IRC channel logs


back to list of logs

<quiliro>i think i found the problem
<quiliro>it was the usb
<quiliro>now it recognizes the hard disk as sda!
<quiliro>grub-install: warning: Sector 50 is already in use by the program `FlexNet'; avoiding it. This software may cause boot or other problems in future. Please ask its authors not to store data in the boot track.
<quiliro>what is the equivalent of sudo grub-install --boot-directory=/mnt /dev/sda
<quiliro>after removing flexnet, i reconfigured but it did not work
<quiliro>i am now init-ing
<quiliro>it worked! after one week
<quiliro>problem with the usb installer
<quiliro>redownloaded and reimaged the usb memory and it worked as a charm
<quiliro>now i would like to make a mirror or hydra
<quiliro>would someone guide me please to start an nginx server in guixsd?
<quiliro>rekado: are you up?
<quiliro>ng0__: who is in charge of hydra?
<davidl>quiliro: lol, nice timing - I just got my first working install as well after about 1 week :-P
<quiliro>davidl: congratulations!
<davidl>quiliro: thank you, and conggrats to you! nginx service:
<lfam>It shouldn't take a week :( Please share feedback about what didn't work on bug tracker
<davidl>lfam: we have both of us I believe.
<davidl>lfam: some part were that old packages had error and needed to be built which made every attempt take a long time and then there were minor issues with transition to 0.13.
<quiliro>lfam: i reported the solution
<quiliro>now i am posting a question about how to mirror hydra for offline installations
<lfam>quiliro: That configuration is not for an offline mirror. It's a caching proxy mirror that passes requests to Hydra and caches them for later use.
<lfam>But, you need to have a connection to for it to work, and each item is only cached when it is requested
<quiliro>lfam: that will not work for me
<quiliro>what can i do about it? can i rsync hydra?
<quiliro>from it
<lfam>I'm not sure... you asked about it on the mailing list in the past, right? Was there a conclusion?
<quiliro>on this chatroom
<quiliro>lfam: i just now sent it to the help-guix mailing list
<lfam>Okay, hopefully we can find a good solution
<bavier>I wish we would somehow pass the make -j value to build-aux/compile-all.scm
<bavier>afiak it's difficult to get at that value from within a Makefile
<quiliro>bavier: does that have anything to do with making an offline server?
<quiliro>how can i download all necesary packages for desktop without actually installing? that could be an option if those packages are reachable by other computers in the network
<bavier>quiliro: something like 'guix system build <desktop-config>' would download/build packages for <desktop-config> but not actually install the system or packages
<quiliro>bavier: /etc/configuration/desktop.scm would be a replacement for <desktop-config> ?
<bavier>quiliro: for example, yes
<quiliro>bavier: how could i use those packages for other machines?
<quiliro>copy /gnu/store?
<bavier>quiliro: you could 1) copy the store contents to the other machines with 'guix archive', 2) use 'guix pack', or 3) set up a 'guix substitute' service and tell your other machines to use it as a substitute server
<bavier>quiliro: the manual details all those options
<quiliro>bavier: thank you. i will read it. your sugestions were very useful
<bavier>quiliro: good luck
<quiliro>bavier: thank you
<bavier>OriansJ: I'm loling reading your 'bootstrapping'
<b11111000000>hi there guix! Why your Icecat is so slow? even on dev tools opening!
<brendyyn>b11111000000: Not sure, can You try testing it with plugins disabled/
<b11111000000>no any plugins except spyblock & librejsx & httpseverywhere - enabled by default
<apteryx[m]>b11111000000: I don't find it slow here.
<b11111000000>3.5s to open <Ctrl-Shift-I> on Core i7 ?
<b11111000000>3.5s to 5s !
<apteryx[m]>I'll try it when I'm back home
<b11111000000>(thinkpad T420 / ssd / intel video )
<b11111000000>Section "Device"
<b11111000000> Identifier "Intel Graphics"
<b11111000000> Driver "intel"
<b11111000000> Option "AccelMethod" "uxa"
<b11111000000> #Option "TearFree" "True"
<b11111000000> Option "XvMC" "true"
<b11111000000> Option "UseEvents" "true"
<bavier>OriansJ: sha256 sums check out on my end
<efraim>sneek: later tell janneke scaleway's armv7 servers don't support guix, I forget which CPU instruction its missing
<sneek>Got it.
