IRC channel logs


back to list of logs

<civodul>lfam: --keep-going just means that it tries to build as many of the derivations as possible instead of stopping at the first failure
<civodul>but if one of the derivations fails, it eventually stops anyway
<lfam>Okay, sounds good :)
<civodul>it's safe :-)
<lfam>I want to run that command on core-updates overnight
<civodul>nice test :-)
<lfam>Yeah, I already did it for several other installations, but not for my user on my workstation
<lfam>I just pushed the libtiff fix to master. Do you think we should also do it as a graft on core-updates, since it will require ~1500 rebuilds per architecture?
<lfam>civodul ^
<lfam>nckx: Cool Unicode "leaving" message :)
<civodul>lfam: yes, i think so :-/
<nckx>lfam: thanks ^_^ Just re-booted my ZNC host running GuixSD.
<lfam>I made a typo in the libtiff commit message. I called the grafted libtiff 'libtiff-fixed' instead of 'libtiff/fixed'
<civodul>heh, that's ok ;-)
<civodul>ACTION -> zZz
<civodul>good night/day!
<ng0>okay, a whole world of static linking could be too expensive for us. dynamic libcs are a different issue
<ng0>also I don't think there will be much gain in whole world static, just more expense at our side.
<ng0>good night
<Apteryx>Anyone using xterm? I'm struggling just to set the history... I've followed advice from the Arch xterm wiki, but it's not working. I have like 100 lines of history, which is very little.
<Apteryx>At least it works when starting it using command line option: "xterm -sl 1024", so my problem probably lies in my .Xresources file.
<lfam>Apteryx: Doesn't xterm come with some documentation?
<lfam>Distro wikis tend to contain obsolete info in my experience
<Apteryx>lfam: Yes, based on what I read I'd need to enable the scrollBar feature by using a XTerm.vt100.scrollBar: true, then this would enable another option which controls the number of scrollback lines: XTerm.vt100.saveLines: 9000
<Apteryx>I put those lines in my ~/.Xresources file, then try starting xterm but I'm still stuck with ~100 lines of history.
<lfam>Apteryx: I'm not very familiar with configuring X, but maybe you need to restart the X server?
<lfam>Or run some command to make it load the new configuration options?
<Apteryx>lfam: I should try that. This would be a bit different that the other options I've played with so far though, these all take effect when I restart any X application (e.g., colors)
<lfam>I've used `xrdb -merge .Xresources` to make changes to .Xresources take effect
<lfam>And those changes had to do with virtual terminals (colors, fonts, etc)
<Apteryx>OK! I'll try that too. Thanks for your pointers :)
<lfam>But I admit that I don't really know what that does. It's been in my shell history file since I started using GNU / Linux distros
<Apteryx>lfam: Great, the xrdb approach already did the trick! I'll try restarting my session now.
<lfam>Great :)
<Apteryx>I actually had the "xrdb .Xresources" line in my .bash_profile. So re-running this should be just as good as restarting my session :)
<sneek>Common_Era, you have 1 message.
<sneek>Common_Era, lfam says: Assuming you are using GnuPG 2.1, you need to install a pinentry program (probably pinentry-tty) and then give the path to your pinentry program as the value of 'pinentry-program' in ~/.gnupg/gpg-agent.conf. Or, use an older version of GnuPG that doesn't require the gpg-agent / pinentry
<Common_Era>Thanks, lfam.
<Apteryx>I ran the following command, hoping to see the colliding dependencies (xdg-mime-databes/shared-mime-info) appear, but they are not listed there: guix package -I | awk '{print $1}' | tac | xargs -I{} bash -c "guix package --show={} | recsel -p name,dependencies; echo" | less
<Apteryx>Is there a guix command I could use to see the recursive dependencies list of a particular package?
<lfam>Apteryx: `guix graph`
<Apteryx>lfam: thanks again :)
<Apteryx>Sorry, another question which I can't seem to answere myself: is there a guix log which would contain the output I saw when I last installed a package?
<Apteryx>And, on another topic, any package recommended for automounting external drives? I'm not using any heavy DE.
<Apteryx>Which shebang should I use for my scripts when using Guix? It seems I can't use the usual #!/bin/bash.
<Apteryx>Answering myself: /bin/sh --> .../bin/bash, so I could use #!/bin/sh, although I had become to prefer the more explicit #!/bin/bash for when I'm really writing bash script.
