IRC channel logs


back to list of logs

<emacsomancer>kmicu: icecat strace log:
<rvgn>Hello Guix!
<sneek>Welcome back rvgn, you have 1 message.
<sneek>rvgn, kmicu says: to highlight kmicu
<rvgn>kmicu I have a notification that there is a message from you. How can I view it?
<rvgn>kmicu I have a notification that there is a message from you. How can I view it?
<au>Hey all. I'm a Guix newb and use it with Trisquel, and when I use "guix pull", I always get locale warnings. Specifically, "guile: warning: failed to install locale" and "bash: warning: setlocale: LC_ALL: cannot change locale (en_US.utf8)". Out of curiosity, will this pose a problem at some point?
<bandali>rvgn, that was the message. they asked you to highlight them. they went to sleep a while ago though, might wanna try tomorrow
<rvgn>bandali Ah I see. Thanks!
<bandali>np :)
<nckx>au: It won't cause a problem with Guix itself, although it's shouldn't happen if you've installed the latest glibc locales as described in the manual.
<rvgn>Hi Folks! Are guix mailing-lists spam resistant?
<au>nckx: Ah, I see. I'll go do that, then. Thanks! :)
<bandali>hey rvgn, yes they mostly are; thanks to both technological measures as well as volunteers that moderate gnu lists
<bandali>though, it’s still possible to retrieve the email address you use to post to the lists, of course
<rvgn>bandali I see. As I am a subscriber to "help-guix" mail-list, I received three emails with same subject "FreeTube". That too, two of them did not had any content. Not sure if it was a spam.
<jje>trying to do "guix system reconfigure /etc/config.scm" and it keeps teeling me "error /gnu/store/...grub-cfg.drv not such file or directory" what should my next step be?
<bandali>rvgn, right. those were apparently deemed non-spam and were let through. the first two empty ones seem like human error
<bandali>(i’m not a moderator, just guessing here)
<nckx>Yeah. It seems like someone unfamiliar with mailing lists or maybe even mail in general.
<nckx>jje: That's… odd. I'd start with a ‘sudo guix gc --verify=contents,repair’ to rule out store damage.
<jje>ok thank you nckx
<Marlin[m]>hi guix
<Marlin[m]>I'm designing a keyboard layout
<Marlin[m]>any tips?
<Marlin[m]>I based it off programmer dvorak
<nckx>Marlin[m]: I noticed. Since you're basically creating a ‘dvp-mod’ with your personal tweaks, I wonder what kind of tips you're looking for.
<nckx>None of your changes mean much to me, but that's fine; it's your keyboard ☺
<Marlin[m]>i might add in more modifiers
<nckx>I see a few emacsisms, so that's your motive
<Marlin[m]>The x and c location
<Marlin[m]>53 02 location
<nckx>Why the X-K swap?
<nckx>Oh, I hadn't even noticed the 5-3 swap.
<Marlin[m]>it's a better position for hitting with the thumb
<Marlin[m]>and i use it as a modifier
<Marlin[m]>not the thumb
<Marlin[m]>the index fingars
<nckx>…X with the thumb?
<Marlin[m]>it's more comfortable
<nckx>You an Esc-X person?
<Marlin[m]>The original position of X on programming dvorak is not good
<Marlin[m]>No, i meant with the index finger
<Marlin[m]>Thumb came to my mind for some reason :P
*nckx sometimes uses pinky-Esc - thumb-X so it could make sense. Not often though.
<Marlin[m]>what i'd like to do is to hit ctrl with the bottom of the hand and press X with index
<Marlin[m]>and that position is pretty good for that
<rvgn>bandali Ah I see. Thanks for letting me know.
<nckx>Hm. Not directly related, but seeing your Esc-as-Caps make me wonder if there's a way to ‘rotate’ those three so Caps→Ctrl, Esc→Caps, and Ctrl→Esc, without writing my own keymap.
*nckx uses ctrl:swapcaps now.
<bandali>you’re welcome rvgn :)
<Marlin[m]>This is the natural position for the bottom row on my keyvoard
<Marlin[m]><nckx "Hm. Not directly related, but s"> i don't like ctrl as caps
<Marlin[m]>I don't use my pinky for ctr,
<Marlin[m]>I use the bottom of my hand
<Marlin[m]>using ctrl for caps with the pinky would be worse
<nckx>That's a great way to do it… on a real keyboard. I use laptops exclusively.
<nckx>Or you're able to do that on laptops, in which case mad propz to you, but I'm not.
<Marlin[m]>on my laptop i do swapcaps :P
<Marlin[m]><nckx "Or you're able to do that on lap"> i am actually
<Marlin[m]>But it's not as comfortable
<Marlin[m]>laptop keys are soft and easy to use the pinky on, but are too flat to hit with the bottom of my hand
<nckx>Plus the rightmost bottom key on my TP is Fn (rebound to compose, as you → might have noticed ♫ ← )
<Marlin[m]>So swapcaps is a good ine in that case
<nckx>Marlin[m]: You don't have to use those <quotes> as far as I'm concerned. I'm just an IRC user.
*nckx pulls up base.lst
<nckx>Combining ctrl:swapcaps followed by caps:swapescape (i.e. ‘trying to be clever’) just gives me two caps lock keys ☹
<rvgn>bandali I just searched for "freetube". It seems like an elite app.
