<alezost>cehteh: sorry, I'm completely lost after all. Apparently it relates to the amount of data, but I don't know what to do about it (I don't even understand if it can be fixed in Geiser or if it's a general Emacs problem or something else) <cehteh>even with insanely high gc-cons-threshold emacs doesnt need that much memory and still fails <alezost>cehteh: btw you can check "M-: gcs-done" to find how many times GC has been performed ***jonsger1 is now known as jonsger
<ng0>but please don't comment on the style yet, there's lots of hate in the patch so far.. I just want it to pass the build after months. <ng0>damn, also the patch must be improved. <ng0>just $CONFIG_SHELL doesn't work <ryanwatkins>Hey guys, must I somehow convert a tarball to a guix package to install it? I am trying to obtain git-annex <ZombieChicken>Any way to know if my emails to the mailing list have been filtered somehow? I've sent a few emails to the mailing list (one was to help-guix and the other guix-devel) and neither have arrived after over 24 hours <adfeno>When building with GCC, which archtecture-related options don't break reproducibility? <jmd>I would have thought that most of them do not. <adfeno>I mean, I have read somewhere that there are options for GCC which break reproducibility <adfeno>Pardon my indirect recomendation of it, but I recall seeing this in a dcoumentation related to Gentoo. <adfeno>From what I gather, it seems -march=[Archtecture] is more reproducible and produces average optimization than the usual -m[Arch abbreviation]. <adfeno>I'll get the reference so that I make sure that I'm not geting confused. <ng0>I've just sent obconf, so openbox will not feel so lonely anymore. <ng0>also I'm slowly making my owned domains obsolete, that's why the commit email and the copyright email currently differ <ng0>my plan is to fade them out over the course of this year <ng0>I'm unhappy with the ", and much more." but I haven't used openbox in years <adfeno>Sorry, I meant `-mtune=` not `-march`. <ng0>civodul: how easy or difficult do you think would it be to write an guix system test for lxdm? Is there one which already tests slim? <ng0>I think slim is so much default that it's implicitly tested with the system tests. I see what I can come up with <ng0>ideally I want lxdm to be an option in addition to slim, because I'm very unhappy with slim <civodul>ng0: we don't have any tests for graphical login managers yet <civodul>essentially using the OCR thing to synchronize <ng0>like you did with the crypt fs test? <ng0>okay, I see what I can do. <civodul>yeah using 'wait-for-screen-text' to catch the prompt <ng0>I also need to write tests (and I think services and more) for secure_delete and plymouth. I don't know which one of those will be harder, for secure_delete I have already figured out that the way I want to integrate it will take some time as my inspiration system uses systemd. plymouth is pre-boot, should be easier. <jmd>Why do they call it "Plymouth" ? <ng0>hm. notmuch-vim (the vim version, or the vim folder in the source) needs ruby-mail. I have no idea about ruby.. All ruby software I tried to package so far failed in ways I do not understand. Do we have anyone working with ruby? Or anyone interested in packaging that gem (mail) ? <ng0>but gem mail looks doable. everything else differed from the expectations of the build system <ng0>isn't the description for package "vifm" much too long? <rekado>ng0: we usually don’t touch package expressions that have already been committed for cosmetic reasons. <rekado>ng0: the description could be shortened when you’re working on the package anyway. (In a separate commit.) <ng0>we've done this before though <ng0>for awesome description was way too long <ng0>but i was just pointing it out, it doesn't bother me <ng0>it seems like something got fixed since september which made bbs100 finally build without a loop :) <dvc>what do people think about bug report mounting "my-home" to /home cannot work again? <sneek>Welcome back dvc, you have 1 message. <sneek>dvc, civodul says: normally Guix can already be cross-compiled the normal way <dvc>I screwed up my guix install - so I'm working on fixing that first, but if the report is legitimate it should be fixed ASAP... Anyone else had difficulties with recognizing ext4 labels? <dvc>I just can't explain why there isn't a backtrace or something from the initrd in the bug report :/ <janneke>if i like a previous emacs /gnu/store/shp26ngj1lmlikl0in8d74i9fply6gri-emacs-25.1 <pareidolia>Does anyone have experience running Nix against the Guix daemon and store? <pareidolia>Some derivations use the builtin:fetchurl builder, which the daemon apparently lacks <jmi2k>Can't xcursor Vanilla DMZ be packaged for Guix because of the license? It's CC-BY-NC-SA. <jmi2k>(well, instead of packaged, read *shipped*) <rekado>NC makes it non-free, though ISTR there are exceptions for non-functional artifacts such as artwork in games. <jmi2k>I love that theme and I'm going to package it for myself, but if I can add it better. Well, if I decide to try, I'll put a warning about the license. In this case, what should I put in the license field? <adfeno>One must question if these types of data are functional or not. <adfeno>I know this is somewhat far stretch but: there are two possible interpretations: <adfeno>(a) It's non-functional in the sense that the user doesn't need the cursor theme to use the desktop.... <adfeno>(b) It's functional because he can't see where his mouse cursor move if he would have no cursor theme at all. <adfeno>Although more input and resarch is need: I would stick to the worst scenario just in case: It's functional. <jmi2k>adfeno: as xorg has a default theme, I think it's not functional (it's not *vital*). I'm not good at detecting those slight details, so discussion is welcome! <jmi2k>what should I put inside license if none matches (if I want to keep the package for me). Is there any placeholder or should I put any license to make Guix not complain? <ng0>i think one example could be openssl <jmd>What's the status of btrfs in GuixSD ? <ng0>last commit i saw it's being worked on <ng0>or was already added <dadinn>I am trying to install GuixSD under KVM but I am not able to boot from the image file in the xz archive <pareidolia>You can launch a KVM/quemu VM directly with that command <dadinn>pareidolia: guix is not set up on the Fedora box, only have KVM on it <dadinn>pareidolia: want to start a new image with virt-install --cdrom image <dadinn>but the GuixSD image is not booting, I get the message "No bootable device" <dadinn>pareidolia: Also "Boot failed: could not read from CDROM (code 0004)" <civodul>jmd: btrfs support landed a couple of days ago :-) <dadinn>pareidolia: actually i am using virt-manager GUI <jmd>civodul: Is it complete? <pareidolia>Anyone using the packaged Nix on the Guix store? <amdmi3>there's problem with guix package list on the website <dvc>jmd: pretty much... it won't currently work for vm's dough... <dvc>well and stuff like btrfs raids is untested and mounting a btrfs subvolume as root fs won't work unless the mount-options are kernel-arguments <pareidolia>When I tried a fresh install yesterday terminal fonts were fugg buttly <jmd>I wish that mkfs.ext? and mkfs.btrfs had a consistent calling interface. <jmd>dvc: How can I retrieve the UUID of a btrfs system ? <dvc>civodul: creating a vm replaces the default file-systems with new ones. this is currently hardcoded to be ext4. then btrfs-progs aren't available in qemu-image and I'd like to avoid adding it if not necessary, and format-partition in guix/build/vm.scm, not sure if that will work for btrfs <dvc>jmd: (partition-field-reader read-btrfs-superblock <dvc> btrfs-superblock-uuid) "/dev/sda1") <dvc>jmd: can you make it a consistent calling interface by adding procedures to guix/build/file-systems.scm in the ext2 and btrfs sections? <dvc>then adding a %file-system-formatters list and a format-file-system procedure or something like that? <Apteryx>Could anyone tell me: In your IceCat "about:config" page, do you have a "general.appversion.override" entry, and if so, of what value? <dvc>civodul: mounting "my-home", wouldn't it show a backtrace if it where initrd related? <adfeno>This is my general.appversion.override <dvc>I believe the issue is before it enters the initrd, or did I make an error in reasoining? <Apteryx>OK. Any idea why we ship IceCat with overriden values? It risks breaking compatibility with some sites. At least XFINITY billing web page didn't work until I set it to '5.0'. <adfeno>Apteryx: Hm interesting question. <adfeno>The XFINITY seems to be evaluating the browser version. <civodul>dvc: right, 'guix system vm' replaces the root file system, but that's fine no? <adfeno>Apteryx: Error due to XFINITY site developers? <civodul>dvc: also, not sure what "my-home" you're talking about :-) <dvc>yes - it just doesn't work if you'd like to use btrfs. there's nothing wrong with using ext4 <dvc>civodul: the bugreport! :) <Apteryx>The check it does is: if("Netscape"==navigator.appName&&navigator.appVersion<"5"&&window.crypto) <Apteryx>I've tested Firefox ESR 45.6 on Windows and it returns an appVersion of 5.0 <dvc>mounting "my-home" to /home cannot work again. <Apteryx>So it seems our setup which does something differently (because of this general.appversion.override thing) <adfeno>Apteryx: I suspect it's an error on part of XFINITY due to general.useragent.override being "Mozilla/5.0 (Windows NT 6.1; rv:45.0) Gecko/20100101 Firefox/45.0", and this one having "5.0" there. <Apteryx>navigator.appVersion and navigator.userAgent are not related apparently (I asked on #firefox). <adfeno>I'll make a quick JS to try reproducing the test. <adfeno>Apteryx: Is that simple line the whole test? <Apteryx>adfeno: Yes, that's the whole test. After it goes on trying window.crypto.random(32) and errors on that. <Apteryx>adfeno: You can press Ctrl-Shift-K to get the web console in IceCat and type this at the prompt (bottom of screen). <adfeno>Apteryx: As far as I can tell.... <adfeno>... It seems to go to the "else" part of the test, which wasn't informed to me yet <Apteryx>adfeno: strange. Could I try your javascript snippet here? <adfeno>The test snippet you shared tells that it will only run if appVersion is bellow 5, <adfeno>So the error must be somewhere else, because this part isn't touched. <Apteryx>This is a string test, so "45"<"5" is true. <Apteryx>Got confused by that before too, until someone pointed that out ;) <adfeno>The developers of that site have everything wrong. <Apteryx>Well it's a hack, but a hack that should work given that the usual appVersion string goes like "5.0 (Windows)" <dvc>civodul: <boot-parameters> and <menu-entry> are essentially the same thing or not? I think we can pass <boot-parameters> to a grub bootloader or an uboot bootloader. Or do you think it would be better to have each bootloader specify their own <menu-entry> record? I think that <boot-parameters> neatly represents a system generation and that a bootloader should take a list of system generations. instead we currently have logic in <dvc>system.scm that converts system generations to "grub specific" menu-entries. <civodul>dvc: yeah there's a correspondance between the two, but hey have different purposes <Apteryx>adfeno: I'll send a mail to guix-bugs and question why we should mess with the appversion in the first place. <dvc>civodul: alternative: Is <menu-entry> something that should be moved to system.scm and common to all bootloaders? <Apteryx>adfeno: I'm suspecting it to be caused by an extension (add-on) since there's nothing suspicious done at the package definition. <ng0>I'll test obconf in a minute or so, I just need to re-configure a test-system <dvc>or alternative2: We pass <boot-parameters> to the bootloaders and they convert <boot-parameters> -> <menu-entry> internally, while users can still specify additional menu-entries in their operating-system definition? <dvc>another small change to the operating-system semantics I'm thinking about and would like manolis feedback on is changing the kernel parameter to take a gexp instead of a package. So it would be used like this: (kernel #~(file-append #$hurd "/gnuhurd.gz")) <adfeno>Apteryx: Doesn't seem so, because the variable name isn't marked as