IRC channel logs

2019-01-06.log

back to list of logs

***alice_ is now known as algrey
<algrey>Hello! I was wondering if anyone knows how to get the touchscreen working on the Thinkpad X200T in guixsd?
<lsl88>Hi!
<lsl88>is ipfs get the right command to fetch a hash?
<lsl88>I mean, fetch the files associated to a hash?
<algrey>Hello again. I asked a question earlier about getting the touchscreen working on my thinkpad X200T and then lost connection to freenode so I'm not sure if anyone responded.
<adfeno>ag
<adfeno>.
<adfeno>a,
<adfeno>(sorry, wrong keys) algrey: have you done this: https://www.gnu.org/software/guix/manual/en/html_node/X-Window.html#X-Window
<adfeno>Particularly, look for "Scheme Procedure: xorg-configuration-file"
<algrey>adfeno, Do I add this to my config.scm?
<adfeno>I don't use GuixSD (only Guix) but if I remember correctly, yes.... Although you do have to tweak it to your liking
<adfeno>Particularly, as the manual points out: You must keep `MachIsKeyboard "on"` for touchpad to work
<algrey>Ok I'll try that out
***rekado_ is now known as rekado
<rekado>lsl88: yes, that’s the right command
<pkill9>why does running `guix build --with-graft=mesa=hello godot` start rebuilding packages? isn't it supposed to just create new packages with rewritten paths?
<pkill9>(i'm only using those packages as a test)
<pkill9>gotta go
<apteryx>emacs-guix problem: symbol's value as variable is void: guix-current-profile
<apteryx>M-x guix p
<rekado>apteryx: there’s a discussion on the mailing list about this.
<apteryx>ok, good to know it's a known issue
<rekado>how did you install it?
<apteryx>from guix
<apteryx>have to logout to test something
<apteryx>this is an interesting blog post: https://www.gnu.org/software/guix/blog/2018/customize-guixsd-use-stock-ssh-agent-everywhere/
<apteryx>sadly I don't see how the .xsession content appearing there should work with SLiM
<apteryx>I've tried it on my machine and it breaks the login
<apteryx>ah, it seems I should add more to it: "Your ~/.xsession is expected to do whatever is necessary to customize and then start the specified desktop environment.". I don't use xsession normally, so SLiM start ratpoison by some other mean.
<apteryx>oh, chmod +x .xsession must help
<asterope>How does "version" field works in packages? Is guix checking somehow if new version is bigger/smaller than the previous?
<rekado>asterope: could you please clarify your question?
<rekado>the version field simply contains a string.
<rekado>there are procedures for comparing version strings that are used by some parts of Guix.
<asterope>sorry for the indirect question. I just have a little problem with versioning my package
<asterope>in short: I need to version somehow a package that follow its upstream's HEAD
<asterope>There are some "stable" releases, but they become outdated very quickly
<rekado>do you use git commit hashes as the version string?
<rekado>that won’t work, because they have no meaningful ordering.
<rekado>in Guix we use (git-version …) with a manually increased revision number.
<asterope>I was worried that just a commit has wouldn't work... thanks
<efraim>I think I fixed gcc-boot0 on core-updates for non-intel systems
<joshuaBPMan>hmm, I was hoping that setting up guix's tor service would allow icecat to access tor hidden services, but apparently not.
<joshuaBPMan>I guess it just allows icecat to browse the web through a tor proxy.
<civodul>Hello Guix!
<efraim>i'm not convinced recutils@1.8 needs bash:include
<efraim>ok, I take it bacj
<efraim>`guix refresh -t gnu` says gnunet 0.11.0 is out
<quiliro>hello
<quiliro>i am thinking of returning to GSD...But I am not sure yet
<quiliro>I wonder if I did the right export of PATH
<quiliro>I always get this message:
<quiliro>export PATH="/home/quiliro/.config/guix/current/bin${PATH:+:}$PATH"
<quiliro>might be needed...
<quiliro>I have /home/quiliro/.guix-profile/bin and /home/quiliro/.guix-profile/sbin on my current PATH variable
<quiliro>Is that good enogh? What is the difference between:
<quiliro>/home/quiliro/.config/guix/current/bin
<quiliro>and
<quiliro>/home/quiliro/.guix-profile/bin
<quiliro>?