<nckx>1337 = gud?
<rvgn>nckx Just asking, have you tried that app before?
<bandali>rvgn, elite app? not sure what you mean. also, i haven’t personally tried it
<nckx>I get... 3 porn sites as first search results but that's just Searx being Searx.
<nckx>rvgn: No, just wondered what ‘elite’ means here, like bandali.
<rvgn>nckx It's
<rvgn>Took me while too ;)
<rvgn>ncks what is bandali? I meant elite referring to UX
*bandali is a who, not a what
<rvgn>There are some screen shots on the site.
<nckx>sneek: what is bandali?
<nckx>Even sneek does not know.
<rvgn>Oh OOPS!
<bandali>oh boy
<nckx>sneek: bandali is a who, not a what.
<sneek>So noted.
<bandali>for youtube these days i use either or mpb
<rvgn>bandali Is it possible to browse videos from mpv??
<nckx>rvgn: Not browse.
<rvgn>I see.
<bandali>rvgn, no but if you have youtube-dl installed you can stream/open youtube urls
<nckx>bandali: This FreeTube thing ‘uses the invidious API’, whatever that means, so I'd just landed on their site.
<bandali> lets you browse too tho
<nckx>It's a... YouTube mirror?
<bandali>kinda. it’s a front-end to youtube, in their own words
<bandali>rvgn, don’t think so. but if you have y
<nckx>So does youtube-viewer, which I use a lot, since I've never cared about a thumbnail (but others do and that's fine).
<nckx>bandali: But it still streams the video data straight from Google?
<nckx>OK, that wasn't clear to me yet.
<bandali>nckx, it seems so
<rvgn>bandali I would like to see the thumbnail of videos. Currently I use "youtube-viewer" but has some bugs.
<bandali>rvgn, try out invidious
<bandali>nckx, it loads videos from * so i’d say yes
<rvgn>bandali I would prefer Non-Web Client. For web, I use
<nckx>rvgn: As a youtube-viewer user, I have to ask: which bugs? Are you still having trouble launching the right video player?
<bandali>rvgn, ah ok. i prefer web for youtube, and invidious was the best i’d found
<bandali>hooktube seems good too
<nckx>Wow. I haven't visited in years. It seems our IceCat has an extension to... something, there.
<rvgn>bandali I see. Yeah, the loading video inside web browser consumes lot of CPU and RAM and freezes the system if I have too many tabs open doing other things.
<nckx>Replace the player with moar knobs. Cool. I like knobs.
<nckx>rvgn: Amen.
<rvgn>nckx xD
<bandali>i’ve found invidious to be super lean, but even still, i fall back on mpv for that exact reason every now and again
<rvgn>bandali I get you ;)
<nckx>Even though this is the definition of circlejerk, I have to add my +1.
<rvgn>nckx you mean on
<nckx>rvgn: I just meant here. What's on lp.o? Some vote?
<rvgn>nckx Yeah +1ing the apps for packaging.
<nckx>Oh, I didn't know that existed?
<rvgn>nckx I missed you msg regarding "youtube-viewer". I filed the bug at
<nckx>Which one might say is indicative of a problem, me being a packager and all.
<Marlin[m]><nckx "It's a... YouTube mirror?"> kinda
<Marlin[m]>it's good because it doesn't use anything from google
<Marlin[m]>it just gets the vids mirrored
*Marlin[m] likes smtube from the smplayer devs
<nckx>Marlin[m]: Sure, but Google still gets a nice list of which videos your IP watched. Which is fine (the same is true for mpv et al), but Freetube's claim of ‘without your habits being tracked’ is IMHO skirting the line of accuracy then.
<nckx>Marlin[m]: Oh, wait, you're saying the video streams *are* mirrored?
<Marlin[m]>yeah, i think so
<nckx>rvgn: Oh, OK, I never don't have youtube-dl installed so have never run into that problem.
<nckx>OK, so 1 person saying they mirror vidyas, one saying they don't, I smell actual work 😛
<nckx>(‘Searching & clicking on things’ is work, yes.)
<rvgn>nckx There is also another person posted in "help-guix".
<nckx>Sheesh peeps yes sometimes not all dependencies are added. It happens. Chill.
<nckx>rvgn: Not at you; something about the tone in that last mail bugs me.
<rvgn>nckx Ah, you mean the last mail on that thread?
<nckx>rvgn: Your last link above.
<rvgn>nckx gotcha!
<nckx>If youtube-viewer doesn't work at all without youtube-dl, it's obviously just a forgotten dependency. Bah. But maybe it's just me & the way I read that mail ☺
<rvgn>nckx Btw, This is the +1ing the app thingy I was talking about:
<nckx>rvgn: Thanks! I'd found that too (after Searx suggested this:, which is... adorable - ‘a C compiler’) but, ehm, call me cynical, but I hope nobody is expecting much if anything from that list.
<nckx>If there is a wandering band of vigilante packagers waiting for the magic ‘+5’ to package something they didn't have the motivation to package before, well, I'm not in it and I don't know anyone who is.
<rvgn>nckx May be not much people know that list exists? xD
<nckx>rvgn: As noted above, it was news to me!
<rvgn>nckx Yeah yeah just saying ;)
<Marlin[m]>i have it bookmarked
<Marlin[m]>i'm gonna start packaging some of it
<Marlin[m]>i do still have to finish my lutris package
<nckx>I guess its audience is people new(er) to packaging who need inspiration - ah, voila, Marlin[m] for example.
