<Apteryx>While doing a "guix system reconfigure /etc/config.scm". Does that mean anything? <Apteryx>Seems like it hung. I'll try restarting the task. ***kelsoo1 is now known as kelsoo
<joshuaBPMan_>Hello, I'm trying to run 'sudo guix system reconfigure /etc/config.scm', but it's taking hours (4 hours last night, and it failed at the end). It looks like it's trying to compile webkit. I suppose I should try uninstalling icecat, and then run reconfigure again. <Apteryx>joshuaBPMan_: I'm reconfiguring my system as well right now. I noly see guix-daemon using all my cpu, but I can't tell what it's doing unfortunately (it's pretty silent). <Apteryx>I don't think icecat uses Webkit, if that helps. <joshuaBPMan_>oh. That's true...icecat uses gecko. Why the heck am I seeing webkit? Maybe some other application requires it. <atw>maybe a GUI mail application or similar? <atw>guix refresh -l webkit or similar might help you find it <quiliro>i gave up on using youtube on epiphany <quiliro>i have a backup of my .icedove directory <Apteryx>quiliro: I just started using Gnus, which is shipped with Emacs. <quiliro>Apteryx: is that the command to install it? <Apteryx>quiliro: It comes builtin Emacs. So more like `guix package -i emacs` <joshuaBPMan_>in my opinion, it is difficult to use it. It's not a point and click email client. <Apteryx>And then inside emacs M-x gnus; but without prior configuration this won't give you a good idea of what it can do :) <joshuaBPMan_>If you don't use emacs, then you probably won't like gnus. <Apteryx>I second that. If emacs is new for your; this is probably going to be a bit overwhelming. <lfam>quiliro: How about claws-mail? It's a graphical mail client <Apteryx>Is there a day of the week where Hydra is susceptible to have more substitutes built? I guess I'm asking which days of the weeks are the Guix developer less active? ;) <lfam>Apteryx: You could plot the Git commits by day of the week <Apteryx>lfam: That would work, yes! Good idea. <Apteryx>efraim: OK! So maybe early Mondays would be a good times to upgrade. <buenouanq>if I want to actually use it to run something, do I want icedtea:out or icedtea:jdk? <lfam>buenouanq: The comments in the package definition (gnu/packages/java.scm line 359) say that the 'out' output provides the Java Runtime Environment <lfam>buenouanq: Look under '~/.config/guix/latest'. It's not a real user interface; I don't think we have a way to list the contents of package outputs yet. Maybe the information should go in package descriptions. <buenouanq>what happens if I cancel an install with C-c before it finishes? <ovidnis>buenouanq: guix is transactional so either the install is successful or nothing happens <ovidnis>any reason in particular stumpwm wouldn't show up as selectable in my login manager? <buenouanq>and what was half downloaded will be garbage collected at some later point? <ovidnis>i dunno, just started using guix a couple days ago <lfam>Half downloaded files will be discarded immediately. Fully downloaded but unused packages will either be discarded when you run `guix gc` or, if you end up asking for them while they are still on your computer, they will be used. <ovidnis>i've got `stumpwm.desktop' in `/home/ovidnis/.guix-profile/share/xsessions' <efraim>lfam: good call on updating openjpeg/fixed to openjpeg-2.1.2, I didn't think of that when I was doing it <lfam>efraim: I hadn't even noticed the new release! <efraim>looks like you forgot the file-name field <lfam>It's weird, when I download the source code with my web browser, the tarball's name is 'openjpeg-2.1.2.tar.gz'. <efraim>i normally pass it to aria2 so I can also download any sig/asc/whatever file also <lfam>aria2 gives the full file name; curl doesn't <lfam>efraim: Does aria2 make it easier to download signatures somehow? <fr33domlover>Hello! Q: I installed X11 via Guix (not GuixSD) but when I run `startx` it fails to get a display. Any idea why / how / where I need to place files / change PATH to make it work? I mean, run X11 from Guix and not from host distro package <efraim>aria2c -Z ftp.url.tld/downloads/file.xz{,.asc} <efraim>-Z for "its multiple files", -c for continue <fr33domlover>(When I run /usr/bin/startx it picks up