IRC channel logs

2020-11-22.log

back to list of logs

<lfam>Howdy
<roptat>wow, either I'm getting good at writting parser, or Android.bp files are super easy to parse
<roptat>talking too fast ^^'
<raghavgururajan>roptat or lfam: Any of you available to push this minor patch-set? 🙃
<lfam>Which one?
<raghavgururajan>44781
*lfam looks
<leoprikler>Someone should write a build system called yawn (yet another wrapper for ninja).
<lfam>Lol
<roptat>I thought it rang a bell... well there's yarn already: https://yarnpkg.com
<roptat>also this yawn, somewhat related to packaging: https://github.com/aclowes/yawn
<roptat>yarn is a package manager for Node.js
<lfam_>raghavgururajan: My only quibble is with patch 2, Update synopsis and description
<lfam_>We should try to avoid subjective judgments or so-called "marketing language", as recommended by the Guix manual section Synopses and Descriptions
<lfam_>These fields should "be factual, mentioning use cases and features"
<raghavgururajan>lfam_ Ah gotcha!
<roptat>looks like I like peg parsers: https://framagit.org/tyreunom/guix-android/-/blob/master/android/build/blueprint.scm :D
<lfam_>If you send an updated patch set, either omitting that patch or amending it to mention more use cases and features, I can push tonight!
<raghavgururajan>You mean the mordern?
<raghavgururajan>lfam_: Few minutes.
<lfam_>I think that "modern" and "clean" are hard to judge. I think we can assume the software is reliable. And keeping privacy in mind is hard to judge without a description of how it does that (for example, does it use the Signal protocol? OTR? PGP?)
<lfam_>It seems like every communication tool claims to keep our privacy in mind these days
<raghavgururajan>lfam_ I agree with the mordern. But the line about clean and privacy, it says it focuses, like in one-going work or objective.
<lfam_>What does it mean to be clean? They washed it with soap? :)
<lfam_>Nobody says their software is dirty :)
<raghavgururajan>I think it was meant to be in contrast to Gajim
<raghavgururajan>No too many buttons or options.
<raghavgururajan>May be I can replace clean with minimal
<raghavgururajan>Would that be okay?
<lfam_>Sure
<lfam_>raghavgururajan ^
<raghavgururajan>Cool!
<raghavgururajan>lfam_: Done! :-)
<raghavgururajan>Is the description okay now or you wanna change to something else?
<lfam_>raghavgururajan: It's good. I re-added the word "graphical" to the synopsis, squashed the last 2 patches, and pushed
<lfam_>I often search for "graphical" so I think it's useful
*vagrantc ponders staying awake or waking up absurdly early to catch guix days sessions
<raghavgururajan>lfam_: Cool and thanks!
<vagrantc>and ... UTC+1 ... i can easily convert to UTC, but UTC+1 makes me never sure if adapting in the right direction
<lfam_>Heh
<lfam_>`TZ=UTC+1 date`
<lfam_>"What time is it now in UTC+1?"
<vagrantc>right ... i have done this :)
<lfam_>I admit I'm still unsure
<vagrantc>$ date --date='2020-11-22 11:59 UTC+1'
<vagrantc>that seems right
<lfam_>It actually gives me a different answer than the "what time is it in time zone" websites
<lfam_>They tell me it is currently 02:35:16 in UTC+1
<lfam_>And that command tells it is currently 12:35:31
<lfam_>So...
<vagrantc>my offset from UTC+1 is nine, so i trust the results of the date command above
<lfam_>Weird that this command doesn't work as expected for me
<lfam_>Sigh
<lfam_>I suppose that I trust the web more
<vagrantc>TZ=US/Pacific date --date='2020-11-22 11:59 UTC+1' -R ... is what i'm looking at
<vagrantc>i like that the tool confirms my basic math :)
<vagrantc>lfam_: maybe your system is not the timezone you are looking for?
<lfam_>I thought that `TZ=UTC+1 date` would show me the current time in that timezone, but it shows me something different than what I expected
<lfam_>I think it handles the +/- operators differently than I expected
<vagrantc>TZ=UTC+1 date -R
<vagrantc>Sun, 22 Nov 2020 00:39:49 -0100
<vagrantc>wat?
<vagrantc>usually when I'm looking at other timezones, i'm looking for a specific time to compare to anyways
<vagrantc>rather than the present
<lfam_>I thought it was 2:40 o'clock in that time zone
<roptat>yep, it is
<vagrantc>note the -0100 ? what?
<lfam_>Oh
<vagrantc>lfam_: i like my method better :)
<roptat> for me it shows the date in UTC
<lfam_>Yes, same here
<lfam_>I find it confusing
<vagrantc>yeah, i've found the TZ variable ... sometimes gives out surprises
<roptat>ah no, it's the same
<roptat>weird
<roptat>UTC+1 is the same as CET
<vagrantc>it appears to invert the TZ=UTC+N or TZ=UTC-N ... i think we found a bug in date
<roptat>well not really, but we set the time to CET which happens to be the same as UTC+1 right now
<vagrantc>really should have specified UTC in the conference ... even though i know for most attendees it will be local time :P
<roptat>well, remove 1 hour for UTC :)
<vagrantc>seems deceptively simple, and i'll admit, it should be easy. yet ... somehow i'm never confident of the results.
<roptat>you're right, I'm never sure I counted that in the right direction
<vagrantc>anyways, unless you're not really at UTC+1, i *think* i know what absurdly early hour i need to be awake at for the release BoF
<vagrantc>now the question turns to ... stay awake or wake up?
<roptat>is that early for you? ouch :/
<vagrantc>3am here ... which i did accidentally practice staying awake till then yesterday... :)
<roptat>waking up early is always more difficult for me than staying awake
<vagrantc>yeah, i think you're right
<vagrantc>so icecat on guix has been tested reasonably well with the BBB instance for guix days?
