>>1275 >no more Python 2.7 Hurray! >speeds up dependency calculation 50-60% Quite substantial. It's a shame a lot of people ditched Gentoo for Arch simply over the Wiki issue; especially since Gentoo did it first and then lost it. But I think it's worth using for Portage alone. It really is the best package manager.

>>1275 I miss Gentoo. I want to reinstall it on one of my machines, but I don't have much time nowadays. Pic related is my old Gentoo install before I wiped it.

>>1275 Not only related to Gentoo, but librsvg is apparently rewritten over time in rust. Didn't even notice that until an upgrade pulls in fucking rust.

>>1284 I think Rustfags have realized that it only takes one or two small projects that are common dependencies to force yourself in as a mandatory install. They're probably hoping that, by convincing packages like this to convert to Rust in some capacity, eventually Rust will just be installed with most distros. Then, once it's universal like Python, Perl, or GCC they can try to get people into it directly. Sucks because it means having to deliberately avoid some software based solely on what it's running. That or use a pre-built binary, but that defeats the purpose of running Gentoo.

>>1293 Is Rust really that bad? At least it's a compiled language that is not slow as fuck, unlike Python. Way too much stuff is written in Python, stuff which shouldn't.

>>1294 The tooling is heavy and slow to the point where it makes C++ compiles look speedy. The community has just accepted this, and while they acknowledge there are performance deficits they don't view fixing them to be a priority. Plus it uses LLVM for compilation, so now you need to have another compiler on your system just to handle that, which brings in another set of dependencies. Ultimately, the problem with Rust is that it's not mature, you can accomplish the same thing in C++ by being careful, and the community is entirely pozzed. People who want to get stuff done and use an effective language just use C++ because there's a ton of libraries for it and newer language features offer a ton of safety while maintaining some backwards compatibility. Rust is an entirely new language that, at best, hopes to just replace C++ and be "more safer". That's a flawed goal in and of itself, because it offers nothing in terms of features that C++ doesn't while the C++ committee is actively working to improve the language, offer more flexibility, etc. Their approach to language-building is fundamentally better. They didn't even take the time to improve the syntax, even though it's what everyone bitches about. They somehow made it worse. This is the hardest thing for C++ to ever change for compatibility reasons, but Rust doubled-down on the worst aspects of the language they're hoping to displace and then proceeded to fix something most people consider a non-issue. In short, the language might perform fine, but why should anyone learn a new language that just aspires to replace another? All of its momentum comes from community hype, and the community is absolutely pozzed. Using Rust is effectively virtue signalling at this stage, and their only hope of getting wider adoption is making it mandatory/common to have Rust as a dependency so that normal people have to use it, too. If people want "correctness" in their languages then they should just stick to Functional. Haskell enforces a similar degree of correctness and it does so via interpreter and with a language paradigm that is more sensible for this sort of thing. Trying to graft correctness onto procedural and object-oriented language is a snipe hunt when you can just write abstraction layers to add safety.

>>1295 Rust bundles attack-helicopters patched LLVM. Compiling it means having two sets of LLVM. Developers are responsible for correctness, not the tool. Unsafe language is just an excuse for bad developers. Much like how attack helicopters need safe landing spaces despite the rough, cruel nature of the world. Undefined behaviors can be checked and avoided. Only developers think about safe or not, computers don't.

>>1294 Start by its name: rust: it expresses deterioration, decay, corruption. Why do retards choose such bad names for their shitty creations? Discord, Mumble, Invidious, Rust, just to name a few. All words with negative meanings. Then there's a shitty gay language named after a stupid TV show. Compare Ruby, Lua, Pascal. Fucking faggots and retards. At least name your shit properly.

>>1299 I'm find with most of those names. If anything, I'm mad that Discord and Slack, two names heavily associated with meme cults and early hacker culture, are associated with the most proprietary, mainstream, shitty software possible instead of something cool and open source. Some languages do just have better names, though. I kind of miss when all of them were just letters, though. C, C++, D, C#, F#, R, etc.

I just reinstalled Gentoo and this is what I end up getting slapped into my face. Fuck Rust Trannies Fuck Gnomefags

>>1307 mask new librsvg

>>1309 Too late. Maybe one day, I will get around this fagrust bullshit.