<bavier1>anyone know what might be causing a gtk gui to fail to display some custom button icons?
<bavier1>well, there are a few other things missing, so it might be that the gtk resources weren't generated properly during the build
<Apteryx>bavier: maybe it isn't finding the pixmap resources where it expects them to be? (usr/share/...)
<Apteryx>bavier: If you start the GUI from the command line, is there nothing of interest being output?
<Apteryx>While trying to find which two installed package might be causing my colliding deps error, I got into some guix graph errors: scm_flush: Broken pipe. I've pasted a report here:
<Apteryx>I can reproduce everytime.
<kadintrooper>Hello :)
<kadintrooper>Whats up
<Apteryx>Oh, maybe my script is too rude and calls guix graph again before it's done.
<Apteryx>kodintrooper, doing well, trying to track down a colliding dependency issue
<Apteryx>How about you?
<kadintrooper>I'm currently installing guixsd in a vm, seeing if I like
<kadintrooper>I've been meaning to do this for a while
<kadintrooper>Might put it on a partition on my hdd to play around with
<Apteryx>kadintrooper: Nice!
<Apteryx>I've been using it for maybe 2 weeks. Got a hard start since I screwed my main install. But it's a good way to learn!
<kadintrooper>Is it bad that besides the whole declarative thing, I really like the logo?
<Apteryx>kadintrooper: Why would it? :)
<kadintrooper>Maybe because it's like, 25% of the reason I'm trying it
<kadintrooper>I'm a sucker for logos
<Apteryx>Well I hope the one who designed it hears you!
<kadintrooper>Is this irc generally quiet? I'm looking through the logs rn lol
<Apteryx>kadintrooper: It depends! It's kind of low volume. But the SNR is good!
<Apteryx>(signal to noise ratio ;)
<kadintrooper>If this vm image works out, would dd'ing the .img to something like /dev/sdx work?
<Apteryx>kadintrooper: Never tried that, but I don't see a reason why it wouldn't work!
<Apteryx>That's something I actually was planning to try for Debian/Hurd install.
<kadintrooper>Apteryx, I'll try it out and let you know if it works lol, installing the system is taking longer than I thought it would
<Apteryx>But still deep in configuring my new system for now :)
<kadintrooper>So I don't really want to have to do this again
<kadintrooper>Though it's not a problem if I have to
<Apteryx>Yeah, it took some time for me too!
<kadintrooper>Hallo prasunanand
<kadintrooper>Welcome back
<Apteryx>How much free disk space does Guix needs to be happy? I got the error: guix package: error: build failed: while setting up the build environment: creating directory `/gnu/store/isapl2x2f3jx84aza40aiz22igs8jxr5-ca-certificate-bundle.drv.chroot/gnu/store/ykzwykkvr2c80rw4l1qh3mvfdkl7jibi-bash-4.3.42': No space left on device.
<Apteryx>I have a 2.4 GB free on my GuixSD partition.
<kadintrooper>How much did you give it
<Apteryx>Total size is 10 GB
<kadintrooper>Uh oh
<kadintrooper>That's how much I gave mine
<Apteryx>Now, you'll be fine. To start.
<kadintrooper>Did you split it into different partitions
<Apteryx>I'm just wondering, 2.4 GB free space seems like it would be good, but when it wants to setup its build environment it apparently reserves a chunk bigger than that.
<Apteryx>Yeah, the drive is 60 GB total. Not that much real estate.
<Apteryx>I'll probably update my install after I get my configs right.
<Apteryx>And give it something more reasonable, like 40 GB
<prasunanand>Hallo kadintrooper
<Apteryx>For now I'm just curious why Guix complains about disk space while I apperently still have 2.4 GB left :)
<kadintrooper>I dunno
<Apteryx>Good night!
<kadintrooper>Good night :)
<Apteryx>Bed time... Good luck with your install!
<kadintrooper>Thank you :P
<mark_weaver>sneek: later tell civodul: We have a grave problem on core-updates. Hydra is generating different derivations than 'guix' on all non-x86_64 systems. Non-x86_64 systems will have no substitutes available. See <>
<mark_weaver>sneek: botsnack
<mark_weaver>ACTION --> zzz
<fr33domlover>hey davexunit! remember that issue I had with terminal color? rxvt and .Xdefaults? I think I found the issue: In Debian based distros, there is a /separate/ package rxvt-unicode-256color that is required for 256-color support
<fr33domlover>I think that's why Guix's rxvt can't do 256 colors
<fr33domlover>It's simply compiled without that feature :-/
<fr33domlover>I'm happy to help fix that - what would be the way to do it? perhaps add a new package for that like in other distros?
