IRC channel logs

2018-12-06.log

back to list of logs

<apteryx>wow, just updated my guix and there are so many grafts; is this normal?
<pkill9>apteryx: it might something to do with grafts being made non-substitutable in this commit http://git.savannah.gnu.org/cgit/guix.git/commit/?id=7bc5657f8434de06d904c36f8960a1b1dddf10ae
<lfam>There are usually tons of grafts. By definition, grafts are only used when many packages are affected
<j3kyl_>hey, guix packages are in the PATH, still Debian Dmenu cant "see" those packages, but terminal emulators can...
<j3kyl_>hey, next-gtk-webkit is list in website packages, but guix package -s cant find it...
<j3kyl_>just installed guix with that guix-install.sh script
***catonano_ is now known as catonano
<reepca-laptop>have we got an input method for hanyu pinyin? I see in ubuntu it's called ibus-m17n, but it seems we've only got m17n-lib.
<brendyyn>ive been meaning to package more fcitx stuff'
<apteryx>lfam: I see! Maybe the new UI made me more aware of them.
<lfam>Yes, previously they were reported as "regular" builds
<apteryx>OK, got my OpenBoardView package to run and load GTK libraries dynamically... It now crashes on a dbus error: ** (openboardview:2958): WARNING **: 23:51:30.265: Error retrieving accessibility bus address: org.freedesktop.DBus.Error.ServiceUnknown: The name org.a11y.Bus was not provided by any .service files
<apteryx>openboardview: imgui.cpp:2509: void AddDrawListToRenderList(ImVector<ImDrawList*>&, ImDrawList*): Assertion `(int64_t)draw_list->_VtxCurrentIdx <= ((int64_t)1L << (sizeof(ImDrawIdx)*8))' failed.
<apteryx>Aborted
<apteryx>Any idea?
<apteryx>ah, the assertion error looks like unrelated to the dbus "warning".
<apteryx>I'll focus my efforts on it
***rekado_ is now known as rekado
<rekado>reepca-laptop: we have ibus-libpinyin, which is what I use
<rekado>brendyyn: AFAIK ibus will be integrated much more tightly with GNOME.
<apteryx>how to deal with git submodules? I tried unbundling it, but it failed (they patched their own version, and it's on some random commit)
<lfam>apteryx: You can set recursive? #t in the git-reference
<lfam>This should go in the manual
<brendyyn>which is probably part of the reason i use fcitx
<rekado>brendyyn: I meant that the other way around: GNOME will have better integration with ibus (and using other input methods might become more difficult). You can still use ibus without GNOME, of course.
<reepca-laptop>rekado: aye, I use ibus-libpinyin as well, which is good for pinyin->characters, but not so great for pinyin->pinyin-with-tone-marks
*rekado builds the i686 disk image
<vagrantc>hrm. libpng test suite failure on aarch64
*vagrantc tries to pull again
<vagrantc>seems like it was an intermittant failure
<civodul>Hello Guix!
<vagrantc>hmmm... still not getting many substitutes for "guix pull" on aarch64
<g_bor>hello guix!
***Server sets mode: +cnt
<vagrantc>the machine i'm building it on is fairly capable, so may as well let it build
<vagrantc>not my little aarch64 vm :)
<civodul>like https://berlin.guixsd.org/xx0zy8cxawzlixklc2kf2hfhzky1l5jr.narinfo
<civodul>heh
<vagrantc>i was hoping to get guixsd on a microSD or something for the pinebook (to avoid the overhead of a vm) but never got it that far
<g_bor>hello guix!
<civodul>hey g_bor!
<civodul>vagrantc: in theory "guix system disk-image" should give you something you can directly dump on the SD card
<vagrantc>civodul: the theory hasn't panned out too well
<g_bor>civodul: I have seen this i686 corrupt iso...
<vagrantc>i guess i could do --no-bootloader and then manually install the bootloader
<g_bor>is that problem reproducible?
<civodul>vagrantc: oh, in theory practice is like theory
<civodul>g_bor: yes, 100%
<civodul>well, it seems
<civodul>though i found that the iso-installer-image test worked recently on berlin
<civodul>so, dunno
<civodul>i'm trying various things but the problem space is still way too large
<g_bor>do we have any working iso after the core-updates merge?
<civodul>yes there's no problem on x86_64
<civodul>it's really just i686
<civodul>which is probably why we didn't notice earlier...
<g_bor>I mean i686.
<g_bor>I got that we are ok on every other system, is that correct?
<civodul>for i686, /gnu/store/ywkib2m7jkmi2qiivaq2zxggs9xpppcv-iso-image-installer.drv from commit 69a64ecd66733e394a3b116b38f67b354dcebcc9 worked well
<civodul>g_bor: the binary tarballs are OK, the x86_64 GuixSD images are ok
<g_bor>hello guix!
