<quiliro>it has been 15 minutes since it downloaded a 1.5 MB package and i see no changes <quiliro>but i can see there is cpu usage going up and down <mark_weaver>I know, but what command are you running that downloaded clutter? <quiliro>guix system init /mnt/etc/config.scm /mnt <mark_weaver>ah, well, that's an expensive operation. maybe 15 minutes is not so surprising <mark_weaver>well, you might go into another terminal, and see whether any processes are using CPU, or what. <mark_weaver>but from another terminal you could investigate more <quiliro>should i ctrl+c and then repeat operation? <cehteh>told ya, don't be religious about free software, see what that leads too ;) <quiliro>is it normal to install guixsd more than 5 hours <quiliro>i have started at 15h30 and now it is 18h45 <cehteh>depends on your hardware and internet connection, moon phase and hydras mood of the day <quiliro>except for 30 minutes that the link was down <cehteh>it shouldnt build stuff or at least not much, something is bad there, that will take a eternity <quiliro>wouldn't it be good to have anyone that compiled it first to offer the binary in a p2p fashion? <cehteh>hydra should do that, but i cant tell much more havent installed it recently <cehteh>it would be nice if the store could be served by bittorrent or ipfs or any other p2p system .. <quiliro>it would be even better that there would always be a binary version of all packages <quiliro>or a way to save them for all installations <quiliro>instead of having them be compiled by each machine <quiliro>or, once a package has been compiled, it could be uploaded to the server <quiliro>so the binary be downloaded by all new installations that need it <quiliro>better even that but in a p2p fashion <cehteh>there should be binary versions .. but the server cant keep that much and is often overloaded <quiliro>but that is no problem in a p2p environment <ng0>i hope to solve the problem a bit with gnunet-fs publish/pull.. though even if it's just adding some pieces and stitching together what exists, it still needs time to realize :) <ng0>quiliro: did you run guix pull before you ran the guix system init? <ng0>yes it does not say so, but that's what i do for brand new isntallations <ng0>run guix pull directly before the system init <quiliro>i stopped the installation and ran guix pull now} <mark_weaver>I think the proper solution is to ensure that hydra keeps a full set of substitutes for released versions until sometime after the following release. <mark_weaver>there are pros and cons to running guix pull from the USB installer before 'guix system init' <mark_weaver>on the plus side, your system is freshly patched from its first boot, with the latest security updates and fixes. <mark_weaver>but, it might not be fully built out, and you might have to do a bunch of compiling <mark_weaver>for example, if you chose to install a desktop system, there's a good change you could end up compiling webkitgtk for many hours from the USB installer <mark_weaver>when using the stable release version for the initial install, we can reliably ensure that hydra keeps those binaries. even if we're not doing it now, that can be fixed. <mark_weaver>and, of course, if you run guix pull, and some build you need is broken, your "guix system init" might suffer a build failure. then you're stuck. <mark_weaver>on balance, I suspect that it's usually better to recommend running 'guix system init' with the release version of guix packages, and then to update from within the installed system. <cehteh>there isnt some way to make an orthogonal snapshot on packages (remember snapshots on CVS?) on hydra? <jlicht>rekado: your proposed recursive importer for cran has made me a really happy camper w.r.t npm imports. Thanks a bunch! <quiliro>guix system error: failed to install grub on device '/dev/sdx' <quiliro>....sbin/grub-probe: error: cannot find a grub drive for /dev/sdX. check your device map. <quiliro> /gnu/store/..............-grub-2.00/sbin/grub-probe: error: cannot find a GRUB drive for /dev/sdX. Check your device.map. <quiliro>guix system error: failed to install gGRUB on device '/dev/sdX' <quiliro>some one have an idea how to install grub correctly' <quiliro>what do you suggest to add to desktop installation? ***grinhaus is now known as wgreenhouse