<vagrantc>i may as well boot to my guix system for it
<vagrantc>been a while since i've updated, maybe should start now... :)
<roptat>yeah, the latest icecat should work
<roptat>modulo disabling some pre-packaged addons maybe
<roptat>not sure about icecat 68
***catonano_ is now known as catonano
*roptat zZz
<roptat>maybe if I fall asleep enough, I can wake up early enough :)
<roptat>see you soon!
<vagrantc>i guess, for me, this will be "guix nights"
*vagrantc also failed to watch anything yet, but should resolve that later tonight
<kozo[m]>mothacehe Hey, I built the iso for guix PBP. I dd'ed it on to the microsd card but the PBP doesn't boot from it. When looking at lsblk, it's a single partition, there is no boot partition. Is there something I had to do?
<vagrantcish>there shouldn't be a boot partition for the pinebook-pro
<vagrantcish>guix doesn't really support a separate /boot
<kozo[m]>OK. Does that mean mothacehe's patch is missing something or did I do something wrong?
<vagrantcish>kozo[m]: did you use the serial console? I don't think it has graphics support yet.
<kozo[m]>Oh, I install it like I would anything else?
<kozo[m]>No, I have not
<kozo[m]>I just got my cable from pine64 the other day actually
<vagrantcish>the upstream kernel doesn't support video output on the LCD screen
<vagrantcish>there's a wip branch that's getting a little stale with patches for a working kernel for the pinebook-pro
<vagrantcish>i think it's still linux-libre 5.8.x, though
<apteryx>vagrantcish: are you staying up for the Guix Day? :-D
<kozo[m]>Ok
<vagrantcish>apteryx: going to try :)
<kozo[m]>vagrantcish How do I go about installing guix using the serial console?
<vagrantcish>kozo[m]: i tried reducing the patchset to something reasonable so it could be merged to guix master, but it still required a lot of hackish patches unfortuanately
<kozo[m]>Gotcha
<vagrantcish>kozo[m]: you connect the serial console to another computer and if you're lucky, the image you booted will show you output
<kozo[m]>OK doke. I shall be patient for that wip-branch
<vagrantcish>kozo[m]: for the pinebook-pro ... something like: screen /dev/ttyUSB0 1500000
<kozo[m]>Thank you for the info, vagrantcish
<vagrantcish>or minicom or whatever, if you're into that sort of thing
<kozo[m]>emacs serial term <thumbs up>
<vagrantcish>i see!
<vagrantcish>TIL
<kozo[m]>It was a glorious day when I found that out. I didn't have to mangle something to get putty anymore
<vagrantcish>well, this is promising ... audio works with icecat ... hopefully it will work with BBB
<vagrantcish>at least, works with the videos to watch
***txgvnn is now known as Guest3994
***Guest3994 is now known as txgvnn`
***txgvnn` is now known as txgvnn
<vagrantcish>huh.
<vagrantcish>I guess the Obvious thing to do is reconfigure my system to UTC+1 for now... :)
<vagrantcish>my guix system was already running at UTC
<vagrantcish>yes, that was clearly the simplest fix for now :)
<ryanprior>I have a few go packages I'm working on that have mutual dependencies, eg each depends on the other. This is not super uncommon in the go world - it basically means that any time you grab source code for one package, you also need the source code from the other. One package in practice, but with separate repositories and namespaces.
<ryanprior>But I'm not sure how to handle it in Guix.
<ryanprior>My thought right now is to have an outer let- form that binds the origin of each package, and then in the package definitions, use those origins as inputs instead of using the actual packages.
<ryanprior>So given mutually dependent packages A and B, the package for A uses the origin for A as its source and the origin for B as its input, but doesn't actually depend on B; and vice-versa.
<ryanprior>Does that make sense? Is that abusing the Guile language features, or is that using them as intended? Is there a better cleverer or more idiomatic way of doing this?
*vagrantcish wonders how long efraim's talk is supposed to be ... seems to cut out at ~16 minutes when explaining vim
<apteryx>GNUtoo: ah, a very interesting talk about what's needed to build Android packages with Guix! https://xana.lepiller.eu/guix-days-2020/guix-days-2020-julien-lepiller-bootstrapping-maven-480.webm
*apteryx zzz
<wleslie>is there a good mu tutorial end-to-end somewhere? even if not guix specific
<wleslie>do we even have just maildir-utils (not mu4e) ?
<wleslie>do people still use openimap or imapfw, or is there something nicer?
<wleslie>sorry, "offlineimap"
<efraim>i use mbsync from isync
<efraim>vagrantcish: there's another copy here https://flashner.co.il/~efraim/build_it_with_guix_scm.mp4
<abcdw>good morning, guix!
<efraim>morning!
<efraim>we start in about an hour?
*vagrantcish waves
<vagrantcish>will at least be back for the release one...
<efraim>DO GOOD NAPPING!
<abcdw>efraim, I think so, but better as roptat or zimoun.
<abcdw>s/as/ask
<bonz060>Morning!
<zzappie>guix morning
<efraim>morning!
<g_bor[m]>hello guix!
<zimoun>Hi! I have launched the Room the Guix Day: https://guixbbb.fosshost.org/b/zim-dmr-hp2-ibu Please join us!
<zimoun>I hope that everything will be fine. :-)
<zimoun>nckx: Are you around?
<bonz060>Just joined :)
***apteryx is now known as Guest26291
***apteryx_ is now known as apteryx
<lalo_salamanca>Guys, what's going on now? Where can we see live events? Any links?
<lalo_salamanca>I'm interested in "From v1.2 to release process"
<abcdw>lalo_salamanca, https://guixbbb.fosshost.org/b/zim-dmr-hp2-ibu
<zimoun>lalo_salamanca: Please https://guixbbb.fosshost.org/b/zim-dmr-hp2-ibu
<iyzsong>party joined!
<zimoun>Does anyone can connect? Any issue?
<iyzsong>no issue here
<vagrantcish>name doesn't resolve for me
<zimoun>efraim: do you have networking issue?
