<alextee[m]>nckx: iirc "[" and "]" don't mean anything in some regex parsers, maybe for python or something, i dont remember <nckx>Maybe you're thinking of ‘(’ (‘extended regex’). [ ] is universal (although how some implementations handle ‘]’ on its own might differ). <nckx>Guile regexps are similar (identical?) to the egrep syntax. <nckx>jackhill: We don't get enough questions about the ‘regular’ syntax already? 😉 <nckx>I totally understand why some languages make regular expressions their own syntactic thing, not strings. <roptat>mh... I'm building the kernel on my arm board, and it doesn't work... it's the second time that I see <defunct> processes owned by a guixbuild user. its parent is make, and it takes one core for nothing. Last time, two processes were in that case and the build had frozen because of it <vagrantc>finally pushed pieces of linux-libre 5.6.x <nckx>…mhw just pushed 5.4.30… <nckx>Strange that he's not answering you. <vagrantc>well, i have effectively changed nothing other than adding linux-libre-arm*-5.6 ... so figured i'd push it <vagrantc>but at least i won't have to constantly rebase every time a new old kernel version gets updated now... *vagrantc rebased 5.5.9, 5.5.10, 5.5.13, 5.6, 5.6.1 and 5.6.2 ... and they're not cheap things to rebuild <raghavgururajan>FOLKS! If you are using Guix System, could you please try install (not just build) linphoneqt from the patch (14-add-linphoneqt-v6) at https://issues.guix.gnu.org/issue/40264 and run the application from terminal to see if you get "Segmentation fault (core dumped)" error? <raghavgururajan>This it to see if the error is reproducible on Guix System, as opposed to foreign distro. <raghavgururajan>nckx Regarding the seg fault, I fixed half of the errors. I have one left. "Core:linphone: also error in pcm_hw.c:1715 - open '/dev/snd/pcmC0D0c' failed (-77)". Can you guide me on what is the nature of this error? Thanks! <nckx>raghavgururajan: What does the ‘groups’ command say? <nckx>(Also, I presume it's an ‘alsa error’… 😉) <nckx>If that command already includes ‘audio’, my wild guess is that *something* is already using /dev/snd/pcmC0D0 and won't let Linphone access it directly. *nckx coughs into elbow and points at PulseAudio. <nckx>I'm not enough of a desktoppy person to know what to do in that case. <nckx>This is a Guix System, rite? <nckx>What does ´ls -l /dev/snd/pcmC0D0c’ say? *raghavgururajan has not got their device back from Technoethical <nckx>I've still not reinstalled Guix on this (Guix System) machine. I'm waiting for myself to finish bcachefs support. <nckx>OK, add yourself to the ‘audio’ group and re-login or whatever's require nowadays. <nckx>Speaking of nowadays, I suspect this should be handled through PolKit somehow, so it might be fruitless. <nckx>That's a very big hammer to bust out at this point but you do you. <nckx>It probably won't break anything but watch out for .config/linphone/foos suddenly owned by root and giving weird errors later. <nckx>I prefer my group suggestion. <raghavgururajan>nckx Btw, this error is platform specific right? My package definition is still good? <nckx>raghavgururajan: Correct. *raghavgururajan is afk to eat dinner <nckx>raghavgururajan: It's *possible* that Linphone is falling back to ALSA because (random example) it's not build with Pulse support because something's missing at build time. But that would be a separate thing. <nckx>Right. I don't know what Ubuntu uses but it's a bit strange that it's trying the ‘older’ ALSA driver first. Still, we should make sure that it works on systems without PA. This isn't wasted effort. *nckx AFK-ish; please ping only on urgent Guixy business. <apteryx>raghavgururajan: I'll try your latest set of patches now <usney>I am using lxde now I switched from gnome. None of my installed guix apps are listed in the menu in lxde. <raghavgururajan>sneek, later tell nckx: The groups suggestion worked. Thank you! But I still get segmentation fault though. At least now with no other errors. <apteryx>Can't apply 12-add-mediastreamer-v2 on master. I'll try to do it by hand. <apteryx>raghavgururajan: really? :-( sorry about this. I thought I had it building here. Let me try again. <raghavgururajan>apteryx No worries! As a backup, I sent another version of patch to spandsp bug thread. <apteryx>raghavgururajan: this just has to do with the tests; I thought I had fixed that by disabling the parallel tests. Seems not. <apteryx>I built it 4 times successfully here with parallel-build? #f. Could you try? <apteryx>mediastreamer, belle-sip, liblinphone and linphone-dekstop? <apteryx>in your patch series it's the other way <raghavgururajan>apteryx Oh yes. I got confused when you messaged "Can't apply 12-add-mediastreamer-v2 on master. I'll try to do it by hand.". :-P <apteryx>I was confused too, I think, that's why it had failed :-D. <brendyyn>i tried copying a patch from the mailing list and its all corrupt with wrapped lines. im trying to fix it up manually <apteryx>raghavgururajan: I'm off to bed. Will give you more feedback tomorrow. *raghavgururajan goes to bed to too <bavier`>I wanna do some guix hacking this weekend; too many other distractions this week <lfam>It's been busy lately. I think that a lot of people are filling their time with Guix <lfam>Apparently but I haven't been paying attention this cycle <lfam>You know that pushing for a release makes bugs appears ;) <vagrantc>hrm. apparently linux-libre@5.6.2 does exist ... with no configuration files or anything. oops. <brendyyn>vagrantc: the patches are on the mailing list <brendyyn>i tried applying it but the patch is corrupt with wrapped lines and stuff. i wasnt able to completely fix it <vagrantc>yes, but i pushed 5.6.2 to master, *thikning* it wouldn't build linux-libre, just the new linux-libre-arm*-generic-5.6 packages <vagrantc>but it also built linux-libre@5.6.2 for x86_64, aarch64, armhf and i686 <vagrantc>brendyyn: sounds like a problem with your mail client <brendyyn>vagrantc: i copied it out of the web interface <brendyyn>how else am i supposed to get it with without subscribing? <brendyyn>i tried viewing the html source in icecat but that was the same too ***sripathroy is now known as nagamalli