<dpg23>ng0: Hey, I was told you were involved with the GNUNet project, can you give a rundown (or links) on how it differs from Freenet or Tor? <dpg23>I managed to get offloading to work now, so thanks everyone for that. <quiliro>i would like to run gnome-control-center <pksadiq>quiliro: try running using terminal. Also enable verbose output: gnome-control-center --verbose <quiliro>error while loading shared libraries <quiliro>something about libavahi-common.so.3: file too short <quiliro>pksadiq: should i guix package -u libavahi ? <pksadiq>quiliro: I'm not a guix user, but a gnome user. Still you can try. ☺ <quiliro>guix package -i gnome-control-center fixed everything <quiliro>pksadiq: now since you are a gnome user...your help is important now.... <quiliro>there is no logout way from gnome...only from xfce <pksadiq>quiliro: what you get running 'gnome-session-quit' command? (You may get logged out if running right) <quiliro>now how can i log out using the gnome interface <pksadiq>OK, click the top-right most menu (with power and sound icons) → your user-name → log out <quiliro>pksadiq: how can i install all of gnome <quiliro>perhaps i am missing some package installed <pksadiq>Hm.. That's weird. Don't know how good guix is for gnome. <quiliro>because i lack also the keyboard distribution settings <pksadiq>I don't know the package name for gnome. And I don't have a guix system to check. May be someone will help you. <quiliro>i only have the shortcut configurations not the distributions <pksadiq>quiliro: what is the output of: gsettings get org.gnome.shell always-show-log-out <pksadiq>Hm.. do this: gsettings set org.gnome.shell always-show-log-out true <pksadiq>If You still don't get the logout try Alt+F2 and type 'r' <enter> <pksadiq>If you can bear keyboard shortcuts, you may also assign some keyboard short cut for gnome-session-quit <pksadiq>go to the above url using gnome web. You will be able to install the extension ***tschwing_ is now known as tschwinge
<efraim>also i'm about to run out of space grafting texlive-texmf-2016 after having timed out twice already <civodul>mark_weaver: we need your faster grafting code! :-) <efraim>/dev/sda8 40184184 39477424 280788 100% /gnu <efraim>hey it finished! real 148m51.416s <efraim>(properties `((max-silent-time . 9600))) ; don't time out while grafting , yea or nay for texlive-texmf <civodul>efraim: there's a command-line option too <efraim>i know, i made a math error and set mine to 20 hours while I was building it, but it seems to be one that regularly times out, especially on nfs mounts <civodul>we also need a modular texlive package <efraim>we'd have to download the full source and then build each piece in its own package and make sure they didn't conflict with each other if more than 1 were installed <quiliro>what gui software can i use tyo install other operating systems by way of virtualization <ng0>i get t he feeling dpg23 is in a different timezone^^ <ng0>later tell dpg23: I'd like to tell you more, but with our possibly differing timezones/availability your best shot is either our (SecuShare) chat on psyced.org (see www.psyced.org for an how to connect) or #gnunet here <ng0>sneek later tell dpg23: I'd like to tell you more, but with our possibly differing timezones/availability your best shot is either our (SecuShare) chat on psyced.org (see www.psyced.org for an how to connect) or #gnunet here <ng0>can i pass more than 1 sentence to sneek? <efraim>if it spills over into a new chat entry you have to "wrap" that one in "sneek later tell" <efraim>and you can pass multiple messages <ng0>i think this one is enough.. psyced handles delayed messages better <ng0>i could break it down to a very minimum and say that GNUnet is politicaly motivated from the very beginning (https://gnunet.org/philosophy), but the differences are not just passed in a 1-2 sentence sneek query. <quiliro>hello...what gui software can i use tyo install other operating systems by way of virtualization? <ng0>quiliro: you could try and package it :) <ng0>but qemu itself isn't hard through for example a bash script to start machines. that's how i do it when i need it <ng0>we need more .desktop files... or gnome doesn't pick them up. xterm has none. <ng0>a general way to generate them would be nice, and