my .xinitrc and loads the window manager) <quiliro>i need to download the latest versions of the substitutes available <quiliro>in order to install them as needed when offline <buenouanq>I tried to use Gnome as a user in no supplementary groups and it was terribly slow and jittery. Top showed everything going through a single cpu. Added the user to wheel, audio, and video, and the problems went away. Normal or important? What would cause this? <cbaines>quiliro, guix build has a --sources option, but I think that's a bit different <cbaines>You might get the right behaviour by just installing what you need in to another profile, to make sure that its in the store <quiliro>cbaines: thank you for your suggestion <quiliro>cbaines: but what i need is not installing but downloading <quiliro>cbaines: plus i do not want to download sources but binaries <cbaines>quiliro, I'm not sure there is much difference between installing in a different profile, and downloading <quiliro>sometimes the latest binaries are not available but i do not mind having older binaries <cbaines>as in I think it would do what you want <quiliro>cbaines: installing might need to compile <quiliro>i do not want to compile because my machine is a centrino duo, which is very slow <cbaines>with Guix, I don't think that you can use "old" substitutes <cbaines>there is either a substitute available for the exact thing you need, or there is not <quiliro>cbaines: i understand guix provides posibility to install older versions of packages <quiliro>i don't think they delete older substitutes when they have new source schemas <quiliro>it would be of no use because the schemas do not occupy much space <cbaines>I don't know of a way to just download substitutes though, without asking Guix to just install the relevant packages <quiliro>should install the latest substitutes available and compile only if new sources are available and --fallback is appended <quiliro>there is the --download-only option in other package managers <efraim>there's a couple of people here who speak/know esperanto <ng0>so I just imported from pypi and got this: <ng0> `(("python--e" ,python--e) <ng0>is this an artifact? <efraim>what could that package possibly be <quiliro>suggestion for emacs newbie besides guix package --install emacs <ng0>reproducible by `guix import pypi flask-wtf` <efraim>oh, I added qtwebkit to pyqt-5.5 and I got the "welcome to calibre" screen, but then it crashed when it went to open <ng0>htgoebel: it could be "e" <ng0>if it isn't e, i'll open a bug <efraim>i'm sure even if we get calibre working again it'll have problems, currently on my machine i have it with pyqt5.5, with qt5.6 and qtwebkit5.7 <roptat>I had icecat crashing very often on my machine, so I tried to fix that <htgoebel>efaim: Mixin versions of qt will be thr reaosn for the crash <roptat>I found that "From firefox-40, using system cairo causes firefox to crash frequently when it is doing background rendering in a tab." according to LFS <efraim>I don't have ng0's 'python--e' package from `guix import pypi flask-wtf' <roptat>so I removed the switch and made it use the bundled version of cairo, and it seems a lot more stable now <htgoebel>efraim: python--e: I can not even download meta-data for package 'flask-wtfrpm' :-( <htgoebel>looks like me guix setup is somehwar mixed up :-( <efraim>it might be git, guix from 20161207.11 imports it alright, git head doesn't work for me <htgoebel>I almost always get this version information. <jmi2k>Hi I finished installing Guix with root encrypted, but I can't boot it. Installation succeeds, no errors, GRUB appears, but it fails and starts a Guile REPL. <jmi2k>I ran guix pull before guix system init, so it should be the latest version available <jmd>jmi2k: Well it was "latest" as of the moment you did "guix pull". <jmi2k>Yes, at that moment (around 10 hours ago) <efraim>htgoebel: AFAICT pyqt needs all of qt to build all the bindings, including qtwebkit, but there's something else up with the calibre package <htgoebel>efraim: are use using 5.7 for all qt and pyqt components in calibre? <efraim>debian builds pyqt with all of the qt modules, pyqt and qt are both 5.7, and as "depends" it only lists pyqt pieces, no