bold. <janneke>has something changed wrt using guix-prettify? it errors out in my init.el <Apteryx>adfeno: Ah, nice to know, I didn't bold entries meant they had been altered by extensions! <adfeno>Let's... I have User Agent Switcher here.... <adfeno>.... I'll see if it changes useragent variable I mentioned <adfeno>Apteryx: Yep... User Agent Switcher changed the general.useragent.override variable... <adfeno>... and in about:config, it was marked as such. <quiliro>adfeno: i am about to report a lot about desktop.scm <quiliro>adfeno: how are you and what are you up to? <adfeno>I'm currently helping Guix find where some packages from Google Code might have moved to. <janneke>civodul: thanks for the :debug cluebat. Also, the separate debug-info file documentation is excellent. However, I cannot get it to work, now I get: <janneke>file /gnu/store/1wsgsls4n6fc7dji8fffbz6ahwk870mc-emacs-25.1/bin/.emacs-25.1-real <janneke>Reading symbols from /gnu/store/1wsgsls4n6fc7dji8fffbz6ahwk870mc-emacs-25.1/bin/.emacs-25.1-real...warning: the debug information found in "/home/janneke/.guix-profile/lib/debug//gnu/store/1wsgsls4n6fc7dji8fffbz6ahwk870mc-emacs-25.1/bin/.emacs-25.1-real.debug" does not match "/gnu/store/1wsgsls4n6fc7dji8fffbz6ahwk870mc-emacs-25.1/bin/.emacs-25.1-real" (CRC mismatch). <civodul>janneke: ooh. could you try with --no-grafts? <janneke>ACTION *really* wants to go on with mes...a broken emacs in not fun! <rekado>janneke: mes is so cool! I also mention it as a bootstrapping project on bootstrappable.org <janneke>currently, i am working on a C compiler that runs in Mes and Guile <janneke>it can almost compile a minimial version of the mes prototype in C <janneke>$1 = (re_char *) 0x2e8985f <error: Cannot access memory at address 0x2e8985f> <ng0>wrt googlecode: is someone here involved in other distros? For libmp4v2 the archlinux maintainer already said they don't know about the new upstream. <adfeno>ng0: Oh my.... that's unfortunate. <ng0>upstream is... silent, only got reply from one contributor who wishes to stay anonymous <ng0>and the situation is basically fork and forget <ng0>so I hope someone knows about a fork which might be usable and reliable for distributions <ng0>the software itself is no problem, and the archive of googlecode is still around <ng0>so until its completely gone it could be added? <adfeno>It's like my failed communcation attempts with khal and colab-discourse-plugin projects. No reply, and it has been months since I sent patches <adfeno>This github.com thing is really killing these projects, now everyone seems to assume that you must have github account to contribute. <ng0>I want to test pagure in some server environment.. they say that it can federate forks/pull requests across instances, or at least that was/is a goal, which is super nice <ng0>we have the redis-service now, so I only have some more services to write for pagure itself <lfam>adfeno: You tried to contribute to khal outside of github? <adfeno>lfam: Yep. Sending the patch to an email address of a developer. <lfam>I'm slightly involved with that project. I'd be happy to submit the patch on your behalf through github. <adfeno>It's actually a tool to help on an issue they had: Support for alarms. <adfeno>I even made a test to see if I could get a reply back (changed the license to GPL), but even sending it second time, got no reply. <lfam>Well, let me know. I rely on khal so I'm always interested to see it improve <adfeno>Do you mind if I paste the script on paste.lisp.org? <lfam>adfeno: They have a freenode channel at #pimutils, by the way. That's the organization that khal is developed within. <ng0>I have a webdav client in my package path, but haven't been able to build it with fixed ssl support I think (or some other issues is still to be fixed).. is anyone interested in fixing it? <lfam>adfeno: Let's take finish discussing it there. <dvc>janneke: can you explain more about how mes works? is it both a scheme interpreter and a c compiler written in scheme? <janneke>dvc: the idea is to minimize [or even remove] the bootstrap binaries <janneke>so primarily it's a very small scheme interpeter, written in hex/assembly <janneke>that scheme interpreter must be able to run a C compiler; eventually we need to compile gcc <janneke>i have prototyped the minimal scheme interpreter in C <dvc>so you are planning on reimplementing gcc's build system using mes? or how would that compile without autotools? I'm clear on the goal - but I'm having difficulty understanding how it's going to be accomplished <dvc>so the idea is that mes and tinycc together can bootstrap all of guixsd? <rekado>dvc: the C compiler in mes might be enough to build a less demanding C compiler. <rekado>GCC is written in C++, so a mere C compiler won’t be sufficient. <janneke>we may have to take several steps to make GCC more bootstrappable <janneke>i'm very open to smart ideas to avoid unnecessary work ;-) <efraim>I assume targeting x86_64/x86 first, and then maybe the other architectures? <rekado>one of the bootstrappable.org projects is to see if GCC 4.7 (the last version to not require a C++ compiler) can be built with a much simpler compiler like tinycc. <rekado>unfortunately, our tinycc package is currently broken in Guix, but we also have pcc. <janneke>efraim: yes, i'm currently using x86 as a first target <lfam>I surveyed the tcc mailing list and repo while fixing build failures in the last core-updates cycle. It seems like a chaotic project. We might have to maintain our own fork. <janneke>rekado: ah nice! finding out that, of course was also on my list :-) <rekado>lfam: there seem to be quite a few forks out there. <lfam>What do you mean by "remove Guix"? <lfam>Remove the full installation, remove the `guix` package that users invoke, something else? <dvc>rekado: thanks! I did. the link to a programmers worst nightmear was interesting... I have only considered this to be a theoretical problem and not one that could plague a real compiler like gcc. <dvc>maybe time to reconsider :) <adfeno>GNUnymous_IRC: lfam asked because you can have Guix for various users. <adfeno>And you can also remove the full installation. <adfeno>lfam: Perhaps its better to answer to both cases. <lfam>GNUnymous_IRC: Assuming you installed Guix on another distro by following the instructions in the manual, section 2.1 Binary Installation, then all users are using the `guix` package in root's profile. As the manual says, "The guix package must remain available in root’s profile, or it would become subject to garbage collection—in which case you would find yourself badly handicapped by the lack of the guix command. In other words, <lfam> do not remove guix by running guix package -r guix." <lfam>I haven't tried this, but it does work for other packages: As root, `guix package --remove guix --install guix` <adfeno>dvc: "Programmer's worst nightmear"? I'm not a programmer, but it seems t be an interesting read. <lfam>GNUnymous_IRC: I'm not sure why you'd want to do this, however. <lfam>If you have a deeper question, you should ask that question instead :) <lfam>GNUnymous_IRC: The manual section 3.2 "Invoking guix package" explains how to search the list of available packages. That is, `guix package --search`. If Guix is not installed, you can view the package list on our web page: https://www.gnu.org/software/guix/packages/ <dvc>adfeno: maybe time to become one? what's the point in all this freedom if you can't exercise it? :) <jmi2k>How do I set the cursor theme (for example, redglass)? I've put "Xcursor.theme: redglass" in my .Xdefaults and it doesn't work. <adfeno>dvc: Eh.... I'm not sure exactly. I do make some shell scripts and some basic JavaScript, but I don't know if being a programmer would be the right work/job for me. <jmi2k>I've done it before, but I just don't know what I'm missing... <adfeno>Mostly because the programmers I know around the city are not very free/libre software friendly. <dvc>well I'm biased. being a programmer is the only work for me :) <rekado>adfeno: you don’t have to see this as work. <dvc>I don't get paid to do programming, I don't think that that makes me not a program. I write programs... <lfam>GNUnymous_IRC: Yes, there are several desktop environments packaged. Many of us use GuixSD on our primary computers <rekado>same here. I hack on stuff, even professionally, but I’m not a “software developer” by trade. <lfam>adfeno: Writing shell scripts is the gateway :) <ZombieChicken>Anyone have any suggestions for a sane and reasonable email provider? Mine is apparently becoming unreliable <dvc>GNUnymous_IRC: haha lol <jmd>ZombieChicken: What do you mean "email provider" ? <lfam>GNUnymous_IRC: That's a vague question. We use Guix to build the operating system GuixSD. But I view Guix / GuixSD as a tool for creating distros. <GNUnymous_IRC>Can I use Guix as package manager in other distro (or in my distro)? <rekado>GNUnymous_IRC: guix runs on every GNU system. It doesn’t have to be GuixSD. <adfeno>GNUnymous_IRC: I do so with my copy of Trisquel. <adfeno>I have the latest versions of most packges that I want. <lfam>ZombieChicken: I'm happy with fastmail.com. I like that their business model is based on collecting money from users, rather than advertising. <ZombieChicken>jmd: Somethign like gmail, tormail, hotmail, ect. that allows IMAP/SMTP access, doesn't mangle mail, allows a user to disable spam filtering, ect. <rekado>yes, Guix works on the Hurd, but it’s still buggy <rekado>you also don’t get substitutes yet <rekado>i.e. you have to build all packages from source. <rekado>I hope that we can build substitutes on our build farm soon. <jmd>I never understood the point of those things. What do they provide that you don't have on your own computer? <rekado>ZombieChicken: zoho.com seems to be a fit to your request. <rekado>jmd: I find it really hard to set up a mail server in a way that actually works <lfam>jmd: I'd like to eventually run all these services myself, but my time is limited. For now, I'll happily pay someone else to carry the pager. <lfam>I understand the trade-offs <jmd>rekado: It takes a little bit of effort. But once it is set up, it is set up. <ZombieChicken>like I tried to send 3 to the mailing list and they never were sent <lfam>ZombieChicken: First messages to guix-devel from new senders require human moderation. <lfam>Is it your first time sending messages there? <rekado>nowadays it seems like major email service providers like to consider everything as SPAM that hasn’t been sent from one of the big providers. <lfam>rekado: I've heard that *if* you set up the various anti-spam sending stuff correctly, it should just work. But that's anecdotal :) <rekado>ZombieChicken: FWIW elephly.net uses zoho.com and I haven’t had any problems with sending to guix-devel. <rekado>lfam: I find this all very confusing. (And I probably shouldn’t because I’m actually a sysadmin.) <ZombieChicken>I'll take a look at my local email setup and see if that might be the problem, though I don't think it should be <lfam>rekado: Yes, I've also heard that it's very confusing <rekado>ZombieChicken: you’re also welcome to try to send me an email (maybe zoho-internal handling is different) <lfam>rekado: It's definitely a conspiracy to centralize email <lfam>;) A joke? Who can tell? <rekado>as soon as there’s a Guix service that makes it easy to run a mail server with GuixSD I’ll host my email locally. <adfeno>I might also have to self-host if my provider goes down.... Or go to Bitmessage :) <adfeno>s/my provider/the provider I use/ <jmd>perhaps somebody ought to get aroud to writing a sendmail service for GuixSD. <rekado>I’ve been wanting to do this for a long time now, but whenever I read up on all the anti-spam stuff I decide that I could get many other things done in the same amount of time. <lfam>rekado: Yes, I'm really looking forward to that. It's hard enough with imperative service management, configuration files scattered across the system <lfam>I run Nixos for my web page on a VPS provider just so I can keep the whole system in one text file. <lfam>Even though I don't really know how to use Nixos <rekado>elephly.net is running on a netbook in my living room, of course using GuixSD. <rekado>(I’m not too worried about uptime; it’s up whenever I need it.) <lfam>GuixSD powers what is certainly my