<fr33domlover>hmmm rxvt-unicode docs say the feature /forces/ 256 color mode, not sure it's required. need to test things
<Digit>ACTION dreams of useflag equivalents
<rekado>does anyone know how to get ao-guile to work? I get "Couldn't find current GLX or EGL context"
<rekado>already tried adding mesa libs to LD_LIBRARY_PATH, but that didn't help
<rekado> don't see much in the strace output
<Digit>"guix package: error: failed to connect to `/var/guix/daemon-socket/socket': Connection refused" why am i getting that? / how do i stop getting that? seems on any install command. guix package -s seems ok still.
<davexunit>fr33domlover: ah that makes sense
<davexunit>now write a patch for the 256 color version!
<fr33domlover>davexunit, i moved to the 'st' terminal a few minutes ago :P
<fr33domlover>it just works, no funny color problems
<rekado>Digit: is the daemon running?
<Digit>ACTION looks in his htop & opens for refresher course
<Digit>thnx. nope, it wasnt. not sure how it stopped.
<Digit>not sure how to start it either (command not found, from trying the example there). this is for my guix in devuan btw.
<rekado>how did you install guix?
<rekado>I have old SCSI disks from 1993 and I'd like to access them in Guix to copy off old data.
<Digit>iirc, i followed the manual.
<rekado>anyone here familiar with old SCSI hardware?
<fr33domlover>Digit, i have Guix-in-Trisquel and I set up an init script for upstart like the manual says
<rekado>I have a SCSI controller card and the disk audibly spins up but the kernel tells me that spinning up times out.
<rekado>any ideas?
<rekado>Digit: if you haven't removed the "guix" package from root's profile you should be able to start the daemon as root.
<rekado>Digit: it should still be installed in the root user's profile
<fr33domlover>rekado, i have no idea but if the hardware is fine, maybe you'll have more luck with an old distro?
<fr33domlover>something from days these disks were common etc.?
<rekado>fr33domlover: I only have GuixSD on all my computers. The only computer where I can install the SCSI controller card is my audio recording workstation, and I won't be able to install anything but GuixSD on it.
<fr33domlover>rekado, do you have any hackerspace around?
<rekado>probably :)
<fr33domlover>a place with computers and tech people :P
<rekado>but I usually find the people at hackerspaces very hard to bear :-/
<fr33domlover>rekado, why's that? (curious, never been to a hackerspace)
<rekado>it's the same reasons why I don't really like "Linux" (sic) user groups or large parts of "nerd culture"
<rekado>maybe it's a Berlin thing.
<Digit>twas a PATH issue. a lil addition to root's .profile, n all should be well now. *installs st, satiating that whim*
<Digit>thnx for the pointers rekado. :)
<Digit>rekado: is that "hard to bear" partly to do with the social akwardness & testosterone poisoning etc that oft accompany geekness?
<Gottox> /win 63
<civodul>Hello Guix!
<sneek>Welcome back civodul, you have 1 message.
<sneek>civodul, mark_weaver says: We have a grave problem on core-updates. Hydra is generating different derivations than 'guix' on all non-x86_64 systems. Non-x86_64 systems will have no substitutes available. See <>
<civodul>mark_weaver: nasty bug that you found!
<bavier>I was wondering why I had to build so many packages on my netbook for my last 'guix package -u'
<bavier>well, that bug is on core-updates, so maybe hydra was just a bit behind on master
<efraim>"error: support for the symbol redirection needed" on glibc-intermediate on aarch64
<civodul>bavier: it might also be the problem that Hydra GC'd too much (~3 weeks ago)
<janneke>what exactly does: guix build: error: build failed: some outputs of `/gnu/store/542lf7n6mizx0kywmv44fs7k6pzw7181....drv' are not valid, so checking is not possible
<janneke>what if i still would like to check?
<amz3>janneke: what command did you run?
<amz3>janneke: if it's guix build try guix build --keep-failed
<janneke>amz3 yes; ah! is see. thanks.
<Apteryx>I started getting a "Fontconfig error: Cannot load default config file" anytime I open an app or run "fc-cache". Any idea?
<rekado>yay, got the old SCSI disk to run
<Apteryx>rekado: Cool!
