IRC channel logs
2025-10-23.log
back to list of logs
<jab>sounds good! You'll probably want to hang out in #irc when you install. The installer may have a few hickups. <yang3>jab Does it only work with thinkpads which you listed? A colleague has lots of older notebooks, but he might not have the ones you listed... <yang3>Or is the only requirement an integrated CD/DVDrom? <jab>well I re-installed the Hurd on my T60. I may need to get a different hard drive. I saw some weird hd0 errors on boot. <yang3>jab I ment that the colleague might have different brands, I could ask him if he has any with integrated CD/DVD Roms, if he can't find Thinkpads listed in his collection (then I'll get it from EBAY) <jab>yang3: I know that the older Thinkpads work. There are some other options laptops that work as well. <yang3>I guess there isn't a list of supported hardware on the wiki? <jab>yes. I'm trying to give you a link. <jab>hurd.gnu.org that wiki is SUPER OUTDATED. <jab>I'll try to make wiki.gnu-hurd.com have an updated wiki tomorrow. <jab>It's silly that we don't have an up to date wiki anywhere at the moment. :( <yang3>ok I can read the text on the cgit, it's fine <nexussfan>yes, we do need an up to date wiki, is sceen wiki always up to date or not? <sneek>Welcome back nexussfan, you have 1 message! <sneek>nexussfan, youpi says: (it's not a one maintainer project :P) : well, look at the commit stats in the past ~2yrs, it's essentially his commits <nexussfan>youpi: Before XLibre, the original author worked on Xorg (official) but when he got banned from FreeDesktop he moved all of his patches to Github creating xlibre. So yes, old commits will be from him <damo22>not sure why you refer to him as "original author", he didnt write bulk of X, as we have seen he can barely write C <nexussfan>damo22: the original author of xlibre, i forgot his username <damo22>if i fork linux im not the original author of linux <damo22>if you fork it and apply N patches then you are the original author of N patches on top of an entire legacy of contributions <tux0r>linus is not the original author of linux either, as it started as a software suite on minix IIRC <tux0r>how many patches make you the author, you say? <damo22>you are the author of what you commit in the history <tux0r>if your code is more than 50% of the fork, are you still not the original author of your fork? <tux0r>sounds like a ship of theseus argument, actually <tux0r>fork a software, replace all of its code over time, yet you are not the author, because the original code wasn't yours <damo22>if i attain a ship, i didnt build it, but i can modify it to make it go faster if i want, but i wouldnt call myself original ship builder <damo22>its still a faster ship, and you can be proud of your changes <azert>Linus is certainly not the only author of Linux, but he is its creator and the leader of the kernel community <azert>He also own the trademark afaik <azert>you could certainly say that that one guy is the creator of XLibre although he is not the creator of X11 and he didn’t author a significant portion of either <azert>so, if I fork Linux and start to work on a software i call Merdix that for instance can transparently share memory and GPUs over several machines to supposedly allow the training of giant models that don’t fit a single GPU. Then I’m the creator of Merdix <identity>i doubt the author of xlibre did anything of interest apart from making a lot of noise and being very xenophobic <kilobug>my feeling is that he is a nazi apologist weaponizing xlibre as a way to grap attention and spread his murderous ideology <sneek>Welcome back kilobug, you have 1 message! <sneek>kilobug, nexussfan says: XLibre actually works AND the maintainers are actually nice and as long as you actually help Xlibre in some way they won't get mad at you <damo22>kilobug: thats a pretty severe description, do you have evidence for that? i dont know him at all <kilobug>lol no that's not me (but then, why would it be so wrong to collect furries ?) <identity>kilobug: because it is a vaguely socially acceptable way to backhandedly say something homophobic <identity>i am yet to see somebody call somebody a furry pejoratively for any other reason. most furries are queers <damo22>i dont care if someone writing free software is gay or not, all i care about is the quality of the software and if the person is rational and reasonable to get along with for the purpose of extending the software. <kilobug>I sure dont care if someone writing free software is gay or not, but having an actual nazi in charge of the maintenance of a software de facto means gays or jews or gispies or disabled or... can't interact safely with said maintainer, which is very exclusive and defeats your "i dont care if someone writing free software is gay" point... indeed, we shouldn't care about people's sexual orientation (as long as <kilobug>we are not