<efraim>I think I got it finally. I'm connected on my phone, was having trouble with microphone and not just audio
<zimoun>efraim: ok
<efraim>echo test wasn't loading, not sure if it's firefox on my phone or my phone being old
<efraim>i'll grab my wife's phone later, her's is better
<efraim>or even her computer!
<apteryx>zimoun: I can't seem to connect
<apteryx>'The site can't be reached'
<zimoun>apteryx: this link https://guixbbb.fosshost.org/b/zim-dmr-hp2-ibu ?
<apteryx>yes
<apteryx>it's strange, it was working earlier
<zimoun>wow! Which browser?
<apteryx>ungoogled-chromium
<zimoun>it should work, maybe try with IceCat from Guix
<zzappie>a was able to hear cbaines in ungoogled-chromium
<apteryx>yeah, perhaps just my network
<jlicht>try resetting your nscd cache if you have 'persistent' name resolution issues
<apteryx>probably something related to broken routes after my ISP reconnects around 3 AM (I had an openconnect VPN active).
<efraim>quick! tell us what the presentation was about! j/k
<vagrantcish>apteryx: you also having name resolution issues?
<vagrantcish>only domains i can't resolve are for fosshost.org ...
<apteryx>vagrantcish: I do! weird
<pelzflorian>me too
<apteryx>The pipe is broken from our side of the Atlantic ;-)
<vagrantcish>ungoogled-chromium and icecat, no difference
<cbaines>The IP is 147.75.35.154
<apteryx>it pings
<vagrantcish>whois data for *.org also ... not working for me ...
<apteryx>cbaines: eh, should 'https://147.75.35.154/b/zim-dmr-hp2-ibu' work? I get a 500 error.
<efraim>Do you use a set collection of packages in the manifest? It might speed up startup time to pre-cache building the manifest, possibly with cron.
<pelzflorian> https://147.75.35.154/b/zim-dmr-hp2-ibu shows error 500 internal server error; i will try to rewrite the /etc/hosts file
<cbaines>apteryx, I think you probably need to send the right Host header
<vagrantcish>wow, been a while since i've edited /etc/hosts... :)
<vagrantcish>cbaines: thanks :)
<apteryx>slapping '147.75.35.154 guixbbb.fosshost.org' in /etc/hosts does fix it
<jonsger>couldn't you use a DNS server like 8.8.8.8 or 9.9.9.9? Shouldn't that help?
<pelzflorian>yes fixes for me
<apteryx>jonsger: it was working earlier; strange
<vagrantcish>apteryx: yeah, same here.
<jonsger>apteryx: so you are using the DNS server of your ISP?
<apteryx>yes, automatic DNS from my ISP
<apteryx>tried in icecat but had no audio; ungoogled-chromium works fine
*vagrantcish is stuck in the guest lobby
<nckx>vagrantcish: You're not listed there :-/
<nckx>Can you reconnect?
<vagrantcish>"Please wait for a moderator to approve you joining the meeting."
<nckx>I did.
<vagrantcish>nckx: i guess now i should ask ... define reconnect?
<vagrantcish>i both tried reloading the page, and closing it and opening it again
<vagrantcish>same result
<nckx>I've approved you twice (from the lobby) but you don't show up in the list of participants.
<mroh>hmm, I can't hear anything, not in icecat nor in chrome.
*vagrantcish wonders if it requires some additional name resolution thing
<civodul>o/
<civodul>apteryx: i had to use ungoogled-chromium too
<nckx>Quite possible. You can look in IceCat's ‘Network’ tab (right-click → ‘inspect element’).
<andreas-e>vagrantcish: I also approved you once.
<iyzsong>i can hear the words using icecat, but can't understand them due to my poor English >.<
<vagrantcish>oooh, now i got unauthorized ... trying the same url from ungoogled-chromium
*vagrantcish tries again
<nckx>Now you're listed.
<andreas-e>Great!
<nckx>I'm using Guix IceCat with no problems.
<jonsger>iyzsong: we need some AI translation...
<vagrantcish>got audio with icecat ... had to enable some javascript for icecat to work
<nckx>👍
*jonsger is out of trouble with firefox ^^
<nckx>Did anyone's BBB page just go entirely black?
<jackhill>nckx: still good for me
<nckx>I still have audio but the UI has vanished 😒
<jackhill>whacky
<jlicht>nckx: Did you full-screen it? It stopped updating the UI for me after that, had to re-enter the room
<iyzsong>mine is working fine.
<nckx>jlicht: No, but I think it did coincide with the change of presentation to ‘Welcome to BBB’ which I see now (had to refresh).
<nckx>Hi all!
<vagrantcish>i guess i shouldn't have logged out ... was trying to re-enable the mic
<zimoun>vagrantcish: are you able to turn on your mic?
<andreas-e>Never leave a running system ;-)
<vagrantcish>zimoun: in theory
<vagrantcish>will we have a hallway track during the break?
<leoprikler>BBB hates me.
<nckx>leoprikler: What's the problem?
<leoprikler>"browser too old"
<nckx>Yech.
<jonsger>leoprikler: which are you using?
<leoprikler>epiphany
<nckx>leoprikler: Does it allow a custom U-A?
<nckx>Setting IceCat's to Firefox has fixed every single such problem for me so far.
<leoprikler>not as far as I can see
<efraim>I'm using some in-between time to review the pijul patch set
<stikonas>BBB doesn't like even old Firefoxes... I think you need at least 78
<jonsger>stikonas: possible
<chrislck>if I particularly want to launch guix version 2.0.9 for testing, is it possible via guix ?
<lfam>chrislck: Do you mean Guile 2.0.9?
<chrislck>yes :-/
<lfam>We don't currently have that version packaged, although we do have 2.0.14 (I assume it's part of the same series and basically a collection of bug-fixes)
<lfam>We do use 2.0.9 as a "bootstrap" binary and it might be possible to launch it from the Git repo but that's not really a first-class use case
<chrislck>ok. I guess it'll be difficult to create a package locally to test 2.0.9?