<rekado>I lifted it up to check if all connectors are tight and then it shortly gave a response.
<rekado>now it's in just the right position to read it out with dd
<rekado>didn't require any software trickery, just the magic touch
<rekado>lots of OS/2 stuff on there
<Apteryx>You've got your own hackerspace ;)
<Apteryx>Could someone with an updated Guix from today or yesterday run "fc-cache" ?
<efraim>its like when i try to charge my phone
<Apteryx>You might need to install `fontconfig'.
<iyzsong>Apteryx: did you just run a gc?
<ng0> can someone tell me what's wrong about for-each here :)?
<rekado>ng0: do you get an error?
<ng0>oh, forgot to pastebin that
<ng0>one moment
<Apteryx>iyzsong: Yes.
<iyzsong>Apteryx: oh, in that case, does `guix build fontconfig --no-grafts` works?
<rekado>ng0: does the file exist?
<rekado>ng0: I wonder: could you do this *before* installation rather than after?
<bavier>ng0: you might need to (mkdir-p ...) the target directory
<ng0>after build would work too
<Apteryx>iyzsong: Yes, it returned: /gnu/store/4fs8dg5jhf99xl3ikl093dn1va4wlv79-fontconfig-2.11.94
<ng0>I create the target directories before
<Apteryx>And if I run: /gnu/store/4fs8dg5jhf99xl3ikl093dn1va4wlv79-fontconfig-2.11.94/bin/fc-cache, I don't get the error!
<bavier>ng0: indeed, I missed that
<Apteryx>How can I make my profile's fc-cache point to that one?
<iyzsong>that's really terrible...
<Apteryx>Should I uninstall then re-install fontconfig?
<rekado>Apteryx: uninstall + re-install is a no-op in Guix
<Apteryx>rekado: Even when grafts were involved?
<iyzsong>you can try 'guix package -i fontconfig --no-grafts', but I guess other packages use fontconfig need to be installed with '--no-grafts' too.
<ng0>the files are not found.
<ng0>but I know they exist in those locations
<ng0>maybe I'm inserting rename-files into for-each the wrong way
<ng0>before install the files aren't in $out
<bavier>makes sense
<Apteryx>iyzsong: When I try to install anything, I get: No space left on device. `df -h` returns: /dev/sda2 9.1G 6.3G 2.4G 74% /
<bavier>Apteryx: is your /gnu/store on a separate partition?
<Apteryx>bavier: Nope, everything is in /dev/sda2
<Apteryx>The other virtual filesystem are all at 0% use.
<rekado>Apteryx: is TMPDIR on the same partition? Are you building things from source?
<rekado>(I get this error when I try to build texlive with only 4GB left)
<ng0> ?: 0 [rename-file "./pulse/50-pulseaudio.conf" ...] why...
<Apteryx>rekado: TMPDIR isn't set (according to echo $TMPDIR). Here's the output of the guix package -i command:
<ng0>it seems I need to specify a full path
<ng0>otherwise it needs up finding /pulse/ for whatever reason
<rekado>ng0: I don't think that's true, but you could of course use (getcwd)
<Apteryx>rekado: If I was to plug an external drive and set TMPDIR to a folder there, I could get out of this situation?
<rekado>Apteryx: in this case I don't think so
<rekado>it complains about not being able to write to /gnu/store
<rekado>TMPDIR is used only to use a different directory from /tmp
<rekado>ACTION wonders if it could a variant of the problem we have on hydra re the .links directory
<Apteryx>Should I try to rollback?
<ng0>(find-files "." ".*pulse\\\\.(la|so)"))) would try to find files with \\ at the end, right?
<rekado>Apteryx: roll back what?
<rekado>Apteryx: rolling back your profile only switches the "current profile" link to a previous version in the store.
<Apteryx>My system to a snapshot where this problem was not occuring.
<rekado>oh, you're in a VM?
<Apteryx>rekado: Sorry, bad wording. Yeah, I might a guix rollback.
<rekado>ng0: you can try this in a Guile REPL
<ng0>which module prodives find-files?
<ng0>i mean, is it guix specific?
<ng0>else i know where to search
<ng0>ah, guix build utils
<ng0>unfortunately the results are correct
<ng0>I just needto fix the search then
<ng0>do we need the .la files?