dating them, at least), but that's why we shouldn't put people who say "lets murder all the gays" in any position of power <identity>the only reason many do not care if someone is gay or not is because the free software community is full of straight white men; every time somebody who is not a straight white man shows up, or goes mask off, suddenly there is a problem <damo22>which is why i asked if you have evidence for such a wild claim <identity>pretending to be colourblind is not going to help with that <identity>damo22: this is not a wild claim: you can go read stuff written in the xlibre repo. i do not think that claims such as «make {x11,america} great again» and actions such as performatively replacing the code of conduct with «404» are defensible, and i did not dig that deep <damo22>that sounds harmless to me, i think youre reading way too much into it. From just that, he sounds more like an inexperienced programmer forking a project that is way too big to handle for one person. <identity>damo22: … you know what dog-whistles are, right? <identity>a dog-whistle is coded language used to signal allegiance to certain groups without saying it out-loud, appearing bening to those unaware; basically a way to say the n-word in polite company. «make america great again» is one example, which is a slogan used by ronald reagan and later donald trump to basically mean «i hate immigrants» <kilobug>damo22: when someone whose motto is «make america great again» is turning his country into a dictatorship and have his allies make nazi salutes, and you say «make x11 great again», it's not innocent... and he didn't remove the code of conduct by mistake, he actually opposes them, and if you search for him, you'll see plenty of messages in which he says things that anyone who grew up in germany, or even <kilobug>in europe, would recognize as nazi apology <damo22>i will be avoiding xlibre because i dont understand this stuff and i dont want to get involved <damo22>plus it appears the quality of code is not good <p4r4D0xum>thank you kilobug , identity for letting me see the issues we have here in hurd <identity>the intent never was to write good code or improve the codebase, the intent was to go to distro maintainers, urge them to package it and then go ham when they say «wtf no» <p4r4D0xum>could you also passionately write some code too? <gnu_srs1>What is the status of x11? Who is responsible and why is it stalled? <gnu_srs1>If the xlibre author did not contribute with important patches, let's continue with x11! <damo22>gnu_srs1: from what i heard, x11/xorg development stopped because the actual people who used to write it realised the design was wrong for modern day desktop and came up with a protocol for desktop compositor to talk directly to kernel drivers, wayland <p4r4D0xum>gnu_srs1: the whole idea about wayland was avoid the mistakes of X11 <identity>gnu_srs1: x11 development is not «stalled», but it is in maintenance mode until wayland becomes widespread enough to just drop x11. there have been many shims made to ease the transition (xwayland, wayland-satellite, an X11 WM that can run with Wayland as the backend is currently in development) <p4r4D0xum>for many years X11 developers found work around the bugs than just fixing them <p4r4D0xum>to the point, that debugging X11 was a nightmare <damo22>it wasnt a bunch of bugs, it was the design <identity>the difference between xwayland and xwayland-satellite is that with xwayland-satellite the WM itself does not need to have any code support X11, xwayland-satellite being a «userland» x11 shim <damo22>if the people writing the graphics stack think wayland is better for free software than xorg, i trust they know best <damo22>i dont like the fact that every client of X knows all the keypresses regardless if its in focus or not, and that cant be fixed in Xorg <gnu_srs1>I've heard that Wayland has some serious design flaws. <damo22>you can have remote desktop display, but not with ssh -X <jab>identity: just an FYI, the best friend I've ever had was an immigrant. He was a retired german CEO. He was the most loving man I have ever met. I could not have asked for a better friend. I am also voted for Trump. <jab>I say that to say that, I want to use Xlibre, because it is a maintained version of X. My understanding up Xlibre is that is has 2,000 more commits than X does. <jab>If the Hurd ever gets wayland going, then I'll use wayland. Because like you, I do believe that Wayland is a better design. <jab>But since, we are stuck on X, I'd like to use a maintained version of it. <jab>A final point, I haven't read up on why people believe this Xlibre guy is a maintainer...but I highly doubt it. Most of the nazis are actually dead. People said that Charlie Kirk was a nazi, and as someone who went to see him go speak, finding out he was killed for speech was stupid. <identity>jab: xlibre is not a