<chrislck>(because 2.0.9 is the minimum supported guile for gnucash and /me wish to test a few @future things)
<lfam>I honestly don't know how difficult it would be. It's easier to make custom packages for Guix than for e.g. Debian, but that doesn't account for any difficutly that may be inherent to Guile 2.0.9 itself
<lfam>It might be possible to define a new package guile-2.0.9 that inherits from guile-2.0 and basically just changes the version and source hash
<nckx>For quick tests you could use guix time-machine with commit 78636f0b8ce0fa7c3483a378ef31832f38a3d19a but all other packages will be from 2016 as well.
<leoprikler>can time-machine safely travel that far back?
*chrislck runs time-machine and will buy BTC. brb.
<lfam>Good luck
<nckx>It's always safe (assuming you don't go reconfiguring/‘upgrading’ your system from inside the DeLorean, just don't do that) but it might fail, sure.
<nckx>My memory doesn't go back that far.
<leoprikler>My intuitive definition of "safe" differed from the technical one once again :(
<nckx>I didn't mean to ‘correct’ you, just to reassure chrislck. ☺
<nckx>It might go crack but it won't go boom.
<chrislck>that's reassuring!
<chrislck>guix time-machine --commit=78636f0b8ce0fa7 -- guile
<chrislck>error: %guix-register-program: unbound variable
<chrislck>'crack'
<lfam>I would try something like this: `guix time-machine --commit=78636f0b8ce0fa7 -- guix environment --ad-hoc guile --dry-run`
<lfam>Nevertheless, I got the same error as you
<lfam>I suspect there are some incompatibilities in Guix's interface between these versions
<leoprikler>wouldn't the correct thing be -- environment instead of -- guix environment?
<lfam>Right!
*lfam is still getting the hang of the time machine
<lfam>But, a lot has changed in Guix since 2016
*nckx is still pulling the past to test ‘guix time-machine --commit=78636f0b8ce0fa7 -- guix environment --{ad-hoc,pure} guile -- guile --version’
<nckx>That's too much --s.
<leoprikler>Sadly guix time-machine can't travel back far enough to make --kill-hitler meaningful :(
<nckx>I'm not going to say anything about 2016.
<leoprikler>--commit-election-fraud
<nckx>Nah, just rebase.
<nckx>guix time-machine: error: Git error: cannot locate remote-tracking branch 'origin/keyring'
<nckx>Right, another cool post-2016 feature.
<nckx>So many of those ♥
<nckx>Although I think that's due to me using ‘upstream’ instead of ‘origin’, not the time machine.
<leoprikler>This does raise the interesting question of whether Guix time-machine will be able to travel back to Corona Days in 2030.
<leoprikler>how do I `git send-email` a patch for the staging branch?
<nckx>It should be. Now that it exists we can mandate not breaking it. That was harder before.
<nckx>leoprikler: Just add [staging] to the subject.
<nckx>See --subject-prefix=.
<mbakke>so I'm trying to use (substitute* ...) in a trivial-build-system, and get "In procedure frame-local-ref: Argument 2 out of range: 1" ... any ideas what's going on? :)
<leoprikler>you got a really really really bad exception
<mbakke>"Wrong type to apply: "foo""
<mbakke>obviously the same works in a guix repl :P
<leoprikler>What exactly is foo hereß
<mbakke>the literal string "foo" :)
<leoprikler>If it's something that could reasonably be a procedure, you probably need to unquote it.
<leoprikler>did you perhaps write (("bar")("foo"))?
<mbakke>(substitute* "/tmp/test" (("foo") "bar"))
<leoprikler>mhh, it's probably not recognizing substitute* as a syntax
<leoprikler>(use-modules (guix build utils))
<mbakke>I'm using mkdir-p from (guix build utils) just above the substitution.
<vits-test>today was fun.
<mbakke>here is a reproducer for the trivial-build-system issue: https://paste.debian.net/1173847/
<mbakke>huh, removing the let binding and using %output instead of out works
<luis-felipe>Oh: https://guixbbb.fosshost.org/b/zim-dmr-hp2-ibu → Server not found
<sneek>luis-felipe, you have 1 message!
<sneek>luis-felipe, nckx says: The Web site banner looks much better now; thanks.
<mbakke>ah, and (use-modules ...) before the let binding works too
<nckx>luis-felipe: The DNS doesn't play well with some continents/countries. You could add ‘147.75.35.154 guixbbb.fosshost.org’ to /etc/hosts.
<nckx>Or use a public DNS server like 9.9.9.9.
<nckx>I'm afraid we can't fix this.
<luis-felipe>Ok, thanks.
<GNUtoo>apteryx: thanks
<simonsouth>nckx: Thanks, I've been having the same problem wrt DNS.
<jonsger> https://dnschecker.org/#A/guixbbb.fosshost.org doesn't look so green
<simonsouth>"It looks like you're using a browser that is not supported". Let's try again...
<simonsouth>Ah, there we go.
***jess is now known as j
***sneek_ is now known as sneek
<nckx>Yes, fosshost's DNS configuration looks quite broken to me, which is suprising.
<jonsger>nckx: do we have a contact of them? to reach them out about this issue
<nckx>jonsger: Yeah, I'm going to double-check and send them an e-mail.
<nckx>Does the delegation look lame to anyone else? I have hobbyist knowledge of DNS but am not a pro.
<nckx>No glue records here.
*janneke finally has tinycc convinced to emit __memmove instead of memmove
<simonsouth>nckx: The problem here seems to be that ns{1,2}.delivr.dev won't resolve, and everything falls apart after that.
<nckx>That sounds like an insignificant detail so it must be crucial.
<nckx>simonsouth: Right!
<nckx>That's because (IIUC) none of the .org NSes return A records for them.
<simonsouth>Hmm.
<nckx>janneke: What's the difference?
<janneke>nckx: terrible tinycc details: tcc adds libtcc1.a at the very end of the linkage line; if nothing uses memmove before, but libtcc1.a does (because it was compiled by tcc and it emits "memmove"), then memmove is an unresolved symbol
<janneke>(adding an extra -lc just puts that in before libtcc1.a)
<janneke>also, libtcc1.a cannot come with its own "memmove", because then memmove is defined twice
<nckx>OK, that makes sense, even to me.