<g_bor>One Ubuntu download is so aggressive on bandwith here that the net became unstable...
<g_bor>lsl88 asked what foreing distro should we use in the video documentation.
<g_bor>Any suggestions?
<g_bor>They already has Ubuntu installed, I believe that might be ok for this type of documentation, given that it is quite popular. WDYT?
<g_bor>another option would be to user for example trisquel or parabola
<rekado>civodul: I don’t see any problems with the raw disk image; building the ISO next.
<rekado>g_bor: does it really make a difference?
<rekado>the video would focus on the steps to install, and I think that’s all done in a terminal anyway
<rekado>so viewers don’t need to know anything about the host distribution, I think.
<brendyyn>How can i install packages that match a search query?
<civodul>rekado: you really need "guix system disk-image --file-system-type=iso9660 -s i686-linux"
*civodul goes to a meeting :-/
<civodul>later!
<rekado>I can reproduce this with the ISO.
<rekado>say it wants to access 1890176 on sda, but the limit is 1886976
<rekado>I note that the file image has 1896603 sectors, so accessing sector 1890176 should work.
<rekado>I also see that this is a real problem with the image file, as I also get an I/O error when mounting the ISO and access the db.
<lsl88>hi guix!
<lsl88>g_bor: yes I thought that ubuntu was fine because of showing the commands in the console, but yes, I was going to add that even I use the console maximized, gnome still shows the panel. but we are not going to screencast, so I guess it is OK. as rekado says.
<lsl88>does someone know how to copy/paste text in a tty? I have been looking for it on the internet and did not find it, unless using tmux, byobu
<rekado>in a tty I don’t know of a convenient way. There’s gpm for mouse selection support on the tty, but I think it’s a bit inconvenient.
<rekado>I think it would be better to do this in a graphical terminal (where font rendering is better anyway).
<lsl88>rekado: I know, but the point is that I cannot manage to open tty's in qemu, while installing guixSD, which is quite strange, or maybe there is a way of setting it that I don't know. It is not terrible, they are just a few commands that I have already kind of remember without looking at the documentation. But I was asking just in case, since I did not know that shitf+pgup for instance allowed me to scroll
<rekado>lsl88: does this work for you? https://askubuntu.com/questions/54814/how-can-i-ctrl-alt-f-to-get-to-a-tty-in-a-qemu-session
<lsl88>rekado: I have already seen that site, but it doesn't work. I cannot test it since when I run the vm, I already have the GUI. And the mouse turns crazy. it starts moving all over the window. My machine is a Lenovo thinkpad, and the f* work with the fn control key
<lsl88>rekado: maybe I could create another VM just to test those solutions and see if any of them work. But for instance, for bare metal installation on my retroPC, the mouse worked for copypasting commands in tty2.
<rekado>lsl88: I just tried it and the sendkey method works fine for me.
<rekado>i.e. first hit Ctrl-Alt-2 to switch to the qemu console
<rekado>and then type “sendkey ctrl-alt-f2” on the prompt
<rekado>then switch back with Ctrl-Alt-1
<lsl88>rekado: I am trying to solve the mouse stuff, i am reading the qemu documentation to see why this happens. Which sendkey are you using?
<lsl88>rekado: sorry I did not see your last messages
<lsl88>rekado: I tried the ctrl alt 1 and 2 but nothing happens. Sorry, by the mouse stuff I meant the mouse getting crazy in the GUI moving all around the window when I click sth
<lsl88>rekado: I managed to open a terminal inside the GUI, but sendkey command does not exist.
<rekado>lsl88: Ctrl-Alt-2 should bring up the qemu console. “sendkey” is a qemu console command, not an executable in the guest system.
<rekado>how did you start qemu?
<rekado>bah, I’ve built the Python bindings for tensorflow, but I cannot use them. I don’t even know why it’s failing.
<lsl88>rekado: sorry for the delay, I solved the crazy mouse stuff :)
<lsl88>rekado: Ctrl-Alt is used for grabbing the mouse, but after my last change, it doesn't work anymore BUT solved the mouse issue
<lsl88>rekado: here is my command, and except for the tty stuff, the GUI seems to work fine
<lsl88>rekado: qemu-system-x86_64 -m 4096 -nodefaults -drive file=hda.img,format=raw,if=none,id=drive-ide0-0-0 -device ide-hd,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0,bootindex=1 -device qxl-vga -net user -net nic,model=virtio -usbdevice tablet
<lsl88>rekado: I am only getting after running that command: qemu-system-x86_64: -usbdevice tablet: '-usbdevice' is deprecated, please use '-device usb-...' instead
<lsl88>qemu-system-x86_64: warning: TCG doesn't support requested feature: CPUID.01H:ECX.vmx [bit 5]
<rekado>lsl88: and what are the contents of hda.img? Is this a GuixSD image?