most important server: the home stereo music library <rekado>no static IP, though, just dynamic DNS <lfam>If that server goes down, everything else in my home must wait :) <ZombieChicken>so apparently my smtpd is barfing on me, which is funny, since it used to work perfectly well and none of the settings on zoho's end seem to have changed... <adfeno>Regarding work/job: To tell you the truth, I'm good at spreading word about free/libre software, and also on countering misinformation. <adfeno>Although I don't have enough money to travel around giving talks. So I generaly talk to people at college instead. <ZombieChicken>back to the stone age and IMAP access I guess. grumble grumble <dvc>adfeno: good if that works for you! I try to convince people that linux it technically superior and that it doesn't matter if they end up developing apps, webapplications or embedded systems, it all uses linux, so you better learn it. but not even that gets me very far - mostly because people are simply lazy (IMO) <ZombieChicken>dvc: People don't think about the servers they work wtih every day of their lives, and learning an OS other than Windows takes actual effort (even if it is a very small amount) <quiliro>dvc: not a question of quality but of freedom <ZombieChicken>and Linux/the BSDs are still seen as overly nerdy these days it seems, <dvc>imagine talking about freedom if you can even reach them on a "it's to your advantage to learn it" <dvc>I think people would think I'm crazy :) (Probably they already do xD) <ZombieChicken>dvc: Are you talking about users or devs regarding that statement? Or sysops, as well <dvc>well they are training to become devs... <dvc>or work in a related field <ZombieChicken>then I'd kind of expect them to run into saner OSes in class <dvc>it's all a M$ shop in Switzerland <dvc>as far as I can tell. my brothers uni has a webapplication for writing and compiling c code sanely without using linux <ZombieChicken>Can't get them on the "Here, this is a free enviroment that you can do your work on (possibly) that doesn't require several hundred dollars to just get your feet wet with"? <dvc>when you buy a laptop you already get a "free" windows license <rekado>oh, there are many gratis applications for that <dvc>they press play and it runs and can submit their exercises online I think <rekado>the price really isn’t that interesting to the people I usually talk to. ***jonsger1 is now known as jonsger
<lfam>Few of my friends are into software development. Talking about "free software" with them is endlessly frustrating. Literally every person interprets it as "gratis". I have to use "open-source". <Apteryx>From what I remember from my last investigation of configuring a personal email server, the problem would have likely been with reverse DNS lookups performed by most mail providers to detect spam, i.e. as long as you don't have a domain with a static IP you're going to get flagged as spam at most places. <lfam>rekado: Indeed! The term is over-loaded. <dvc>I tell people that I like compiling software myself. and that google and banks don't use software they don't compile themselves so why should I =P <lfam>But, I can fix my terminology sooner than I fix this bug ;) <rekado>lfam: in German I can use “freie Software” and people don’t usually assume that it means gratis <Apteryx>I don't want to pay hundreds of $ per year just to get a "static DNS", and renting one of those virtual servers pretty much defeats the reason I'd like to host my own server (keep control over my data). <lfam>Orwell specifically wrote about the sorry state of the word "free" in English <adfeno>It's not a bug in english.... It's with the native speakers mostly. <ng0>Apteryx: i think bh? bhnet? something ipv6 offers that functionality <lfam>It's hard to distinguish between the language and the people that use it <adfeno>Dictionaries define 0 price as "gratis or "free of charge". While "free" is for "freedom" <ng0>I've really forgotten the name.. something with b <ng0>there's also smaller DNS registrars offering dyn dns <ng0>oh, it was about static ip.. never mind then :) <lfam>adfeno: As a native Enlish speaker I can assure you that the "gratis" connotation is much stronger than the political connotation, at least in the USA <ZombieChicken>lfam: fwiw, I think my first forays into libre software started with me trying to find a reasonable text editor for Windows that didn't cost me money <ZombieChicken>rekado: Can't use a static IP proxy that forwards your mail to your home server? Or get a static IP from your ISP <adfeno>Check de Blanc's talk "F as in Freedom" to know more about our SEO bug. <rekado>ZombieChicken: my problem is primarily with sending email and being sure that it’s accepted on the other end. <ng0>their plan is to offer ipv6 for free so they can pressure everyone else into offering it <rekado>ZombieChicken: can’t get a static IP from my ISP unfortunately. <ZombieChicken>rekado: Which is why I suggest maybe looking into a proxy with a static IP. Just route your stuff through there <adfeno>The talk is available at media.libreplanet.org no need to look at YT. <efraim>I'm thinking of setting up my mail server locally and using a DO droplet as a relay <ZombieChicken>ACTION awaits the day when decentralized servers are the norm and not the execption <lfam>Hi efraim! Did you get your RK3399 board yet? <lfam>ZombieChicken: Don't wait, start working towards it now :) <efraim>ordered it, gave them my mailing info <efraim>fixing up the last bits (i hope!) with lumina, and then its time to finally fix the aarch64 stuff <lfam>efraim: Cool, I'm interested to hear how it goes. It looks like the most powerful aarch64 dev board that's actually for sale <ZombieChicken>plus part of the problem with anything decentralized is getting people to use it. Chicken and the eggg and whatnot <ng0>and early adoptors providing and working towards such :) <ng0>6 core cortex-a72 :O <ZombieChicken>plus finding sane systems that aren't x86/x64 is a bit bigger problem atm imo <ng0>I looked up what's rk3399 <jmi2k>What's the opinion here about the RPi? Should I expect a GuixSD version for it? <ng0>it's not yet entirely