<janneke>so, i'm having tinycc emit __memmove, and will be adding __memmove to libtcc1.a
<janneke>...but trying to do this, tcc has been just segfaulting on me for half a day or so...
*nckx sent fosshost a friendly suggestion to fix their DNS.
<nckx>Hm, I appear unable to write a sentence that doesn't sound like I'm petting an evil cat in my moon-base.
<nckx>It was friendly! I promise.
<leoprikler>"That's a nice substitute, you have there, Mr. Bond. It would be very unfortunate if something happened to it."
<lfam>"Taking out the garbage" is usually a nice thing to do
<nckx>I'm just saying all they have to do is apply my exhaustively documented steps and we'll have full root acc^W^W^Wtheir DNS will be much more robust next Guix Dayz.
*nckx AFK.
<jlicht>"Anway, here's wond^H^H^H^Ha PEG parser" :-)
<eth01>Hi folks, we did have some issues with our anycast dns network (fosshost here) I believe the issue has been resolved
<leoprikler>I find the gstreamer.scm ⇔ video.scm loop extremely disturbing
<eth01>I've asked our network admin to investigate
<eth01>apologies o/
<mroh>eth01: Thank you very much!
<eth01>I have also replied to Tobias' email ^ np
<eth01>i believe issue was fixed yesterday, it's unclear if issue is still ongoing..
<luis-felipe>Hmm, it doesn't work for me though...
<jonsger>is someone using waybar and can tell me which font is required to show those symbols down their? I still have some "unicode boxes" left...
<raghavgururajan>Hello Guix!
<raghavgururajan>This site can’t be reached. guixbbb.fosshost.org’s server IP address could not be found.
<raghavgururajan>Oh already discussed
<raghavgururajan>never mind
<eth01>^^ What happens if you do a trace route please PM me the results
<abcdw>civodul, example of transparent import for python in nix: https://www.tweag.io/blog/2020-08-12-poetry2nix/
<raghavgururajan>> eth01‎: ^^ What happens if you do a trace route please PM me the results
<raghavgururajan>me?
***j is now known as jess
<raghavgururajan>[1] https://dnschecker.org/#A/fosshost.org [2] https://dnschecker.org/#A/guixbbb.fosshost.org
<raghavgururajan>Oh PM, sorry.
<luis-felipe>raghavgururajan: What command did you use? I'd like to try
<raghavgururajan>luis-felipe: I did not use any command. The line "This site can’t be reached. guixbbb.fosshost.org’s server IP address could not be found." was shown in ungoogled-chromium.
<luis-felipe>raghavgururajan: I mean, for the traceroute
<raghavgururajan>luis-felipe: I used the website dnschecker.org
<ryanprior>I can't access it either. Open Jitsi room for anybody who wants to chat: https://meet.jit.si/CO2dbbyp/GuixHackersWhomstCantGetIntoBbb
<luis-felipe>$ traceroute -w 3 -q 1 -m 16 guixbbb.fosshost.org
<luis-felipe>traceroute: unknown host
<raghavgururajan>eth01, luis-felipe: Now I used some commands. Here are the results for each: https://paste.debian.net/1173867/
<ryanprior>I can't reach fosshost via dig or nmap
<jonsger>ryanprior: does adding `147.75.35.154 guixbbb.fosshost.org’ to /etc/hosts.
<jonsger>help
<jonsger>?
<raghavgururajan>Based on the results from dnschecker.org I posted above, it seems DNS values are not propagated properly. Some places resolve, some do not.
<ryanprior>jonsger: yes I can connect by adding that to my hosts file
<jonsger>ryanprior: fine, it's a DNS setup problem. Its known to fosshost and they are aware of it now :)
<raghavgururajan>> ‎jonsger‎: ryanprior: does adding `147.75.35.154 guixbbb.fosshost.org’ to /etc/hosts.
<raghavgururajan>Thanks a lot! It works.
<zimoun>janneke: I have tried Guimax but the guix.scm failed, sadly.
<janneke>zimoun: oops
*janneke wonders what version of guile-gi that's stuck at
<zimoun>janneke, I have just checked it couple of days ago, when speaking about GUI. And now, that GUI pops up again, it remembers me that I have not reported. :-)
<janneke>zimoun: np, thanks!
<leoprikler>Any Rust folk here, how long does it usually take to build Rust on Guix?
<leoprikler>[single version, not full bootstrap]
<efraim>6 hours? I don't remember
<raghavgururajan>dannym: Hey o/
<PotentialUser-86>Greetings ! I was wondering if there ever was a talk about creating a Guix wiki page, similar to the ArchLinux one ? And would there be any complications that I didn't think about other than just maintening the website ?
<zimoun>Thanks everybody for participating! I have learnt a lot and refill with new ideas and energy.
<dannym>raghavgururajan: Hi :)
<divoplade>Hi :)
<divoplade>I'm trying to push my git repo over SSH to my guix server
<divoplade>However, I get: bash: git-receive-pack : commande introuvable (this is run on the server)
<civodul>abcdw: thanks for the pointer!
<divoplade>So I have installed git as my user on the server, and set up .bash_profile
<divoplade>So when I connect to my server through SSH, I can run git-receive-pack, because .bash_profile set up the paths
<divoplade>But with git push it does not work
<divoplade>It seems that git push does not bother to read the server's .bash_profile
<divoplade>(the remote* .bash_profile)
<efraim>We just lost power, I'm out
<vagrantcish>hah, i just started producing power ... transmission cost is probably prohibitive, though
<Formbi>hi
<Formbi>I'm trying to set up guix-publish on a Guix System machine and it doesn't wanna work
<Formbi>I have «(service guix-publish-service-type (guix-publish-configuration (port 2137)))» in config.scm
<Formbi>and I get connection refused when trying to connect
<cbaines>Formbi, assuming you're using the default host option, that's to listen on localhost6
<cbaines>*localhost
<cbaines>You'll need to change that if you're trying to access it from other machines
<Formbi>mhm
<dannym>Apropos UI usability: I just added a channel and did guix pull and I get:
<dannym> guix-androidhttps://framagit.org/tyreunom/guix-android.git d822f9d
<dannym> amd file:///amd e0ff18c
<dannym> heads https://github.com/daym/heads-guix.git d27623e
<dannym> guix https://git.savannah.gnu.org/git/guix.git 1ca0c34
<dannym>Note: missing spacing in first line in original.
