<person123>can someone help walk me through partitioning my disk for an encrypted installation? <fogbugz>how do you manage gnome desktop settings in a reproducible way? using gsettings schemas? <OriansJ>person123: how detailed do you require assistance? The minimal involves setting up a boot partition, a luks volume and a grub config to automount your encrypted volume ***cstrahan_ is now known as cstrahan
<Stream>i have two monitors and display manager places login field in the middle of both monitors, i tryed to disable one monitor but it doesnt change anything, how to fix it? i need some kind of config files or it is done in configuration file? <Stream>yes i use default one, gona try other now. i had also problem my main monitor is at right side but in linux default is left, which is why i changed them but when i did it messed crosshair and even worse later, now i cant start xfce and gnome because i changed monitor positions, i just login then it gets stuck and returns me again to login screen unlogged <str1ngs>off topic, is there away to use startx? I like to not use a DM at times. <str1ngs>just installing xinit is not enough, I don't think <wigust>str1ngs: Why is it not enough? As I remember it is. But you need to specify a window manager in ~/.xsession <str1ngs>startx looks for the X binary relative to the startx path <str1ngs>I even added xinit xorg-server to system config. hoping that would help. but I think it deference's symlinnks <wigust>Stream: I tried a GNOME with GDB, but in Guix generated QEMU virtual machine. It's not ready to use as I remember from mailing list discussion. I think it will be easier to configure Slim to auto-login if you don't need multiple users to login than mess with login managers. <Stream>the second problem is more about changing the sides of monitors idk how to describe it, but its like i got main monitor right and secondary left but in linux defaults are inverted, so when i try to change defaults both in xfce and gnome first problem with coursor appears and after some time i just cant login into them at all, i think its not because of slim <Stream>the coursor icon and the coursor itself are in different positions <str1ngs>sounds like mouse is having a existential crisis ***edouard__ is now known as EdouardW
<fogbugz>wigust: I'd be very interested to get plain startx running. I tried months ago but it didn't seem to be possible. <wigust>fogbugz: Does it cannot find find a 'X' binary of xorg-server? <fogbugz>wigust: i vaguely remember that was the case <fogbugz>i need to try again. For me there are just two hurdles to adopt Guix as a daily driver (I'm already using it, but these make the experience painful). The minor one is no startx. The major one is IceCat. I like the project, but it's not very actively maintained or well documented. <Stream>what is the problem with icecat? it works on my system <fogbugz>Stream: yeah, it works on mine too. But the project is not very actively maintained, ships with very old plugins, lots of the configuration patches are outdated. I wonder where it's going with the new Firefox 57. Besides, making it work for e.g. WebRTC at meet.jit.si involves patching tons of things. <rekado_>fogbugz: we also have epiphany and eolie. <civodul>m-o: looking again at commit 00e39b2ea4, it wouldn't hurt to keep the EFI System Partition, would it? <m-o>civodul: the problem is, when a partition with esp support is detected by intialize-hard-disk, who calls install-efi which is broken on arm32. <m-o>civodul: but that maybe the problem to solve instead of removing EFI partiton though <fogbugz>rekado_: yeah, epiphany works great. But these days, lots of things like (open) webrtc only works on chrome or firefox. Chrome is pretty bad privacy wise. Even Chromium has some blobs. Firefox is good, IceCat is much better. I just wish it was well maintained. <civodul>m-o: ah ok, i didn't know about the ESP detection logic <civodul>rather, i think i forgot about it :-) <numerobis>Hi! Has anybody written a systemd service to start GNU Shepherd, by any chance? <linix1>did anyone manage to get openvpn and networkmanager working? <linix1>I tried setting a nm plug ins directory environment variable (found it in networking.scm) and it didn't work <linix1>Also I tried adding it as a plug in in my configuration (as described in networking services in the manual) but then I get a scheme error <dvn>anyone else having trouble with haunt? <dvn>i get a backtrace when i try to build <dvn>ERROR: In procedure scm-error: <dvn>ERROR: no code for module (system reader) <dvn>though I have guile-reader installed <civodul>dvn: that's with "guix package -i haunt"? <dvn>i'm beginning to think it's something to do with my scm though <civodul>haunt propagates guile-reader so that should be fine <apteryx_>Hello Guix! What was the command to list the environment variables again? <pkill9>to list all environment variables? `env` <apteryx_>I mean; the environment variables that Guix requires to be set for its packages, something like 'guix package --search-paths' (memory just came back ;) <apteryx_>I somehow missed in in the output of 'guix package --help' the first time I looked. <dvn>civodul: no it's not my code <civodul>dvn: it's not your code? what do you mean? <dvn>i thought maybe haunt was just giving me confusing errors because my skribe was bad or something <dvn>but i tried it with an upstream haunt site, and it still fails the same way <rekado_>I have a linking problem with a package using sdl-union <rekado_>/gnu/store/76dhmyga4iajwfdg142llm3lwq9dz4m4-sdl-union-1.2.15/lib/libX11.so.6: error adding symbols: DSO missing from command line <rekado_>looks like I must link with libX11 even though this package doesn’t use it. <wigust>dvn: What are your