<brendyyn>vagrantc: i dont really understand how im supposed to get get the patch properly then? <brendyyn>looks like you'll have to change your password... <vagrantc>on some web archives there's a way to get an mbox file, that you can then pass to "git am" <brendyyn>alright. i find that interface really bad. I type in linux 5.6, and it shows all these unrelated bugs from 2013... <vagrantc>guess someone also wants wireguard and exfat enabled.. <nagamalli>I Tried two updates to understand updation and adding packages. Hence i tried two simple packages. <nagamalli>brendyyn: I had to login to someother PC and messed up <Veera>bavier`: I had chat with civodul about deadline nearing for Outreachy Contribution. <brendyyn>ok it applied and im building the kernel now <Veera>bavier`: he sent a irc chat msg to you yesterday about my xplanet patch review <bavier`>Veera: oh, whoops, didn't get that, house had a power outage <bavier`>Veera: great, I'll check your latest <bavier`>didn't realize it was part of Outreachy ***sripathroy is now known as nagamalli
<raingloom>oh damn. so, i patched guix/scripts/system.scm annnd it looks like guix is now trying to bootstrap everything from zero??? <raingloom>intensional store would come in handy right about now ;_; <brendyyn>im on linux 5.6 now. it fixed my external webcam issue. <brendyyn>i was wondering, how can i make my laptop fan speed more consistant? i find when im at 20% cpu usage it is very inconsistant, turning on and off all the time instead of remaining steady. <smithras>brendyyn: I remember reading that for lower cpu usage, disabling hyperthreading can prevent some of the fan and power spikes, but I don't remember where I read that and I don't have any personal experience with it <brendyyn>hyperthreading seems like a good feature though. although linux kindly notifies me my hyperthreading implies some side channel vulnerabilities ;/ <brendyyn>sensors shows my cpu sitting just under 60C most of the time. I suspect that is a cut off limit for changing fan speed <bricewge>wnereiz: Look like it come from python-pep8, it has been months since it had some changes, the patch is still in the repo and I can build it. <brendyyn>smithras: i just thought good fan control was something linux should do automatically. <bricewge>wnereiz: Maybe try cleaning your guix cache, but I'm really not sure it will help. <brendyyn>sulr70: does it still warn you about guix pull even after you ran it? <sulr70>brendyyn: There is no warning after `guix pull` <Veera>brendyyn: you may set a suitable cpufreq governor for kernel <Veera>brendyyn: powersave or conservative <Veera>brendyyn: it keeps cpufreq as low as possible; if the setting is performance cpu would change frequency frequently leading to ups and downs in cpu temp <Veera>brendyyn: Though I have heard some MB's are at fault. <brendyyn>the cpu is in use since im video streaming <Veera>brendyyn: controlling fan speed is not good idea. as it changes as per cpu temp <Veera>For newer Intels there is now pstate cpu frequency controller. <Veera>See dmesg output and you can set pstate to a fixed % of use. <Veera>min and max cpu usage can be set for pstate <brendyyn>that sounds like it is completely avoiding the problem of fan control <smithras>I think what Veera was saying is that directly controlling the fan speed is dangerous without also controlling the cpufreq <Veera>Since most of time it is dormant I set to min 1 max 1 <brendyyn>i wasnt planning to directly control anything. i found a tool called nbfc but its not in guix <Veera>smithras: yep; if somehow fan speed is reduced and cpu temp increases; linux kernel automatically down throttles cpu freq <Veera>I have not tried reducinf fan speed though <Veera>say the there is unbalance, then cpu temp would reach max threshold for safe use; down throttling would happen frequently. <smithras>seems like nbfc should be fairly straightforward to package, guix already has the mono tools <Veera>It is better to keep cpu temp less as much as possible. for longer cpu life <brendyyn> available cpufreq governors: performance powersave <Veera>perhaps in kernel build option conservative was not enabled <Veera>may be newer ones don't provide conservative <brendyyn>Model name: Intel(R) Core(TM) i7-4710MQ CPU @ 2.50GHz <Veera>See I am not CPU HW guy with details of each CPU <Veera>features, compatibility with kernel <brendyyn>im actually curious what the minimum fan speed is maybe its just going between 0 and 1 and cant going any slower in the first place <Veera>I don't know of guix; but in other kernel pstate can be selectively enabled instead of cpufreq <brendyyn>sensors doesn't even show the fan speed in the first place <Veera>no cpu fan speeds are setted in rpms <Veera>I have not tried that myself. some os's allow it. <Veera>may be lmsensors package shows the setting <Veera>with pstate you can precisely set the cpu power required for your app and set the max accordingly <brendyyn>it seems the ondemand setting is working well <brendyyn>all i want is things to work normally by default <dhanvanthri>Hey guys, im trying to overwrite the filesystem created by rhe graphical installer (to reduce iter time for ebcryption), however even after partprobe and umointing everything, lsblk keeps howing the old table. please sned help ***Veera is now known as Veera_