<efraim>sneek: botsnack
<janneke>morning Guix!
<sneek>janneke, you have 1 message.
<sneek>janneke, efraim says: scaleway's armv7 servers don't support guix, I forget which CPU instruction its missing
<efraim>janneke: i have to run, but i think it was missing neon
<quiliro>saluton guix
<janneke>hi quiliro
<quiliro>janneke: hu!
<janneke>efraim: ah...althogh gcc/binutils support it, guix only supports it partly
<janneke>efraim: having a name to go with it `scaleway' already help, thanks
<efraim>janneke: scaleway has aarch64 boxes now too I believe
<janneke>efraim: yes, just chatting with a friend
<janneke>efraim: do we have bootstrap binaries already, or do they need to get built?
<janneke>i'll be getting an aarch64 login soon
<efraim>They're built, I have an unofficial binary install on or you can build from source
<efraim>I also have substitutes being served from but I don't have the key posted anywhere atm
<reepca>Hello from GuixSD! \\o/
<mekeor>Hello *wave*
<reepca>Just need to figure out how to make it realize I have two monitors, then either remap left shift to caps lock or finally break out the soldering iron and fix left shift and I should be set!
<efraim>janneke: i uploaded the odroid's guix key to
<janneke>efraim: very nice, thanks!
<efraim>in an effort to build stuff, i'm running a variant of 'guix package -A | grep -f1 | sort | uniq | shuf | xargs ./pre-inst-env guix build'
<rekado_>I’m trying to build GCompris for armhf (because a friend wants to run it on a RPi3). We still need to teach the perl-build-system how to cross-build things.
<efraim>other option is trying to write a systemd unit for curiass
<janneke>so what's up with this .guix-profile/share/emacs/site-lisp/guix.d directory?
<janneke>how do i refresh/update my load path to use my newly installed emacs package?
<janneke>ehh, other than manually pushing the new directory to my load-path...
<efraim>qtbase uses bundled harfbuzz :/
<janneke>efraim: yay, bundling!
<janneke>how do i hook-up bitlbee with google hangouts?
<janneke`>gnome-screenshot -w -d 5
<janneke`>** Message: Unable to use GNOME Shell's builtin screenshot interface, resorting to fallback X11.
<janneke`>Segmentation fault
<janneke`>ACTION cries
<janneke>i know, i know, we're attempting to fix this madness..but hey, is a screenshot too much to ask?
<rekado_>I found that some GNOME tools have a few more runtime assumptions that makes it difficult to use them.
<rekado_>e.g. gnome-maps fails on systems that don’t use GNOME.
<rekado_>epiphany’s e
<rekado_>epiphany’s adblocker doesn’t work if the browser isn’t started in the correct environment
<rekado_>for screenshots I use scrot
<janneke>rekado_: yeah... and the worst thing is: I've been using GNOME since 0.8 or so and /I didn't care/
<janneke>only now that i've abanded the GNOME wagon, it starts to hurt -- silly me
<brendyyn>I think the adwaita-icons-theme-3.24.0 substitute on hydra is corrupt
<janneke>rekado_: scrot...never heard of it -- thanks!
<ng0__>if an application suddenly decided to go +r only on extracted files (used to be different 2 years ago) and I need to apply a patch to the readonly Makefile, is there a way to do this and still use the patch funtion in the (source)? chmod o755 after unpack is too late
<ng0__>maybe it's an error on my side.. the patch I wrote months ago used to work
<ng0__>now it is rejected and the error says the Makefile is readonly
<rekado_>ng0__: you could maybe chmod in a snippet…?
<ng0__>oh snippets
<ng0__>thanks :)
<mbakke>janneke: "maim" is another good screenshot utility :)
<ng0__>I was finished with these 2 packages I wanted as a screenlocker, but it keeps acting funny. I think screenlockers need to be used in our screenlocker service, rght?
<janneke>mbakke: thanks!
<janneke>ACTION just packaged emacs-disable-mouse!
<ng0__>oh. there was a change after 2 years in the Makefile, that's why it's rejected :)
<ng0__>updating my patch fixed it
<ng0__>if I wrap a binary and another package which must use this binary should definitely use the wrapped one not the original.. there is my mistake :)
<ng0__>ah I already use the wrapped one. hrmm
<ng0__>if I no longer respond I managed to lock myself out again
<ng0__>there must be something with PAM..