GUILE_LOAD_PATH and GUILE_LOAD_COMPILED_PATH? <dvn>GUILE_LOAD_PATH /run/current-system/profile/share/guile/site/2.2 <dvn>GUILE_LOAD_COMPILED_PATH /run/current-system/profile/lib/guile/2.2/site-ccache:/run/current-system/profile/share/guile/site/2.2 <wigust>dvn: Those are only system profiles <wigust>dvn: If you installed guile-reader in your user profile, you need to have "$HOME/.guix-profile/share/guile/site/2.2/". <wigust>ACTION doesn't know why this is not handled automatically <wigust>dvn: So, something like export GUILE_LOAD_PATH="${GUILE_LOAD_PATH}:$HOME/.guix-profile/share/guile/site/2.2/" in your .bashrc <wigust>haunt exports some GUILE_LOAD_PATH, but seems missing path to reader <dvn>okay, makes sense. the explanation. i don't get why that isn't automatic <wigust>dvn: "cat $(which haunt)" will show that it's just a wrapper and currently missing path to reader. <fogbugz>What's the equivalent to Nix wrapProgram in Guix, wrap-program? I have a new package working for LyX (a LaTeX editor), my first Guix package. I'm just missing to wrap the binary so that calls to python are made to the python input, not resolved dynamically. But I'm somehow . Any sinot doing it properlymple example code from another package I can look at? <wigust>dvn: haunt package needs some tweaks to work without manual investigation. I wonder is it common for most of Guile packages? Maybe we need to have a "export GUILE_LOAD_PATH=…" in /etc/profile? <wigust>fogbugz: wrap-program is for inserting environment variables. You probably want patch-shebang or substitute* <fogbugz>wigust: thanks. Is this documented somewhere? <wigust>fogbugz: It's in guix/build/utils.scm. Also grep for examples. <Gamayun>fogbugz: Great that you've got LyX working :) <fogbugz>Gamayun: oh you use LyX? It's for my dad and his mates. They typeset papers in LyX instead of learning LaTeX! <wigust>fogbugz: Fill free to ask here or send a mail to mailing list. We will help as much as we could. <Gamayun>I used to at uni, before learning a bit more LaTeX... Was good as a gentle way into it. :) <fogbugz>Gamayun: cool, yes its a great gateway to LaTeX <lfam>mb[m]1: I just pushed some changes to the grafted replacements of libarchive and libxslt. Should they be ungrafted on core-updates or left as-is? I'm not sure if the low-level packages are frozen on that branch yet or not <mb[m]1>lfam: gcc was recently modified on core-updates, so feel free to merge and ungraft. <lfam>mb[m]1: Okay, I'll do that shortly <lfam>mb[m]1: I merged master into core-updates and push <fogbugz>D GNOME looks anworks great, except for a few bugs. It's two versions behind upstream, so it would be great to upgrade to get some bugfixes. Is it out of date because it's not popular in GuixSD? <efraim>We have the meson build system so that's not holding us back anymore <efraim>I think we just need someone to update all the packages and reconfigure and make sure it still works nicely <fogbugz>pkill9: sorry typorry Screen and a mobile are not a good combination, it was a typo <fogbugz>Yeah, in any case, just to get my point through, GNOME is now on 3.26, Guix has 3.24 packaged. <fogbugz>It works great, but there are some bugs that have been fixed on 3.26, thus i was wondering whether people simply use Xfce or something else. <cbaines>Has anyone had success with putting the store on a separate device? I've just tried, but grub can't find the bzImage file <cbaines>I've had a look at how Guix generates the grub configuration, and it seems to make the assumption that the store is on the root mount point, if I'm interpreting it correctly <fogbugz>efraim: efraim: ok, i had missed your message about meson <pkill9>cbaines: i there may be a compile flag to make Guix look for the store in a different location <pkill9>or maybe you can pass an environment variable <cbaines>Hmm, I've found the boot-parameters record now, which appears to have something to do with the device used... <cbaines>I've got as far as determining that device file systems will be silently ignored when attempting to use a separate store, but I'm having trouble joining that up to the generation of the grub entries <cbaines>I've managed to generate a more promising grub config by switching to using a UUID for the store device, but Grub can't find that either yet... <civodul>cbaines: you're getting broken grub.cfg's? <cbaines>Well, I think I'm trying some unsupported things <cbaines>I think I'm having problems with the mount point stripping now, as I'm trying to mount a subvolume containing the store <civodul>this is supposed to work, there's an installation test for that <cbaines>I think it's the addition of the btrfs subvolumes that are causing the problem <cbaines>the mount point stripping code isn't aware of the relevant mount options, and there also isn't a way to configure a different prefix <kmicu>cbaines: thank you for investigating. I have had Nix store on a subvolume. It worked w/o any problems so there must be a fix for GuixSD too. <cbaines>kmicu, I've got it working now, the two changes I had to make were: using a uuid, rather than a device for the file system, and tweak the stripping of the mount point prefix <pkill9>for now i will just use the binary download <rekado>civodul: I have an old profile which can no longer be built since the addition of man-db <rekado>problem is a broken symlink in an old python package <rekado> /gnu/store/vcx1n5nj4gr52xx5m6gvi7zrwngy06s3-python-2.7.11/share/man/man1/python.1 points to a non-existant file; this causes “man-page->entry” to fail. <rekado>I think broken packages in a profile should not prevent one from upgrading individual packages in that same profile. <rekado>my proposal is to filter the list “man-files” returns <rekado>to make sure the files actually exist