***Veera_ is now known as Veera
<raingloom>umm, how do i execute a monadic action like this? (with-monad %store-monad (begin copier-to-greg )) <raingloom>where copier-to-greg is the result of (copy-closure) <raingloom>this is all in the context of guix/scripts/system.scm <raingloom>ah, looks like there is a (run-with-store store) ***apteryx_ is now known as apteryx
***Veera__ is now known as Veera
<str1ngs>guix-vits: what is your guix describe? <kmicu>Hi brendarn what laptop is that? An Ideapad by any chance? <kmicu>brendarn: thank you. Generally the fan control is in BIOS and accessible through ACPI. Changing governors has no direct impact on fan speed. <kmicu>(Ideapads have a kernel module for that. system76 engineers should know what to do, maybe you need to migrate some kernel patches from Pop! OS.) <brendarn>kmicu: oh ok ill have a look. i wish this stuff was just build in to linux *kmicu uses echo NUMBER > /sys/devices/pci0000:00/0000:00:1f.0/PNP0C09:00/VPC2004:00/fan_mode for VPC2004 ACPI driver ’cuz Linux kernel is a monolithic garbage fire. So deprecating support for it and moving to Hurd is welcome. 😺 <highvoltage>was that announcement just an april fools joke or what? I tried to install a package from the iso and just got the wishful thinking link <guix-vits>str1ngs: ./pre-inst-env guix describe: e8a1c5af8421910d80e6bf4ecc6440e9200c76b1 <kmicu>highvoltage: Linux kernel deprecation is a joke part; support for Hurd is for real. <str1ngs>guix-vits: is that guile-gcrypt issue resolved now? <janneke>okay, got first prompt with our hurd image! <rekado_>sneek: later tell brendyyn I’m sorry you find issues.guix.gnu.org “really bad”. Could you please send constructive criticism as a reply to the thread “Brainstorming features for issues.guix.gnu.org” on guix-devel? <janneke>but hey; civudul fixed the big problems, of course <janneke>civodul found that we are using the build system's mig, which is 64bit and generates 64bit stubs *rekado_ can’t connect to ci.guix.gnu.org <janneke>rene (see guix-devel) has helped with /libexec symlink (is that a hack or a fix?) <kmicu>Getting first Emacs with hurd image is the time I can migrate to Hurd as my daily driver. <rekado_>“Your system administrator has blocked your computer or device. Please contact the system administrator.” <rekado_>that’s when I visit the HTTP variants of our websites <rekado_>I think this is actually the MDC network blocking me <Franciman>i was reading that guix system is thinking to switch to gnu hurd <Franciman>instead of working on gnu hurd we could be working on plan 9 <rekado_>Franciman: are *you* working on plan 9? <rekado_>Veera: I just looked over your v3 of mtpaint. You seem to have missed a few of my questions. <rekado_>Veera: it would be good to add a comment as to why that commit is needed (because there are no official releases), and it would be good to try building with GTK 3 instead of GTK 2. <Franciman>rekado_, I am not currently, but I could. Also, plan 9 seems much more polished than Unix <rekado_>Franciman: you’re welcome to try porting Guix to Plan 9. <Franciman>I didn't mean to push anybody, I was just proposing <kmicu>I propose that others should work and port Guix to Redox OS (Rust) ’cuz I like it :) <Veera>rekado_: The developer says you can help testing with gtk3; it is not stable <rekado_>Veera: then please add this as a comment. <rekado_>Veera: preferrably with a link to the developer comment <rekado_>Veera: because otherwise we might be left wondering why gtk+-2 is used when it seems that gtk+ should be working just fine. <Veera>rekado_: did you read my just previous post to that patch; i have explained these points. <Veera>rekado_: I will add comments to definition <rekado_>Veera: I only see four messages in that thread: your three patches and my review of v2. <xelxebar>Instead of doing `guix install` for user-specific packages, I'd like to declaratively manage the user profile. <str1ngs>xelxebar: you can use a manifest file and pass -m to package <xelxebar>Just curious, but where do people generally keep their manifest files? <bricewge>guix-vits: It fails here too. I'm suspecting efa578ecaece67366b4b0e2266de7c2faaa4ae54 *guix-vits neither i'm :) <mbakke>hm, so it looks like inferiors are broken right now? <ATuin>hi everyone, is there any attempt of integrating gradle builds with guix? <ATuin>how are we supposed to deal with all the suff gradle is downloading? <ATuin>ahh i will take a look, thanks <ATuin>actually i'm using the "more" repo frm Julien locally :D <ATuin>trying to build kotlin which was failing <civodul>roptat: i'd encourage you to move compcert to a separate channel <apteryx>mbakke: eh, looks like my son passed here <roptat>civodul, you're right, it's probably the only propriatory software in there... I'll do that shortly <roptat>ATuin, most of my packages in guix-more are failing :/ <roptat>including kotlin (I was focusing on kotlin@0, which still fails, but has a very low count of error messages) <ATuin>anyway, lot's of things to read and learn so it's good to have it <ATuin>i managed to build kotlin manually from the source of a guix build that failed <ATuin>but ofc it donwloads lots of stuff in $HOME/.gradle <roptat>kotlin@0 is the last version that can be built without kotlin <ATuin>all the deps basically, guix should take care of those i guess since probably some of them are needed at runtime <roptat>gradle is not important, we can always build manually <roptat>yes, that's what I have in guix-more <roptat>at least, what I tried to have, but some of them are probably wrong <roptat>or not built correctly, because I was using the ant-build-system <ATuin>it was failing because it was missing build.xml <ATuin>i guess it can be