<amz3``>after guix pull the errors are gone <amz3``>for some reason, installing guile-bytestructures triggers the download of bootstrap binaries <jdz1200>I think I figured out the answer to my grub-efi on bare-bones issue... Needed to add "bootloaders" to "use-package-modules". Woot <rekado_>can no longer decrypt the hard drive :( <rekado_>need to pull an image of the disk and reinstall <reepca>Hmm, a package I'm trying to get working copies a source file and tries to write to it, which the permissions aren't set to allow. How might one control this? <ng0>Haunt is pretty amazing if you use it the right way. I'm still discovering the many possibilities of it. <reepca>jlicht: Yeah, it's looking like I'll have to add a phase to make all the source files writable. Just wondering if there was already something existing that could do something like that. <reepca>Hm, what's the proper way to make certain guile modules available for a build phase? <jlicht>reepca: gnu-build-system has imported-modules <jlicht>reepca: Hmm, scratch that: it seems that this functionality is not exposed currently <sneek>quiliro, you have 3 messages. <sneek>quiliro, rekado says: Have you tried doing as the error message suggests? It says to specify the location of qmake by configuring with the “--qmake” flag, giving the location of the “qmake” executable as an argument. <sneek>quiliro, rekado says: But first you should check that “qmake” is actually available at build time. What package provides it? Is that package part of the native-inputs? You can use “find /gnu/store -name qmake” to find the name of the package providing “qmake”. <sneek>quiliro, rekado says: “qmake” is part of the “qtbase” package, which is in the inherited native-inputs already. So you may just need to tell “configure.py” about it. <rekado_>quiliro: actually, this turned out not to be the problem <rekado_>quiliro: I got it to build and started openmolar <rekado_>quiliro: there's just one remaining problem (something to do with the PYTHONPATH variable) <rekado_>quiliro: unfortunately, my file system got corrupted and I cannot fix it, so I need to reinstall GuixSD <quiliro>rekado_: oh!...great and terrible news at the same time <quiliro>do you have the text of the changes you made? <rekado_>it's on a different machine which I cannot access without the keys that I need to restore from my disk first <rekado_>quiliro: I'll try to send this to you tonight <quiliro>rekado_: or maybe we can continue construction of python-qcintilla from where we were <rekado_>unfortunately, I don't have the time to do this now <quiliro>rekado_ i will be connected until tomorrow 11h00 utc ***jonsger1 is now known as jonsger
<reepca>argh... emacs leaked 6GB somehow <quiliro>reepca: what did you do to have to restart emacs? <reepca>I don't know, I closed all the non-essential buffers and ran garbage-collect and added up the totals it presented. Only about 130MB was actually in use according to that. <quiliro>Aftere removing Tor, i get a bunch of messages in /var/log/messages about Tor: <quiliro>Jul 15 14:40:27 localhost Tor[1]: Problem bootstrapping. Stuck at 5%: Connecting3C538F44038BBFA077587DD755 at 194.109.206.212:443) <quiliro>when the wireless card is not connected <quiliro>and when it is, something of the sort of: <quiliro>Jul 15 15:10:19 localhost Tor[1]: Tor has successfully opened a circuit. Looks like client functionality is working <quiliro>i also have font problems, i see weird characters like: ͡° ͜ʖ ͡ <quiliro>or is it too late down where you are? <joshuaBPMan_>hello, maybe this is a dumb question, but I'm having a hard time trying to figure out how to sort through all of my guix email. On thunderbird, I've set up a throw all of guix-devel@gnu.org into one folder...BUT I still get guix emails in my regular inbox. Is there some way to put all of guix emails into one folder? Also I am trying to get more familiar with using email in Emacs, but it can be such a pain to set up. <cehteh>ACTION uses claws-mail and lets popfile sort his emails with the imap backend, thats close to a private secretary, learns from mails i sort manually, diverts them into 20 or so folders <cehteh>i even thought about training it to recognize trolls :) but didnt do so yet <quiliro>joshuaBPMan_: i have gnus...but i have not found out how to sort in folders yet <cehteh>but its written in perl, i think it is portable for almost any OS <joshuaBPMan_>quiliro: I suppose in gnus you can use gmaine... I think that is the server name. <quiliro>i used to do it in claws mail....but i magrated to gnus <joshuaBPMan_>yeah. I'd like to try email in emacs, because I think it would be pretty cool. I've tried it in the past, with mixed results. It seems like gnus has the most features, though it can be a bit slow. <cehteh>i use multiple clients (mobile phone, laptop, pc) .. i really prefer when the instance sorting the mails is somewhere else <cehteh>and so far i didnt liked the email in emacs approach :D <cehteh>too much to learn, blocking emacs sometimes <cehteh>anyway sorting emails with filters on the client is a bit impractical because you have to have one dedicated client for that which must be always runnin, or syncronize these filters or other things <cehteh>the only think i do is that i let the client delete everything older than half a year in the trash <jdz1200>Is it possible to tell `guix pull` to only update things that are pre-compiled upstream? <cehteh>ACTION curses about that too .. some 'tagged' releases would be nice, even if automatically generated <cehteh>lemme guess you wait for icecat to compile too? :) <apteryx[m]>Has anyone had success with WebRTC and our build of Icecat? <jdz1200>Not yes, cehteh, but I can see how that would be a problem lol <jdz1200>I'm thinking `guix