not construct them each time they are missing. <quiliro>ng0: hi...thank you for your response...i am still learning how to use guixsd and guix.....so i cannot package yet......could you help me with a link that teaches how to boot a usb containing my own guixsd boot in order to run it in qemu_ <ng0>usb i don't know, just direct images <civodul>efraim: what's the status of libarchive/cmake on core-updates? i lost track of it :-) <efraim>cmake fails that one test with cmake 3.4.3, 3.5.0 and 3.5.2 <efraim>i haven't yet tried a different version of libarchive <efraim>there's also cmake 3.6.0rc2 but I assume that it wouldn't make a difference <efraim>the error is CMake Error at CMakeLists.txt:75 (message): <efraim> LibArchive_FOUND is set but version number variable LibArchive_VERSION is <civodul>that's an error when building libarchive, not when building cmake, right? <efraim>sure, I can get it started in a few minutes, but it takes me about 2 hours to build cmake <ng0>quiliro: imo understanding is optional and comes through learning, for me this worked while I started looking at other packages and pick an easy one which wasn't there yet, or even just look at the .scm files to see what it all does :) <ng0>if the qemu documentation doesn't cover it, the gentoo wiki does :) <ng0>or I can pastebin my script, which isn't general purpose yet <ng0>not included: how to create a disk <quiliro>downloaded gnu-system-demo-0.6.qcow2.xz <quiliro>will be able to boot with qemu-system-x86_64 ? <ng0>understanding is mandatory, damage not covered. <ng0>not that iot damaged something here, just saying :) <quiliro>will gnu-system-demo-0.6.qcow2.xz need to be decompressed previously? <davexunit>I think you need to be more descriptive. what exactly doesn't work. html5 is a big specification and the various browsers support different features <davexunit>and webkit certainly supports html5 features <quiliro>ng0: get htis message from your script: Could not access KVM kernel module: No such file or directory failed to initialize KVM: No such file or directory <quiliro>it is strange because lsmod show kvm present <civodul>quiliro: what does "ls -l /dev/kvm" show? <ng0>don't know, can't help. this was a script written on gentoo with my own kernel. <ng0>i noticed we don't have libgnomekbd, or am i wrong? <ng0>gkbd-keyboard-display command does not exist <ng0>i assume it's part of libgnomekbd <Illya>How can I run guix in virtualbox? <Illya>Ah. Got it working by reordering the disks <efraim>libarchive/test/test_write_format_gnutar_filenames.c:58:2: note: use option -std=c99, -std=gnu99, -std=c11 or -std=gnu11 to compile your code <efraim>anyone know how to set this? configure flag? <efraim>i'd rather set it than disable the test <civodul>efraim: presumably #:configure-flags '("-std=c99") <civodul>#:configure-flags '("CFLAGS=-O2 -g -std=c99") <civodul>i'm surprised libarchive's build system doesn't do it already <efraim>the other options are what? c89 and c03? 1y? <civodul>ACTION hopefully unbroke evaluation on 'master' <quiliro>Could not access KVM kernel module: No such file or directory failed to initialize KVM: No such file or directory <efraim>... and I even chose a font where zero has a dot in it. -02 doesn't give the same results as -O2 <pksadiq>quiliro: have you enabled virtualisation in BIOS? any debug message with: dmesg | grep -i kvm ? <quiliro>it is diabled by bios...i don't find where to abilitate it <pksadiq>quiliro: look into your bios settings. You will find it somewhere <quiliro>pksadiq: thanks, will reboot and be back then <civodul>the interface to create Hydra jobs sucks somewhat <quiliro>pksadiq: the bios was in fact disabling virtualization <quiliro>Could not access KVM kernel module: Permission denied failed to initialize KVM: Permission denied <pksadiq>quiliro: try again: dmesg | grep -i kvm ? any other debug messages <quiliro>kvm: Nested Virtualization enabled kvm: Nested Paging enabled <quiliro>my command is qemu-system-x86_64 -enable-kvm -m 4096M -hda '/media/cesar/Descargas/trisquel_7.0_amd64.iso' <pksadiq>quiliro: Now its guix's turn. Some else should help you :) <davexunit>quiliro: your user needs to have permission to access KVM <davexunit>running as root is a quick way to get that to work <quiliro>davexunit: how do i add my user