autogerenerated somehow from the sources <roptat>the ant-build-system autogenerates a build.xml when you pass the #:jar-name argument <roptat>I'm not even sure kotlin@0 is the right target: even though it's the last that can be built without kotlin, I doubt it can build the newest kotlin directly. I'm not sure if there is a chain of kotlin versions we can find that eventually build the latest <roptat>if you can find out what's wrong with kotlin@0, it will be very helpful :) <ATuin>aha so your approach was to use kotlin@0 to build the rest <ATuin>i will try to build that one now then <roptat>civodul, (and others) question about my maven patches: I add a few files and change java-utils.scm, which means a rebuild of the java ecosystem. I also change or add ~200 packages. Do you think it makes sense to put them in separate files (currently in my work I have maven-bootstrap.scm and maven-plugins-bootstrap.scm). Separating them can probably help see which produce the fake maven repo, and which do not yet. Opinions? *rekado *still* builds guile-emacs <roptat>ATuin, yeah, more recent versions require an existing kotlin binary, so if we want to boostrap kotlin, we either need another compiler (we would need to write it because there's only one kotlin compiler) or build from older versions <ATuin>i see, looks like the easiest path is to build it using an old version <ATuin>i will read what kind of magic gradle does to build <roptat> the versions of dependencies I use to build are probably too recent, but I'm not sure what is missing exactly <roptat>but I don't remember precisely, because I haven't touched it in at least 4 or 5 months <bricewge>What is the reasoning behind putting error messages to /dev/console in Guix? <ATuin>mmm java-jdom-for-intellij is exported as java-jdom since it inherits that package right? so that means that when using the cli i need to use "java-jdom" <mbakke>derp, I must have messed up my bisect, it points to 7884fc434f5d4a11b1dd566553b83fae09c44fad as the first bad commit :P <brendos>seems i no longer have permission to edit network manager connections as my uesr <civodul>roptat: like rekado, probably separate files if possible <mbakke>what the ... now I can't reproduce the inferior problem any longer, perhaps because I'm already past the step that failed <civodul>roptat: but you can already email guix-devel to explain what you have if you want <civodul>bricewge: where do you see /dev/console being used? <mbakke>changing to a different commit makes the error visible again <civodul>mbakke: what problem are you hunting? <mbakke>it works on a commit from ~2 weeks ago, but not current master <mbakke>any use of inferiors in manifest will trigger the problem <mbakke>looks like the problem is with cloning the channel <mbakke>i'm using both --no-grafts and -n in my test <mbakke>no wait, in the message on help-guix it fails on compute-guix-derivation <mbakke>I've probably messed up my bisect again, stopped it too early :P <civodul>mbakke: yeah it has to do with build handlers, with-store, dynamic-wind, etc. <brown121407>hi! do you know how I can get man pages for things like printf and other C functions? <civodul>mbakke: could you send the details to bug-guix so we can keep track of it? <civodul>brown121407: hi! try "guix install man-pages" <mbakke>civodul: that was the plan, wanted to find the broken commit first though <mbakke>but I have to go soon, so I'll create the bug first <bricewge>civodul: When a shepherd service write to stderr it goes to /dev/console <mbakke>civodul: bug reported, I'm going AFK for a while <bricewge>civodul: BTW efa578ecaece67366b4b0e2266de7c2faaa4ae54 (the one fixing --with-commit) broke `tests/guix-build-branch.sh` <dhanvanthri>Hello guys, I finally got guix up and running, however I'm having the issue I was afraid of during installation; To recap, even though my net is stable, the installation kept failing, I turned to manual install and resumed installation after every failure by rerunning guis system init, after I got into the system, guix pull failed due to the fact that glibc was not installed, which was one of the packages that hung during installation. <dhanvanthri>Guix pull in general was failing due to "network issues" but I just kept rerunning it untill it was done, but when I tried to install emacs, the glibc thing happened. I did manage to install glibc and get emacs running (that's where I'm on irc from), but I'm afraid of further errors/crashes due to the spotty install <dhanvanthri>Please put my literal 48h install process to an end <3 <bricewge>dhanvanthri: What is current error that you got <bricewge>The current installer is known to crash, it's better if you can manage to build your own installer ISO until a new a version of Guix is released. Which should be soonish. <pkill9_>you shouldn't need to install glibc on it's own, maybe you mean glibc-locales? <dhanvanthri>bricewge: I managed to overcome it, just by installing the missing package, but it WAS: guile: warning: failed to install locale <dhanvanthri>Yeah, guix pull failed like 8-9 times, and I had to keep just rerunning it till it was done. It's not a problem with the system itself, but some type of weird error with the server from which I'm retrieving these files. <dhanvanthri>I had couple terminals in the background constantly pinging the 2 servers from where my packages where getting downloaded ci.guix.gnu.org and git.savannah.org, and they didn't experience ANY packet loss during these failures... <dhanvanthri>This is literally the same problem (as far as I can tell) that plagued my install process for the last couple days. The packages would stop downloading and not resume, even though my connection to the server seemed uninterrupted. <dhanvanthri>And