<rekado_>after waiting for over 2 months I got a replacement for my broken librem laptop today … but it doesn’t boot from my hard disk <dadinn>I use it specifically for Debian bootstrapping, it creates the root filesystem, and is quit configurable... got bored doing Arch-wiki manually essentiall <rekado_>this model is different in that it has a TPM <rekado_>I can boot our installer from USB just fine <rekado_>but when I select any system generation in GRUB it just freezes <rekado_>do any librem+tpm users here have an idea what might be wrong here? <vagrantc>i though the purism machines did a detection but not forcing anything... <sneek>Welcome back vagrantc, you have 1 message! <dadinn>the poing about the init-instroot script, is I have added a new feature yesterday which allows automatically installing executable dependencies, whith an extensible way to configure it for multiple distros (not just Debian) <dadinn>So I am thinking I will adapt it to Guix, Arch, and redhat... wherever a bootstrap installation method makes sense <vagrantc>klys: don't see what the issue is... looks like you built successfully? <klys>vagrantc, I have a couple pastes for you then <vagrantc>klys: could you use the recommended paste.debian.net instead? <vagrantc>that first just looks like a problem building rust... <vagrantc>which happens ... unfortunately very often <klys>I was told there's no way to get icecat from guix on 32-bit intel. <klys>well, I'm running a 32-bit kernel <klys>I'd rather not use 64-bit software here <vagrantc>unfortunately, some things are just not working :/ <vagrantc>sounds like you're saying it is a 64-bit capable machine ... why run 32-bit? <klys>that's nobody's business <vagrantc>fwiw, i'm hardly skilled with guix ... i mostly use and develop on debian and guix has been an interesting experiment ... so i might not be the best person to ask guix related questions <klys>besides I only have 3 GiB ram <rekado_>can/should we cut the bootstrap path short on i686-linux? <dadinn>hmm, is LVM still not supported on Guix? <dadinn>vagrantcish: I can't find an lvm2 package :/ <vagrantcish>i have: lvm2 2.03.09 out gnu/packages/linux.scm:3290:2 <nckx>The variable is also lvm2. <dadinn>maybe i need to do some update? How do I do that with guix? :P <rekado_>dadinn: also be sure to take a look at the manual <vagrantc>but it's been in guix quite a while .... <nckx>vagrantc: Eh, scratch that. That's my own LVM crap^Wstuff. <vagrantc>i've been using it at least since august of 2018 <dadinn>rekado_: you mean boot off LVM not supported is because it is not supported in the system config? <nckx>Tried to get booting to work but turns out that would better be done by someone who's actually used it. <nckx>dadinn: Yes. And there's no underlying code to mount it etc. <vagrantc>it pretty much just needs "lvm vgchange -ay" from the initrd and a few changes in grub... <rekado_>it’s probably not *that* difficult; requires some initrd changes etc <dadinn>nckx: I can see that there is a ZFS package now... I am assuming booting from it is also problematic <vagrantc>they're pretty much just like any other block device once initialized <vagrantc>e.g. there's no running daemon or anything <dadinn>vagrantcish: isn't lvmetad some sort of daemon? <nckx>dadinn: No booting from ZFS either, indeed. <vagrantc>i've been running lvm on debian machines for many years ... no daemons running here <dadinn>As mentioned earlier I have these bootstrap scripts for Debian, and now I am trying to migrate them to Arch, and Guix... I have spent years on getting this installation process for LUKS, LVM, ZFS, UEFI, etc running spotless for Debian, so it might be relevant. <dadinn>vagrantc: you mean the lvmetad? I know it can be disabled, because it's often a PITA <vagrantc>dadinn: seems to not be used in a default debian isntallation, which suggestions to me it's optional <dadinn>I am getting an error with `guix pull`, it says "no matching pattern" <nckx>dadinn: lvmetad isn't enabled by default. <nckx>dadinn: It says a lot more than that. Could you paste(.debian.net) the rest? <dadinn>nckx: I am in a VM in the live environment, possibly the disk went full, after a failed try to install zfs kernel modules :/ <nckx>dadinn: Reporting that unhelpful error would still be helpful. A match error is never the right thing to throw. Unfortunately there are a lot of those left in Guix… <linka>hey. there's a process that doesn't want to be terminated that prevents computer shutdown. do you know how i could debug that? <nckx>Hi linka. Who says so? What does it say? <dadinn>nckx: Oh, sorry, I have rebooted now the pull seems to work... all I remember is it was saying "no matching patter" #f, can't recall what was in the stack-trace :( <linka>sorry, i don't have access to that screen now. i'll try later <dadinn>I have tried to install the lvm2 package, it seemed to work, but I still can't find the pvcreate,vgcreate,lvcreate executables... why is that? <dadinn>ah, there is an lvm2-static package... what is the reason for them being separate? <dadinn>also, whenever I call guix, I see a gulie: warning: failed to install locale ***xMopx is now known as XmopX