standalone qt modules <efraim>when I remove .config/calibre I get the welcome screen, after I get past that I get the 'ImportError: No module named QtWebKitWidgets' again <efraim>i'm going to toss all of the qtmodules in pyqt5.6 and switch calibre to qtbase and pyqt5.6 and see if that gets me anything <htgoebel>qtwebwit ist no longer part of qt 5.6 and 5.7. Sou you need to create python-pyqtwebkit, I assume. <htgoebel>efaim: In Debian, calibre requires python-pyqt5 (5.7), python-pyqt5.qtsvg and python-pyqt5.qtwebkit. <htgoebel>We do not have pyqt5.qtsvg and pyqt5.qtwebkit yat. <htgoebel>Debian is maintaining its own clone of pyqt. <htgoebel>That's ugliy. We need to analyse the this clone to find out what they have gone. <jmi2k>How can I list files in the Guile REPL when booting (it's all I have, I can't boot Guix)? <jmi2k>I want to see what devices are available in /dev <htgoebel>Our source archive includes QtWebKit and QtWebKitWidgets. Maybe we simply need to add qtwebkit as input? <alezost>sneek: later tell ovidnis so you installed "sbcl-stumpwm:bin" in your user profile, right? I think that SLIM looks for .desktop files only in a system profile, so you need to install stumpwm globally, i.e. try to add (packages (list sbcl-stumpwm "bin") ...) to your system config <efraim>rather than copy tons of lines of code, with just qtbase configure took 100 seconds on my machine to decide which bits to build, adding everything I could find made it take 200 seconds to decide which to build, so it looks like we need to add tons of inputs to pyqt <efraim>I added everything except for qtwayland <efraim>i g2g, i'll check back later to see what my machine built <jmd>Hmm. Our master branch seems to be totally broken atm. <jmd>guix system reconfigure gives: "guix system: error: corrupt input while restoring archive from #<closed: file 0>" <buenouanq>I was able to successfully do a fresh install a few hours ago. <quiliro>any idea of how to download the latest available substitutes without falling back to sources? <quiliro>i mean the case when latest source file has no available substitute for that version, to download the latest susbtitute that _is_available <htgoebel>quiliro: AFAIK this is not possible. guix only knowns the "current" version - with "current" being what is defined "now" in gnu/packages/*.scm <efraim>htgoebel: based on what I can see, calibre now works <efraim>i'll prepare a couple of patches to send in to the mailinglist <htgoebel>quiliro: No, it's part of to guix program itself. <quiliro>htgoebel: i can see there are several versions of some packages available <htgoebel>If you have GuixSD installed, it should be somewhere in /share/guile/site/*/gnu/packages/ <quiliro>so why can't it be the same if the the substitute for the update is not available? <htgoebel>If there are several version available (guix package -A ...), then you may install a different version. <htgoebel>But guix does not fall back to older versions if there is no substitute availalbe <quiliro>is there a way to download all substitutes? <htgoebel>(In the long run substitutes should be available for all releases anyway.) <quiliro>every time i want to install guixsd i have this problem: i do not have a connection but once every 3 days and my machine is old so it compiles veeery slowly <htgoebel>quilrio: I can't tell. I assume there is no way to download all. You would need to know what is available on the server, and AFAIK there is no such thing like a "list" command. <htgoebel>IC. Do you have a second machine, may be old? <quiliro>one is a laptop and the other a desktop <quiliro>i average 24 hours of nonstop installation for guixsd <htgoebel>You could install guix (not GuixSD) on the second machine, run "guix publish there". <htgoebel>Then on the machine running GuixSD, export the store to the other machine. So you'll have some kind of a cache on your second machine. <htgoebel>You can then use --substitutes-urls to use this cache. <htgoebel>You'll still have to build packages where there is no substitute at hydra yet. <quiliro>so i cannot stay many hours at the internet site <htgoebel>I do not understadn why you need to build that much stuff <quiliro>it is not too much but it is a