the graphical installer would hang if this happened, but luckily, not the cmdline install process. <dhanvanthri>I was advised to submit a bug report, but I'm basically new to being a contributing member of free software and was wondering exactly how to do that. <dhanvanthri>Anyway, I was wondering exactly where to put my .emacs file now that I managed to get onto the guix system, it doesn't seem to work when the .emacs file is in my user's home dir <aadcg>I was wondering about the interplay between use-package and the emacs packages installed with guix. In particular, it seems that those emacs packages are autoloaded at init. So, is it possible to defer those autoloads? Thanks. <ATuin>mmmm roptat i need some help to understand the ant-build-system <roptat>ATuin, what do you need to know? <ATuin>the lower in the build-system is defined as ant-build (function) <ATuin>which internally defines another function (builder) that calls ant-build (itself?) <ATuin>let me search the line of code <ATuin>there, what is calling there exactly? <ATuin>i'm looking for the best way to modify the build.xml in the package intellij-130 to add some flags to the builder <roptat>you don't really need to understand these details, but as you can see, this is all in a quote, so it's not interpreted yet <ATuin>yeah i get that part, that's executed when the derivation is done i guess <ATuin>ok, now it makes sense, it was just that they had the same names, but different values at different time <roptat>which in turn applies build phases in order <dhanvanthri>Guys what are the browsers available in guix by default? I'm not able to find firefox, and I browsed the packages from a different computer, but browser only returned qutebrowser <usney>I am using lxde now and none of my guix apps are listed in the menu like they are in gnome 3. <bricewge>dhanvanthri: You have IceCat in place of Firefox and ungoogled-chromium instead of Chrome. <bricewge>dhanvanthri: To report your bug just send an email to bug-guix@gnu.org. <dhanvanthri>bricewge: Just a casual writeup? Or is there a specific process/format? <bricewge>There is no format as is but it needs to be clear enough if you want it to be solved. <dhanvanthri>bricewge: Wrt icecat, it seems super out of sync with the latest firefox, maybe I'll switch to degoogled-chromium <rekado>dhanvanthri: I don’t think it’s a problem with the server from which you downloaded substitutes from, because that one is know to work fine. <rekado>dhanvanthri: it is probably your network or perhaps a transient problem that got cached by nscd. <bricewge>Thin to send the log of the failure, explain what was you wanted to do when you stumbled upon the bug, what went wrong, ... ***OriansJ` is now known as OriansJ
<dhanvanthri>rekado: Well, now that I got everything running, I'm gonna just submit a bug report and call it a day <dhanvanthri>rekado: I mentioned the bg terminals pinging those servers, since I assumed that meant my connection wasn't spotty to them <bricewge>dhanvanthri: Yes unfortunately, IceCat is based on Firefox ESR. <dhanvanthri>bricewge: Why is firefox not in the default repo? Is firefox not freedom respecting? <lfam>dhanvanthri: There's a trademark issue with the "firefox" name, and there are some FSDG issues with firefox that are too hard for us to correct for every firefox release, so we have to base icecat on firefox ESR <dhanvanthri>lfam: I understand, this is why debian is able to install it, since their distribution guidelines aren't exactly the same as the gnu project's guidelines. <dhanvanthri>I mainly ask since I got a lot of people to abandon chrome for firefox, and if it was actually not freedom respecting I'd be sad :'( <lfam>It's basically okay. These are very fine points of disagreement with Debian <rekado>dhanvanthri: what are “those servers”? By default we download binaries from ci.guix.gnu.org — and you can’t ping it. <dhanvanthri>rekado: I tried pinging it again now, and it defintely failed <rekado>ci.guix.gnu.org is configured to not respond to pings. <rekado>during the installation you only need git.savannah.gnu.org if you’re running “guix pull”. <rekado>and when that isn’t responding once ncsd might cache that failure, which can be very frustrating. <dhanvanthri>rekado: I really don't know what ncsd is, or how I'm using it in this process <rekado>nscd = name service cache daemon <rekado>it’s a part of glibc and caches domain name lookups (among other things) <dhanvanthri>Unrelatedly, I'm trying to swap caps and control, but the usual method of adding it to /usr/bin/setxkbmap is out of the question. In general, how do I find out what to do in situations like this? <rekado>does this work for you? –> setxkbmap us -option caps:ctrl_modifier <civodul>bricewge: fixed the --with-commit test in 3cea55078857ecd8ca67fd7cf4eaebb13fb6d9d3, thanks! <rekado>the firewall of the datacentre where ci.guix.gnu.org is hosted ignores ICMP <rekado>so… the reason for the guile-emacs build to fail is that it is terribly slow <rekado>dhanvanthri: ncsd caching failures too aggressively does feel like a bug, though, and I wish we could change that behaviour. *rekado hasn’t investigated this issue much <pkill9_>i hate pulseaudio, when it works it's fine <rekado>pkill9_: your problem appears to be unusual <rekado>pkill9_: have you been able to get at any more detailed logs? <dhanvanthri>rekado: Yep. Relatedly, I just noticed that glibc wasn't installed at all on my computer. I just installed it <pkill9_>nah, i thought the log i provided was the most detailed <rekado>dhanvanthri: that’s a misunderstanding. <pkill9_>i can't rmemeber what conclusion was drawn from it <rekado>dhanvanthri: glibc is obviously installed because all of the packages link with it. <pkill9_>oh