to the group authorized to use kvm <ng0>bsd-2 and bsd-3 are okay. original bsd not? I have to check if this bsd is new bsd <pksadiq>quiliro: is kvm listed in groups: groups | grep kvm . If not, use adduser: # adduser $(id -un) kvm <ng0>though it's also mirrored and probably packaged by debian, so netcat by openbsd will be probably alright <quiliro>-bash: adduser: no se encontró la orden <pksadiq>quiliro: may be useradd/usermod? see then man pages <efraim>civodul: libarchive built successfully with -std=c99 <efraim>i added a note, next to test cmake <quiliro>they are not installed and cannot find the package that contains them <pksadiq>quiliro: I think there is a shadow package that provides usermode. I think this should be installed by default. try sudo usermod <quiliro>nano /etc/group`would not work ither <bavier>quiliro: were you ever able to get your wifi card working? just curious. <quiliro>thank you for asking.....what a good memory <bavier>quiliro: oh, sorry to hear about your machine <quiliro>bavier: do you happen to know how to modify group members in guixsd? <quiliro>i am trying to run qemu as user but it will not unless i am root <bavier>quiliro: section 7.2.5 of the manual <quiliro>and i cannot find adduser, usermod or vigr on guix <bavier>you can declare the groups of a particular user in the system config <bavier>quiliro: e.g. I have "kvm" in supplementary-groups in my config <davexunit>civodul: just ran into the nondetermninistic module compilation bug while trying to build a newer guix snapshot package :/ <davexunit>does someone have to purge the bad thing from the cache when this happens or something? <civodul>davexunit: which Guix commit is it from? <civodul>410 comes from Hydra (the software) itself, not from nginx <civodul>that's what we get when the thing isn't in store <davexunit>I'm using an "old" guix whilst trying to bootstrap a machine <civodul>it means that these binaries have vanished from everywhere <wingo>should i backport the split-and-decode-uri-path thing to 2.0? <wingo>maybe peope are relying on + being incorrectly decoded to space <wingo>it is correct for application/www-url-form-encoded or whatever <quiliro>finally i could boot as user with qemu <quiliro>but i could not install because i have not defined something i don't know <davexunit>civodul: it's weird how it thought there was a substitute available when there in fact was not. <davexunit>quiliro: why are you booting the installation image with qemu? guix can generate guixsd VMs <civodul>davexunit: i just realized that Hydra always returns 410 for nonexisten /nar, instead of 404 <civodul>yes, not a problem in itself, just surprised that it's 410 and not 404 <civodul>and the problem is that we have a narinfo in cache for this thing, when in fact the nar itself is missing <civodul>but nginx cannot be taught that there's a connection between the two <quiliro>davexunit: i am not booting a guixsd installer inside guixsd .... i am booting a parabola installer in guixsd <quiliro>qemu-system-x86_64 -enable-kvm \\ -cpu host \\ -drive file=/media/cesar/sistemas/parabola-mate-2016.06.01-dual.iso,if=virtio \\ -netdev user,id=vmnic,hostname=guixvm -device virtio-net,netdev=vmnic \\ -m 1024M \\ -monitor stdio \\ -name "Guix VM" <quiliro>gemu-img create disc_space_image.qcow <janneke>i rebase on master and cannot run guix environment ... anymore? <quiliro>qemu-img create disc_space_image.qcow 5G <janneke>guix: environment: command not found <civodul>janneke: works for me on v0.10.0-1472-gab23fb8 <janneke>civodul: thanks...how is this even possible, i'm confused <civodul>it could happen if there's some error while executing module top-levels <civodul>while loading (guix scripts environment) <janneke>OK, that takes away most of the magic, thanks :-) <janneke>oh, I have my grub changes and gptfdisk, mcron+rottlog wip <janneke>just wanting to remove ps+pdf from the merged nasm <janneke>I must not reconfigure without grub changes or it will fail :-) <paroneayea>civodul: I wonder how open you would be to bundling davexunit's (ice-9 json) with guix for now. <davexunit>paroneayea: my intention was for jlicht to use it to unblock himself and keep making progress, and then later undo it and use guile-json. <civodul>paroneayea: not against, but people already have