<brendyyn>quiliro: so there is the guix package manager it's self, installed to .config/guix/, and then .guix-profile, which are all the packages installed by guix for your user
<quiliro>brendyyn: oh... so i must include both in my PATH?
<brendyyn>yep
<quiliro>brendyyn: cool...thanks
<brendyyn>and you musn't install guix to your profile
<quiliro>yes....i did 'guix package -i guix'
<brendyyn>you shouldn't do that
<brendyyn>remove it with guix package -r guix
<brendyyn>and then ensure your PATH is set so that the guix in .config/... is used
<quiliro>is having 'source $HOME/.guix-profile/etc/profile' sufficient?
<quiliro>i have removed guix now
<quiliro>with 'guix package -r guix'
<brendyyn>you can think of guix as like a special package that's treated separately from the rest. if you have guix in your profile, when you run 'guix pull' to update guix, it will put the updated guix in .config/... but the old guix will still be in you profile, and depending on how PATH is set, this old guix may still overshadow the new one, and so you'll discover you are still on the same version even after
<brendyyn>succefully updating
<brendyyn>then if you run 'guix package -u guix', this will not work either, because you'll use the same old version to update... to it's self. so you must use the guix in .config/...
<brendyyn>you can run 'which guix' to check if you are using the right path
<quiliro>/usr/local/bin/guix
<quiliro>i installed guix when emacs-guix would not run Meta-x guix
<brendyyn> export PATH="/home/quiliro/.config/guix/current/bin${PATH:+:}$PATH"
<brendyyn>and run guix pull
<quiliro>after 'guix package -i guix' it worked well
<brendyyn>i think should make it work
<quiliro>nothing to be done
<quiliro>i just did guix pull about 30 minutes ago
<quiliro>before removing guix
<quiliro>so now it will not have anything to update after I ran the commands you suggested
<brendyyn>should be ok once the PATH is set right
<quiliro>so the PATH was correct before?
<brendyyn>guix pull just updates guix its self. its like running Pacman -Sy
<quiliro>yes
<brendyyn>or like pacman -Sy; pacman -S pacman
<quiliro>but it did not include /home/quiliro/.config/guix/current/bin
<quiliro>and it updated guix correctly
<brendyyn>you ran guix --version to check the change?
<quiliro>since it had nothing to do after running the commands suggested by you a moment ago...
<quiliro>will do now
<quiliro>guix (GNU Guix) 0f763db448758c8b7b9ae91d6870f75fc0ff987b
<quiliro>is that good to go?
<brendyyn>yep
<quiliro>so the PATH was correct before then?
<brendyyn>but its /usr/local/bin/guix?
<brendyyn>i think so
<brendyyn>but confused why its using that /usr path apparently
<quiliro>/home/quiliro/.config/guix/current/bin/guix
<brendyyn>ok thats correct now
<quiliro>now it is....before it was /usr/local/bin/guix
<quiliro>but i will test by closing the terminal and opening again
<brendyyn>export GUIX_LOCPATH=$HOME/.guix-profile/lib/locale
<brendyyn>i have that too
<quiliro>which guix
<quiliro>/usr/local/bin/guix
<quiliro>those PATHs should be included in the manual for foreign distros
<quiliro>brendyyn: you use a foreign distro?
<quiliro>I use Debian temporarily
<brendyyn>i believe it is in the manual somewhere but its not so obvious to find it all
<brendyyn>i have both.
<quiliro>which foreign distro?
<brendyyn>Arch
<quiliro>double boot or vm?
<brendyyn>two different computers
<quiliro>cool
<brendyyn>i use them as one via synergy
<quiliro>synergy
<quiliro>?
<quiliro>i will investigate
<quiliro>do you offload?
<brendyyn>I did setup offload
<quiliro>nice
<brendyyn>for that there is a bit more setup
<quiliro>yup
<brendyyn>you neeed /etc/environment set
<quiliro>i will test it some time
<brendyyn>like
<brendyyn>GUILE_LOAD_PATH=/home/b/.guix-profile/share/guile/site/2.2:/home/b/.config/guix/current/share/guile/site/2.2
<brendyyn>GUILE_LOAD_COMPILED_PATH=/home/b/.guix-profile/share/guile/site/2.2:/home/b/.config/guix/current/share/guile/site/2.2
<brendyyn>atleast, that works so i just left it
<brendyyn>then ssh keys without password, and guix archive --generate-key on both
<brendyyn>anyway, im going afk now. night.