yea, that my audio device was sleeping <rekado>dhanvanthri: it’s just somewhere in /gnu/store, not in some profile <pkill9_>but i see there is a bluez-alsa package, so i will probably just use alsa for now atleast <rekado>pkill9_: I’m using bluez-alsa on a headless server <rekado>I use it to send mpd output to bluetooth speakers. <bricewge>guix-vits: civodul fixed the failing test <rekado>I don’t think it’s a good idea to use it on a system with a desktop environment that is built around pulseaudio <pkill9_>i thought desktop environments can use either alsa or pulseaudio? <pkill9_>when pulseaudio dies for me, all my applications just revert to using alsa i think <pkill9_>unless you mean a full DE like gnome <dhanvanthri>I just tried the T60 keyboard layout, and it didn't auto detect/configure my volume keys. <dhanvanthri>Sorry, there was more of a sentence that I cut out, I was wondering if there was some auto way to get em working <dhanvanthri>I added the pulse-audio item to the xfce panel, but still nothing <bricewge>dhanvanthri: I think this is more a problem specific to your desktop envrionment. <bricewge>Start xev, press the keys you are taking about <dhanvanthri>Yo, every single time I install a package I get the error: "guile: warning: failed to install locale" <bricewge>The event will indicate you the key name <bricewge>If it starts with XF86 like, XF86VolumeUp your Desktop Environment should catch up <bricewge>dhanvanthri: If I remember correctly this error should also contains what you can do to fix it <bricewge>dhanvanthri: Check the documentation of your DE about multimedia keys <dhanvanthri>bricewge: I've also exported the appropriate things as indicated at the end of any install command <dhanvanthri>bricewge: However only the full string works, trying to us guix package --search-paths -p packagename fails <bricewge>Sorry I don't remember enough about this issue to be helpfull <dhanvanthri>dhanvanthri@dh-x200 ~$ guix package --search-paths -p glibc-locales <dhanvanthri>dhanvanthri@dh-x200 ~$ GUIX_PROFILE="/home/dhanvanthri/.guix-profile" <dhanvanthri>dhanvanthri@dh-x200 ~$ . "$GUIX_PROFILE/etc/profile" <dhanvanthri>dhanvanthri@dh-x200 ~$ . "$GUIX_PROFILE/etc/profile"^C <dhanvanthri>dhanvanthri@dh-x200 ~$ GUIX_PROFILE="/home/dhanvanthri/.guix-profile"."$GUIX_PROFILE/etc/profile" <dhanvanthri>dhanvanthri@dh-x200 ~$ guix package --search-paths -p glibc-locales <rekado>why do you use “-p glibc-locales”? <rekado>I also don’t understand your paste. <dhanvanthri>rekado: I just assumed it meant package, since when I ran it without an argument it said guix package error <rekado>dhanvanthri: I don’t know what commands you have run. <rekado>I wonder what made you think to use “-p” at all. <rekado>I suggest taking a look at the manual <dhanvanthri>rekado: The error message literally calls for that invocation :'( <dhanvanthri>rekado: Sorry not the error message, the post install message <rekado>I still don’t understand your paste. I don’t know what these lines mean and what you hope they accomplish. <dhanvanthri>rekado: I don't understand what I'm doing, and I should be more careful clearly <lfam>If the error message does say to use -p we should look into that <lfam>dhanvanthri: This issue with the locales is known and we are working on it. It's a bad user experience problem <rekado>lfam: it does mention -p but only for the profile that was used <rekado>when the profile was called “glibc-locales” it’s going to print what dhanvanthri pasted. <rekado>so the user error happened earlier already <lfam>Basically what you need to do is install glibc-locales for your user and whoever runs the guix-daemon (probably root). You also need to correctly export the variable GUIX_LOCPATH for both your user and where the guix-daemon is run. How do you start the guix-daemon? <rekado>(see display-search-path-hint in (guix scripts package)) <lfam>dhanvanthri: That long message I sent was for you ^ <dhanvanthri>lfam: I switched on my computer and presumable it was there <lfam>It's Guix System? Or Guix on another distro <lfam>Okay then you don't need to install glibc-locales or do anything <dhanvanthri>lfam: This is what I was saying from earlier, I kept quitting and rerunning the guix system init command during initial setup since the installation kept failing due to (apparently) ncsd caching my failure or something <lfam>That's not related to this problem <dhanvanthri>lfam: I did already install glibc-locales too, post system install <lfam>Uninstall it. That's handled automatically on Guix System <dhanvanthri>lfam: Well, guix package --search-paths still gives me "guile: warning: failed to install locale" when I run it <dhanvanthri>lfam: Also the same error when I try to install any program. The programs run and install, but it's kinda scary to see this everytime <lfam>You can ignore it for now, it's just a warning. That's not as bad as an error <Veera>install glibc-locales warning msg: is not that important <Veera>installed a month earlier it is still there <lfam>You shouldn't have the warning on Guix System but why you do is a mystery <Veera>I use GNOME and other DM , it is not a noticeable problem <Veera>like rekado said I will try it. <Veera>I was afraid to try something in this new system <lfam>dhanvanthri: To help debug it, you can try `realpath $(echo $GUIX_LOCPATH)` <mbakke>dhanvanthri: have you run 'guix system reconfigure' since you installed Guix? The error will occur if the system uses a different version of "libc" from what your pulled 'guix' does. <dhanvanthri>mbakke: Big rip, I did after install, but not after pull <dhanvanthri>lfam: "/gnu/store/1lcniyxkxkh8g73zvh2gpbccvl6ggna7-locale-2.28" <lfam>That confirms it, mbakke is correct <mbakke>dhanvanthri: 'sudo -E guix system