***catonano_ is now known as catonano
<PotentialUser-21>I get the "corrupt input while restoring archive from socket" error when I try to install/upgrade anything. <apteryx>mbakke: perhaps benign, but there are lots of directories under Ruby's ext/ sources. <luhux>Why does an error in some services make shepherd also error? <luhux>I am trying to define a game server service, but something went wrong, causing all other services of the container to fail to start <apteryx>luhux: are these user services or system services? <apteryx>OK. Is it easy to reproduce? I also wouldn't expect the whole shepherd to go down. <brown121407>Hi Guix! If I need to package a Rust crate and we already have let's say version 0.3.4 and I need 0.3.5 specifically would it be ok to name the new variable rust-<crate-name>-0.3.5? <brown121407>Also, if I have pulseaudio-service-type in my config (and the pulseaudio package installed) shouldn't that make herd status ouput something about pulseaudio running? ***nikita_ is now known as nikita`
<NieDzejkob>brown121407: why do you need 0.3.5 specifically? they should be compatible. You could update the 0.3 package if it's just too old... <NieDzejkob>dadinn: lvm2-static is for use in the initrd. the lvm2 package does provide the binaries. Do you have ~/.guix-profile/sbin in your path? Maybe you need to reload the profile? (run the commands that guix printed out or start a new login shell) <brown121407>NieDzejkob: I need 0.3.5 because I want to package something that depends on exactly on that version. I would update our 0.3 but I don't want to risk to break other packages that may depend on 0.3.4. Is there any way to check if updating it would break other packages? <brown121407>NieDzejkob: The "directories" crate (3.0.1), which needs "dirs-sys" at 0.3.5 <NieDzejkob>oh, it doesn't need exactly 0.3.5, just 0.3.5 or higher <NieDzejkob>the easiest way to test a bumped package is to build its dependents <brown121407>Is there a flag to build all the dependents or should those be gathered manually? I'm looking through the manual but I can't find such a thing <NieDzejkob>roptat: java-hamcrest-library fails to build, could it be related to the maven-build-system you pushed? <NieDzejkob>brown121407: for rust packages, grep... you don't need to bother with transitive ones, so rust-dirs suffices. I'll push a bump <roptat>NieDzejkob: hamcrest has a weird race condition or something, so the rebuild of the world might trigger it. My build farm had no issues with it for instance <roptat>Oh the java world, which is not so big <roptat>but ci doesn't seem to have any hamcrest-library build even scheduled for some reason... <NieDzejkob>the search is wonky. It sometimes shows up as failed for me <roptat>NieDzejkob, anyway, running multiple times, I sometimes experience a failure on my end too <roptat>I might have a solution, running with --rounds=15 to be sure <PotentialUser-30>hello there! I've got an issue with the GNU Guix TUI installer, it returns to the beginning of the installer when I select either options manual, guided, or guided encrypted, at the partitioning phase, and I can't see any error log, nor find an error log, do you know how I can debug this issue? Thank you! <greenrd>I have again tried to email a bug report to bug-guix@gnu.org and again my bug report does not appear in the web interface, and nor do I receive a bug acknowledgement email <greenrd>I have set up SPF in my DNS for my domain, so hopefully the third-party mailserver I am using for my domain shouldn't look too spammy <nckx>greenrd: It can take a while. When was this? <nckx>Unless it's been several hours I wouldn't worry yet. <nckx>You're not stuck in moderation, at least. <greenrd>I guess it was the SPF that did the trick then <greenrd>I should have done that years ago, thank you for making me do it! :D <brown121407>I set up my packages from a manifest. If I do `guix package --delete-generations` and `guix gc`, why does guix need to redownload all the stuff when I use `guix package -m manifest`? <brown121407>This is pretty annoying considering that it means that every time I run the garbage collector I have to redownload the whole 2.61GiB of texlive-texmf again. <roptat>the most likely reason is grafts <roptat>guix needs the ungrafted package to know if it needs to apply grafts, but the garbage collector will only keep the grafted version <roptat>so the garbage collector removed the ungrafted texlive, then guix needs to download it before it can determine that it already had the grafted version and uses it <NieDzejkob>brown121407: thanks for the rust-directories patch! If I may ask, what's the motivation behind that particular package? Are you planning to submit a non-library package that uses it? <brown121407>NieDzejkob: Yeah, I was thinking to try to package topgrade <NieDzejkob>brown121407: Ah, nice! I was mildly confused since usually these Rust packages show up as entire patchstacks ;) BTW, you might be interested in the improved importer in bug#38408 <apteryx>mbakke: hello! Is cherry picking a couple commits from core-updates back to master alright? <NieDzejkob>apteryx: if they don't cause a rebuild themselves, I think the only concern is the merge <alextee[m]>er, i'm trying to update zrythm but i just noticed we patch a file. i added a trademark policy that says altered versions need to change the name and logo. im not sure if this counts as an altered version <alextee[m]>should i rename the package and mark this as deprecated or something? <NieDzejkob>alextee[m]: what's your goal with the trademark policy? <alextee[m]>NieDzejkob: to protect the identity of the project/product <NieDzejkob>do you *want* distros to all create a separate, free trademark because they need to patch the xdg-open path? <alextee[m]>obviously i dont mind with minor patches like distros do but i don't think there's a legally viable statement i can make that allows reasonable modifications <roptat>alextee[m], if you agree with the content of the patch, you could also give guix a special permission :) <roptat>(you're the trademark owner, right?) <alextee[m]>roptat: i'd me more than happy to, if that's acceptable <brown121407>NieDzejkob: I indended to make a patch series but I didn't know when I would get the time to get every package done so that's why it ended up like that. Thanks, 38408 looks nice. <NieDzejkob>the issue with that is then Guix is no longer free software, since if you modified the zrythm package definition, you couldn't redistribute it <alextee[m]>right, unless you modified it so that it doesn't include the trademarks, which is trivial to do <alextee[m]>i added a build flag to change the name that appears in the UI, and there's 4 image files to be replaced <NieDzejkob>trivial? I'd say it requires significant creative effort to design the new logos <alextee[m]>i think i should start shipping "alternative icons" with the tarball in future releases <NieDzejkob>I feel like your trademark policy is achieving the precise inverse of its goals <mbakke>apteryx: sure, as long as they con't cause a huge rebuild like NieDzejkob mentions (are there non-rebuilding commits on c-u?) <nckx>It's still free software, but it's a pain to package and maintain, so we'll change the name so's not to be forced to ask upstream's permission for every patch, so fewer users will find & use it. Counterproductive at best. <mbakke>apteryx: I like to use the '-x' flag when cherry-picking between branches to provide a hint in case it causes a merge conflict <NieDzejkob>alextee[m]: what situation are you worried about, exactly, that you're trying to mitigate with this policy? <alextee[m]>idunno, i mean this policy is even implied in some countries <alextee[m]>anyway, i don't think i'll change it unless there's some magic legal speak that allows reasonable modificiations <alextee[m]>im fine with changing the name/logo in a patch, it's still recognizable <nckx>alextee[m]: Which ones? I'm not aware of *any* country where a reasonable effort to provide the original product (which is freely distributable) can be shut down through trademark law. <nckx>And no, patching something it to work on $distro is not a meaningful modification. <alextee[m]>US i think. if you sell something under the name and logo, you're the trademark owner. unless you add a license that can probably void the trademark like a CC license on the logo, then you can make other parties stop using it because its your t rademark <alextee[m]>wordpress says something along the lines of "it's ok as long as you do it in a way to promote wordpress", but it still requires permission from what i understood <NieDzejkob>that's the basics of how trademarks work. I don't think that's how it interacts with freely-licensed software, though. What's your source here? <alextee[m]>anyway, does it matter? i don't think i'll change it. it looks reasonable to me and many others i asked. <NieDzejkob>Did the people you asked have background in software licensing and trademark law? <nckx>alextee[m]: I won't comment on Wordpress since I'm not familiar with it. But you're mistaken if you think your trademark claim means you own the brand or can prohibit a patched version of your software to use the name ZRythm. You can't. <nckx>You can do so if the patched version is detrimental to your brand; not merely because it is patched. <NieDzejkob>If there were no negative consequences in adopting the trademark