<nckx>I just hope new users don't overestimate how likely their favourite package is likely to be packaged just by posting it in a list (or sending it to the ML, which happens). But oh well.
<nckx>Marlin[m]: Oh! You're the lutris person! How's it going?
<rvgn>Marlin[m] Can I suggest you this:
<nckx>rvgn: ‘A small caveat...’
<nckx>That sounds like the author knew what they were talking about, but you say it's certainly not the case (on Guix at least)?
<nckx>I'm very willing to add it as an input but only if it's actually needed.
<rvgn>nckx I see. The videos I tried did not open even without youtube-dl. May be no lower quality formats?
<nckx>Hmm, I just ‘guix remove’d youtube-dl and can watch My crush on GNU Guix with youtube-viewer just fine.
<nckx>rvgn: Yeah, maybe if the videos are old/480p/whatever-only?
<rvgn>nckx I am sure about the issue :D
<nckx>Which Guix videos are unlikely to be.
<nckx>rvgn: I'm not doubting your word, just not able to reproduce it. Do you have a search term/video titile I could plug into youtube-viewer?
<rvgn>Music Videos??
<nckx>My randomly input music videos are likely to be very different from yours. ☺
<nckx>But bingo, Dmitri Shostakovich - Waltz No. 2 (by TheWickedNorth) fails to play.
<rvgn>nckx Search "Best Pop Songs" and look for the video "‎New Pop Songs Playlist 2019 - Billboard Hot 100 Chart - Top Songs 2019 (Vevo Hot This Week)"
<nckx>rvgn: Thanks.
<rvgn>nckx The channel will be "Playlist US-UK Music" and thumbnail will be "Ariana Grande".
<rvgn>nckx :)
<nckx>rvgn: I'm using the terminal version, but I got one.
<nckx>It's a pretty bad failure mode. It just fails without any message.
<nckx>I was tempted to keep youtube-dl optional if youtube-viewer printed a nice ‘to support older videos, please install...’ message, but this sucks. Patching it is!
<rvgn>nckx I see.
<nckx>It seems like that slackbuilds link I posted was inaccurate or outdated after all (and it wasn't that old). Sigh.
<rvgn>nckx Thanks for testing it. I have updated the bug report by replying to that thread.
<nckx>rvgn: Oh, I'm already working on it.
<nckx>But thanks ☺
<rvgn>nckx Oh wow! That is great, thanks a lot. :)
*nckx .oO Don't delete 1000 mails in mu4e if you need to do any text editing in emacs. You will not be doing any text editing in emacs.
*nckx listens to some nice music videos instead.
<rvgn>nckx Could you be able to ping me when you are done with the patch? Just would like to upgrade and try :)
<emacsomancer>nckx: I love mu4e but it's surely the emacs 'app' that eats the most cycle
<nckx>emacsomancer: More than 3 hours of my 1day 14h uptime (and if you think ‘huh, that's actually less than expected, on average’ you are a true mu4e user™).
<Marlin[m]><nckx "Marlin: Oh! You're the lutris p"> i gotta get webkit and gnome bindings for python-gobject... or something like that
<nckx>rvgn: Patch pushed & mail mailed.
<nckx>Please close the bug if this solves your issue.
<nckx>Marlin[m]: Oh, right, that's why I'd repressed those memories. :-/
<rvgn>nckx Thanks I will guix pull and upgrade the app now.
<nckx>Anybody know how to view the ‘raw’ mail message in gnus? C-u g doesn't do it.
<bandali>nckx, try v v?
<nckx>‘verdict: no signature found’
<nckx>bandali: My bad, I should've mentioned this is gnus'(s?) ‘article view’ used to view mu4e messages.
<nckx>Although it is gnus.
<nckx>It is also mu4e.
<bandali>hmm, i’m not sure? in gnus-summary-mode, v v invokes gnus-summary-show-raw-article
<bandali>don’t know if that works in mu4e or not
<nckx>C-u g returns ‘This is a pseudo-article’. Problem is I don't actually grok gnus (I just prefer its view to mu4e's) so I don't know what that's telling me.
<nckx>M-x gnus-summary-show-raw-article does the same. So at least mu4e isn't messing with the bindings.
<bandali>sounds like an issue on the mu4e side then?
<bandali>i’d never tried C-u g, but i just tried it and it seems to work in gnus
<nckx>I thought ‘pseudo-article’ might be a well-known ‘oh, so I need to do X instead’ thing in gnus circles; I guess not.
<nckx>bandali: With mail? OK, then it's indeed a shortcoming on the mu4e side. Thanks!
<bandali>yeah i tried on a regular mail message with gnus’s nnimap backend
<bandali>nckx, m
<nckx>Shame, apart from that annoyance, gnus' message view is perfect for me.
<bandali>maybe try looking through mu4e-view.el’s source and see if you find anything relevant?
<nckx>More intuitive than mu4e's at least.
<bandali>but yeah article-mode is *awesome*
<nckx>bandali: Sure, but that's not something that'll fit into tonight's schedule ;-)
<bandali>as a side note, i went from mu4e to notmuch-emacs to gnus :)
<bandali>nckx, haha
*nckx almost found the best of 2 worlds...