reconfigure your-config.scm' should fix it in that case <lfam>That's the old version of the locales <mbakke>dhanvanthri: I think the installer writes it to /etc/config.scm <apteryx>raghavgururajan: resuming my testing of your linphone patches <mbakke>though you can place it anywhere <alextee[m]>im trying to guix pull and i get this: guix pull: error: symlink: File exists: "/var/guix/profiles/per-user/alex/current-guix" <alextee[m]>and i still can't build anything in the guix repo: guix build: error: icu4c-CVE-2020-10531.patch: patch not found <alextee[m]>are you supposed to configure/make it? i was able to build before without doing so <alextee[m]>i think i did sudo -E guix pull before for some reason <alextee[m]>hmm a lot of things b0rked: guix describe: error: failed to determine origin <raghavgururajan>sneek, later tell apteryx: Let me know if 11 (mediastreamer) is all good to be pushed. :-) <rekado>alextee[m]: check the permissions <rekado>alextee[m]: can you show us the full output and the command you ran? <alextee[m]>rekado: i rm -rf'ed /var/guix/profiles/per-user/alex/current-guix and now pull works <alextee[m]>im searching for fogpad for example but nothing shows up after guix pull finishes <alextee[m]>you mean the output from trying to build something? ***jonsger1 is now known as jonsger
<dhanvanthri>Sorry guys, had to restart to get stuff working right, did someone say something about guix system reconfigure needing sudo? <rekado>alextee[m]: be careful with “rm -rf”; current-guix is a symlink only. <rekado>if you’re on Guix System you don’t need to worry because /gnu/store is remounted read-only. <rekado>but you can cause damage on foreign distros if you end up deleting store items. <rekado>try “hash guix” and then check “type guix” to be sure that you’re actually using the pulled guix <rekado>dhanvanthri: “reconfigure” needs to change the system (and a GRUB entry) so it needs to run as root. <alextee[m]>guix is /gnu/store/cbchyxvyvxncz14pv1kaxlj7r5vmnl17-profile/bin/guix <rekado>it should be a link such as /home/rekado/.config/guix/current/bin/guix <rekado>because that’s a link to /var/guix/profiles/per-user/alex/current-guix <rekado>which in turn is a link to the current generation of the current-guix profile <sneek>Welcome back jayspeer, you have 1 message! <sneek>jayspeer, zzappie says: sorry I didn't say clear enough. I meant that you can use 'guix environment' to isolate 'ruby gem' installation. Without altering your profile. take a look at https://paste.debian.net/1138220/. will create isolated HOME where you can install arbitrary packages using gem. To get rid of libstdc6++ problem you cab install gcc-objc++@ 6.5.0 in that profile <rekado>which is a link to the profile in /gnu/store <rekado>when your “guix” points directly into the store you won’t benefit from the symlink magic that “guix pull” uses to replace itself. <dhanvanthri>My caps and control reset on reboot, and since I reboot I can´t find the command that fixed it :´( <rekado>let’s unlearn the reflex to reboot :) <rekado>alextee[m]: let’s try to figure out what’s wrong first. <rekado>does .config/guix/current exist? <rekado>dhanvanthri: setxkbmap us -option caps:ctrl_modifier ? <rekado>alextee[m]: does .config/guix/current/bin/guix exist? <alextee[m]>lrwxrwxrwx 2 root root 56 Jan 1 1970 /home/alex/.config/guix/current/bin/guix -> /gnu/store/wc4wlvqy4xlq6r964ff50xql4a5mdhnv-guix-command <rekado>alextee[m]: how about readlink /home/alex/.config/guix/current <rekado>alextee[m]: the date is for reproducibility <alextee[m]>readlink says: /var/guix/profiles/per-user/alex/current-guix <jayspeer>I've finally managed to install guix on a *real* hardware (after embarasingly long time); this in turn allowed me to finally test emacs-guix (I don't know why it refuses to work on foreign distro, perhaps I was doing something wrong), is anyone else using it? it seems to build every package, but I have enabled substitutes <rekado>when you run “/home/alex/.config/guix/current/bin/guix describe”, do you get the expected output? <rekado>because in that case all that is wrong is PATH <rekado>PATH should have $HOME/.config/guix/current/bin at the first place. <alextee[m]> /gnu/store/cbchyxvyvxncz14pv1kaxlj7r5vmnl17-profile/bin:/gnu/store/cbchyxvyvxncz14pv1kaxlj7r5vmnl17-profile/sbin:/gnu/store/99h3nsyj1fnay0fwamzgzlk67gszf2kv-glib-2.60.6-bin/bin:/home/alex/.guix-profile/bin:/home/alex/.guix-profile/sbin:/run/setuid-programs:/home/alex/.config/guix/current/bin:/run/current-system/profile/bin:/run/current-system/profile/sbin:/home/alex/.gem/ruby/2.5.0/bin *alextee[m] checks his .zshrc <rekado>get rid of the /gnu/store entries. <rekado>they probably got there because you may have source’d an etc/profile file without setting GUIX_PROFILE first <rekado>the etc/profile file will use GUIX_PROFILE when it is set; otherwise it will use the target in /gnu/store <rekado>this means that with GUIX_PROFILE your variables get updated automatically as your profile changes <alextee[m]>this may explain the warning about GUIX_PROFILE i keep getting every time i install something <rekado>but without it you will be stuck at the exact version you sourced <alextee[m]>i do this at the top source ~/.guix-profile/etc/profile, but i am not setting GUIX_PROFILE anywhere before that (not sure what else i should check besides .zshrc) <rekado>just set GUIX_PROFILE to $HOME/.guix-profile first <rekado>it may seem redundant but the presence of the GUIX_PROFILE variable makes the difference <rekado>if you’re curious you can take a look at any profile’s etc/profile file <alextee[m]>i now finally understand the hint! hint: Consider setting the necessary environment variables by running: <rekado>yeah, you may be tempted to just merge the two