<ng0__>I'll send the two patches, I don't have an idea about pam
<ng0__> if anyone wants to fix the PAM part.
<brendyyn>I'm getting this error insalling GuixSD ...grub-2.02/lib/grub/x86_64-efi/ doesn't exist
<brendyyn>I don't think I'm even using efi though
<ng0__>is it a fatal error?
<brendyyn>Yeah, it fails to install grub. I think the reason may be that I initially had a GPT partition, but I converted it to MBR and now maybe the grub installer is reading the disk and detecting the wrong things
<brendyyn>I'm not sure how to fix it. If I have to reformat and reboot I'll have to spent another day or two downloading substitutes again
<ng0__>in case you really need to do it, I'd go with an X-less bare-bones config. Depending on your connection it is done in less than 30 minutes
<brendyyn>I guess that is wiser as a first step
<brendyyn>cfdisk shows everything setup correctly though
<wigust>brendyyn: I think this error not related to MBR/GPT. I had this problem yesterday. I fixed it by adding bootloaders in use-package-modules. But you will probably get another error because of MBR.
<wigust>Also I added (grub grub-efi) in grub-configuration, but as I read after it can detect efi automatically, so maybe don't need this.
<ng0__>this is only necessary if you use efi, and you only need the bootloaders package-module if you define a grub package
<ng0__>with bios you can just have the device grub is on and that's it
<ng0__>you could I meant
<brendyyn>didn't work
<wigust>Also in recent `guix pull` I need to transpose (grub grub-efi) and (device "/dev/sda").
<brendyyn>wigust: I'm not using efi thoough, but the installer seems to think I am and tries to install it
<wigust>brendyyn: may be specify (grub grub)?
<brendyyn>wigust: That's the default
<brendyyn>I may need a way to set --target
<brendyyn>Heh, I ended up installing it manually with grub-install --target=i386-pc --root-directory=/mnt /dev/sda
<brendyyn>I guess I'll have to do this everytime I reconfigure
<rekado_>brendyyn: please don’t get used to doing this. If it is a real bug (and not a problem of misconfiguration) it would be best to fix this, so that other users can benefit.
<rekado_>brendyyn: please report a bug with as much detail as possible.
<janneke>i pushed some WIP things to version-0.13.0
<janneke>To git+ssh://
<janneke> 013e1e853..7a0efa77a master -> master
<janneke> 8e8200916..bc2a4daec version-0.13.0 -> version-0.13.0
<janneke>ACTION has pushed revert commits for the two inadvertedly commits pushed to 0.13
<j-r>Is the proper place to send patched that add new packages?
<wigust>How to get working completion and find reference in /etc/config.scm? I do find-file, than switch to guile scheme, run-guile. What do i miss?
<rekado_>janneke: oops
<rekado_>j-r: yes
<rekado_>janneke: maybe we can even reset version-0.13.0 to its previous state. But I’d rather wait for civodul to comment on this.
<mbakke>brendyyn: it sounds like your system has booted in UEFI mode
<mbakke>grub tries to autodetect the platform, even if you specify the "bios" grub in the configuration
<mbakke>enabling legacy bios in the firmware settings should fix the problem, methinks
<mbakke>you can check whether you have booted in UEFI mode by testing if /sys/firmware/efi exists
<wigust>Ah, to get /etc/custom.scm working need to geiser-compile-current-buffer ^_^
<davidl>just ran guix system init with --fallback and it's now building LibreOffice from source.
<mbakke>davidl: I recommend starting with a minimal system declaration to make the installation process faster
<mbakke>also, libreoffice is something that typically belongs in a user profile, not system-wide :)
<davidl>mbakke: ok.
<davidl>mbakke: but isn't LibreOffice supposed to work with a subtitute? I don't understand why some substitutes won't install.
<efraim>on my aarch64 board i'm thinking of running 'until guix package -A | grep -f1 | sort | uniq | shuf | xargs ./pre-inst-env guix build; do git pull && make; done'
<mbakke>davidl: Libreoffice is a large package with *many* dependencies. Each time any dependency is updated, a new substitute will be built on Hydra.