<nckx>bandali: Heh. Maybe. Who knows where I'll end up. I really like mu4e's actual, er, ‘mail engine’/workflow though.
<nckx>Can't see anything beating that.
<nckx>(Famous last words, I know.)
<Marlin[m]>gnus is da best
<nckx>Halp I'm being ganged up on by gnus.
*nckx searches their entire mailbox at real-time speeds to scare them away.
<nckx>OK, for once I agree that singular they can be confusing.
<nckx>rvgn: U watchin vidyas happy?
<bandali>gnus has pretty good search, though not as great as notmuch
<bandali>but one can actually make gnus use notmuch as a search backend
<rvgn>nckx It works now. Thanks a lot :) How can I close the bug?
<rvgn>nckx vidyas??
<nckx>Oh. Sorry. I meant videyas. On the tubes.
<rvgn>nckx The video that didn't work for me before, works now. :)
<nckx>rvgn: I've closed it now. To close bugs yourself, just send your ‘Thanks! This fixes foo’ mail to <bugnumber>-done@... instead of <bugnumber>@. Simples! ☺
<nckx>rvgn: Great.
<nckx>At least I finished one thing tonight.
<nckx>bandali: Temptress.
<rvgn>nckx Thanks
<bandali>nckx, i believe every emacs user has to at least try gnus at some point ;)
*nckx glad someone else tactfully answered The Pinger and also that they're not the only ones who sleep too little.
<nckx>bandali: Why's that?
<bandali>nckx, i just feel like it’s so nicely integrated into emacs
<nckx>If configuring *any* new emacs client in emacs weren't a big tin of meh, I probably would have by now.
<bandali>i mean, in one sense, it literally is: it’s developed in the emacs repo haha
<bandali>and i find it does everything i need and then some
<bandali>but i definitely agree it requires time investment
<bandali>but like, for instance, its article-mode is the best things i’ve ever used
<bandali>and using different mail backends is pretty trivial (imap, maildir, mbox, …) though i mostly only use imap due to speed
<bandali>and for one thing, its article-mode is the best message viewer i’ve ever used
<nckx>bandali: Does the IMAP backend support IDLE? That's one thing of the several that would be required to make me switch.
*nckx .oO This bandali seems to be sending mixed messages about article-mode.
<bandali>nckx, hmm, i don’t think so? i pull my mail in batches
<bandali>using getmail and isync (mbsync)
<bandali>but refreshing is just g, so you can press that and fetch new mail if you connect directly to your main imap server
<bandali>nckx, ha, why mixed message? :p i thought it was all praises
<nckx>bandali: I use mbsync too. And I just ‘benchmarked’ a fetch: 20 seconds :-/
<nckx><bandali> but like, for instance, its article-mode is the best things i’ve ever used
<nckx><bandali> and for one thing, its article-mode is the best message viewer i’ve ever used
<nckx><bandali> but yeah article-mode is *awesome*
*bandali lols
<bandali>didn’t even realize i wrote that twice (or more :p)
<bandali>goes to show how much i like it ha
<bandali>nckx, 20 seconds for syncing all directories in one account?
*nckx *sighs* Yes, I know...
<nckx>du -hs .mail
<nckx>1.3G .mail
<nckx>Maybe that's not ‘normal’? 🤷 I dunno.
<nckx>Maybe it's mu's fault?
<nckx>Indexing all my mail, pfft.
<bandali>i’m a bit confused; is that 20 second fetch inside mu4e? from local maildir?
<bandali>or is that by mbsync syncing mail from remote imap?
<nckx>time mu index -m ~/.mail
<nckx>real 0m0.289s of which 92.16% CPU
<nckx>So, no.
<nckx>bandali: That's my ‘refresh command’, which is ‘mbsync --all’
<bandali>ah. it’s been a while since i last used mu4e
<bandali>i see
<bandali>my mbsync takes a while too, but it depends on volume of mail, number of mail folder, and latency (ping to server)
<bandali>but once mail is on my computer, reading it is blazing fast. i think i first tried setting up gnus with maildir, but it was kinda slow
<bandali>then i set up a local dovecot imap server and connect gnus to that. blazing fast
<nckx>So that's fine as a background command, but makes ‘just hitting g with your muscle-memory when looking at your mail’ not an option.
<bandali>right, makes sense
<bandali>fwiw, my setup is similar to this:
<rekado>I still use offlineimap and it’s incredibly slow.
<rekado>I wanted to switch to mbsync but I remember trying this some years ago and switching back due to compatibility problems that are related to the IMAP server configuration at
<bandali>rekado, feel free to ping me if you decide to try again. i’d be curious to see if it works out now, or if not if i may be able to help
<nckx>bandali: Thanks. Does gnus talk directly to Dovecot, or does it create a duplicate maildir?
<bandali>when i switched from offlineimap to isync i didn’t have much email, so i probably didn’t notice much slowness. but isync is pretty fast
<bandali>nckx, cheers. the former i believe
<nckx>I thought about adding lucene to my dovecot server but, well, I have mu4e and a full local copy anyway so what's the point... Using dovecot+lucene locally is a clever solution to not using mu.