lines, but it results in slightly different behaviour <rekado>a bit unfortunate, but using the variable is again due to reproducibility <rekado>we can’t embed /home/alex/.guix-profile (or whatever location you set for “guix install”) in the store. <rekado>so with the GUIX_PROFILE trick the etc/profile files are reproducible and work for any profile location. <rekado>(and they have a usable default, which is to set the variables to the exact /gnu/store locations) <rekado>glad we fixed this without a reboot :) <vagrantc>nothing exploded with my 5.6 linux-libre change thus far? it looked like it was building some things i wasn't expecting <alextee[m]>uhm, i get this now: guix describe: error: failed to determine origin <alextee[m]>type guix says: guix is /home/alex/.guix-profile/bin/guix <alextee[m]>hint: Perhaps this `guix' command was not obtained with `guix pull'? Its version string is 1.0.1-15.0984481. <rekado>according to (guix scripts describe) this happens when the command does not have a git repository <rekado>alextee[m]: I wonder if your deletion may have screwed things up a bit <alextee[m]>oh so here is all the source code for the commands, nice. this will be helpful in the future <rekado>the symlink forest is robust, but when disturbed by manual deletions it can become weird. <rekado>I wonder if you could get out of this by running “guix pull” again <rekado>well, we could try to fix it the hard way <rekado>what does readlink /var/guix/profiles/per-user/alex/current-guix say? <pkill9_>rekado: good news, i found a very clear indication of the pulseaudio problem: in the log: "I: [pulseaudio] core.c: We are idle, quitting..." <rekado>what does readlink /var/guix/profiles/per-user/alex/current-guix-1-link say? <alextee[m]> /gnu/store/134rl42vv8p0igk3jcd2jzsyp823wn48-profile <alextee[m]>readlink that says nothing, that's the end i guess <rekado>so… perhaps try “hash guix” again <rekado>in Bash “hash guix” will tell it to update its cached location for “guix” <rekado>since “/var/guix/…guix describe” worked for you earlier I wonder if maybe it just uses the wrong guix <rekado>when you use the absolute file name, does that work? <rekado>i.e. /home/alex/.guix-profile/bin/guix describe <rekado><alextee[m]> type guix says: guix is /home/alex/.guix-profile/bin/guix [22:02] <rekado>it should be .config/guix/current/bin/guix <rekado>~/.config/guix/current/bin/guix describe does work, doesn’t it? <rekado>earlier you had /gnu/store/cbchyxvyvxncz14pv1kaxlj7r5vmnl17-profile/bin:/gnu/store/cbchyxvyvxncz14pv1kaxlj7r5vmnl17-profile/sbin:/gnu/store/99h3nsyj1fnay0fwamzgzlk67gszf2kv-glib-2.60.6-bin/bin:/home/alex/.guix-profile/bin:/home/alex/.guix-profile/sbin:/run/setuid-programs:/home/alex/.config/guix/current/bin:/run/current-system/profile/bin:/run/current-system/profile/sbin:/home/alex/.gem/ruby/2.5.0/bin <rekado>without the /gnu/store things that would be /home/alex/.guix-profile/bin:/home/alex/.guix-profile/sbin:/run/setuid-programs:/home/alex/.config/guix/current/bin:/run/current-system/profile/bin:/run/current-system/profile/sbin:/home/alex/.gem/ruby/2.5.0/bin <rekado>this suggests that at some point you ran “guix install guix” <rekado>because only then would it find “guix” in /home/alex/.guix-profile/bin <rekado>in any case: /home/alex/.config/guix/current/bin should come first <rekado>you can check later with “guix package -l” to see when it was added (and if it’s still there) <rekado>part of it comes from source $GUIX_PROFILE/etc/profile <rekado>but it won’t put .config/guix/current/bin there <alextee[m]>oh we said GUIX_PROFILE=/home/alex/.guix-profile, and then according to /etc/profile it just appends /bin as the first path <pkill9_>rekado: do you run pulseaudio as a system service? i suspect it may be refusing to respawn because i run pulseaudio as a user <pkill9_>because i get this error in my /var/log/messages: "User-configured server at unix:/tmp/pulse-socket, refusing to start/autospawn." <rekado>pulseaudio-aware applications should “activate” pulseaudio automatically <rekado>pkill9_: who owns /tmp/pulse-socket? <rekado>can your user account write to /tmp? <pkill9_>yea it can, and /tmp/pulse-socket doesn't exist <pkill9_>i tried removing ~/.config/pulse, but i'll also try rmeoving ~/pulse <rekado>alextee[m]: right, so in your shell init files you should arrange for export PATH=$HOME/.config/guix/current/bin:$PATH at the end <rekado>alextee[m]: on Guix System this should really be done automatically, but it’s possible that we aren’t doing the right thing when zsh is used, dunno. <alextee[m]>that might be it. im getting a lot of these things with zsh it seems <alextee[m]>i'd use bash but it has horrible defaults and im not aware of any easy way to make it more usable lol <alextee[m]>i expect when i type "he" and press up to find things starting with "he" only in the history. this is the main reason i use zsh <alextee[m]>while zsh "just works fine" out of the box, (except on guix) :D *alextee[m] goes back to packaging <rekado>alextee[m]: yeah, I use Emacs M-x shell for that :) <rekado>it’s still bash, but it behaves rather different <rekado>it’s just too inconvenient to leave it once you’re inside <NieDzejkob>alextee[m]: looks pretty zyn-fusion-specific, I'd go with audio.scm <NieDzejkob>you could also consider giving it its own file if it's more than a handful of packages <alextee[m]>it's just this, and then zynfusion, so i guess audio and music <pkill9_>pulseaudio isn't reading ~/.config/pulse/*.conf <pkill9_>i did an strace and it only tries to access /etc/pulse/daemon.conf <alextee[m]>looking in my "sents", can't find it. im pretty sure i did it before <rekado>send mail to <number>-done@debbugs.gnu.org <alextee[m]>i sent in the zynfusion patches and a few others