policy, I'd probably do so myself for my software. But there *are* negative consequences: distros have to do extra work to provide your software, and won't give you any advertising, since users will be getting your software under a different trademark. In this situation, having this policy to prevent the situation of "idunno" is at best irrational. <alextee[m]>then i can just update the package and it should be fine. i asked a debian packager and they told me that it would be ok if they patched it after fetching the original source or something like that <alextee[m]>im not 100% clear on what counts as an altered version and i dont want to define it <alextee[m]>that's why im asking about what i should do the package :P <alextee[m]>NieDzejkob: i doubt most intended users learn about a DAW from their distro <NieDzejkob>A vague policy harms both sides. You get uncertainty as to whether you get to enforce it, the users get uncertainty as to whether what they're doing is actually legal. <NieDzejkob>they could google for free software DAWs, and then search in their distro to check whether each is packaged. They'll try the ones that are available first <alextee[m]>still covers this "if the patched version is detrimental to your brand" <NieDzejkob>if we rename the package, they will try to make do with audacity or whatever first <alextee[m]>NieDzejkob: but i sell distro packages so i have commercial interest in the ones i provide <alextee[m]>there's debian 10, ubuntu 18/20, archlinux and fedora 32 atm <alextee[m]>which covers most cases, so it's not a problem of finding a distro that supports it <alextee[m]>and 100% free software distro users like on guix and parabola can easily tell where their software comes from , even under a different name <alextee[m]>anyway, please let me know if you have a suggestion of what i should do with the package. i don't mind leaving the same name with the patch or changing it <NieDzejkob>That's a curious business model. What if I started providing my own packages of your software, for debian, ubuntu and fedora? <alextee[m]>NieDzejkob: you can do that unless you change it (to point to your website for example) <hendursaga>The latest blog post is pretty cool! I always wonder why I miss hearing about such amazing, powerful software like Ganeti. <NieDzejkob>alextee[m]: Hmm. What if I published the ones you are offering for a price? <alextee[m]>NieDzejkob: you can still do that. the trademark policy explicitly allows you to redistribute unaltered versions <alextee[m]>there's a start window that tells you about the copyright/trademark owners, there's links to the zrythm website all over, and bug reports point to the main bug tracker <alextee[m]>it would even benefit the project if you do that <alextee[m]>besides the sales are mostly for supporting full time development, not for profit <alextee[m]>some people might not care, but many of the people who throw money there (and at ardour) are doing so to support development <NieDzejkob>alextee[m]: I don't think you need to encode what you want into legalese. You could just say that "changes adjusting Zrythm to specific environments and/or operating systems are allowed", for example <alextee[m]>NieDzejkob: hmm that doesn't sound bad. i'm trying to think if that statement can be abused somehow (put yourself in the feet of some company that wants to exploit a project for profit) <alextee[m]>maybe add a "however any other changes including but not limited to functionalities, messages presented to users and hyperlinks are not allowed" <NieDzejkob>why do you want to prevent adding functionalities? This makes any github fork opened to PR a feature illegal <NieDzejkob>companies couldn't keep the code to themselves anyway because it's AGPL3, so... <NieDzejkob>oh. perhaps "specific software environments" would work better <alextee[m]>NieDzejkob: this is not about the software code, it's about the wordmark and the logo <alextee[m]>i think it's best to borrow from existing trademark policies, approved by lawyers instead of attempting to add that wording <alextee[m]>that's why i wish there were other examples out there, or if the FSF made an article about what you should and not do when using trademarks in software <alextee[m]>i took most of the wording in that policy from other ones, like debian's and mozilla's <alextee[m]>hmm that patch is not really needed. i think we only added it to avoid adding xdg-utils as a propagated-input <Bryophyllum>\o Question: Is Guix System capable of restarting services under certain circumstances on system reconfigure at the moment? If not, how do I know, which service requires a restart? <Blukunfando>They’re listed in the ‘guix system’ output. If you don’t see any mention that any particular service couldn’t be restarted, you’re done. <Bryophyllum>Blukunfando: Do you mean the `guix system reconfigure` output below the line of text that says that the bootloader was installed? <alextee[m]>is it possible to install packages from a file rather than from channels? <apteryx>alextee[m]: sure, 'guix package --install-from-file=your-package.scm' ***dadinn is now known as dadidd_