<ng0>i believe my for-each is wrong
<ng0>but i have no idea how to fix it
<bavier>ng0: we usually install la files
<ng0> (rename-file file (string-append pualib "/" file))) I think this takes "file" from the (find-files), so "file" is the file to be renamed and the the second string (string-append ...) is where it will be moved to
<ng0>this was functional before I moved it to for-each
<ng0>I think rename-file complains that there is no target location
<rekado>ng0: have you tried (basename file)? If FILE is a full path this would make sense
<ng0>what do you mean?
<rekado>say FILE is /gnu/store/something/bla/foo then what is (string-append pualib "/" file)?
<ng0>if I specify basename, file gets separated and is only the filename, nothng more?
<rekado>I don't understand the question
<rekado>ng0: have you tried "basename" in the REPL?
<ng0>I'm no longer in the repl, I've applied this to the package
<rekado>I really recommend asking the REPL, because it gives you better answers than IRC
<ng0>and better help than the error messages of guile
<ng0>even if I used the repl, I have no idea what you wanted to express. use (basename file)? and outside of repl, what then?
<bavier>ng0: (string-append puaconf "/" (basename file))
<rekado>I hoped you'd understand the problem by answering the question I posed above ;)
<ng0>the error "messages" of guile are a bit frustrating at this point. I should just learn more guile before continuing. I thought using guix as a practical application to learn guile step by step after reading some books would work out, but deciphering more than the readable parts of the error messages would really help
<davexunit>the level of confusion you have about guile is perplexing given how long you've been writing guile code
<ng0>my mind moves in strange ways, I might assume some things to be more complicated than they really are, approaching them in a maze like way
<rekado>I find it helps to ask yourself clear, simple questions and move on from there.
<rekado>like: what does my code actually tell Guile to do.
<rekado>then test that in the REPL
<ng0>for example, I know what this could mean but not what exactly it does mean: ERROR: In procedure rename-file: Invalid cross-device link
<davexunit>if you did a search you'd see that "invalid cross-device link" is an error corresponding to a system call
<davexunit>not guile specific at all
<rekado>you cannot create a link across devices.
<ng0>odd enough I did not do this knowingly here
<civodul>ng0: /tmp and /gnu/store are different file systems ("different devices") in the build env, for example
<civodul>so you can't link(2) from one to the other
<ng0>i see
<ng0>so this renaming has to happen after 'install
<ng0>before 'install I would have to do it in /tmp and /gnu/store
<ng0>moving files from /tmp to store
<ng0>I think I need to do this more focused. I'm mentally half on the train already which I need to catch in 20 minutes
<ng0>okay, thanks for your help :)
<ng0>I made it succeed. I just had to find in out, not "pulse" or "."
<ng0>one of 2 find-files is now fixed, I'll play a bit with this in repl when I'm back later
<ng0>bye o/
<jje>after doing "guix system reconfigure" grub is telling me the is no root partition (dev/sda2). and yet it boots after a few seconds delay. can anyone help diagnose this?
<rekado>running "./pre-inst-env guix environment --pure guix --ad-hoc autoconf automake" with latest master on x86_64 I get no substitute for Guile 2.0.11 and when building locally the tests fail.
<rekado>1 failure
<rekado>FAIL: fluids.test: fluids are GC'd
<lfam>rekado: It works for from 1a86b671cef2 (gnu: Add ghc-gnuplot.), but I already had the built Guile, apparently
<lfam>*works for me*
<rekado>hmm, I'm switching to for substitutes now
<rekado>hmm, building gettext theer
<rekado>how odd
<rekado>(that's my precious central installation at work)
<rekado>shouldn't have started this right before planning to go home
<lfam>I read this with great interest:
<lfam>Looking forward to having that with Guix
<efraim>Isn't it cddl?
<lfam>Hm, a test time-out for python-pillow on armhf:
<lfam>efraim: Oh, is it? I didn't look at the license yet
<lfam>The `bcc` program that I assumed is what needs to be built (but I didn't look closely yet) is Apache 2.0
<lfam>I thought we'd need a 4.9 kernel, maybe with a changed configuration, and the bcc program
<civodul>lfam: that looks really cool, indeed
<civodul>though the whole idea of BPF is a bit scary at the same time
<lfam>I admit I don't understand what will happen "under the hood" to make these tracing capabilities work
<civodul>BPF is a bytecode interpreter running in the kernel
<civodul>so it allows user land to inject code into a running kernel...