<mbakke>it's likely the substitute from the 0.13.0 install is already garbage-collected
<mbakke>unfortunately we don't have the storage capacity to keep substitutes available for very long
<brendyyn>mbakke: It doesn't exist
<mbakke>brendyyn: libreoffice is currently queued on Hydra, so we should have substitutes in a few hours:
<brendyyn>mbakke: I don't think my system even has UEFI
<brendyyn>Plus, I installed GuixSD 0.12 on this before on another drive
<mbakke>oops, I meant davidl in the previous message!
<mbakke>brendyyn: so /sys/firmware/efi does not exist?
<brendyyn>mbakke: correct
<mbakke>yet grub tries to install for EFI..very odd!
<davidl>mbakke: ok, thanks.
<mbakke>brendyyn: can you email with details about your system and configuration?
<mbakke>efraim: lol, why not :)
<mbakke>efraim: what kind of storage do you have attached? that's gotta use some disk space. maybe toss in "guix gc"? :P
<brendyyn>mbakke: Sure
<mbakke>brendyyn: were you able to boot into the new system?
<mbakke>after the manual grub-install
<brendyyn>mbakke: Yes I'm running it now
<mbakke>brendyyn: okay, great :)
<mbakke>do you get the same problem when running reconfigure?
<brendyyn>I haven't tried it yet. guix pull is taking half an hour to compile scheme files
<mbakke>brendyyn: Okay. Yes, the compilation times after the switch to Guile 2.2 are pretty heavy, hopefully we can serve substitutes for it soon
<davexunit>yeah, the compiler does more optimization... which means more time compiling.
<davexunit>I'm excited to see a 'guix pull' that can deliver a pre-built binary.
<brendyyn>What part of guix causes most of it? gnu/packages?
<mbakke>brendyyn: yes, especially large modules such as python.scm
<rekado_>we should also split up compilation, so that we have multiple guile processes compiling things
<brendyyn>Do package definitions really need much optimisation?
<mbakke>there is some discussion here:
<davexunit>no, they don't need much optimization.
<davexunit>there's been discussion about compiling the modules in gnu/packages with all or most optimizations turned off
<brendyyn>Ok, well looks like plenty of more experienced hackers are already on top of it then ;)
<davexunit>though I think if 'guix pull' is changed to deliver already compiled things then it doesn't matter so much.
<rekado_>I think we need to clean out the bug tracker soon.
<apteryx[m]>I managed to hang my system with Guile :o
<apteryx[m]>Maybe a change I did in the Guix sources; guix environment guix would trigger a memory heap problem and crash. Did this a couple times and the whole system came to a halt.
<apteryx[m]>Had to manually teboot
<apteryx[m]>Yep, it was a change I did... Hmm.
<brendyyn>Anyone here tried packaging icedove?
<lfam>On a single-core of Xeon E5-2676 with 1 GB RAM, `guix pull` needs to hit swap on a spinning disk, and 10000 seconds is not enough for it to complete before the "max silent-time" mechanism kills `guix pull`. Retrying with 100000 seconds timeout and max silent-time.
<lfam>The swap is really killing it
<rekado_>lfam: have you tried enabling zswap?
<janneke>yeah, swap and you're gone
<lfam>rekado_: Good idea. Maybe that will be enough to get me there without swapping
<lfam>I could also work out of Git, but that's kinda giving up :)
<rekado_>lfam: we could avoid this by not compiling everything with the same Guile process.
<rekado_>but to avoid interpreting too much we need to sort the modules
<rekado_>and I don’t know how to do this in a generic fashion.
<janneke>rekado_: mailed you and civodul about my goof-up
<rekado_>janneke: yes, I saw it.
<rekado_>janneke: I don’t think it’s a big deal. Since the branch is frozen, maybe we can just reset it and pretend nothing happened.
<janneke>rekado_: that was what i was hoping...:-)
<janneke>rekado_: do yo have a way/setup to make such mistakes impossible/very hard to do?
<janneke>i have been used to the habit of pushing without naming a branch: git push origin, gitlab, whatever...i guess i should stop that
<lfam>I think it should be fine to delete the branch and re-push the old state, right? It seems very unlikely that anybody tried to deploy the branch in the meantime...