<rekado>just make sure that file evaluates to a package value <roptat>alextee[m], also if you have a channel but don't want to guix pull it, you can use -L /path/to/channel/checkout <alextee[m]>yeah that's what i use mostly when testing packages <nckx>alextee[m]: Have you talked to any qualified legal experts also familiar with Free software? It's a pretty unusual niche (you own some part of the name, but can't influence supply of the underlying product). <nckx>Just out of curiosity, because I find this interesting, nothing more. <alextee[m]>nckx: i would but i dont really have the funds to <alextee[m]>and last time i asked licensing@fsf.org they said i had to pay a large amount <nckx>alextee[m]: Eh? Really? For advice? <nckx>Pro bono sure has a different meaning over here :-/ <alextee[m]>yeah, idk, maybe they changed the way they do things <alextee[m]>i talked with a friend of my sister's who works with copyright and trademarks in general and they were very confused about me trying to use a CC-BY-SA license on a trademarked logo <alextee[m]>which suggest not to do that and use a trademark policy instead <alextee[m]>this part: Allowing anyone to reuse or modify your logo or trademark as a matter of copyright could result in your inability to limit use of your logo or trademark selectively to accomplish those purposes. Applying a CC license to your trademarks and logos could even result in a loss of your trademark rights altogether. <blackbeard[m]>alextee: hi! Trademarks are a very specific area, basically the law is designed to punish you if you use them. <alextee[m]>blackbeard: yeah, it looks like giving explicit permission for specific things is the only way to go about it, otherwise you risk trademarks becoming diluted <nckx>Even then, you're expected to actively defend the trademark. Don't underestimate that. It can be a full-time job (some free-software/open-source maintainer said it basically was, for them, but I regret to say I forgot whom). <blackbeard[m]>Because you may have a trademark for a specific region and activity <blackbeard[m]>And I may have a similar trademark for a region that you are not registered in <alextee[m]><nckx "Even then, you're expected to ac"> i want to make zrythm my full time job anyway and spend many years on it (one of the reasons i want a registered trademark) so i'm prepared for that <nckx>But I agree with blackbeard[m]: this is a very specific legal area, much more than copyright, which is relatively uniform. And we know how simple the latter is… 🙂 <nckx>alextee[m]: OK, that's valid (and cool!), as long as you're aware. <blackbeard[m]>Basically it is up to you to say. "That person is using my trademark unlawfully " <alextee[m]>nckx: who knows, maybe in the future there could be a zrythm foundation! <alextee[m]>having a very obvious notice seems like a good way to deter most cases of infringment though <blackbeard[m]>Let me see if they have some guide that may be useful for you <alextee[m]>blackbeard: re: the regions, there's the madrid agreement which can extend to multiple countries too <alextee[m]>i live in the UK, and i've seen an application for that, but it says it requires you to have a registered mark in the UK first <PotentialUser-30>hello! it seems that without running `fc-cache -rfv` it causes severe font breakage across several applications especially IceCat - could GNU Guix run that automatically maybe? <alextee[m]>that's cool! are you planning to focus on free software? blackbeard: <alextee[m]>you should make a business out of it. i'd pay you for advice <nckx>blackbeard[m]: How'd your thesis go? <NieDzejkob>PotentialUser-30: that's a long-standing bug, #26877 <PotentialUser-30>Also it would be interesting if GNU Guix's gnu-build-system could make use of ccache by default <PotentialUser-30>but I realize build outputs are already big enough you probably can't also store intermediate build artifacts, though, with file system de-duplication it could help that <blackbeard[m]>I was reading a paper on the first million of websites by Alexa rank <blackbeard[m]>It is so bizarre to me that the data protection laws are not designed to stop something like that <PotentialUser-30>European ones are but in practice many people violate it blatantly and regulators don't have infinite time <blackbeard[m]>Basically data protection laws say you can't "transfer " data without Owner consent but you can do a "remision" (Spanish ) <blackbeard[m]>So many companies can just say this is not a transfer, it is a remision and that's legal <PotentialUser-30>GDPR disallows dark patterns in data choice