<civodul>well now we need a BPF compiler back-end for Guile
<civodul>i'm sure wingo already has thoughts about this :-)
<lfam>Yes, sounds risky... but of course complexity and utility tend to increase together
<civodul>right, but the monolithic architecture necessarily increases the risks here
<lfam>Well, although I haven't used the Hurd yet, I hope that working on Guix helps the Hurd project tangentially :)
<phant0mas>it does :-)
<lfam> <> Looks like our linux-libre-4.8-x86_64.conf is only missing the optional CONFIG_HAVE_BPF_JIT=y
<civodul>yes well, i wonder whether grsec has that enabled :-)
<phant0mas>hey civodul can you have a quick look on this long forgotten patch
<phant0mas>so we can start those hydra jobs :-)
<phant0mas>I have applied it locally but I forgot it's not on core-updates
<civodul>phant0mas: i just replied: LGTM!
<phant0mas>civodul: thanks, I will reply to your latest email asap
<phant0mas>do we have uucp group in guixsd?
<lfam>phant0mas: Not on my system
<phant0mas>or any group that controls access to /dev/tty devices?
<phant0mas>I want to be able to access ttyACMx and ttyUSBx
<phant0mas>without sudo
<lfam>The tty group?
<phant0mas>somehow I missed that :-/
<phant0mas>thank lfam :-)
<phant0mas>thank you
<davexunit>phant0mas: is this for AVR stuff?
<phant0mas>davexunit: yes, I am moving with my home project :-)
<davexunit>I'd also like to use my regular use to flash AVRs :)
<davexunit>phant0mas: oh yeah, you are automating some home stuff?
<davexunit>I'd be interested to hear progress there.
<davexunit>maybe I can borrow an idea or two.
<phant0mas>actually I am renovating the whole house and the plan is to automate everything
<phant0mas>davexunit: I will report any progress on twitter :-)
<davexunit>phant0mas: cool :)
<davexunit>I'm particularly interested in how the various devices will communicate
<phant0mas>the important parts through ethernet cable, the non important with low power, low distance wireless modules
<phant0mas>the various devices will only implement mechanisms, the raspberry(aka brain) the policy
<phant0mas>smaller vector of attack this way
<davexunit>might I recommend ditching the raspberry pi for something like the beaglebone black
<davexunit>rapberry pi requires lots of nonfree firmware in order to run
<phant0mas>the problem is I already have a couple of raspberrys laying around
<phant0mas>davexunit: let's move this conversation to twitter :-)
<detrout>are there good guix related twitter accounts to watch?
<lfam>Raspberries this late in the season?! Wow, lucky ;)
<davexunit>detrout: no not really
<amz3>can you point me to a tutorial on how people sign guix's commits please?
<amz3>detrout: guilelang
<phant0mas>amz3: if you already have a pgp key you could just use git commt -S
<lfam>amz3: See the option '--gpg-sign' to `git commit`. It's respected by other Git commands as well
<bavier>lfam: I just picked a pint or so from the bushes in my backyard ;)
<lfam>There's also the Git configuration option 'signingkey' where you can specify which key to use
<lfam>bavier: I found a few stragglers last week here on the East coast. A rare treat in the cool fall weather!
<davexunit>ACTION either has a raspberry or blackberry bush in his yard
<davexunit>it didn't produce any fruit this year, and I've only owned the property for a few months, so I don't know which it is!
<lfam>Heh, that will be a good surprise either way!
<bavier>MN is breaking records for longest growing season right now
<lfam>Yeah... warm year
<lfam>We should bring raspberry pies to the next Guix meeting ;)
<bavier>lol :)
<amz3>which guix meeting?
<lfam>Well, you've inspired me. Off to the cafe for some raspberry jam
<lfam>amz3: I don't know... FOSDEM?
<amz3>I have a lot of common fig this year
<amz3>but no apple and no pear :(
<lfam>Oh, figs! One of my favorites
<amz3>It's the first year the tree gives something, the problem is that i prefer black fig
<bavier>love figs
<lfam>amz3: I guess you'll just have to plant another tree :)
<amz3>something I miss a lot from my home country, is
<amz3>it's called barbary fig in france
<phant0mas>I will bring cretan raki to the next Guix meeting :-)
<phant0mas>amz3: I think we call it φραγκοσυκα here
<amz3>I am happy today because I made good progress on libgit2 bindings
<amz3>we are only missing diff/delta bindings to make a gitweb replacement
<amz3>that said it's very advanced work
<amz3>it's misssing all the procedure to edit a repository
<amz3>pull/merge/clone/create commit
<amz3>pull is not event a single procedure in libgit2
<amz3>pull is actuallly several procedure
<amz3>in the sens that you can not always pull --rebase something
<amz3>the thing is that in guix it is in the normal case, the case where the user doessn't point it's repository to a dev repository, it's always possible to do git pull
<amz3>yes, git pull = git fetch + git merge
<rekado>I’m still forced to build Guile 2.0.11 from source.