blob free <dvc>we don't have a armv6 port <ng0>seems that they are working towards it (RPI) <dvc>or which rpi are you talking about? <dvc>rpi1 uses armv6 rpi2 armv7 and rp3 armv8 <efraim>suse built an aarch64 version for the rpi3 i thought <jmi2k>it's not a problem as I can install Guix on top of another distro. It's interesting to mix reproducible systems + SD cards <jmd>Uggh. It's annoying that guix/ui.scm defines N_ to do somthing highly unconventional. <dvc>if you don't mind using nonfree software you should be able to do (package (inherit) (source (origin "linux"))) and install guix on the rpi guix system build --no-grub your operating system and then manually boot guixsd from the uboot console. You'll also need to delete a bunch of x86 modules from the initrd, but this should basically work. <ZombieChicken>If you're installing GuixSD onto an ARM system and having to remove x86 code from the initrd, I think that's a bug <jmi2k>dvc: I prefer installing Guix in another distro, but thanks ;) <dvc>jmi2k: that won't help - the other distro is just using binary blobs <lfam>ZombieChicken: We don't support GuixSD on ARM out-of-the-box yet <jmd>247W would be a very small space heater. <adfeno>There might be hope for ARM support using Beagle Bone Black or Libre Tea computer Card. But, it's just a suggestion. <lfam>adfeno: We are working towards it! Help wanted :) <ZombieChicken>How well does qemu handle running ARM code under x86 (other than running it very slowly)? <adfeno>lfam: I intended on buying one of these ARM-thingys, but Big Mother has her pockets closed. <jmi2k>Has anyone successfully changed its cursor theme? I really don't know what is happening. Any ideas? <lfam>Servers use a lot more power than "consumer" computers <ZombieChicken>lfam: I was under the impression POWER systems were massive power hogs compared to Intel's offerings, which is why I said that <lfam>That's the reason that people are interested in powerful ARM CPUs for server usage: electricity is a major cost when running a server <lfam>Yes, I've heard that too, but I'm not an expert <ZombieChicken>POWER8s on a Tyan board seem to pull 190W/247W (withouth/with turbo) <adfeno>lfam: Electricity and runniing server: Don't mention it, reminds of my fight against a person sending pings of death to my last server. <adfeno>Just kidding actually, but the attacks happenned. <lfam>You'd need to decide on a computing benchmark and compare it to some other chip that offers similar performance on the benchmark <efraim>can I pass a USB device to wine with our wine? <dvc>civodul: cloudflare isn't evil! they take security and privacy of their users very seriously and it's an awesome free service... <dvc>it's not immoral to make money... <dvc>and it runs nginx which is free software... <lfam>They make it difficult to access web pages anonymously through Tor <dvc>ah, it's the first I've heard about that. but they protect their customers freedom of speech no matter what from what I've read... <dvc>anyway see you guys! was fun... :) <ZombieChicken>lfam: iirc, Cloudflare is looking for a way to allow Tor access w/o worrying about the abuse. <pareidolia>I am looking for a package that provides i965_drv_video.so for accellerated video decoding. In Nix it is called "vaapi-intel" is there an equivalent I overlooked? <ng0>ZombieChicken: they already do, but the problem is that they default to locking tor out <ng0>yet they celebrated themselves with fixing the issue :/ <lfam>pareidolia: I'm not sure, but perhaps the libva package? <pareidolia>lfam: It installs only dummies when I do guix package -i livba <ng0>civodul: I've gone through old threads, and I maybe will give torbrowser another try once the two other browsers I'm workin on are finished <lfam>pareidolia: Sorry, I don't know anything about this area. Hopefully someone else has some useful advice <rekado>xfontsel tells me “Error: Aborting: no font found” <ng0>do you think we could/should add the tor lts version however they end up deciding on the RFC discussion they have right now about this? Or do we stick with just most current release? <rekado>I already installed font-terminus (a bitmap font) and ran fc-cache -f <lfam>ng0: Can we build a Tor browser that remote hosts cannot distinguish from the Tor project's Tor browser? I thought that making sure Tor browser users were all indistinguishable from each other was the point of using Tor's build. <ng0>we should advise that the build is no official build, as the gentoo (not portage) build does <lfam>For example, Debian doesn't build their own; they just provide a shim to allow users to install it directly from the Tor project <lfam>It integrates directly with Tor <albertoefg>lfam: that seems like the best approach so even if the iso is old you still have latest tor <ng0>they are upstreaming what's possible <ng0>some already landed in current releases <ng0>but we don't package recent firefox, at least not in master <lfam>ZombieChicken: That makes it less likely that users will screw up their Tor configuration and lose the benefits of Tor <ng0>ZombieChicken: plus one or two tor browser extensions <ng0>you can do all kinds of wrong use with tor <lfam>I'd say it's hard for the people for whom using Guix is hard <ng0>it#s really easy to mess it up <rekado>to use Tor is it sufficient to tell the browser to connect to the SOCKS proxy our tor-service provides? <lfam>The Tor browser makes Tor accessible to people who aren't "techies" <ng0>you need a socksport <ZombieChicken>ng0: Are you talking about doing wrong things while browsing while using Tor, like enabling Javascript on a bunch of questionable sites? <ng0>ZombieChicken: for example, yes <lfam>albertoefg: Consider the case of political dissidents. They are usually not technically skilled. <ZombieChicken>Ah. I see that as more of a browsing problem and not a Tor usage problem <lfam>And yet they have a dire need for internet privacy and anonymity <ng0>I would really consider backporting the already merged patches of meta tor, but I think it's easier to locally have an up to date firefox <albertoefg>yes I do what i say if you use tor you probably take the time to learn how