guile-json no? <janneke>civodul: ab23fb8 is fine...learned something interesting...let's find my silly typo/rebase ting <paroneayea>though we could also drop an external dependency in guix by using ice-9 json too :) <paroneayea>civodul: yeah, jlicht is running into problems with the json + gexp stuff <davexunit>paroneayea: I think we should do that migration once (ice-9 json) is fixed up and accepted into guile proper <civodul>but it would be good to discuss the guile-json/gexp thing <quiliro>qemu-system-x86_64 -boot d -net nic -net user -cdrom /media/cesar/sistemas/parabola-mate-2016.06.01-dual.iso -hda parabola.qcow -m 5G -localtime <davexunit>ACTION thinks he thought of a decent CLI for saving/loading profiles with 'guix environment' <paroneayea>I've been finding that every time I upgrade guix, I do another "guix environment" on my mediagoblin guix.scm <paroneayea>and then I have to wait a few hours to be productive again. <paroneayea>I really want something more persistent that still is built off of a declarative list of packages. <davexunit>the initial work to support it was completed awhile ago <paroneayea>and the hours thing is usually the fault of latex ;) <davexunit>my thinking is that something like this would be nice: <davexunit>--save would make a .guix-environment profile in $PWD <davexunit>re-using an existing environment would then be something like: <davexunit>(I think there's a name clash with 'load', but you get the idea) <bavier>we already have the idea of a 'profile' for 'guix package' which will instantiate a profile at the named file. it would be nice if 'guix environment' could play well with profiles created in that way <davexunit>bavier: 'guix environment' already creates profiles <davexunit>they just aren't registered as GC roots or anything <davexunit>it's more of a UX issue that anything. under the hood it's all the same. <bavier>I'm think like "drop me in a container using this profile that I've already created with 'guix package'" <davexunit>guix environment --profile=$HOME/.guix-profile --container <davexunit>the question there is: how does it compose with other things? <davexunit>guix environment --profile=$HOME/.guix-profile --container --ad-hoc emacs <davexunit>that would be an environment that builds a new profile based on an existing one and adds emacs to it <davexunit>guix environment --profile=$HOME/.guix-profile --container --ad-hoc emacs --save=$HOME/.guix-profile <davexunit>I think there's some things to be thought through here. <davexunit>should saving an environment also store which options you chose? <ng0>i packaged netcat by openbsd. shall I let it collide with other netcats, or just copy the binary+manpage to nc.openbsd ? <ng0>shall i let o.o grammar ouch <ng0>well you get the question <ng0>why I ask is, the distros I know package it either to colide or rename it <ng0>the resulting binary is "nc" <ng0>i keep forgetting that. sorry <davexunit>a user likely wouldn't install two version of netcat in their profile <ng0>doesn't nmap depend on a specific netcat? <davexunit>I don't know, but if it did then nmap would refer to that one. <davexunit>or rather nmap *should* refer to that one via an absolute path. <davexunit>whether or not the package actually does that is another story, but that's the expected thing. <torrhal>installed all deps but compiles with error <ng0>urgh. how to import the secret keys with gnupg >=2.1 ? <ng0>this new model is weird <efraim>i use tab completion a lot with gpg-2.1 <efraim>acutally I've always used tab completion a lot with gpg <efraim>i've actually gotten very familiar with the man pages for gpg and ffmpeg-all <ng0>gpg: error computing keygrip <ng0>I'm familiar with the old sec/pub model, but i have never migrated a new one <ng0>ah right.. that's the ssh part <efraim>i think the only thing I haven't tried is downgrading libarchive from 3.2.0, and I'd rather not do that with all the patches it had before <efraim>I'm going to just comment out the libarchive test and build past cmake and see how it goes <ng0>no patch today it seems. need to fix torsocks first. <rekado>tomorrow I have an appointment to flash libreboot on my X200S. <rekado>I see that my GuixSD system does not install a file like this. <rekado>Will I need a separate partition for /boot with a copy of