<bandali>ha. i’d look to try and replace lucene with good ol’ notmuch if i find some free time. but lucene has been pretty good so far too
<bandali>brb gotta grab me dinner
<nckx>bandali: Smakelijk, and thanks for forcing me to think about alternative MUAs. o/
<bandali>nckx, thanks :) and anytime \o feel free to ping me or drop by #gnus if you decide to try it out or run into something with it
<rekado>IIRC the problem had something to do with IMAP message IDs and it led to lost messages :-/
<rekado>that’s why I’m really wary of trying again
<rekado>re mu4e: I briefly considered going back to notmuch but then I remembered that it doesn’t let me delete emails.
<rekado>I get a lot of mail that is not spam but that I know I don’t want to keep or ever search again.
<rekado>(work email about service disruptions, talks, journal clubs, announcements of free left-over conference food, etc)
<bandali>i see
<bandali>rekado, re deleting mail: have you seen ?
<bandali>fwiw, gnus does a wonderful job of deleting or moving around email automatically
<bandali>so you’ve got to give it a shot ;)
<rekado>I tried gnus long ago and I found it confusing to configure.
<bandali>it definitely has a lot of bells and whistles :)
<bandali>but for me personally so far i’ve found it’s very much been worth the time i invest into setting it up
<bandali>if anyone’s curious:
<bandali>to try and get back from this OT discussion, i’ll just one more thing about gnus, and that’s that i’ll be happy to help around on #gnus if anyone needs it
<bandali>*i’ll just say
<rekado>bandali: thanks for the offer. One day I might take you up on it.
<bandali>cheers rekado
<rekado>scipy is broken since the staging merge
<rekado>the reported problem is that the theme for the HTML documentation isn’t found.
<tune>having a hellish time with my guix system config right now... apparently I made a change that wasn't valid in the past and the hints aren't enough for me to fix it
<tune>/etc/config.scm:72:20: error: remove: unbound variable
<tune>hint: Did you forget `(use-modules (srfi srfi-1))'?
<tune>I'm thinking perhaps the hint is just *wrong* and misleading me. sometimes it also says rfr or something and "lists". I've tried adding to use-modules and use-package-modules and they're just not recognized. problem in my services block I think
<tune>last change was adding some lines someone gave me to disable gdm from %desktop-services%
<vagrantc>tune: if you posted your config to a or something, it'd be easier for people to help
<rekado>tune: have you added (use-modules (srfi srfi-1))? That’s where “remove” is defined.
<tune>rekado: when I added those to my existing use-modules section it didn't seem to work
<tune>one moment I'll get my config pasted and such
<tune> here's my config
<tune> here are a few of the errors I get
***MinceR_ is now known as MinceR
<efraim>TIL my macbookpro 3,1 has free wifi drivers. Makes me wonder why I bothered with the macbook 4,2
<civodul>Hello Guix!
<thomassgn>Hello! anyone familiar with the grub rescue interface?
<civodul>hello thomassgn!
<civodul>you mean the command line?
<efraim>i try not to have to play with it, what are you trying to do?
<thomassgn>Ah, I've just rebooted after 'guix system init ...' finished successfully and it seems the GRUB "prefix" is wrong, so it can't enter "normal mode" Been looking at the GRUB docs, and think I'll have to try rerunning guix system init from a live disk or something. I have boot on a USB and root fs on a raid10. :) I guess I've messed up somewhere :)
<thomassgn>I've commented out a whole lot in the config - reason for the rootfs move from usb to raid was lack of space for store/builds. I'm not sure what would go wrong to make GRUB not run right, but here it is
<thomassgn>I'll fiddle with this throughout the day, if anyone takes a look I'm real happy for feedback :)
<thomassgn>Oh, the grub error is "error: attempt to read or write outside of disk 'hd3'." (I'll stop filling the channel now :)
***rain1_ is now known as rain1
<rekado>tune: I see in your paste that you haven’t added the recommended use-modules line.
<civodul>janneke: gcc-mesboot-wrapper and gcc-mesboot have a different 'version' field, is that expected? 4.7.4 vs. xb
<civodul>4.7.4 vs. 4.9.4
<civodul>it's probably harmless tho
<kmicu>That applies to Guix too so sharing it here
<kmicu>sneek: later tell rvgn to highlight kmicu for an explanation for guix pull messages about manually exporting envars.
<sneek>Will do.
<rekado>scipy is broken. Trying to fix it but the error isn’t very helpful.
<rekado>it fails building the docs.
<civodul>rekado: looks like you still have a "guix gc -D" loop running on berlin, no? :-)
<civodul>or maybe two?
<civodul>perhaps we should stop them because that takes the "big GC lock", and thus everything waits periodically for it to be released
<rekado>yes, I’m still running the loops
<rekado>should I make them sleep after each round?
<rekado>stopped them; we’re now at 12TB free
*rekado tries to upgrade scipy to 1.2.2
<civodul>rekado: yes, if you could insert a 'sleep', that'd be better
<efraim>Maybe even add an mcron job for it ;)
<civodul>yes, though i'm not sure we should do it at all
<civodul>after all, if we don't do that, the GC will eventually run (there's an mcron job), and it'll find that it can delete all these things and perhaps a few other things as well
<civodul>the problem is "the few other things"
<civodul>but i'm not entirely sure this is a problem
<g_bor[m]>Hello guix!
<g_bor[m]>What do you think about packaging fork-awesome?
<g_bor[m]>Istm that we are currently stuck on Font Awesome version 4, because v5 is non-free, and it looks like it won't be any soon.