maintained version of X11, it is just a political stunt with some really bad code attached <jab>identity: can we check the xlibre repo? Lunduke told me that Xlibre on its first release, had 2,000 additional commits on top of X. If that's true, then wow! Sounds like more activity in Xlibre. <jab>Also, time will tell. Redhat was sponsoring the development of X for years. It seems like Xlibre and X.org do NOT have corporate sponsorship. So probably wayland will win in the end. <youpi>jab: if those 2000 commits are just crap, it's no good :) <damo22>number of commits dont necessarily mean better code, eg i can make 1 commit to gnumach that will break the kernel <kilobug>jab: you can't judge software quality by number of commits, from what I've heard most of those commits are lame, even undoing each other, and the overall quality is very bad, and it's all for the same person... but that's not really the main issue, the main issue is that the guy is a nazi apologist, and xlibre is a just weapon he uses as a pr stunt <youpi>jab: I *very* often see people forking a project at do a lot of commit only to change the codestyle and arrangement <youpi>which is a lot of commits for nothing actually <identity>most of the commits are re-applying commits that were reverted upstream due to numerous issues. other commits are code style and thinking that ^ is exponentiation and not bitwise-xor. not only are they are openly nazis, they are also bad programmers <gnu_srs1>xorg was rewritten rather recently to use a modern build system. That was a major effort. So what hinders xorg to be slimmed down to be more efficient? <damo22>im sure there was a reason they had to start from scratch <identity>sometimes nothing *is* better than something <damo22>why do you need a proxy server for the screen? <damo22>maybe it makes sense if you are on a mainframe sharing a bunch of terminals over the network <youpi>gnu_srs1: the main concern of xorg is that it lacks isolation between applications <youpi>and that cannot be fixed without changing the protocol <kilobug>from what I understood, Xorg still has a huge "technical debt", it's a very old project that saw many changes in paradigm and technology and had to constantly adapt, so it's heavy-weight and cumbersome in many ways (and XLibre isn't changing that), restarting from scratch does make some sense... <youpi>so better restart from scratch, and also leveraging the knowledge about rendering layers <youpi>I wouldn't say that Xorg has that much of a technical debt, the code is good <youpi>but it bears the legacy protocols <kilobug>but then I don't know if wayland is really much better, and I just don't feel like having to change my habits and tools and window manager and relearn everything for no real gain, so I'm still using xorg for now... <youpi>in terms of rendering pipeline etc. it's definitely better <Arsen> 11:12:09 <damo22> gnu_srs1: from what i heard, x11/xorg development stopped because the actual people who used to write it realised the design was wrong for modern day desktop and came up with a protocol for desktop compositor to talk directly to kernel drivers, wayland <Arsen>waylands improvements don't come from "talking directly to kernel drivers" <kilobug>I'm neither fully pro or against wayland, I'm mostly lazy to do the switch and waiting to see how things go ;) <Arsen>they come from how events and images are exchanged <youpi>notably because it was defined precisely by the people that were developing xorg <Arsen>(X programs also directly talk to the kernel for GL, of course, and X no longer uses any userspace drivers, that's not been the case for ~10 years, and has started to been phased out for something closer to 20)\ <kilobug>I've heard stuff like if the window manager dies in wayland, all clients die too... that's not the case with X, I can restart or upgrade or even change my window manager without restarting everything, if it's true it's definitely a regression, but sure not that much of a big deal either <identity>last time i switched to X11, it was a death by a thousand cuts. not only are there theoretical issues, but there are practical issues that i have to deal with as a user, like endless artifacting <Arsen> 11:45:16 <kilobug> I've heard stuff like if the window manager dies in wayland, all clients die too... that's not the case with X, I can restart or upgrade or even change my window manager without restarting everything, if it's true it's definitely a regression, but sure not that much of a big deal either <Arsen>if X11 dies, everything dies <Arsen>with wayland, if kwin_wayland dies, programs can (and do) reconnect <kilobug>if Xorg dies, everything dies, yes; not if the window manager dies <kilobug>but then maybe it was a mistake or I misunderstood the point <Arsen>there's no reason to conflate the window manager process and display server process in wayland, but in practice they are