<quiliro>brendyyn: night! :-)
<quiliro>morning here!
<quiliro>sleep well
<quiliro>ping
<adfeno>hi there quiliro
<adfeno>Also, #guix, does anyone know a good email filtering or splitting rule to place Debbugs replies in a general "bug guix" email directory?
<adfeno>I thought I could use List-id as a variable, but if someone replies to me directly and puts the bug's email address in Cc, then my email client doesn't know that it's related to Guix, only that it came from debbugs.gnu.org (which also happens to have other projects using that bug tracker)
<_tibbe>adfeno: I use `mu`s guile support to just filter cc, bcc and to for the debbugs mailing list address.
<adfeno>_tibbe: However I want to point out that the replies are sent via Cc to [bug number]@debbugs.gnu.org, and other projects also use debbugs.gnu.org, thus just by the email address there is no way to filter/split it correctly. Unless one goes to wild chase of all open bugs related to Guix and adds each to a regular expression match based on that domain name.
<efraim>I can check what I use(d) when I get home
<adfeno>OK, no problem :D
<adfeno>Hm, strange that guix build icecat fails if add --enable-debug and --enable-debug-symbols options
<adfeno>_tibbe efraim: Please forget my last question regarding filtering/splitting messages related to bug-guix. Apparently the message that came to my inbox didn't come with all the header fields *normally* set when someone puts [bug number]@debbugs.gnu.org in the To field. Instead, the person put my address there, and the Debbugs one in Cc, thus causing the issue.
<reepca>rekado: ah, turns out I didn't write my test variation of the package quite like the original. I fixed that, here are the new environment variables the flake8 build reports: https://paste.debian.net/1058919/
<reepca>(previously it contained python2-python2-configparser in PYTHONPATH, got that fixed)
<reepca>still fails just like the original, of course.
<amz3>hey #guix
<apteryx>hey
<leungbk>i tried asking this a few min ago unauthenticated, so i don't know if this actually reached the channel:
<leungbk>when i try running "gpg --full-generate-key" and fill everything in, i get "gpg: agent_genkey failed: No pinentry Key generation failed: No pinentry", whether or not i have pinentry installed
<apteryx>leungbk: try installing pinentry?
<leungbk>i installed it and get the same msg
<leungbk>pinentry-program /home/brian/.guix-profile/bin/pinentry-gtk-2
<leungbk>is what my gpg-agent.conf looks like
<leungbk>i wrote that in after seeing a similar issue in the guix mailing list
<pkill9>leungbk: did you kill the gpg-agent already running before you changed gpg-agent.conf ?
<leungbk>thank you, that was indeed the issue
<apteryx>looking at our pinentry programs, I'm wondering why we have pinentry-emacs and emacs-pinentry. Do I use the former to use the later?
<apteryx>Do I need*
*apteryx is reading about it here: https://lists.gnu.org/archive/html/help-guix/2018-03/msg00283.html
<THFKA4>anyone else getting "no code for module (guix build utils)" during a guix pull in the last couple of days?
<rekado>THFKA4: this is a non-deterministic bug in Guile.
<rekado>you may need to use a different version of Guix to run “guix pull”.
<THFKA4>thanks! i'll give it a try
<amz3>can somehow install e3 and tell if it starts please?
<joshuaBPMan>amz3: just a moment. What is e3?
<apteryx>a WM
<joshuaBPMan>oh. grr. ok. I guess I can still do that for you. haha
<joshuaBPMan>what's cool about it?
<joshuaBPMan>I'm personally using the awesome window manager, and it works pretty swell.
<pkill9>the e3 package is a text editor apparently
<apteryx>ah.. then I'm confused :-)
<reepca>did you mean i3? i3 is a window manager
<apteryx>must have mixed enlightment and i3 ;-)
<apteryx>joshuaBPMan: it should be easier to test at least :-D
<amz3>sorry, for the noise, i debugged the situation
<amz3> http://lists.gnu.org/archive/html/guix-devel/2019-01/msg00067.html
<amz3>related convo
<amz3>there is a version of the linux kernel that break e3 the editor
<apteryx>weird
<amz3>what the status of offloading substitutes to IPFS, CDN etc...