<sneek>I'll keep that in mind.
<rekado>sneek: what is Istm that we are currently stuck on Font Awesome version 4, because v5
<rekado>sneek: what is Istm that we
<sneek>Last time I checked Istm that we is currently stuck on Font Awesome version 4, because v5 is non-free, and it looks like it won't be any soon.
<rekado>sneek: forget Istm that we
<rekado>sneek: silly bot, have a botsnack
<g_bor[m]>sneek: botsnack
<kmicu>Just to clarify for folks before mornin’ coffee/mate: g_bor[m] wants to package Fork Awesome not Font Awesome.
<g_bor[m]>it looks like a maintained project, and debian choose to patch some packages to use it, since they also cannot upgrade.
<efraim>i wonder if we could automate some resize2fs magic for VMs
<efraim>armbian's is just some shell magic, same fdisk -> partprobe -> resize2fs as vm-image.tmpl suggests
<civodul>efraim: the VM image we ship looks like it has a large hard disk, even though the image itself is small
<civodul>with qcow2, we can make the hard disk arbitrarily large at no cost, AIUI
<demotri>g_bor[m]: Fork Awesome, looks promising, we should investigate :-)
<rekado>now scikit-learn is broken because of the scipy upgrade…
<civodul>the rabbit hole
<rekado>I’ll upgrade it to the last version that supports Python 2.
<rekado>version 20.x is the last with Python 2 suppotr.
<civodul>perhaps we can drop python2-scikitlearn soon, then?
<rekado>we may need to purge some packages from bioinformatics.scm.
<civodul>rekado: in mumi, is there a way to select specifically "guix-patches" or "guix"?
<rekado>civodul: it’s currently hard-coded to be either.
<civodul>i was looking at replacing references to on the web site with references to
<rekado>you can use for *any* GNU debbugs bug number.
<demotri>civodul: Thanks for fixing that docker-bug so quickly :-)
*rekado –> afk
<civodul>rekado: yes
<civodul>but i wondered if we could have a list equivalent to that of
<civodul>a "tag:patch" filter would do
<efraim>civodul: i'll take a look at it, I was making a custom image to upload to THE CLOUD
*efraim does wavy hands
<civodul>the cloooooouud!
<civodul>you can pass --image-size=100G then
<htgoebel>rekado: Re. PYTHONPATH: For which of the many use-cases do you want to switch to GUIX_PYTHONPATH?
<htgoebel>Shall it point to the place where python is actually installed (say: to the profile) and thus contain a single path?
<civodul>new post!
<efraim>It wouldn't be a problem to make an image that appears larger than the space allowed to the VM?
<civodul>a problem for whom?
<civodul>the result is just that "df" will report a lot of free disk space, and that you can actually use that space
<efraim>For me, if the vm thinks it can have 100 gb and I can use say 50gb
<efraim>I know I can make a smaller qcow2 image, but just one would be nice
<civodul>the only limit then becomes the available size on the underlying storage, on the host
<Marlin[m]>mornin guix
<civodul>the worst that can happen is that, in the guest, you'll get ENOSPC even though the disk is not full per "df"
<civodul>hello Marlin[m]!
<cbaines>efraim, I recently packaged cloud-utils, mostly as it includes a `growpart` script
<Marlin[m]>hey guix
<Marlin[m]>I might put porting or at least some games from itch to guix
<bavier>hi Marlin[m]
<Marlin[m]>Maybe i should port some free games, as itch also has proprietary ones
<Marlin[m]>I'll start off with mindustry
<Marlin[m]>Personal favorite of mine
<Marlin[m]>i think i'll put some reminders on the description about the game being for sale on
<Marlin[m]>It's a pay what you want one
<str1ngs>lzip should mean faster downloads for substitutes that are not cached on the server I assumed?
<brendyyn>lzip is what is being used for the cache
<str1ngs>gotcha, thanks
<str1ngs>the IPFS information is even more interesting though :)
<str1ngs>and exciting.
<xavierm02_>How do I get the name of a service?
<xavierm02_>In guile
<str1ngs>guix system search will search services
<str1ngs>oh, you meant in guile sorry
<xavierm02_>Alright I found: It's service-type-name
<xavierm02_>Is there some way to define ∘ such that (define display-service-list (∘ display (map (∘ service-type-name service-kind)))) is accepted?
<xavierm02_>Right now I have to eta-expand: (define display-service-list (∘ display (lambda (x) (map (∘ service-type-name service-kind) x))))
<sebboh>I have been struggling to guix to boot on my desktop at home (currently I am not there). Eventually I decided (wrongly) that my motherboard couldn't boot from my SATA SSD.. So I fell back to another OS for troubleshooting, and then I accidentally successfully installed and booted Debian on that same drive on that same motherboard.
<sebboh>so.. could somebody paste bin their `sudo find /boot` from a working UEFI guix system? Maybe the output of `sudo fdisk -l /dev/sda` or whatever? I guess I don't really know the right questions to ask (which is probably why the deb installer could make it work when the guix installer (at least partially "me") could not.)
<nly>Hi sebboh
<sebboh>hi nly
<nckx>sebboh: Not sure what you're trying to do (that makes 2 of us) but:
<nckx>I did drop the sudo since I didn't see the point.
<rekado>htgoebel: no, it should be used as the replacement for PYTHONPATH in etc/profile and in shell wrappers.