<Formbi>>The host (and thus, network interface) to listen to. Use "0.0.0.0" to listen on all the network interfaces.
<Formbi>cbaines: does this mean I should use «(listen "0.0.0.0")»?
<mbakke>Formbi: yes
<Formbi>thanks
<Formbi>and how should I specify additional URLs in the daemon configuration?
<cbaines>Formbi, additional URLs for what?
<Formbi>when I put «(substitute-urls (cons* "192.162.1.100:2137" %default-substitute-urls))», guix was saying «192.168.1.100:2137: invalid URI» and crashing
<samplet>jlicht: Is your work on the NPM binary importer available publicly?
<Formbi>192.168.1.100:2137*
<rekado_>WIP: https://elephly.net/downies/ode-to-one-two-oh.{txt,ogg} – any volunteers for adding some comping / soli / lyrics / vocals?
<cbaines>Formbi, it needs to be a URL/URI, so you need to add something like http:// at the start
<Formbi>ah, of course, thanks a lot!
<Formbi>also, my mpd service stopped working recently
<mbakke>rekado_: groovy! :D
<Formbi>could I look somewhere else than /var/log/messages and /var/log/debug?
<roptat>oh nice :)
<vagrantcish>rekado_: i can't believe it's not written in scheme!
<rekado_>vagrantcish: it was mapped out in Lilypond first; hope that counts :)
<vagrantcish>heh
*vagrantcish wonders about composing a verse consisting exclusively of typo/grammar/spelling fixes fixed in the last release
<janneke>rekado_: that's so cool
<abcdw>Thank you everyone for a great day. See you soon. o/
<Formbi>ok, now the publish works like a charm ^^
<Formbi>thanks!
<vagrantcish>ok, i've got a new package description: https://paste.debian.net/1173890/
<vagrantcish>it nearly parses, even
<zzappie>rekado_: Wow what time signature is it??
<nckx>(Tobias here) eth01: Cheers for the prompt reply! Other than that, AFAIK, the experience was great. Thanks for hosting everything.
<raghavgururajan>cbaines: I found out the issue regarding Gajim and Python. Python2's enum34 is interfering with python3.
<vagrantcish>the ultimate collage for http://issues.guix.gnu.org/44675
<eth01>nckx: awesome, thanks. I believe the dns issue was merely an intermittent issue, unable to replicate in UK and other core locations we operate from, but I agree there is an intermittent issue somewhere in our global infrastructure :)
<nckx>I'm in Belgium and got consistent (negative) results.
<raghavgururajan>eth01: Whats the IP for home-page or core-domain fosshost.org
*nckx 's been entertaining the last few hours, perhaps it's improved.
<eth01>raghavgururajan: 104.198.14.52 for our home page, for your bbb instance it is 147.75.35.154
<raghavgururajan>eth01: bbb session has ended. Just wanted to read about fosshost. :-)
<eth01>fosshost.org :)
<raghavgururajan>eth01: Is it currentlt part of LibreHosters Network?
<lfam>Thank you eth01 for supporting us! I was really impressed with the BBB service
<vagrantcish>yeah, BBB handled the BoF sessions pretty well ...
<raghavgururajan>yeah, BBB was damn good eth01. Thanks!
<vagrantcish>between being able to use the mic and also the text chat
<vagrantcish>doesn't allow multiple presenters very easily, though
<eth01>raghavgururajan: we were accepted onto the initiative of the librehosters network - but it's not something we hang our coat on
<vagrantcish>though that could be fixed by sharing an etherpad or gobby window for shared collaboration on editing text
<raghavgururajan>eth01: Cool!
<eth01>its possible to use an etherpad on bbb with some customisation
<vagrantcish>eth01: cool
<eth01>and thanks for your feedback ^^, if there is _anymore_ we can do to help guix, we already help with some stuff for @cbaines, let me know!
<eth01>we're in #fosshost on freenode
<roptat>what's the issue with fosshost? resolving DNS? connecting to the machine?
<nckx>roptat: They are/were missing glue records at the root.
<roptat>ah
<nckx>I'm still not getting any (quick foreach query of org root servers).
<nckx>Anyway, I'm sure competent people are on it ☺ See y'all later.
<raghavgururajan>Is Andrew Tropin here?
*nckx AFK.
<roptat>they left 20 minutes ago
<raghavgururajan>Oh okay.
<raghavgururajan>In their presentation, I wonder what 'module system' and 'implicity' means.
<jonjitsu[m]>I'd like to install guix package manager/build system in an existing ubuntu system and I do not have enough disk space on / for the store. I was thinking of bind mounting a directory from another partition at /gnu and then running the setup script. Will this cause any issues?
<rekado_>zzappie: it’s 5/4 throughout
<rekado_>(the chorus is the story of a contribution: sharing as others have shared, evaluating the patch, procrastinating on bugs… then countless recompilation in pure environments; and the first letters are always R E P L)
<rekado_>vagrantcish: that’s a great idea! :)
<rekado_>zzappie: it’s 5 because ‘ode to one two oh’ is five syllables; and it’s a quasi-palindrome
<zzappie>rekado_: ha, for some reason I could't recognize it :) Because there are accetns that make it feel like it's combined signature
<zzappie>very cool
<zzappie>I was collecting songs in 5/4 at some point
<rekado_>for me the danger with 5/4 is to fall into the familiar 3+2; then everything sounds like it should be ‘Take Five’. So I syncopated everything
<zzappie>Yeah It's true
<cbaines>raghavgururajan, OK, that makes sense. I guess having gajim not inherit the PYTHONPATH would avoid this.