slow machine <quiliro>libreoffice and gnome usually have no susb <htgoebel>May downloading all required source help you? Thus you could just do that at the internet site and build at home (hopefully) <htgoebel>but libreoffice and gnome do not change that often, do they? <quiliro>perhaps downloading all source without compiling will help <quiliro>then i can buiild at home what i need to install <quiliro>"In the future, we want Guix to have support to publish and retrieve <quiliro>binaries to/from other users, in a peer-to-peer fashion. If you would <quiliro>like to discuss this project, join us on <guix-devel@gnu.org>." <quiliro>htgoebel: do you know how to download the sources without compiling? <quiliro>i think the sources when compiling sometime require to download more sources <htgoebel>For a single package its "guix build --source". But unfortunalty this does not give you the dependencies :-( <quiliro>perhaps there is the possibility to feed a file to guix build <quiliro>and include the list of guix packages <htgoebel>the sources *while* compiling *never* requires more source. But guix may need to download and compile dependencies <quiliro>so if i download all sources then...everyting is solved because all sources include all dependencies <quiliro>htgoebel: so how to download all sources? <quiliro>on pidgin, when i click on a link, there is an error (translated from spanish): Could not open URL. The browser command <<xdg-open>> is not valid. <quiliro>2 hours installing calibre and it is not done yet <htgoebel>quiliro: "guix graph -t bag" should give you a list of all (recursive) dependencies. <htgoebel>guix graph libreoffice | | grep "label =" | sed -e 's/.*label = "//' -e 's/".*//' | sort -u <htgoebel>correct one: guix graph libreoffice | grep "label =" | sed -e 's/.*label = "//' -e 's/".*//' | sort -u <htgoebel>Then you can feed this to guix build: …… | sort -u | xargs guix build --source <quiliro>htgoebel: libreoffice does not have a particular desktop dependency <htgoebel>Hmm, when testing this, it fails. Mayby some more tweeking is required. <htgoebel>quilrio: You can pass several packages to guix graph <quiliro>htgoebel: so what is the correct command? <quiliro>oh....it is the version of the package to be removed <quiliro>how to remove it without removing all "-" <quiliro>only remove everything after the last "-" <efraim>Guix build foo --sources=transitive downloads the dependencies recursively <htgoebel>efraim: The help is a bit terse on this :-( <quiliro>htgoebel: thank you for the lesson with grep and sort and guix too! <quiliro>problems should never be solved twice <efraim>There should be a Guix build command to download the substitutes that are available for building a package/system <efraim>it might be --sources, i can never remember if all or package is the right one in that case <quiliro>guix build: error: build failed: some substitutes for the outputs of derivation `/gnu/store/wm8ii88dwwzm6h08hf8xrz9w5k66c5fh-glproto-1.4.17.tar.bz2.drv' failed (usually happens due to networking issues); try `--fallback' to build derivation from source <quiliro>guix build calibre --sources=transitive <quiliro>so that means that this downloads substitutes when they are available and source where they are not? <efraim>sources=transitive means to download the source tarballs to build everything up from the base <quiliro>oh...so linux-libre is needed to install calibre <quiliro>is ther a way to make pidgin open links? <tux__>Hi ,just started reading about guix. how do i express systemwide preferences /do white-blacklisting of packages.e.g. exclude pøtteringware(avahi,pulseaudio,systemd), replace openssl with libressl block qt etc ? thanks <cbaines>tux__, with Guix, "systemwide" preferences just have to affect Guix when you are configuring your system <cbaines>if you want to exclude certian packages properly, that requires checking that they are not in the build graph, and replacing them if they are <cbaines>I don't think there is an easy interface to this yet though <efraim>There are currently no plans to create use flags like gentoo, but you are able to alter package definitions to use or not use whatever packages you'd like <tux__>@cbaines: thanks, eg