<rekado>nothing fancy
<rekado>my goal is merely to keep Guix from breaking non-Guix Python applications.
<sebboh>nckx: at first glance, that is what mine looks like
<sebboh>nly: yours is different than mine. You don't have a /boot/efi.
<sebboh>Thank you both!
<nly>Oh right, I don't use EFI, hope it helps
<nckx>sebboh: My bootloader form's also super not fancy:
<nckx>sebboh: This couldn't be an NVRAM thing or a Restricted Boot hijack?
<sebboh>nckx: hm, my current config.scm has some experiment with legacy/non-UEFI stuff in it, but I think my previous was the same as yours except I didn't have a timeout form.
<sebboh>nckx: I wonder that. Does anybody in here have to add something from a third party to get guix to boot on hardware? nckx: I've never had any trouble booting debian on this hardware, but for all I know they do something for me automatically. (I don't use any special non-free installer, though.)
<nckx>sebboh: Can't really say. My most recent laptop (this one) is from 2012, I think that's pre-Restricted Boot nonsense.
<sebboh>aha.. "shim x 64" Does that mean anything to anyone? the working/booting debian install has /mnt/EFI/debian/shimx64.efi
<nckx>sebboh: I vaguely remember something about Ubuntu needing to use a Microsoft-approved ‘shim’ to boot on such machines.
<nckx>sebboh: Yeah, that.
<htgoebel>rekado: Does this contain a single path element only?
<htgoebel>(Sorry for asking, I lost track of guix development and ATM can't remember all details.)
<htgoebel>I'm asking, since if (GUIX_)PYTHONHOME would be easier to implement and more appropriate.
<nckx>I wasn't sure if Debian'd been reduced to that too.
<nckx>sebboh: All I can say for sure is that Guix provides nothing like that at all.
<nckx>If that shim GRUB isn't completely gimped, you might be able to chainload the real (Guix) one from it.
<nckx>And then you're free.
<sebboh>I will attempt to determine if my motherboard actually depends on this "shim" and if I can disable the feature that depends on it.
<nckx>sebboh: If your UEFI set-up tool says anything about ‘secure’ boot, disable it.
<nckx>(Not sure how much you know about UEFI, apologies if that were obvious ☺ )
<sebboh>Meanwhile, my brother-in-law's next motherboard (TBD) will probably require this secureboot stuff. (Right? a desktop computer from 2021 will have secureboot or successor?) So, guix should have a ticket about that if it doesn't already.
<sebboh>nckx: is my UEFI set-up tool the same as my "BIOS"? ("hit DEL to enter setup"?)
<nckx>(It's misleadingly called ‘secure’ to trick average users into thinking it's a good thing.)
<nckx>sebboh: Yep.
<nckx>sebboh: I *think* Microsoft currently forces manufacturers to enforce restricted boot if they want Windows 10 support, but it's not my area of expertise.
<nckx>They might have backtracked from that 🤷
<sebboh>ok. No, I checked for UEFI options and found very little. Meanwhile there is a GUI (actual GUI, you drag points on a 2D plane! rather than click up and down arrows or buttons..) for managing the fan power vs sensor temp profile.. it's a gaming PC I got a good price on. Made by MSI.
<efraim>sebboh: I'm running guix system on a macbook 4,1 which I converted from a debian system so my /boot might not be so helpful:
<nckx>sebboh: Have you tried searching for <model name> disable ‘secure’ boot?
<sebboh>nckx: no, I tried <model name> boot UEFI sata SSD (because I was having a bit of an X vs Y problem..)
<nckx>I'm not finding much good news about using the shim to load a real GRUB (I mean: arbitrary non-MS-approved codez! Oh noes!) either ☹
<nckx>But I don't really know what to search for either.
<sebboh>oh. Chance brought me to this URL: That is not the same motherboard as mine. However the screenshot looks the same. AND.. look where they are in the menu.. under a menu option called "windows", they have this disable secure boot option.. OK, well, my bios has a Windows menu option, which I have never opened. ;) yet!
<nckx>I had a laptop once that hid a very generic option (I don't remember which, but it was something like enabling USB boot protocol support) as ‘Boot Windows XP’.
<nckx>PC firmware truly is the ‘hey kids, wanna see a dead body’ of software engineering.
<sebboh>If Secure Boot dies out (or can always be disabled), it will because buyers make returns when they find a secure boot that can't be disabled. And that depends on medium sized businesses using bespoke bootloaders. Yeah, or XP. :)
<nckx>Oh, and my previous machine could only see a file exactly named ‘boot/bootx64.efi’. I had to rename my grubx64.efi to that and it suddenly worked. Stuff like that. Not a great tip, but if you're really stuck...
<sebboh>To quote from the debian installer... "some BIOSes are buggy!"
<nckx>sebboh: By ‘enforce’ above I meant ‘force it to be enabled, without an option to disable it’. But that might have been FUD. I hope it was.
<nckx>sebboh: If you're in a position to make businesses make the right decision, and do, I salute you ☺
<sebboh>yeah. I believe that the majority of smartphones have some kind of smartboot "enforced" like that. Paradoxically, it is the more secure and more expensive flagship phones that allow the user to disable it. (Because of returns, I bet!)