<raghavgururajan>cbaines: You mean in the source-code or in the package-definition?
<cbaines>raghavgururajan, the gajim process inheriting the PYTHONPATH value from the environment, which could be done through the wrapper.
<cbaines>raghavgururajan, it looks like the Gajim uses the PYTHONPATH for extensions though, which this would break
<raghavgururajan>cbaines: Yeah extensions would break.
<cbaines>raghavgururajan, Not having python2-enum34 in your profile would help a bit, do you know why that's there?
<raghavgururajan>I think claws-mail has input python-2.
<raghavgururajan>When I uninstalled it. Gajim OMEMO started to work.
<cbaines>raghavgururajan, it doesn't propagate it, so it won't end up in your profile
<brettgilio> https://ci.guix.gnu.org/search/latest/image?query=spec:guix-master+status:success+system:x86_64-linux+hurd-barebones.qcow2 is giving me a bad gateway
<lafrenierejm>What's the process for using packages provided by custom channels in `guix system reconfigure`? Trying `sudo -E guix system -L ~/etc/guix/system reconfigure ./etc/guix/system/desktop.scm` gives me "no code for module".
<cbaines>lafrenierejm, -L should come after reconfigure I believe
<brettgilio>Yes. cbaines is right lafrenierejm reconfigure is the command for system. Your syntax order is wrong
<lafrenierejm>Got it. Thanks, cbaines and brettgilio!
<raghavgururajan>cbaines: Hmm. Let me see what else I removed.
***sneek_ is now known as sneek
<roptat>what's the license for our documentation?
<roptat>gfdl 1.3 or 1.3+? invariant or no invariant
<roptat>?
*jonsger confuses cuirass and publish service -> ends in bad configuration :P
<eth01>if anybody had issues earlier with the dns for guixbbb.fosshost.org please can you run this command: dig +trace A guixbbb.fosshost.org and upload to a pastebin?
<civodul>roptat: gfdl1.3+, with no invariant sections
<civodul>is there an ambiguity here?
<civodul>it's been a great day, people, we should have more of these!
<civodul>perhaps we could have live informal discussions focusing on one topic from time to time
<brettgilio>^ maybe on mumble?
<brettgilio>Id join
<civodul>yes, why not
<brettgilio>Id love it
<roptat>no ambiguity on the invariant section actually, because that would be nonfree, but yes, it's not clear if it's 1.3 only or or later
<civodul>like for GPL, if you don't specify a version, then "or any later version" applies
<nckx>eth01: https://paste.debian.net/plain/1173902
<vagrantcish>civodul: let's not forget back cover texts and front cover texts, now!
<raghavgururajan>GFDL?
<raghavgururajan>GNU Free Documentation License
<roptat>well, don't we specify 1.3?
<raghavgururajan>Oh I missed your following message
<vagrantcish>rekado_: if you have any idea how to incorporate into your ode, this is my typo collage in the form of a package description: https://paste.debian.net/1173890/
<raghavgururajan>roptat: Yeah, 1.3 should be mentioned. https://www.gnu.org/licenses/fdl-1.3.html#addendum
<civodul>vagrantcish: not the slightest invariant thingie, i promise!
*vagrantcish shakes head at the whole FSF promoting clearly non-free license thing
<civodul>yeah well, that was a sad episode for free software
<civodul>not so much the license itself, but rather the stubborn divisive behavior of the "project leader"
<vagrantcish>it's the not correcting it after so many years that's the big issue
<vagrantcish>right
<civodul>but hey, it's behind us! (except for Emacs)
<vagrantcish>hah!
<civodul>quite a few maintainers did the right thing in the following years
<vagrantcish>i think GCC and maybe glibc still have that problem
<civodul>yeah, they're not adults yet in that respect ;-)
<roptat>ok, I had a call with someone that manages fedora's weblate. We've set up something at https://translate.fedoraproject.org/projects/guix/ (not sure if you can see it) that temporarily uses a repository that I manage at framagit; hopefully we can migrate to the savannah repo as soon as it's accepted there, but discussions seem to have stalled...
<civodul>roptat: yay, looks nice!
<civodul>it's ok to host on Framagit if Savannah is not welcoming the repo
<nckx>eth01: Earlier someone pointed out <https://dnschecker.org/#A/guixbbb.fosshost.org>. I don't usually take much stock in such sites, but it did resolve all the other domains I've tried 100%.
<nckx>Why is Weblate controversial?
*nckx has wanted to ask this for ages.
<vagrantcish>ok, so i've packaged guix for Debian, now i need to package dpkg and apt for guix...
<Formbi>what does «BoF» mean here?: https://guix.gnu.org/en/blog/2020/online-guix-day-announce-2/
<vagrantcish>birds of a feather ... some saying "birds of a feather flock together" ... essentially an informal conversation between people sharing a common interest
<civodul>vagrantcish: wasn't efraim looking into packaging dpkg?
<vagrantcish>civodul: i'd be curious
<civodul>anyhow, that sound like the Obvious Next Step
<vagrantcish>hah, i misread that as obnoxious! :)
<civodul>i mean, an ambassador would definitely do that
<vagrantcish>but yes
<civodul>:-)
<janneke>lol
<Formbi>so is there even a point to include those informal conversations in the public blog post?
<vagrantcish>it may be crazy, but at some point i want to explore using guix to build "proper" debian packages
<civodul>it's not; back in the day Nixpkgs had stuff to build packages in other distros, using checkinstall mostly
<civodul>Formbi: those discussions were open to anyone
<civodul>that's why they're on the schedule
<vagrantcish>Formbi: sometimes the ideas from BoFs lead to ideas worth exploring in more depth, so i'd say it's worth mentioning if someone took notes worth summarizing
<civodul>s/anyone/everyone/ even
<nckx>Formbi: These were in the ‘main room’ so to speak, TBH I doubt anyone noticed a difference.