iwant to compile curl only against gnutls (no openssl) there would be no make.conf-alike, i had to use one of two options ->a) at runtime : export GUIX_BUILD_OPTIONS="" b) package definition ? <quiliro>i installed xdg-utils and the error went away but it would still not open the browser <jmd>Can someone explain to me this code from gnu/tests/ssh.scm ;; Enable TCP forwarding of the guest's port 22. <jmd> (make-marionette (list #$command "-net" "user,hostfwd=tcp::2222-:22"))) <jmd>What is the "guest" ? <tux__>@jmd just guessing ...that looks like portforwarding from qemu host to guest -to ssh into the guest when there is only usernet <jmd>But what is the "guest" ? <jmd>Does that mean a guest account? <jmd>In which case it doesn't really make sense. accounts don't have ports. <rekado>jmd: the guest is the virtual machine system. <tux__>you ssh localhost:2222 and arrive in the guest_vm at :22 <jmd>So guest is the marionette instance ? <tux__>its a trick when the guest has no full net, idont know anything about guix ,i wager it is used as a better isolated chroot .therefore you ssh into it ... quite nice <tux__>as a safer build system ... again just guessing <quiliro>if i am on trisquel and install guix, how will the packages that are already installed work? <jmd>they'll work just as before. <quiliro>jmd: if i install packages with guix they will be installed with all dependencies regardless if they are already installed in trisquel or not? <jmd>guix will ignore what is or is not installed in the host system. <quiliro>i reffe that the dependencies of the package i am installing with guix are already installed <jmd>quiliro: Because it is impossible for a guix package to be dependent upon a package in the foreign distro. <amz3`>How can I work around this, maybe I could use the git repository? <amz3`>also it recommends to use --fallback but there is no --fallback option to guix pull <amz3`>I already retried a couple of times <amz3`>I am trying to update via giit <quiliro>perhaps try to --install that package <quiliro>guix package --install module-import-compiled <amz3`>it's a pseudo package there is not --install for it I think <amz3`>I will try to build everything from git and see what happens <jmd>guix pull --fallback wouldn't really make a lot of sense. <amz3`>build the package from source <jmd>guix pull isn't about building/downloading packages. <amz3`>I just installed guix from binary and did guix pull and i tried to download new binary packages <quiliro>how do i know if guix is downloading sources or substitutes? <quiliro>jmd: is it possible to distinguish a source download from a substitute download? <jmd>quiliro: If it's downloading from mirror.hydra.gnu.org then its a substitute. Otherwise it's a source. <jmd>In general, you don't care what it's downloading (except perhaps for debugging purposes) <quiliro>ok then: substitutes come from hydra and sources come from other servers and have .drv files? <jmi2k>Does anyone have installed successfully GuixSD with full root encryption? I'm stuck. <quiliro>i haved suggested trisquel users to use guix as package manager <amz3`>guix environment guix is compiling all the things :( <jmd>quiliro: Substitutes come from hydra (or whatever substitute server you have configured). Anything else being downloaded comes from elsewhere. Note that some substitutes are also sources. <quiliro>jmd: how to make a compilation of all packages? <jmd>You want to compile ALL packages in Guix locally? <jmd>That'll take about a month unless you have a very fast machine. <amz3`>git pull --rebase is very handy while working on packages <quiliro>jmd: my idea is to compile in a powerful computer where there is great power and download all substitutes from there <jmd>That's certainly a noble if ambitious plan. Probably the best way would be to set up your own hydra instance. Unfortunately I don't really know how to do that. <quiliro>perhaps i can ask a university to host it <quiliro>what do you think would be the requirements of the server? <amz3`>quiliro: I think you'd rather ask the mailling list or search the mailling list to get started because that a question that has already been asked <amz3`>guile-next is not up-to-date :/ <amz3`>for