>polkit no longer works with consolekit So my options now is instead of just using something that is only somewhat cancer i now need to install either full on aids or just aids-lite. I'm just going to ignore updating for another few months and hope it gets better.

>>1331 i got consolekit to work, had to unmask a useflag in order to get consolekit. I don't give a shit if it's 3 years inactive it's better then systemd and still allows me to use polkit for VMs so fuck you gentoo devs.

Had anyone have problems with the Tor Browser tarball from the official site on Gentoo? The GUI is so unresponsive for me.

>>1275 Finally Python27 dies! A python -> python3 sym-link will probably never be accepted because all the scripts which believe "python" to always be a 2 flavor.

how does masking and unmasking work? I've been using gentoo for two years and still don't really understand it

>>1353 tbb works fine for me

>>2593 >mask package <portage will warn you when you try to install it, nag you over and over if it's a manditory dep but if you mask it you are guaranteed to never install it >unmask package <it installs normally I don't understand why you're confused. unless you're speaking about amd64 keywords which is mostly just there to separate experimental/not so tested builds from the stable builds

>>1331 Why would you use any of these *kits?

>>2769 mostly being a lazy shit half wanting to use kvm ( virt-manager pulled in polkit and i couldn't be assed at the time) This only cemented my original opinion that gui applications are the spawn of satan and using authorization apis is brain cancer retarded. i did export my virt-manager config as a qemu one-liner so i don't really give a shit anymore Consolekit used to be alright but i suppose with the prevalence of systemd cancer it wasn't ever going to survive. it's mostly unmaintained now which is why they later removed the ebuild entirely. The funny thing about what happened was them removing the useflag from polkit created a circular dependency from hell where polkit/consolekit kept pulling eachother in and i had no way of telling it to stop. removing consolekit ebuild only magnified it because it started pulling in polkit without existing and so the only way to remove it was to change polkit to using systemd/elogind which i was never going to do. then in my desperate attempt to salvage my install i somehow broke PAM modules leaving me unable to login. i'm kind of dumb. i was using polkit/consolekit as a bandaid for my stupidity kind of like how fags use systemd. my story is a tale of why you shouldn't be a lazy shit when using gentoo.

So I'm sure as anyone who uses qutebrowser knows, it relies on qtwebengine to function, which already took forever and a day to compile unless you had a pretty fast computer with plenty of RAM to spare. I've not seen much discussion over the issue with trying to compile qtwebengine 5.15.2, in which OOM reaper kills the compile after a while, in my case that's half a day wasted trying to compile this because I'm not compiling this on a fast computer, yet I had been able to compile it on the same hardware before. Emerge does not show an actual error when the compile dies, because the compile itself hasn't failed, but dmesg shows that it's been killed because it ran out of memory, which even though the hardware isn't that fast did not make sense that it could fail with 8gb available. The problem is that qtwebengine is bundling the ninja build system in it, and ninja has been the cause of OOM even on systems with a lot of memory because ninja will detect how many cores you have, and add 2 to the parallel jobs option, disregarding your MAKEOPTS argument. I have seen no workaround to this ninja build system steathily killing the compile halfway through beyond actually disabling all but 1 core so that it doesn't exhaust memory by trying to allocate for {cores}+2 threads by running: # chcpu --disable 1-3 I'll be compiling this garbage for the next few days thanks to this crap.

>>2920 that's why I don't use qt shitware. qutebrowser is clearly not worth the pain of compiling qtwebengine and all the other qt-dependencies. I fail to understand how a supposedly "minimalist" browser can take longer to compile and at the same time require more heavy dependencies than bloated browsers like palemoon et al.

>>2922 Personally I like having the vim keybinds and being able to just set config options for just about anything, like default darkmode on all websites or having all new instances start in private browsing. I'm going to compile surf and see how it is but as I had already read, surf doesn't have tabs so that will take getting used to. Surf still relies on webkit2/gtk+ so I don't know how much of an improvement it's going to be. It's a fucking shame qutebrowser relies on Qt.