<vagrantcish>i basically want to use guix-daemon's containerization, networkless installation, etc. for building debian packages
<rekado_>vagrantcish: thanks! I think I’ll have an easier time using the correct words that these correspond to, though :)
<vagrantcish>rekado_: that would miss the point, perhaps :)
<vagrantcish>rekado_: it was a somewhat absurd proposal ...
<nckx>vagrantcish: Are those the most common?
<raghavgururajan>> nckx‎: Why is Weblate controversial?
<raghavgururajan>Hmm?
<vagrantcish>nckx: those are all of the ones since the last release, i think ...
<vagrantcish>nckx: that i submitted
<vagrantcish>oh, i missed a "permits to"
<rekado_>kinda hard to sing ‘commnd’, or to enunciate the difference in ‘overide’ vs ‘override’ :)
<nckx>raghavgururajan: I dunno, it seems like everyone exchanges knowing glances whenever it comes up & I'd like to know why, or if it's my imagination.
<raghavgururajan>nckx: Ah! I came across it for first time today.
<vagrantcish>rekado_: maybe it could have a "director's cut" version with the original, and the "edited cut" with the "fixes"
<rekado_>the harder problem is to carve out some space for a vocal line; the bass really dominates the verse
<vagrantcish>rekado_: but attempting to pronounce spelling mistakes is part of the fun :)
<rekado_>true!
<nckx>It looks free as in freedom.
<raghavgururajan>nckx: My first reaction on seeing "web-based" was "Oh boy!". I think it is better to have web-interface to a software, instead of making the software itself web-based.
<vagrantcish>what was crazy is i don't think i even changed the word order ... it was basically a reverse-order of what i got with: git log v1.1.0.. --author=vagrant --extended-regexp --grep='spelling|grammar|typo' --patch
<vagrantcish>removing duplicates
<civodul>wait wait wait, are we talking about the release song?...
<vagrantcish>civodul: trying to!
<civodul>ah ha!
<rekado_>civodul: here’s a draft: https://elephly.net/downies/ode-to-one-two-oh.{txt,ogg}
*civodul listens
<rekado_>currently it’s just A B A B; I think it’s fine to just do A B A B A (fade out), so no more music needs to be written
<rekado_>(that would also be a palindrome, like the title)
<civodul>rekado_: ooooh, you rock!
<vagrantcish>maybe we can change the typoed package description into poetry instead of music
<civodul>damn it, you folks are awesome
<rekado_>vagrantcish: yes, it can be a whole package of free culture!
<rekado_>*packages (sic)
<civodul>the lyrics are wonderful
<civodul>LOVE IT
<rekado_>:)
<civodul>guix install guix:song
<raghavgururajan>rekado_: The music is nice! what is it?
<raghavgururajan>*for?
<civodul>the title says it all :-)
<rekado_>I probably won’t be able to change anything about it before the release (busy day tomorrow); but I can upload all files tomorrow morning for a dilligent remixer
<civodul>i'm running "make release" BTW, so everything should be ready within a few hours
<rekado_>brb
<civodul>rekado_: cool!
<raghavgururajan>ODE 1 2?
<civodul>do we have a "dilligent remixer" here?
<civodul>raghavgururajan: an ode to 1.2.0, the version that's to be released tomorrow
<raghavgururajan>Ah!!!!
<vagrantcish>ok, poem form: https://paste.debian.net/1173905/
*raghavgururajan puts that music on repeat mode
<nckx>raghavgururajan: OK, I thought it was just an additional Web interface to .po files for those who wanted such a thing, not a replacement.
<civodul>vagrantcish: this is... terrible? terrific? but so representative!
<rekado_>vagrantcish: excellent!
<lfam>Haha I love it
<rekado_>we could also do A B C B A, with C being a break for a recitation of the poem
<nckx>vagrantcish: Damn it that's actually good.
<nckx>In a bad way.
<civodul>cwaziness, can't wait for the final cut
<vagrantcish>i give a shoutout to lintian.debian.org
<maav>rekado_: wow :)
<lfam>And great job on the music rekado. Nice tune :)
<rekado_>thanks, y’all!
<civodul>maav: any experience translating poems? :-)
<maav>thank *you* :)
<maav>hehehe, not many
<maav>some song
<civodul>this one is kinda hard to translate, because of the idioms
<nckx>Smooth rekado.
<vagrantcish>how do you properly translate a typo?
<civodul>yeah, tricky!
<vagrantcish>you have to mangle it in the spirit of another language?
<civodul>how would you translate "allows to" also
<civodul>and then, the translation would have to be approve by the poet
<civodul>but what if the poet doesn't speak that language?
<vagrantcish>civodul: so, basically, you can do whatever you want now because "my release is compiling" ?
<maav>the translator is a traitor say italian people, you cannot really translate but you create another interpretation 0:)
<civodul>vagrantcish: yes, you guessed it! :-)
<civodul>ah ha!
<vagrantcish>i admittedly changed the word order in the latest version
<civodul>the artist's touch
<lfam>It makes me happy to see all these other forms of expression around Guix :)
<civodul>interestingly, "libray" sounds like French "libérez", meaning "liberate" (imperative)
<civodul>coincidence?
<civodul>i doubt it
<vagrantcish>oh, could do a translation into one language, and then have someone translate it back
<vagrantcish>ideally clean-room
<civodul>clean-room translation
<jonsger>vagrantcish: isn't it then reverse engineering ^^
<vagrantcish>jonsger: that would be running diffoscope on versions with and without the translations embedded into the relevent .go files
<jonsger>hm, libreoffice doesn't feel stable on wayland...
<vagrantcish>wow, i think the number of packages in guix has grown by over 50% since i started...
<phil>hello all, I accidentally wiped my .bashrc :/. I was wondering where I could find the default .bashrc that comes with a fresh installation of guixSD
<civodul>phil: it's in /etc/skel/
<phil>@civodul thanks
<jlicht>samplet: sure, next week. And it's mostly your work ;-)