some reasons I am compiling guile-next instead of using the substitute <Apteryx>Nice, the invisible scrollbar handle issue I had in IceCat has been resolved. <jmd>It there a hook I can use whenever a network interface is enabled? <ng0>hacking session result of today, I have pagure. I only need to write a service for it now <ng0>just posted the patches <ng0>11 python packages, if anyone wants to review <ng0>fedora should use an valid certificate <jmd>ng0: I love that word: "only" ! <ng0>we "only" want to "change" the world :) <sovra1>I'm trying to patch a scm piece of code to get wi-fi working: patch --backup-if-mismatch -p0 < networking.scm <sovra1>but it returns an error saying: can't find file to patch at input line 5 <sovra1>then asks to add the patch file manually but on that line there is just this: @@ -21,7 +21,9 @@ <sovra1>I'm new to this program and I don't know how it recognize the diff output and what @ stands for. <ng0>v2 of pagure ocming soon, fedora offered a different urlk <lfam>sovra1: Can you share the patch file? <lfam>sovra1: Okay, assuming that you have the Guix source code in ~/guix, and the patch is at ~/patch, you can try applying the patch like this: `cd ~/guix && patch -p1 < ~/patch` <lfam>But, the patch doesn't apply to the latest master branch anyways. Maybe the patch already been integrated into the Guix master branch? <lfam>sovra1: Yes, the changes were (at least partially) made in commit 2e328698248 (services: Move polkit to (gnu services dbus).) <lfam>sovra1: The part of the patch that changes (gnu services networking) wasn't applied to the master branch <lfam>Try removing everything from the patch file above the line that reads "diff --git a/gnu/services/networking.scm b/gnu/services/networking.scm", and then re-try the command I gave above <amz3`>I have this error when running guix: warning: failed to install locale: Invalid argument <lfam>amz3`: It should just be a warning that doesn't cause any problems. But we can still fix it :) <lfam>How do you start the guix-daemon? <lfam>amz3`: Make sure that glibc-locales is installed in root's profile, and then add this line to the systemd service file that you use to start the guix-daemon: <lfam>Environment=GUIX_LOCPATH=/var/guix/profiles/per-user/root/guix-profile/lib/locale <lfam>You might be able to use the smaller glibc-utf8-locales package, which uses a small subset of glibc-locales. It depends on your needs <lfam>amz3`: The glibc-utf8-locales package includes de_DE.utf8, el_GR.utf8, en_US.utf8, fr_FR.utf8, and tr_TR.utf8. It's for testing, but some of us (ab)use it to save space since it meets our needs <jmi2k>Does full root encryption work? After two days, I can't get it working, and I don't know the problem. Please help! <Petter>jmi2k: Can you paste the configuration you're using. Maybe we can spot something. <jmi2k>It installs fine, grub runs and ask for the password, but boot hangs because it doesn't find the LUKS partition. <jmi2k>About the installation: a GPT disk with three partitions: 32 GiB FAT, 128 MiB BIOS boot and the rest ~100GiB LUKS root partition. UUID are all correct (I've checked them a lot of times :-)), and as I said, it install without errors. <Petter>I'd try using the device name instead for source of mapped-devices. <lfam>Hi :) Remember I said there would be bike-shedding ;) <lfam>I'll try using your patches to re-install my Syncthing today <lfam>And I'll offer a review, although I'm not the best person to review the build system code <Petter>Nice. I haven't heard of any other attempts at building it yet. <sneek>paroneayea, you have 1 message. <lfam>paroneayea: I want to update gpgme soon :) <lfam>sneek: Have this moldy botsnack I found under the sofa <paroneayea>lfam: so I guess the new module it should use is python-gnupg? <paroneayea>lfam: it looks like the new release is tagged in git, but the assword page isn't updated. <lfam>Yeah, I get the impression that the Debian Git repo is where it all happens now <paroneayea>also I'm not totally sure which dependency we use <paroneayea>lfam: I've asked in #assword@OFTC what the status of the release is. <lfam>paroneayea: Yeah, I guess we have to update gpgme and assword together <jmi2k>Peter: I