the Guix-generated grub.cfg available as /boot/grub/libreboot_grub.cfg? <anthk_>hello, is there a way to use gdm instead of slim? <ng0>rekado: for me libreboot with guixsd on t400 just worked as I do not yet encrypt the harddive <ng0>harddive \\o/ harddrive... <rekado>I hope this will work for me too. <ng0>after setup you get a grub which in turn can select other grubs <ng0>so maybe the file in question is only needed if you do it yxourself and don't use a default libreboot image <ng0>but that's just my very short experience :) <rekado>I expect not to be online for a couple of days after flashing. <rekado>the X200S has a small SPI flash chip on the bottom side of the main board with really tiny solder points, making it quite difficult to flash. <rekado>I hope that my attempts to wire the pins to the top side of the main board were actually successful. <ng0>i heard the x series are sometimes easier to flash, at least with connectors, not soldering, than the t series <ng0>no disassembling of every part <davexunit>rekado: did you need to solder onto those teeny pads? <davexunit>I have great difficulty trying to hand solder anything that is surface mounted <rekado>it's not even like a regular SMT component you normally encounter. <davexunit>curious if you used any special technique to help accomplish the task <rekado>I noticed that the pin pitch is the same as the distance of wires in a flat ribbon cable, so I soldered the ribbon cable to the pins. <rekado>I kept it really short, just long enough to lead the pins to the other side of the board. <ng0>i would've prefered to solder on the t400, but i don't trust my soldering enough. <davexunit>rekado: how did you solder such small joints without connecting two by accident? <rekado>using a ribbon cable is easier because the cable stays in place once you fix a single pin. <rekado>my soldering station is pretty good, but the tip is much too big and no longer heats up evenly, so it was quite challenging. <rekado>I hardly used any solder near the pins, so it's mostly reheated from what was already there. <rekado>less solder = smaller likelihood of bridges. <rekado>but I think it may have been smarter to remove the chip completely, solder wires to the pads, lead the wires to the other side, and place a new bigger chip there. <rekado>the top side of the board has enough space for a flash chip. I wonder why they had to put it on the bottom side. <rekado>I now have an ugly header glued to the top side, making it a little easier to access the pins. <ng0>if it is where i suspected it to be, i'd say less disassembly time <rekado>yes, only the keyboard and palm rest need to be removed now. <rekado>you can even power on the laptop in this state, without much difficulty <ng0>where as t400 and others are disassemble: everything. <rekado>yeah, I had to disassemble everything on the X200. <rekado>tried to skip a step and accidentally broke a flat ribbon cable in the process. <rekado>made the sound card disappear; but I was able to repair it, surprisingly! <rekado>ACTION meant the X200S above, not the X200 <ng0>sorry, my getmail/sendmail is stuck behind. <rekado>ng0: when you do "guix system reconfigure", does it break things and "require fragile workarounds"? <ng0>i used to use coreboot in a broken state, now i have libreboot, but I need to set it up somewhere else (or virtual) to test the patch, which is why i did not test it back then <ng0>last i talked to jookia, they had a working full disk encryption branch with this <ng0>but we (guix) were not really accepting towards some patches at this point, or discussion has lead to this. <ng0>or so i was maybe told <rekado>I don’t understand. I think the patches just needed review and modification. <ng0>i don't know. it was not my focus, and what jookia told me is blury to me now <ng0>I lack devices to review them appropriately <rekado>well, I hope I can review them after tomorrow :) <ng0>that would be good :) <ng0>hm.. would someone like to fix torsocks-2.1.0 up? it fails testing, and I seems to not get in this moment why <emyles>hi, is "guix deploy" a thing yet? I can't find much info. <ng0>gentoo and other distros apply patches to mutt which are logical but are illogically rejected by upstream. may I add some of them to our mutt, or should this stay local?