often the same process, I will grant that <Arsen>but, again, clients can reconnect more easily w/ wayland <identity>kilobug: it is more of a problem of wayland as it is implemented, not with the protocol, frequently the wm and the display server are the same program <Arsen>(gtk and qt do that transparently) <kilobug>oh, I see, thanks for the clarification <gnu_srs1>so how much effort is needed to change the protocol xorg compared to creating something new, like wayland, that have design flaws. <Arsen>would you prefer if wayland was called X12? <jab>in fun news, my T60 is still being a pain. I need to try to install a new hard drive. Also, does the installer detect SSDs ? <Arsen>because in effect, it is X12 <jab>Arsen: I believe there actually was something called X12. :) I don't think any code was ever written. <Arsen>I know the xlibre reactionary originally spammed forums about X12 before forking X11 and fixing typos in order to call it xlibre <jab>huh. I learned something new today! <jab>also, one thing that I want to try out today... <jab>is to use x11 forwarding over ssh. <kilobug>ssh -X is very useful, and so is Xpra, I wonder if there is a wayland alternative to that <Arsen>because X11 is only network transparent in name, it's kinda horrible - modern toolkits draw large pixmaps very often <jab>I want to physically at me T400 OpenBSD machine, but I want to ssh into a Fedora talos II power9 machine. <Arsen>X11 assumes that's all over a fast link and doesn't compress anything <jab>is that just as simple as "ssh joshua@talosIPAddress" <Arsen>ergo it's actually horribly slow with programs written since "one window per program" became common practice <identity>kilobug: again, RDP is likely to work just fine <Arsen>you don't even need startx - it uses your local X server <Arsen>kilobug: waypipe is quite good <Arsen>it's dedicated to remoteing so it does compression on the pixmaps it sends <gnu_srs1>X11 has/had? compression over the network. I just don't remember the name. <kilobug>but isn't RDP much more like VNC, allowing to access a whole desktop remotely as one big window ? ssh -X and xpra gives you just one software running remotely but behaving as if it were just a local program, with its own window(s), etc. <Arsen>I've only ever used RDP "rootfully", I don't think you can use it "rootlessly" <Arsen>damo22: oh that probably does help a lot <Arsen>I didn't think of it at the time <kilobug>xpra is actually very neat, it's like screen (or tmux) but with graphical app <identity>kilobug: if you want that, waypipe should do, though i did not try it <kilobug>you can attach/dettach from various computers and it keeps running, I don't know if there is a wayland equivalent to that <Arsen>kilobug: seems that the Xpra folk want to make it work on wayland? <kilobug>Arsen: looks promising indeed, thanks :) <Arsen>(not aware of any other alternatives though, to clarify) <Arsen>I kinda wanna add a more native wayland backend to emacs, the current one relies on GTK3, and GTK3 1) lacks the reconnect logic, and 2) can't open multiple screens or disconnect from screens (if you attempt that, it crashes) <Arsen>1) is fixed in GTK4, 2) isn't AFAIK <etno>Arsen: the video of Daniel Stone is great, thanks! <Arsen>where is the gsync api documented? <Arsen>ACTION is trying to understand the lowlevellock.h implementation for mach <almuhs>do you know any currently-in-use trivfs translator? <etno>I don't have the mail setup required to send a patch, sorry <Arsen>can someone get me the sysnames of a hurd glibc build (from config.log) please? I don't have a toolchain currently set up for hurd <Arsen>hm, hurd uses ints as pthread_t? <Arsen>hmm, how I might be missing something, trying to reuse HTL I get tlsdesc.h failing to build because: <stdin>:9:78: error: invalid use of undefined type ‘struct pthread’ <Arsen>... and I don't see any hurd tlsdesc.h or similar logic to avoid this step <Arsen>aha, but there is a tlsdesc.sym! <Arsen>hm, this seems like a problem, sysdeps/x86_64 is searched before sysdeps/managarm <youpi>you can make the Depends file do what you want <youpi>I don't remember what's supposed to be between sysdeps/<arch> and sysdeps/<os>, but you can always add a sysdeps/<os>/<arch> <Arsen>I don't see an equivalent in the hurd sysdeps dir, though <Arsen>yeah, I have tried the latter also, but it's also after sysdeps/<arch> in the path <Arsen>... which seems odd, I'd expect the opposite, so I suspect I screwed something up <Arsen>there we go, needed to adjust the base_os computing logic to handle the new port <Arsen>hm, generally, where is mach documented? <Arsen>(currently looking to understand __thread_get_state) <Arsen>aha, found it in the reference manual, but it refers to the codebase, time to pull from savannah at a snails pace <Arsen>ah, I see, here it's used to access %fs on the thread ID, which, in turn, contains the pointer to the TCB