Took almost no time at all to compile surf once I realized I had to compile gcr-3 with the gtk flag, immediate issues I noticed: >actually uses more cpu time if scripts and plugins are enabled for websites like this one >cannot handle embedded videos like uploads here >cannot handle html5 video players >streaming video from sites like cytube is completely non-functional I'll have to see if I can get plugins working.

Looks like installing the right gst plugins fixed the webm and mp4 files not playing, as well as embedded jewtube video on surf, must still be missing a gst plugin because html5 video isn't playing. Zooming is pretty fucky though, it magnifies everything on the page instead of just increasing font size.

>>2930 Jealous Windblows-fag here, from what I understand, Suckless software like that is designed to work in tandem with other (ideally suckless) software. Something like Tabbed would probably remedy that, as well as providing a tab system for any program that can and is allowed to use it.

>>2960 There is a tabbed browsing solution for surf, not that it really matters much if there's tabs, in my case there's so few sites I even bother to visit now that tabs are largely irrelevant. Bookmarks are more important and I still have to see what the options for that are, there is one thing I forgot about and that's the ability to create keybinds to run scripts on pages in qutebrowser, which I had just one to run youtube-dl on urls from invidious. I'm sure I could do something similar in surf but it would take more work to do that since I don't think surf doesn't have vim keybinds or the ability to tag urls with key combinations.

>lxde being removed 2021-04-14 that's pretty gay.

>>3286 That's because it doesn't have an active maintainer right? It seems like every year there's a bunch of changes being made that end up fucking my setups. >xdm being retired forcing changes to be made to existing init scripts this broke my laptop initially even though i didn't use xdm to start X, I always used xinit/startx from a console depending if I wanted to conserve battery or not, eventually I figured out I couldn't use xinit at all and had to use startx >removed some old video driver I need for an IBM thinkpad because it was too niche I never did get X working on that thinkpad after that >xorg-server now uses logind by default, can still use suid by opt-in only I'm getting pretty tired of this.

>>3288 As far as i know LXDE being unmaintained is mostly a non-issue seeing as the devs stopped working on it because there really wasn't much more they could do with it. Not out of the fact that it was bloated, just out of a fact they did so well with it there wasn't anything left to do aside very very minor bug fixes. So instead of sticking it out they decided to rewrite the whole thing but ignore all the nice agnostic factors that made LXDE great. LXQT isn't even that good compared to LXDE, the amount of deps required to manage is the obvious thing. It seems like gentoo maintainers have been going on a killing spree with anything that goes unupdated for a year. even if it really doesn't need updates. What annoys me the most about them removing LXDE is that they are removing agnostic tools along side it like lxappearance and pcmanfm. which unlike the other parts of the DE were you could somewhat make an argument to it's removal, i fail to see the reasoning behind removing those. even when "unmaintained" all lxappearance really does is serve as a nice gui tool that creates and edits xconfigs. lxde as a whole doesn't really do much that warrants high maintenance. >It seems like every year there's a bunch of changes being made that end up fucking my setups. >I'm getting pretty tired of this. at least they can't delete local ebuilds or unoffical overlays. I'll probably have to switch to using an overlay in the future because of this shit.

>>1299 Because it's the only way many of them are even capable of being honest with us about how fucking garbage their software is.

Anyone able to have a purely JACK audio setup working? Of what was actually able to use JACK for the audio server it was fantastic, but I always ran into the issue of not enough of what I used would support it. Meanwhile everything adopted pulseaudio in a heartbeat, which I always see errors in dmesg for.

>>3363 I don't think enough software supports jack to get pure only jack setup to work. closest setup I've seen is using alsa loopback devices with apulse.

>>3336 >It seems like gentoo maintainers have been going on a killing spree with anything that goes unupdated for a year. even if it really doesn't need updates. Talking from the future here, but LXDE didn't get removed. That said, being unmaintained upstream isn't just about getting new features, it's about getting support when code eventually rots and becomes unable to compile/function properly on modern systems. LXDE currently requires very few patches, but I'm sure if a bug or two appear that aren't trivial to solve, they'll consider it for removal again. You're usually free to contribute to keeping it alive, if you can fix the issues. Same shit happened with opentmpfiles, as the latest version had significant bugs and had to be masked, nobody stepped up for over a year to either fix those or release a new upstream version, so they decided it wasn't worth keeping around.