<lfam>With a note to guix-devel for posterity
<janneke>lfam: that would be nice
<lfam>Unfortunately, even with zswap enabled, the system must swap
<lfam>%Cpu(s): 0.7 us, 3.0 sy, 0.0 ni, 0.0 id, 95.9 wa, 0.0 hi, 0.0 si, 0.3 st
<lfam>This probably rules Guix out on all those dinky but popular armhf boards with 1 GB RAM
<lfam>They typically lack fast I/O interface that could provide a reasonably fast swap
<rekado_>it’s the same on my i686 system with 1GB RAM
<rekado_>even on the i686 with 3GB RAM I’m running out of memory sometimes when building Guix.
<lfam>I already had decided for myself that I wasn't interested in them because I'm don't care about hardware that can't reasonably build its whole world, but people do like those things because they're cheap.
<lfam>Thankfully cheap armv8 boards are starting to hit the market with >1G RAM
<lfam>Well, I'm going to see if 100000 is enough for this Amazon VPS
<lfam>100000 seconds, that is
<lfam>I probably won't do it again, though
<ng0__>is there a way gexp can be used in the account creation in a service to make the user- and group-name configurable? I have tried something but I think I will leave gexp out of this for now and make the names not configurable
<ng0__>right now I have something like this: (define (darkhttd-account config) (let* ((darkhttpd-user (darkhttpd-configuration-user config)) (darkhttpd-group (darkhttpd-configuration-group config))) #~(begin (list (user-group (name #$darkhttpd-group) (system`#t)) etc
<ng0__>I don't understand gexp completely yet
<ng0__>I just think that names shouldn't be static, but all services make them static
<ng0__>removing this for now, but I'd be curious for a solution
<bavier`>ng0__: I recall at least one service that currently makes the username and group configurable
<ng0__>maybe I picked it up there
<DoublePlusGood23>Where do you define new package "files"? i.e. `gnu/packages/foobar.scm`
<bavier`>ng0__: postgres
<DoublePlusGood23>ng0__: thx
<bavier`>wait, no
<bavier`>the guix-daemon service lets you change the build slave group
<bavier`>cuirass service has configuration options for user and group
<bavier`>ng0__: ^
<ng0__>ok, thanks
<ng0__>well I improved it. It no longer starts. No starts, no failures. Everyone happy :D
<janneke>wow tcc, gcc, you really never did think about bootstrapping eh?
<bavier`>they didn't think about the *reproducible* bootstrapping part
<bavier`>bootstrapping was a one-time hat-trick of sorts
<janneke>trying to compile tcc.c i cannot imagine it started off so complex, elaborate, difficult
<janneke>and i'm starting to wonder, maybe i should try to find the first tcc that was able to compile gcc...or try to target a very early gcc
<janneke>i don't know...
<bavier`>janneke: have you looked at 8cc?
<janneke>bavier`: haven't heard of it, let alone looked...
<lfam>Bah, openssl 1.0.2l fails to build on core-updates
<janneke>bavier`: my only real target is: compile gcc or compile guile...could 8cc help there?
<bavier`>janneke: it's a C11 compiler that targets x86-64
<bavier`>but I think it requires a C11 compiler to build, which might be harder
<janneke>what is C11?
<bavier`>the C 2011 standard
<janneke>ah, C99+12
<janneke>"8cc is a compiler for the C programming language. It's intended to support all C11 language features while
<janneke>keeping the code as small and simple as possible." -- /me like that
<rain1>it does seem like it'll have to work by compiling a sequence of increasingly comprehensive c compilers
<bavier`>but yeah, ideally the compiler should support more language features than are required to build it, so that it can be used to ratchet up
<rain1>ive been making slow and steady progress with x86 stuff
<lfam>openssl build failure on core-updates:
<janneke>8cc does not compile gcc, afaics -- it sounds as a great effort though
<janneke>rain1: nice -- i did make quite some progress on compiling tcc, only now i'm a bit discouraged
<rain1>oh, how come?
<janneke>constructs like: char child_path[4096], *child_name;
<janneke>if (foo) printf("tcc: using '%s'\\n", child_name), fflush(stdout);
<janneke>no braces, but a comma -- i mean, why?! do i really want/need to support that while bootstrapping
<rain1>I suppose C has a really really long tail
<rain1>in terms of th ework to do to cover all these unusual features
<rain1>you know, i wonder what compilers are able to build GCC
<janneke>sure, and i know that, but avoiding these seems easier than actually using them?