didn't try without UUID, because I'm installing it in a USB and device paths change. But I'll test it in a VM and see what happens. <Petter>jmi2k: Yeah, I was only thinking to see if this is where the problem lies. <buenouanq>jmi2k: if when you get this working, a explicit easy to follow write up would be very appreciated <Petter>Should there be a * here? (file-systems (cons* (file-system <sovra1>lfam: thank you it works ;-) it was just a folder problem! <Petter>I don't have one, but I only have one (file-system). <lfam>You need asterisk if you are cons-ing more than one thing <Petter>Oh ok, then that's not the problem. <jmi2k>buenouanq: Fine, I don't have any problem in explaining the process if that helps :D <ng0>question. does anyone know why Sree Harsha stopped with the gnunet-fs guix task? At least it seems so after reading through the most obvious crosslinked threads from 2013 - 2014 on gnunet-devel and guix-devel <ng0>if not, I'll send Sree Harsha an email asking about what went wrong <ng0>we had some brainstorming today and came to similar results as the people involved in the threads from 2013 - now <ng0>ng0, casually asking questions about a 3 years timespan :D <lfam>You should try replying to the threads; the right people might not see this IRC chat <ng0>at least ludovic was involved in the discussions, and maybe this can be chatted about tomorrow. <ng0>otherwise I try to process the threads and CC Sree and other people <ng0>our brainstorming process was independent from the threads, then I went through 3 years of threads and saw, hey.. they had good ideas, the right ideas, what happened? <buenouanq>I see gnunet as being one of the most important free software projects currently in existence - I really hope they can pull it off. <ng0>what we today went through involves gnunet-fs, gnunet-identity, and gnunet-gns all which are independent of the known to be sometimes broken (vpn, cadet) applications <ng0>so it must be doable <ng0>lfam: it's a bit unfortunate that I can#t be at the reproducible build gathering the next 3 days afterwards <lfam>ng0: I hope they keep good notes so we can read what they learned <ng0>lynX is there from us of secushare, but I wasn#t able to find someone for more than 2 days to watch for my cat <ng0>could've brough the cat to my friends family though.. but I didn't think of that before driving here <ng0>one more message and we have 31k messages in guix-devel in total <lfam>civodul: I just finished testing the staging -> master merge by reconfiguring and rebooting my GuixSD system <lfam>I also updated all my packages on a foreign distro. Seems to be working :) <civodul>you can trigger an evaluation of master right after that <amz3`>which would me a new hydra build of all the things? <civodul>lfam: good thing, thanks for the work! <civodul>lfam: we're almost on time compared to the initial schedule :-) <lfam>civodul: Next time I'll be more aggressive about starting the initial staging evaluation on schedule :) <lfam>We could have been done on time if I had done that <lfam>amz3`: What do you mean? <civodul>yeah, though we were having disk space issues <civodul>but thanks to mark_weaver this seems to be fixed now :-) <amz3`>lfam: I can't guix pull, and fail with the laste derivation which has importer in its name <amz3`>anyway, it's time to go do bed. I 'll retry that tomorrow <civodul>but maybe the machine is loaded and it failed to serve the CSS file in a timely fashion <lfam>civodul: Hm, can you start the new evaluation then? <ng0>i'll try to test the rust build system v2 this month, but I'm not so sure about my schedule <ng0>StoreDir: /gnu/store <lfam>civodul: How to reproduce? <civodul>ah no, it's my shitty hotel network, grrr <civodul>i wish 'guix substitute' could use tor <lfam>I can't believe the hotel blocks TLS <lfam>That's total malpractice <ng0>flixbus and bahn wifi are fun too <ng0>replacing certificates :D <civodul>i don't think they care that much about best practices <lfam>It's far from "best", IMO, although I know what you mean. It's near the baseline of acceptable. <lfam>How can guests check their bank account safely? Use their web-hosted email accounts? <civodul>at least certificate verification fails