to use it <lfam>I don't see a reason to _not_ make privacy easy <lfam>I mean from our perspective <lfam>In response to albertoefg's question <lfam>Obviously some people have different values <lfam>No, the "we need surveillance" argument. <lfam>I don't think that the thousands of people who work directly on the USA internet surveillance system would agree that they have nothing to hide. <ng0>but there are very few people who try to find a balance and communicate to politicians about this <lfam>I think they would be very upset if they lost their privacy <ZombieChicken>ng0: I think half the problem is that politicians are technically incompetent <ng0>you mostly have hardliners. I see myself on the anti-surveilance end as well, but I know people who are in balance in this <ZombieChicken>the other half is they follow marketing as well as your average person does <ng0>those who are doing the youbroketheinternet project <lfam>I disagree. The whole problem is that people choose cynical disengagement from political problems, thinking they can solve these issues with a purely technological approach. <lfam>The people who hold political power sacrificed their other goals to gain it. <ZombieChicken>Is that because someone told them the tech approach would work? Is it because they are idiots and don't understand computers don't make things easy? <ZombieChicken>Why things are the way they are isn't exactly a simple cut-and-dried single issue <lfam>I don't know where you live, but in the USA the left and the tech-liberty people are both decades into a cynical relationship to government. They refuse to get involved. <lfam>We don't have a political left at the national level anymore <ng0>and here (most of) the political left went into decline for technology decades ago <ZombieChicken>I live in the US. I just generally assume that Congress simply doesn't understand anything and the three letter agencies rather enjoy the abuse of power <lfam>ZombieChicken: *People* have to dedicate their lives to political work if they want to have an impact. Congress is people who did that <ng0>I meant to say, they don't see any value in using or combining politics with technology <lfam>It's not for everyone, but the left has failed to encourage people to get involved, IMO <ZombieChicken>No. Congress is made up of idiots who happened to sell themselves to the masses to line their pockets <lfam>I disagree wholeheartedly. What we hear about Congess and what they actually do in their work is quite different. <lfam>They do line their pockets; I agree with that. <lfam>But you are exhibiting the cynical attitude that I described. <rekado>there’s more to politics than just Congress <lfam>It's actually quite easy to have an impact at the local level, which is where careers start. Most local elected officials don't face challnegers at election time. <ZombieChicken>Yeah. Remind me how much power a local government has again? <lfam>An energetic person who is willing to *dedicate their life* to politics can gain public office in a medium-sized USA city within a decade. <albertoefg>lfam: i thought this was a technical discussion not a moral one :p <lfam>Local government has almost total power over its citizens' property and freedom <lfam>Local government sets land policies and controls the police <lfam>In the meantime, people with different values who aren't cynical have taken over the federal and state governments. <lfam>Cynicism about politics = acceptance of status quo <ng0>I think we should build torsocks from the git commit which made the current version. a tarball in a developers sub account on torproject.org always feels odd <ZombieChicken>maybe acceptance that the status quo isn't likely to change without serious upheaval <lfam>I think we aren't going to convince each other of anything, and in the meantime we are basically spamming the #guix channel. If we meet in person someday we can continue this discussion :) <lfam>We disagree on what the issue is <ng0>see gitweb.torproject.org (or was it on github now?) <civodul>dvn: sure :-) my only problem with CloudFare is that it makes the life of Tor users very difficult <ng0>I'm afk now. system finished and I can soon report on obconf working or not, but I just realized I've forgotten the food I prepared for the 2nd time in 4 hours <rekado>civodul: sorry, I didn’t notice this. I only started using Tor today (thanks to tor-service) <civodul>dvc: sure :-) my only problem with CloudFare is that it makes the life of Tor users very difficult <dvc>civodul: sorry for coming on too strong before... I have to respect other peoples opinions. yes that's a valid cirtizism <civodul>rekado: that's ok, i didn't worry (or know) about CloudFare until i used Tor ;-) <dvc>I am all for the war against crippled software and hardware... <civodul>you have to their captcha once in your life to understand why that makes people angry ;-) <dvc>my primary goal is enabling people, if that means that they have to use binary blobs then that's unfortunate, but my primary concern is enabling users... <rekado>I just installed stumpwm and I’d like to hack on it via SLIME from within Emacs. I read that I need swank. <dvc>and I think I'll end up writing a SaaS application one day so I get touchy when I think someone considers that immoral... :) <rekado>does anyone else here use stumpwm with Emacs? <dvc>anyway, I think we agree on many things which is good :) <rekado>dvc: when a SaaS(S) provider uses free software that’s good for them, but it doesn’t mean that the users benefit from it. <dvc>yes, but I have to make a living somehow. A donate button on a webpage isn't going to work for me... And I think SaaS businesses should give back and not only take. GPL is for that right? <rekado>ACTION thinks that “SaaSS” is one of the few FSF terms that are descriptive instead of awkward <rekado>I find it curious that many have an expectation to make a living by writing software. <rekado>recently we had an FSFE meeting in Berlin and that was a big topic in the discussions <rekado>programmers can make a living when they are hired to hack on free software, adding features or customising it. <rekado>that’s part of