popups, consent must be explicit, etc.. <PotentialUser-30>Google or Facebook makes so many different switches to opt-out when all people want is just no tracking at all with a single no-nonsense switch <pkill9>PotentialUser-30: i was thinking that, have it store parts of ocmpilation <pkill9>didn't know there is a name for it though (ccache?) <blackbeard[m]>Some people do, they want ads that they are interested in. Hey some people even enjoy ads <pkill9>i think it would be good to be something you enable <nckx>pkill9: ccache is an implementation. It caches (g)cc calls: if you symlink gcc to ccache, then call gcc twice with the same arguments/files/environment, it will return a cached object file. <PotentialUser-30>pkill9: it's a compiler wrapper that can store compiled objects by hash, it's a more generic way to deal with caching than use any build system's specific thing <PotentialUser-30>Nobody explicitly asks for ads, they may not complain if they are forced onto them but they wont ask <nckx>What's certain is that plenty of people prefer targeted ads over ‘irrelevant’ ones. <pkill9>it's like how they claim to be offering 'personalised ads' as if people were asking for htem in the first place <nckx>Many people are. Just not the vast majority. <nckx>It's a false choice, but that's the honest answer you'll get in the current reality. I know many people who opt in, let alone wouldn't opt out. <nckx>Bah, now I'm hopeless too 😛 <nckx>blackbeard[m]: Does your thesis explore their perspective by any chance? <lispmacs>i was trying to figuring out if I could package Arduino IDE or not. I don't know anything about ant but probably could figure it out. However, in the README it says that the build processes downloads and compiles associated arduino libraries <nckx>I find them fascinating if incomprehensible. <lispmacs>will that somehow conflict the guix package or security model? <NieDzejkob>lispmacs: you'll have to configure/patch the build process to use sources provided by guix <lispmacs>NieDzejkob: I'm guess that is the reason nobody else has packaged it yet <nckx>lispmacs: Worse (=better): It's downright impossible to download anything during the build process. <nckx>Not what I'd expect from a legal thesis but good. <nckx>blackbeard[m]: It wasn't meant as criticism at all! I'm sorry if it sounded that way. <NieDzejkob>man, why must the apostrophe be so close to the enter key :< <nckx>Yeah, civodul's just trolling our muscle memory at this point. <LeonLainDelysid>I have installed Guix on a raspberrypi but its processor is too slow to build the sources itself. I was told before that there is an option to tell guix to ask another more powerful machine on the network to run the build jobs itself. Does anyone know how I can do that? *nckx awaits ‘guix git pull’. <NieDzejkob>LeonLainDelysid: are you referring to offloading? <NieDzejkob>nckx: hey, that could actually be useful! atomically pull and authenticate... <nckx>LeonLainDelysid: It's a bit involved to set up but a wonderful feature (my laptop hasn't built a package in years). You can even set up an x86_64 box to build for arm. <nckx>That's a separate step but equally almost-magical. <LeonLainDelysid>That would be the idea. Because my building machine runs in 64bit but the raspberrypi is on arm <nckx>LeonLainDelysid: 64bit ARM (aarch64) or 64bit x86? <nckx>I think aarch64 can just build and run armhf binaries natively. <LeonLainDelysid>It would be my normal everyday laptop or desktop which are much more powerful and build the guix binaries in the blink of an eye <brettgilio>Hey all, I sent my new GPG key to replace the one I lost control of awhile back. Still waiting for confirmation to resume contributing. Any word? <nckx>brettgilio: Hi! You're welcome to contribute through guix-patches@, of course, and I can add your key to the keyring branch. *nckx reading the mails so far. <brettgilio>Thanks nckx, and happy to be back to help guix continue to be the best system ever! :) <pkill9>what do people think of wraping the share directory? <pkill9>I want to add the share directory for inputs that need to be propagated + installed into the main profile so that things like dbus find it, since they typically only search for the share directories specified when the user is logged in <pkill9>which package provides qml and qt quick? <pkill9>it's easier than i thought to make packages, i keep thinking htere's a huge amount of work to do to get it compiling on guix, but it's always quicker htan i previously thought <pkill9>i keep thinking some esoteric thing is going to go wrong <vagrantc>hrm. new test failure trying to update diffoscope *vagrantc doesn't understand how to debug pytest stuff inside guix-daemon <mbakke>vagrantc: pytest does not print any test output by default, you can use 'pytest -vv' for verbose execution ***XmopX is now known as xMOPX