pull` at scale causing duplicative work. <joshuaBPMan_>if I have a tiny patch to change the documentation, do I send it to guix-devel or guix-patches? <rain1>which OS can I build and use guix on? it doesnt' work on alpine <rain1>i gave it a try and it built but would not run <rain1>not quite true, it runs ok but it failed building gcc I think with guix package -i <jonsger>joshuaBPMan_: I think "From, To, CC or BCC" contains "guix-devel" should work well <lfam>rain1: If it doesn't work, can you make a bug report? :) <lfam>Thanks jlicht, I appreciate it <lfam>quiliro: That's because Tor is a system service, controlled by root. <rain1>lfam as root I run guix-daemon --build-users-group=guixbuild --disable-chroot, now I do guix package -i which & it wants to compile gcc 5.4.0 <rain1>now ill just wait for that to run and paste the error <quiliro>how can tor be up on root and not be installed? <rain1>x86_64-guix-linux-gnu-ld: cannot find -lstdc++ <rain1>I don't know why, but on possible reason this might have happened is because alpine is using musl libc not gnu libc <quiliro>i have to stop tor every time i turn the machine on <quiliro>since guix always requires compiling, it a bummer <rekado_>All this time ... and I still cannot fix the problem <rekado_>grub-install fails with "unkown filesystem" <rekado_>so I cannot reinstall without reformatting the disks <rekado_>already tried testdisk to recover partition tables but to no avail <rekado_>this is frustrating because it looks like only grub complains about this <rekado_>I can unlock and mount the disk just fine after booting from USB <rekado_>none of these are available and they likely wouldn't help <quiliro>if it can be mounted, why do you say it is corrupt? <quiliro>if it can be mounted, MBR is not damaged <rekado_>it's a LUKS partition with a single ext4 file system in it <rekado_>In grub (on the SPI chip) I can unlock the LUKS container, but I cannot access (crypto0), because Grub says "unknown filesystem" <rekado_>that's also why I cannot pass control to the Grub on the encrypted disk <rekado_>but when I boot from somewhere else and unlock the LUKS container I see that everything is fine <rekado_>I can read all files and fsck doesn't show any problems <rekado_>I have /boot/grub on the file system inside the LUKS container <rekado_>as I said, the contents of that inner file system appear to be okay <quiliro>"The /boot partition needs to be unencrypted." <rekado_>grub-install -v shows me that it complains about the unlocked mapped device file to be missing, but it is clearly there <apteryx[m]>rekado_: Has this setup (encrypted /boot) worked for you before on GuixSD? Or is it something new you are trying? <rekado_>apteryx[m]: I've been using this up to this mornin <rekado_>when my computer froze and I had to power it off the hard way <rekado_>booted from USB, unlocked the LUKS container, ran fsck on the inner file system <rekado_>recreated the partition table, still won't boot <apteryx[m]>OK. That sucks :/. Apparently Grub's crypto module is doing something differently than what cryptsetup luksOpen would do, since the later still work, right? <quiliro>"I think it maybe possible after putting some LUKS and LVM modules in core.img which will be placed on GPT partition, but the rest (kernel and so on) can be located in encrypted /boot on same partition <rekado_>quiliro: I don't think this applies. <quiliro>maybe something is missing in your new grub <reepca>I'm pretty puzzled. There's a profile somewhere referencing this thing I want to delete, but I can't figure out what profile it is. It's not under /var/guix/gcroots/profiles/per-user and it's not a system profile. <lfam>Is it a guix environment currently in use? <quiliro>rekado_: did you get an answer on #grub? <lfam>Or a build that's currently runing <rekado_>reepca: the Emacs memory leak you mentioned can be prevented by configuring Emacs with a certain option. I just can't remember it. <reepca>lfam: I don't think so - any environment should have been closed out of when I restarted emacs. And top doesn't show guix-daemon as having any children (which I think would indicate a build running, but not sure) <lfam>reepca: Have you tried grepping in /var/guix for the store path of the item in question? <quiliro>rekado_: what is the exact error message? <reepca>... I'm silly. I had it a binary from the path running. <reepca>which actually makes me curious - does "guix gc" include "the output of lsof" or similar in the list of roots? <nckx>I feel silly too... How do I mark a file-system as optional so it doesn't break the boot? <nckx>I've tried (flags '(no-fail no-auto)) (needed-for-boot? #f) to no avail. <nckx>I've never needed to do this before but it can't be this hard... <nckx>reepca: That's a damn good idea, actually. <reepca>Hm, I need to get the output path of a package to be passed to #:configure-flags... but if I understand correctly, the output path is a hash of, among other things, the builder. So I can't really know it until build-time. So is there a way to add it at build-time without replacing the entire configure phase? <reepca>nckx: indeed, I wouldn't have thought of it. Gives me respect for the authors of the daemon that has been a bit lacking after a lot of time reading the source. <nckx>reepca: (assoc-ref %outputs "out") should work there as well, I thought. Doesn't it? <reepca>nckx: By "there" do you mean in the package definition and not at build-time? <nckx>reepca: Maybe I misread your question, but (arguments `(#:configure-flags (list (string-append "--mystuff=" (assoc-ref %outputs "out") "/stuff")))) definitely works. <reepca>Huh. That's interesting. Are the argument definitions evaluated at build-time then? <reepca>ah, I seem to have failed to notice that little '