<rekado>civodul: I’m building the ISO again without compression and with an explicit “-pad”.
<lsl88>rekado: yes, I'd better use paste in debian the commands I ran
<civodul>rekado: "-pad" adds extra space in the image?
<civodul>on my side i'll try to have the ISO-building-VM run x86_64 code
<rekado>it should be enabled by default (when mkisofs is used), but I don’t know if that’s true for xorriso / grub-mkrescue
<rekado>it adds a few kbs to avoid IO errors
<civodul>oh that looks like exactly what we need
<rekado>unfortunately, that didn’t help :(
<rekado>I get the I/O error even before booting has finished.
<civodul>you tried "-padding 10m"?
<rekado>no, just “-pad”, which is a mkisofs option
<civodul>oh
<civodul>i'm still confused about xorriso, mkisofs, and all that
<civodul>the xorriso info page mentions "-padding"
<civodul>but you're right: grub-mkrescue does "xorriso -as mkisofs"
<rekado>according to grub-mkrescue we can pass xorriso arguments after “--”.
<rekado>so it’s worth trying with “-padding 10m”
<civodul>by chance that's what i was trying :-)
<civodul>in (gnu build vm)
<joshuaBPMan>hello, does the package emacs-all-the-icons? not install the icons for you?
<lsl88>rekado: Here is what I am doing https://paste.debian.net/1054650/
<joshuaBPMan>Because it seemed that I had to install said icons. Is there some way that we could modify the all-the-icons package to have the icons as their inputs?
<joshuaBPMan>*propagated-inputs
<rekado>lsl88: the warning about CPUID.01H:ECX.vmx means that your CPU does not support virtual machine extensions.
<rekado>this results in very poor virtualization performance.
<lsl88>rekado: that did not happened with vBox, it is quite strange. Should I add more cores, check sth in my BIOS, and that kind of stuff? or do you believe it is useless?
<civodul>rekado: victory!
<rekado>civodul: YES!
<rekado>civodul: what did you do?
<civodul>it's a bit sad that the conclusion of all this is a kludge, but it's so cool that we're done :-)
<civodul>rekado: "-padding 10m"
<civodul>then i waited
<rekado>heh, okay
<civodul>i mounted the image and it's alright
<civodul>i booted it and it's fine too
<civodul>so i guess we're all set
<rekado>awesome!
<rekado>lsl88: you could run lscpu and look at the “Flags” line
<rekado>lsl88: if it contains “vmx” then the host CPU does have virtual extensions, but it isn’t available for qemu for some reason.
<lsl88>rekado: I believe the same happened to Gabor with Ubuntu. Wrote: This is the result of Ubuntu configuration of their qemu package.
<rekado>you may have to use a different variant then
<rekado>I think they may also have “kvm” or “qemu-kvm” and so on
<rekado>or… you could install qemu from Guix ;)
<joshuaBPMan>does guix not support MIT license? I'm not seeing it in the licenses.scm file.
<rekado>joshuaBPMan: there are different licenses that are commonly called “MIT license”
<rekado>joshuaBPMan: a common one is the Expat license.
<joshuaBPMan>ahhh.
<yrk>joshuaBPMan: http://www.gnu.org/licenses/license-list.html#Expat
<joshuaBPMan>Well, I guess the atom file-icons is expat then.
<joshuaBPMan>At least it's the same language as the expat
<joshuaBPMan> https://github.com/file-icons/atom/blob/master/LICENSE.md
<lsl88>rekado: it has the flag. Could that be because I did not uninstall vBox? Of course I am not running both Hypervisors at the same time. And I promise to use a GuixSD T-shirt over my clothes in Brussels and I am telling everybody about guix. I am still running guix over another VM just because even Ubuntu breaks out of nothing, I like trying commands and I am still learning and I don't want to break my machine if I make a mistake. I don't m
<lsl88>nd uninstalling vBox and running another foreign distro for going on playing with guix. WDYT? I can imagine your answer ;)
<rekado>I don’t think it’s because vBox is installed.
<rekado>I’m afraid I don’t have enough experience with Ubuntu to figure out why the virtual extensions won’t work.
<rekado>I only have one guess:
<rekado>maybe your account needs to be in a certain group to use kvm
*rekado needs to leave now
<lsl88>rekado: OK, I will search for that on the internet. I don't like Ubuntu for the freedom stuff and for other reasons. I would never recommend it for a server, for instance. But I need it for other stuff.
<joshuaBPMan>Hello, I'm a little confused. Why does the package emacs-matrix-client specify propogated inputs like "a, dash, f", but not "a@1.0, dash@0.5, f@.03.2"?
<joshuaBPMan>Isn't the point of guix that you explicitly specify dependencies versions?
<roelj>joshuaBPMan: If you mean the package definition, then the "a", "dash", and "f" point to Scheme variables (which are a specific version for a given git commit of Guix).