<rekado>Does anyone know why that is?
<rekado>(and the tests fail, so I cannot make it succeed)
<rekado>that’s latest master
<rekado>ah, I know why
<rekado>I don’t seem to have guile gnutls
<janneke>rekado: /me too
<rekado>huh, no, even with gnutls I don’t get a substitute.
<rekado>building guile again — and failing.
<lfam>rekado: What command can I use to reproduce the failure?
<rekado>lfam: I only do this: ./pre-inst-env guix environment --pure guix
<rekado>it’s just latest master
<rekado>last commit is fdc45ab1b0e3bca9c4991f90614a0177f98c30d8
<rekado>(your imagemagick fix)
<rekado>with “--no-grafts” Guile doesn’t seem to be built
<lfam>How can I request a rebuild of the Guile that fails? I already have it built in my store
<lfam>I'm not sure which Guile it is so I don't know how to refer to it with --check
<rekado>if you already have it “guix build guile” will give you the latest guile, no?
<lfam>Is it failing to build /gnu/store/4j2xhmx5ja1f6dh5y86qcm2rln40jjl4-guile-2.0.11.drv ?
<rekado>I’m trying to get into a guix environment without grafts, then “make clean”, “reconfigure” and “make”.
<lfam>That's what I get for `./pre-inst-env guix build guile -d`
<lfam>From the commit you specified
<rekado>I don’t know. I cannot see the derivation (no scrollback)
<lfam>Okay, I'll try `./pre-inst-env guix build guile --no-grafts --check`
<rekado>downloads on that machine are very slow :( So weird. I cannot blame everything on NFS.
<rekado>with “--no-grafts” I’m not asked to build Guile
<rekado>now it’s just downloading geiser, emacs-minimal, etc.
<civodul>it's /gnu/store/4j2xhmx5ja1f6dh5y86qcm2rln40jjl4-guile-2.0.11.drv that's not built on, for some reason
<civodul>you can see that when running "guix build guile@2.0"
<civodul>same with --no-grafts
<civodul>i think that happens because still has some of the outputs of guile@2.0, but not all of them
<civodul>e.g., it lacks the "debug" output
<rekado>it’s fine with “--no-grafts” because presumably I already have that guile.
<rekado>wha, I’m going crazy
<rekado>I have a file called “string” in my “guix” repo clone
<rekado>when running “make” somehow this file is consulted
<lfam>What's in the file?
<rekado>when it’s present the compilation of “nix/boost/format/libformat_a-free_funcs.o” will fail
<rekado>lfam: it’s grep output
<rekado>In file included from ./nix/boost/format.hpp:23:0,
<rekado> from nix/boost/format/
<rekado>./string:1:66: error: too many decimal points in number
<rekado> Binary file /gnu/store/7j4bx7rri4csxpz2ar2nphy7phgk8pkh-rcas-web-0.0.0/share/guile/site/2.2/rcas/utils/report.go matches
<rekado>that’s nuts
<rekado>moved it away and now the daemon compiles
<rekado>what a day
<lfam>Huh? Is that a well-known file name for `make`? What's going on there?
<rekado>ah! #include <string>
<lfam>Strange! I made the same file with some garbage and got a different failure
<rekado>in nix/boost/format.hpp
<rekado>does the compiler search the cwd first?
<lfam>I thought you had to use #:include "foo" to get this behavior
<lfam>But I'm not familiar with C++
<rekado>that’s what I thought, too
<lfam>Is this the relevant documentation? <>
<rekado>bah, looks like I cannot use Ao (ao-cad) on my X200s. Needs OpenGL 3.x
<lfam>How about KiCad? That's basically ready AFAICT
<lfam>Ah, I see that ao-cad is Scheme-y
<lfam>I had to cancel that Guile build for now. I'll restart it later
<rekado>KiCad seems to be EDA software. I just need a tool to create a drawing.
<rekado>(for EDA software I use “pcb” and “geda-gaf”, also using Scheme)