<rain1>I believe it in written in C++ now
<janneke>yeah, we need to target 4.7 or so first
<bavier`>janneke: a tcc fork might be useful
<bavier`>with trivial rewriting to avoid some of those niche requirements
<janneke>bavier`: yes, i am in effect starting that...however small -- trying to see if tcc devs are on irc but it's awfully quiet
<rain1>it's definitely a tough problem, but something to consider is: even if we cant compile gcc itself all of the work done is valuable and anybody can come build on top it
<rain1>especially the C compiler in scheme, I think that is an incredible project which has indepnedent value
<janneke>yeah, that's a nice perspective
<janneke>thanks, that too :-)
<rain1>pcc might be a good compiler to consider too, as part of the bootstrap chain
<janneke>just wondering atm, should i fork tcc where it's now, or some earlier version, or....what else is out there?
<janneke>so a bit discouraged to push hard on tcc now
<janneke>and i'm happy, heard already of two new compilers: 8cc, pcc :-) ...
<rain1>thats a tough question, i don't know a lot about tcc. perhaphs a good short term thing would just be to compile a sample set of c programs - for example cat, ls, etc. core utils
<rain1>I started a wiki for notes on bootstrapping
<rain1>and of course anybody is welcome to sign up and use it..
<rain1>ive gotta run but i hope to study mes in detail soon and then ill have som emore questions ^^
<janneke>rain1: thanks for the headsup!
<wigust>What is a proper way to add file in /gnu/store/...-xorg-server-1.19.3/share/X11/xorg.conf.d ? Pathing source of xorg-server?
<wigust>I found xorg-configuration-file #:extra-config in documentation, but i'm not sure is what i want.
<wigust>Also I found example, but doesn't help too.
<wigust>I want to make like in
<quiliro>what has happened with ?
<ng0__>ACTION assumes space goblins to random questions
<quiliro>ACTION is interested in PragmaOS
<ng0__>what has happened is a very broad question
<ng0__>what is happening is that a system doesn't write itself over night..
<quiliro>ng0__: how can people contribute and where?
<ng0__>currently it's like slackware, there is no central repository for example. I apparently still suck at documentation: there is a very crude 'todo' bugtracker on, the best is the mailinglist though. shockingly little mail, won't flood your inbox :) I have almost everything on the todo list up for grabs in my repo
<ng0__>I move stuff to a new server tomorrow or tonight, so then there'll be a new location for my repository.. there's an outdated mirror on iirc
<quiliro>oh...that is good news!
<ng0__>icedove / thunderbird is one big application which needs to get done
<ng0__>in case I failed to link it in a good way:
<quiliro>I think that the web page is confusing....and i do not think i am the best man for the job...but would like to contribute...preferably on something i can do well
<ng0__>what exactly is confusing? my bottomline for websites is static, small, displays well enough in lynx.
<quiliro>i don't mind that
<ng0__>organization of links in it could be better... working on it
<ng0__>it's basically just tell me what you think you can do and just try to come up with the best. You could take a look at it in a few days, so that I can link my personal repos which will then be on .. the mixture of all kinds of repos is confusing to me
<ng0__>an anti-feature of stagit :)
<quiliro>it is not easy to find what i want to do...i have visited the site twice and both times i get a sensation of not knowing what to do...i do not remember exactly how..but i remember feeling that way...i think a web site should lead you to doing something...and i cannot find a way to provoque that ....but i am thinking
<ng0__>Ok. I try to think about it and fix it. I know how it should look like, but I always hated webdesign in highschool
<rekado>rain1 janneke OriansJ bavier: You’re all welcome to use the mailing list and website to coordinate work or post about status updates.
<rekado>if you want to publish anything there just send me a patch to the website sources and I’ll apply it.
<ng0__>quiliro: it might take a while until I can fix it, but I have a rewrite in Guile again, because I am about to write additions to haunt and bindings for page compilers I used
<rekado>there are some people subscribed to the list, who are not here on #guix, so feel free to use that as a way to reach more people who are interested in bootstrapping.
<ng0__>so the issues is that the "A" of AIDA in design is missing
<quiliro>the web problem is not about beauty
<ng0__>(which is "ACTION"
<rekado>(sorry, OCD.)