<apteryx>pinentry-emacs and emacs-pinentry would be obsolete according to emacs-pinentry's author: https://lists.gnu.org/archive/html/emacs-devel/2018-01/msg00264.html
<apteryx>I'm trying the suggested modern way to achieve Emacs buffer GPG passwords input: (setq epa-pinentry-mode 'loopback)
<apteryx>amz3: I don't know it, but talking for myself I haven't had time to hack on gnunet integration yet
<apteryx>oh well, doesn't work with magit
<quiliro>adfeno: thank you
<adfeno>Hi there quiliro, you're welcome
<adfeno>However, I want to know what is the reason for pinging someone? (If it's just for greetings, please ignore this question)
<quiliro>i think my afair with guix is going serious...i hope my partner will not be jelous!
<adfeno>Hehehehe
<quiliro>adfeno: i had no messages. so i thought i was disconnected. that is why i sent a ping. to know if i was getting through
<janneke>apteryx: oh no, i'd hate to lose emacs-pinentry again
<quiliro>on other news, i think the installation manual for foreign distros misses some paths to add
<adfeno>Hm, which ones?
<quiliro>export GUIX_PROFILE="$HOME/.guix-profile"
<quiliro>source "$GUIX_PROFILE/etc/profile"
<quiliro>export PATH="$HOME/.config/guix/current/bin${PATH:+:}$PATH"
<quiliro>this solved all my problems
<quiliro>for the users...i have not checked if the root needs something else...i think it is ok....if it is not broken, don't fix it
<quiliro>i see we have guix-0.16.0-8
<quiliro>it would be nice to have a hacker ready GSD....nevertheless a hacker should be ready with anything....the more disassembled, the better
<janneke>hmm, waiting for locks or build slots...
<janneke>restarted local and build host daemon, to no avail
<janneke>ah, the remote guix-daemon didn't kill all its children with herd stop guix-daemon...
<janneke>samplet: i have now three simple awk test passing and created a minimal failing test that we must pass in order to support autoconf >= 2.62
<quiliro>i get squares on icecat...i guess fonts are not recognized
<quiliro>but i already added them
<quiliro>failed to create cairo scaled font, expect ugly output. the offending font is 'DejaVu Sans 9
<quiliro>font_face status is: file not found
<apteryx>janneke: the situation is so confusing (and judging by the thread here, I'm not alone: https://lists.gnu.org/archive/html/help-guix/2018-03/msg00310.html).
<reepca>sneek: botsnack
<sneek>:)
<reepca>oh good, you're there
<janneke>apteryx: yeah, i figured that much .. good to know we're not alone
<apteryx>pinently.el was incorporated into Emacs 25, only to be removed from Emacs 26; it now stands alone as emacs-pinentry. Its job seems to be so that any request to gpg goes through Emacs (with a dependency on pinentry-emacs?). Meanwhile, a new configuration was added (epa-pinentry-mode) to Emacs, that should *just* work regardless of the pinentry flavour installed as it uses a loopback mode directly exposed from
<apteryx>gnupg itself.
<reepca>sneek: later tell quiliro: have you tried 'fc-cache -f'?
<sneek>Will do.
<janneke>apteryx: ok, that sounds like good news...
<apteryx>*but* the epa-pinentry-mode will not redirect the requests to gpg that originated outside of the Emacs gpg library calls (epa-*), so it's useless say with Magit.
<reepca>so with the exact same sys.path the exact same python fails to import a module that it can import fine outside of the build environment, and that exists in the build environment as well. Any python experts have any insight into this witchcraft?
<apteryx>so, long story short, I think for all the cases we still need pinentry-emacs + emacs-pinentry installed, along with the allow-emacs-pinentry option in ~/.gnupg/gpg-agent.conf
<janneke>apteryx: ow, ..does not with Magit that's wicked!
<adfeno>sneek: later tell quiliro: See https://debbugs.gnu.org/cgi/bugreport.cgi?bug=33996
<sneek>Okay.
<adfeno>reepca: ^
<adfeno>sneek: later tell quiliro: Also, I'm trying to debug this by building IceCat with debug symbols enabled, but Mozilla Firefox's Cargo/RustC build system is giving segmentation fault due to some "invalid memory reference" when compiling a specific file (something like "encoding_c", if I recall).
<sneek>Okay.