what I’m doing for a living: add features to free software. <rekado>I’m not saying that other means to make money are evil <ng0>ah, last thing before I'm away, I think setting your own themes with the way we install awesome doesn't work. <dvc>dream job then :) working two days on my own SaaS application and spending the rest hacking on free software is my utopia :) <rekado>I just think it’s curious that there seems to be an expectation that earning money by writing software justifies, for example, to sell proprietary software modules to a free core. <rekado>ng0: I think comments like have more effect in the bug tracker. Here they just get lost in the noise. <dvc>I think that free software enables small companies to compete with large ones, which leads to a distribution of wealth <rekado>at the meeting some argued that what I’m doing (adding features to free software) is evil, because I’m making it harder for the software company to sell their “enterprise” versions. <dvc>haha lol that's crazy =P <rekado>their thought was that by competing with them I destroy their motivation (or their funds) to continue developing the free core. <dvc>well, if they are the main developers of the free core I can understand the sentiment. Taking someone elses free software and simply selling a proprietary extension is a different story <dvc>but they have other things they should focus on - like good customer relations <dvc>that's what makes a business survive not necessarily the product <dvc>a customer wants someone to care and solve their painful problem. how it is solved matters less to them. customers that just want cheap or free are usually "bottom feeders" and customers that will drop you as soon as rekado writes a free extension :) <adfeno>dvc: About CloudFlare: It requires non-free JS automatically. <dvc>as I said, I'm against crippled stuff. non-free JS runs in a sandbox. it doesn't hurt me or restrict me... <adfeno>.. The page also is editable, and is meant to provide ways to pressure CloudFlare on changing the situation, or is meant for us to find a way to write a GreaseMonkey userscript to provide same functionalities as the non-free JS. <adfeno>But it's still a computer program. <adfeno>A computer program that you run, via webbrowser. <adfeno>And, a computer program that most users run automatically. <adfeno>They don't have a yes/no dialog. <dvc>but the js itself isn't that useful even if it is free. and somewhere there needs to be a way to earn a living. that's were I draw the line... but opinions differ <dvc>also - you can inspect the js. I use the compcert compiler which is saddly only open source and not free... <dvc>and you can disable js if you don't want it right? that's a browser feature <rekado>dvc: the browser feature is a little coarse grained <rekado>dvc: librejs aims to help in that regard, though in practise it’s a little too restrictive by default <rekado>ideally, though, we wouldn’t have to use librejs <rekado>it would be cool if web thingies allowed for alternative implementations by defining user-facing APIs. <rekado>using a browser nowadays is like using Emacs with lots of packages but without being able to actually inspect or customise the packages. <rekado>I’m a big fan of the idea of .js and the promise of taking the web serious as a service platform. <dvc>well any good webapplication provides some form of API nowadays which is used by all clients, web, mobile or other <dvc>so in theory someone could write a different client if they feel strongly about it. <rekado>unfortunately, that’s still pretty rare when considering the bulk of web applications out there. <rekado>web developers often take APIs for granted, but rarely pass the advantages on to users <dvc>ok - if I develop a webapplication it will have a documented interface. who am I to restrict others from using my service in ways I haven't imagined? :) <rekado>I think one of the reasons why there isn’t really a community of developers of alternative web frontends for web services is that few even consider that the JS they run in their browser is very much like a regular desktop application. <rekado>it’s a bit odd that we’ve come from JS as a scripting tool to add annoying snowflake animations to writing full client-side applications, but most of us treat these applications as if they had to be taken as they are <adfeno>rekado: I don't know if this can be proven to be worldwide misconception (by society, that is): But I agree with your message about most people thinking that JS isn't client-side software. <dvc>a desktop application can do real damage dough. a js application is much more restricted. and if someone takes the time to write a different frontend, they can just aswell also write a free backend or not? <adfeno>dvc: There are other ways to make money, you don't need to make such compromise as to help spread non-free software. <dvc>it's not spreading non-free software. it's solving a painful problem for a target audience in a way that doesn't restrict or hurt customers but enable them. free or nonfree has nothing to do with it as my customers don't care and I don't care for people who are not my customers (at least in this hypothetical case) <dvc>and it's all sandboxed in a webbrowser <dvc>but anyway. enough about nonfree js... <cehteh>how do i remove a system-generation? <dvc>also adfeno, I'd be targeting businesses in some niche not anything you'd ever be interested in using :) <dvc>cehteh: what does guix system reconfigure --help say? :) <dvc>cehteh: actually I'm not sure it's possible currently, I think that I saw something mentioned about this lately <dvc>I'm a guix user not guixsd yet... <cbaines>cehteh, I don't think there is a convinient command yet, but I believe all that needs doing is deleting the right file in /var/guix/profiles/system <cehteh>would that remove the associated kernel too? <cbaines>Once the link is removed, the gc process will be free to remove any unused packages <cehteh>well first i try from the system-generations emacs buffer, it lets me marks things for deletion <cehteh>but it doesnt look like it recreated the grub menu of course, i guess a reconfigure fixes that