<nly>thank you for packaging sbcl-next :)
***ChanServ sets mode: +o civodul
***civodul changes topic to 'GNU Guix | https://gnu.org/s/guix/ | videos: https://gnu.org/s/guix/blog/tags/talks/ | bugs and patches: https://issues.guix.info/ | paste: https://paste.debian.net | Guix in high-performance computing: https://hpc.guixsd.org | 0.16.0 is out! https://gnu.org/s/guix/blog/2018/gnu-guix-and-guixsd-0.16.0-released | This channel is logged: https://bayfront.guixsd.org/.well-known/logs/'
<civodul>it's here! https://gnu.org/s/guix/blog/2018/gnu-guix-and-guixsd-0.16.0-released
<civodul>\o/
<davexunit>congrats on the release, folks!
<kmicu>ヽ(*^▽^)/
<pkill9>٩(◕‿◕。)۶
<mbakke>Aw nckx had 3 commits more than me in this release :-)
<lsl88>:-D:-D:-D
<enderby>hi, i just read the latest blogpost that says the default substitute server has changed to ci.guix.info, do I need to change anything in my guix config to start using it?
<enderby>I had setup guix a while back with step 7 https://www.gnu.org/software/guix/manual/en/html_node/Binary-Installation.html#Binary-Installation
<vagrantc>is it feasible to upgrade a binary guix installation?
<enderby>me?
<vagrantc>i'm asking for my own interests
<enderby>ah
<vagrantc>enderby: regarding your question, you may need 'guix archive --authorize < /PATH/TO/NEW/KEYS' and edit guix-daemon.service to point to the new servers
<enderby>vagrantc: icic, thanks will look into it!
<g_bor>hello guix!
<g_bor>congrats on the release!
<rekado>release party!
<emacsen>Hi. I'm a guix noob, and trying to get it going. I'm on a Ubuntu system and I used the install script. My installation keeps complaining about locale, even after I run the locale installation and env variable commands as it asks me to
<g_bor>emacsen: most probabl you need those command to be executed from the root profile also, and then restart the daemon.
<emacsen>g_bor, gotcha. Okay, let me try that
<g_bor>but it is just a warning, it should not cause problems...
<emacsen>g_bor, is that all programs or just the locale package?
<g_bor>emacsen: only the locale package.
<g_bor>and the environment variable
<emacsen>g_bor, okay tried that. still getting the warning. Trying to get the other programs I want working in the meantime
<g_bor>emacsen: this locale stuff can be very tricky to get right...
<g_bor>but aside from the warning it usually does not do any harm
*emacsen is gonna try to get hexchat and audacity working
<emacsen>and mumble.
<g_bor>Now that we have a new release, I will try to get a powerful hardware working tomorrow, and have a look at the current state of the locale stuff.
<emacsen>g_bor, it may be indeed no big deal but as a user, when you finish the installer and are greeted with an error, it's jarring
<davexunit>it's a tricky thing
<davexunit>and the issue goes all the way down to glibc
<emacsen>brb, gonna use the guix hexchat :)
<davexunit>we have a use-case that the glibc devs don't think is worth addressing
<emacsen>okay, that didn't work. sad sad
<dustyweb>hi
<dustyweb>what's up
<g_bor>hi!
<Sleep_Walker>it seems that nss-certs has some weird sum https://ptpb.pw/AiGb
<Sleep_Walker>especially for this package it should be fixed :)
<lfam>Sleep_Walker: Try using --fallback in the meantime
<Sleep_Walker>ack
<lfam>Sleep_Walker: I think this is the bug report: https://bugs.gnu.org/32942
<Sleep_Walker>I see
<Sleep_Walker>thanks!
<Sleep_Walker>sorry for the noise
<lfam>It's not noise :)
<efraim>any interest in an IETF RFC guix channel?
<lfam>What would that mean?
<efraim>A guix channel with the thousands of rfcs they've put out :)
<efraim>in .txt and .pdf format
<lfam>Oh... I think you meant like an IRC channel :)
<jonsger>Sleep_Walker: I have update guix in devel languages, but I'll wait until guile-2.2 lands in Factory and after that I'll submit the update...
<Sleep_Walker>jonsger: good!
<apteryx>efraim: that would be interesting!
<efraim>the hard part now is figuring out how to format the input from shell scripts :)
<efraim>i might have to drop the pdf for now, it's not downloading reliably
<lfam>Has anyone experienced ext4 filesystem corruption on GuixSD since updating to linux 4.19? <https://bugzilla.kernel.org/show_bug.cgi?id=201685>
<lsl88>emacsen: hi! I faced that issue before and happlily solved it :) i am guix pulling over my qemu machine now, so my RAM is hating me. I have to go now but later I share in paste debian how I solved :)
<emacsen>lsl88, awesome thank you