightly snappier. You may use somewhat less RAM, too. Probs not much difference in overall performance tho (besides what the custom kernel can provide) - unless a given package is so new/obscure/poorly-maintained that it hasn't yet gotten inline asm for the performance-critical parts.
>>101561160Hardware compatibility.
The use flags are pretty limited. firefox has like 4 use flags. You can make a smoll kernel too. But overall its a waste of time.
Space savings because you only have the code necessary for your processor and system.Maybe some performance optimization but unless you're an autist you won't care.
>>101561160psst.https://www.gentoo.org/news/2023/12/29/Gentoo-binary.html
It's fun and autistic
>>101561160the time wasted tinkering and compiling outweighs the benefits for any end userfor a corporation it might be worth it if there is a large fleet of homogeneous machines and a dedicated person or team who know what they're doingit's also neat if you are still only learning but after that i'd say it's a waste of timeofc neet trannies with nothing but free time and no useful thing to spend that on will of course disagree
>>101563092Sure, but who the hell wants to use Gentoo stable on their desktop?
>>101561160The ease of patching things manually when necessary and not having systemd or anything by lennart poettering on my machine.
>>101561526>See a patch you want? No need to wait for it to trickle down to your distro's repos, just apply it yourself.it's been like 10 years since i cared this much about a feature or a patch
>>101563262stop doing all the drugs. Your response has nothing to do with what I posted.
>>101562014Gentoo has a problem right now with systemd cucks removing use flags all over the placeI recommend just learning to live with the fact that you're gonna have to abuse EXTRA_ECONF in /etc/portage/env more
>>101563271OP asked for the advantage of compiling from source with a reference to Gentoo. You linked to news of Gentoo's support for binary packages. The article cites a bunch of desktop packages (KDE, LibreOffice), but notes that these binaries are only being built for Gentoo stable. I questioned who would want to use Gentoo stable on the desktop. What am I missing?
>have a stable system>want one piece of software a bit more updated>casually compile itI know, I know, flatpaks do exist but just saying.
>>101562014>You can make a smoll kernel tooThe trouble is kind of the same on all distros. Assumed one spends an hour deselecting stuff under device drivers.
>101563312He asked advantages of compiling everything from source. The conversation is much different between everything and somethings. Being a desktop install or not doesn't have any impact on the advantages.
>>101563276>systemd cucks removing use flags all over the place gentoo has fallenno seriously how do we stop this systemd virus from spreading? I hate how much devs bend the knee to this shit
>>101561160Did it once. To put it into simple words for someone who never did it: everything runs incredibly fast
>>101563400systemd was inevitable.
>>101563400There's nothing we can do. Gentoo at very least makes it very easy to ignore whatever cucked shit upstream is doing because you can just throw a copy of the ebuild in your own repository and patch it. The worst case scenario would be eclasses getting cucked, but even that can be dealt with by just overriding them.
>>101561160i hate ebuilds so much
>>101563400Systemd is one of the best things in Linux. You get something far more reliable and easy to configure ordering for in a dynamic manner (e.g. if X service is set to "After" Y service, and you then make a new service and set it in the "After" for Y service, X service will still start at the correct order in the boot chain, and so will everything reliant on X service be shifted to the new boot order).Besides this, easy in-depth logging, e.g. journalctl with various flags. Being anti systemd is one of the weirdest things I've ever heard honestly.
>>101565250nigger
>>101565250this. I can understand anti-wayland because it actually affects the end user experience, but being anti-systemd in 2024 is retarded
>>101565250>reliablelollmao even
>>101566194You find it unreliable? How? My stuff boots in the designated order each time, avoiding race conditions etc. because the services are started dynamically based on dependencies.If you fucked something, you have probably made a bad dependency (e.g. network device relying on something that only succeeds to deploy when there's a network connection).
It makes upgrading your CPU more fun
>>101561160dunno but I use gentoo because I want a "stable" rolling release and no other distro offers it. debian testing gets frozen after awhile, arch is basically untested garbage, etc.
>>101566258>My stuff boots in the designated order each time, avoiding race conditions etc. because the services are started dynamically based on dependencies.This is like a rustard saying memory safety is a novel feature.>>101565371Systemd is a microsoft owned and operated universal backdoor whose purpose is remote hardware attestation of every commercial linux machine, to break the GPL with tivoization. Lennart Poettering has made several blog posts about using TPM to control, sorry, "secure" your machine.>latesthttps://0pointer.net/blog/brave-new-trusted-boot-world.html
>>101566258Context, I use openrc and skip sysvinit by using openrc-init directly.Openrc handles dependencies already while abstracting away the complexity that traditional init scripts would have, so I prefer it. systemd offers zero control that shell scripts offer. Unit files are objectively terrible unless your needs are simple; they offer zero extensibility, and if you NEED to script, then you need to use wrapper scripts.For what should be an init system, systemd does too much and is clearly politically motivated to control the entirety of the low-level userland on Linux. This is unacceptable in my opinion - especially given the developers' incompetence in handling severe bugs. They've destroyed people's home directories in like three ways thus far, and a handful of genuine security issues have been ignored and closed as wontfix. I can't trust it or the developers. It's that simple. I'm not even going to get into all the complete shit code they regularly try to submit to the kernel which gets shot down.Also, contrary to what the developers say, systemd is not by any stretch faster at booting systems up. From what I measured, openrc with rc_parallel flipped on is like six seconds faster in boot times, so whatever boot time advantage it supposedly offers is a bold-faced lie.
Filtering out retards like you.
>>101566442Not sure if this is what you mean but if you mean booting up scripts, it's a core function of systemd. All of my custom services launch scripts which themselves handle complex tasks.It is possible to control several scripts from within the service file (e.g. ExecStartPre/Post), or else, you just do it the usual way and launch the second script from within your main script.I assume you mean something else though. Yes it is a problem when bugs are refused a fix. I think it is quite secure though as it's used on enterprise-grade distros, by serious companies with lots of data at stake. Where there are major issues, is it systemd causing the issue or an incompatibility of the distro which is trying to implement it?
>>101567072I actually bought a book on it when servers I work with implemented systemd. It taught me a lot of the advanced features in systemd, which is where the benefits become more apparent. You could find the info for free but it was pretty cheap and had a lot of info in one place.