<sebboh>nckx: I am not but I will keep it in mind, trust. :)
<nckx>Not all phones? To be honest, and this is just a sad indictment of the current world, I thought it was even worse than that.
*nckx #notallphones
<sebboh>I think this is the source code for shimx64.. Egad, what license is that? Beyond that, remember, the mechanism by which this works could be described as PKI or DRM, so.. look at this: ... I think this is WRONG. Actually, what I said earlier about disabling Secure Boot is wrong. It should be always on, but *I* and noone else should be able
<sebboh>to provide the root of trust for devices I own.
<sebboh>In fact my device should not pre-trust Google or Microsoft out of box. Users should have to opt-in to that.
<nckx>sebboh: Yeah. Which is why ‘Secure Boot’ is a misleading propaganda term that shouldn't be used ☺
<nckx>It is exactly that: DRM at the boot firmware level.
<sebboh>What should it be called? The trust chooser? "Senator, you shouldn't let microsoft pay IBM to lock my kid's computer's trust chooser to microsoft!"
<sebboh>wait, am I reading that URL right? "restricted boot" is an FSF-favored term? I guess I'll queue the article itself up on my pocket Google.
<nckx>But sebboh, this is the competition of the free market doing what's best for consumers, you leave your big government out of this.
<nckx>sebboh: Well, that does actually describe its purpose and function, unlike ‘secure’.
<sebboh>oh, that's right, I should not employ senators as foils in my rhetoric. Honestly there's just too much "well actually all that stuff you used to think was is not" to keep track of.
<nckx>¯\_(ツ)_/¯ ikr
*nckx peeks whether they can afford free hardware yet. Nope.
<rubic88>If running an mcron job as a user, is it necessary to specify the full path to the program to be executed (or alternatively, source the user's profile to get the execution PATH)?
<rubic88>user == non-privileged user, I mean.
<nckx>rubic88: mcron should not reset a user's PATH.
*rubic88 dumping my envvars to verify ...
<nckx>Ah, the safe insulation of ‘should’...
<nckx>And you're saying that the environment whence you execute mcron has a different $PATH?
<nckx>Where are you starting mcron?
<rubic88>I'm running mcron as a service from config.scm and have a job defined as #user:rubic88 (the job runs, but only if I use absolute execution paths.
<rubic88>I mean within the script that mcron is executing.
<nckx>rubic88: ...but you said you were running mcron as a non-privileged user...
<nckx>If you're running it from your system configuration, of course it will only have the system path.
<rubic88>nckx: Makes sense. If I want to run it as a non-privilged user, how would I do so?
<nckx>rubic88: If you use ‘system’ (as opposed to ‘system*’), that will actually spawn a shell for you so you can use bash syntax (including sourcing whatever you want and even running su...)
<nckx>rubic88: I thought you were running it from within your ‘user session’ (and how you'd do that presumably depends on what you use, e.g. .xsession, GDM, ...)
<nckx>rubic88: The manual mentions ‘Additionally a USER keyword argument can be supplied to use a different user than the one defined in 'configuration-user' global variable.’
<rubic88>nckx: I see. Will look for that verbiage.
<nckx>Is that what you're using? Because I've never done that.
<rubic88>nckx: Nope. I thought specifying #user:rubic88 would run the jobs as a non-privileged user.
<nckx>(Presumably you'll still have to set PATH, I'm just trying to figure out how you're trying to launch commands, but unfortunately my people need me & I must go.)
<rubic88>nckx: Thanks, bye.
<nckx>rubic88: Sorry I wasn't of more help. Good luck. o/
<ngz`>Hello folks. I have a genuine question: can Guix use inputs' static libraries in a package definition, or is there a limitation?
<rvgn>Hello Guix
<sneek_>rvgn, you have 1 message.
<sneek_>rvgn, kmicu says: to highlight kmicu for an explanation for guix pull messages about manually exporting envars.
<rvgn>kmicu Just highlighting as you asked.
<arkhan>greetings, can be activated from emacs guix environment, that is to say to have a behavior similar to python-virtualenv
<rekado>htgoebel: no, it’s one entry per package.
<rekado>ngz`: yes, you can use static libraries.
<ngz`>rekado: OK. Thank you. Is there any gotcha to pay attention to?
<ngz`>rekado: (I'm thinking about bug#36043 for reference)
<rvgn>kmicu Sorry, got disconnecting for a while.
<ngz`>For example, does the inputs need to be a native-input if I want to use the ".a"?
***atw` is now known as atw
<rekado>ngz`: no, the .a file just needs to be provided. Sometimes it’s in a different output.
<nullsix>hi, how can i have 'startx' to work, dont want a graphical login manager
<g_bor[m]>Hello guix!
<g_bor[m]>Istm that some software from the wishlist is already packaged.
<g_bor[m]>One that I believe is done is GNU Jami. Can I remove it from the list, or the packaging isn't complete yet? If something is missing I am willing to help, but it seems tob
<g_bor[m]>e ok...
<rvgn>kmicu Are you available to discuss about the envars?
<xavierm02_>Hey. Is there a reason why modify-services exists, but modify-operating-system doesn't?
<nckx>nullsix: (service special-files-service-type `(("/bin/startx" ,(xorg-start-command))) ; worked for meee
<xavierm02_>I'm having trouble using set-field: (set-field my-os ((operating-system-host-name) "QWE")) yields "error: source expression failed to match any pattern".