<ng0__>urgh.. so there is this package. it's been 14 years since a release. someone picked it up and applied *some* fixes most distros were carrying around but the person made no commit for 2 years now (it is mostly legacy maintenance).. I have a few more fixes this needs to carry. strategy a: we carry the fixes in the package and I try to make the new maintainer pull them in or strategy b: I use my repository (and newer release tarball) of it.. what sounds
<rekado>“b” is only an option if it carries commitment.
<rekado>you would effectively become the maintainer and have the responsibility to apply security fixes
<ng0__>it is basically only for the build system. I could ask one of the tails or whonix people if they have more interest / people to maintain it
<rekado>and it would be your responsibility to keep on hosting it.
<ng0__>i know
<ng0__>tails uses it (as long as I haven't implemented their hook its the only system which uses this applicaiton in core system), so I'll ask them. for now I'll rebase my patches and include them.
<ng0__>oh, I just saw I made 1 cosmetic-functional fix (use prefix in more places) and the rest is used by systems whic hcan not / don't want to access or use the kernel module. the first one can be a substitute, the second one needs to be included as a patch in guix.
<ng0__>otherwise it doesn't build
<reepca>GuixSD isn't detecting my second monitor - it's on, and it's the same as the other one, but xrandr only reports one display. I'm using i3-wm. Should I be looking at the xorg service to see what to tweak?
<rekado>reepca: don’t know. This works for me without tweaking the xorg config.
<rekado>reepca: I usually only have one display, but when I connect a second one (e.g. a projector) xrandr does detect it as it should.
<rain1>nice one rekado, i am signing up
<rekado>rain1: yay!
<rekado>rain1: there isn’t any traffic on that mailing list yet.
<rekado>rain1: but you’re free to make it your own
<rain1>>Build GCC 4.7 with a smaller compiler
<rain1>yes this could be an important thing..
<rekado>the project page is *very* vague
<rekado>and it’s not clear at all who (if any) is working on these things.
<reepca>Hm, I'm using an AMD A10-7850k for graphics, worked well enough under ubuntu (don't think it used any proprietary stuff).
<rain1>how can i submit a commit ot you ? I added a link to the mailing list from the front page
<reepca>Is xrandr supposed to give up-to-date output? I unplugged both monitors and ran the command and it gave the same output...
<rekado>the idea is that *we* (the bootstrapping community) do these things together, but I’m not clear on who that community is :)
<rekado>rain1: thank you! You can just send email to the mailing list or to me directly.
<rain1>the logo is really cool :D
<rain1>i sent my mail
<rain1> <- wow! good work
<slyfox>when does core-updates usually gets evaluated on hydra?
<slyfox> reports '2017-03-31 20:04:59' as last evaluation time. I wonder what the occasion was
<rekado>slyfox: maybe the occasion was freezing core-updates :)
<rekado>we don’t build core-updates constantly
<rekado>because changes there require rebuilding the world
<rekado>it would be silly to do that regularly, because all the work would be undone with the next commit.
<slyfox>but when core-updates gets frozen does it get frequent rebuilds at that time?
<rekado>that just means we won’t push to the branch for a while
<rekado>while it’s frozen we can let hydra build it for us.
<rekado>this uncovers a couple of problems, usually, which we then try to fix, followed by another evaluation.
<rekado>and so on
<rain1>my goodness gcc is big..
<slyfox>so when you are this fixup mode is every core-updates evaluation started off manually?
<rekado>slyfox: yes, we usually start evaluations there manually
<rain1>janneke: I think that mes solves a very important general problem for bootstrapping any language (not just gcc)
<rain1>from looking at here basically any language haskell or similar, to start the bootstrap you need either an implementation of it in C, or you need an interpreter to start off the self hosted version
<rain1>and writing interpeters for languages is many times easier in scheme than C
<slyfox>running them might be a challenge though :)
<rain1>it's rather sad about LCC compiler (used in quake3)
<rain1>I asked the authors to release it under GPL or something, but they aren't able to because of the book publishers
<rain1>yeah you know you do all that work writing the thing yourself and then somobody else says you dont own it. that happens a lot in publishing.
<rain1>im just having trouble with PCC so I started thinking about that
<bms_>What are you doing that you can't use GCC?
<rain1>we're trying to bootstrap gcc from simpler compilers
<rain1>I was just hoping i could answer the question "can pcc build gcc 4.7" but its difficult