<lfam>I guess I'll just skip that test. I don't feel like creating a multi-line substitution for a brittle test that will be fixed in the next Go release. <kyamashita>Does libjpeg usually come with a pkg-config file? ***mark_otaris is now known as Guest34703
***Guest34703 is now known as mark_otaris
***mark_otaris is now known as Guest76322
***Guest76322 is now known as mark_otaris
***exio4 is now known as hacker
***hacker is now known as init
***init is now known as n4x
***n4x is now known as exio4
<pksadiq>when installing guixSD I get the following error: guix system: error: build failed: some substitutes ... '/gnu/store/...-module-import-compiled.drv' failed ....; try `--fallback' to build derivation from source. If I do --fallback, every package seems to be installed from source. How can I install only the failing package from source? <paroneayea>pksadiq: hm, are you using the usb installer, or? <pksadiq>installing via 'guix system init ...' as said in the manual <paroneayea>pksadiq: hm, ok, so that *should* have the substitute server authorized <paroneayea>pksadiq: I seem to remember that someone found that the garbage collection on Hydra had been a bit overly-aggressive, and some packages got garbage collected from the release that maybe shouldn't have <paroneayea>pksadiq: anyway I'm not sure. Are they packages that previously were pulled down via subistitues? <paroneayea>if it's "I'm seeing a lot build now", it could be because you're now building a bunch that can be built now that you passed in --fallback <paroneayea>I'm going to bed, not sure if my comments are helpful at all. good luck! <pksadiq>paroneayea: previously I have had success installing the same way (before a week or so). Now I got a new (ver old actually) system which I can use soley for guix <pksadiq`>I got disconnected. Did I miss anything? ***alberto is now known as Guest96596
***Guest96596 is now known as albertoefg2
<ng0>is there some special setting vim users use to get what .dir-locals.env in guix.git checkout does? occasionally I use systems where there is vim instead of emacs <ng0>how do I put this "config-file" to practice when I want to extend the torrc? <ng0>i know how the tor service is constructed, but i still don't know how to extend it as I've never done this before <ng0>is it similar to openssh-service then? <ng0>we should really document this.. " It is passed config-file, a file-like object, with an (...)" this assumes that everyone knows how config-file should be constructed. <ng0>config-file should be in the index, i could only find "file-like objects" <rekado>“config-file” is just a variable name. <rekado>“The ‘local-file’, ‘plain-file’, ‘computed-file’, ‘program-file’, and ‘scheme-file’ procedures below return “file-like objects”.“ <ng0>that is not clear from reading the description of tor-service in the documentation <rekado>that’s what I get when looking up ‘file-like objects’ in the index <rekado>it’s the usual format for a variable name. <ng0>what I mean is, you should be able to find out how to extend a service without searching through the index <rekado>are you saying you should be able to extend a service without reading the manual? <ng0>... I mean the manual should link more <ng0>let's say I look up tor-service and I have no idea about guile. I find tor-service.. then in the next step I will have no clue how to add custom values to the torrc <ng0>from a users perspective <rekado>it says that the service takes a file-like object. Looking up what a file-like object is you see that it can be a number of different things. It’s not a Guile thing. <ng0>hm.. maybe it's just me. bouncing back and forth between gentoo and guixsd systems <ng0>okay I've read it again <ng0>I still think there should be more links in the documentation <ng0>Ok, it is clear. But linking would be an improvement <rekado>yes, a reference to the explanation of file-like objects would be good. <ng0>hm.. we start no SocksPort by default? <ng0>at least the service doesn't mention it <jmd>The new release of guile-ncurses does something wierd. I can't package it. <ng0>I have to correct my statement: it is not clear. I've tried to use the file /etc/torrc where I have additional content for the service, but all the ways I've tried to use (local-file "/etc/torrc") did not work <efraim>I thought they were waiting for november 2nd <ng0>can someone give me an example how to use local-file in a system config? I've searched and tried long enough.. <efraim>sneek: later tell csanchezdll I reverted the expat commit on my aarch64 branch and I was able to build the static-binaries for the replacement aarch64 tar I wanted, so I can try that on the actual hardware in a few hours <roptat>hi, I'm trying to add two packages, and their license is a bit confusing <roptat>just a list of the mentionned licences? <rekado>roptat: in the first case I’d set the license to GPL (as our ffmpeg package) and add a comment above the license field. <rekado>roptat: in the second case I’d use a list of licenses, again with a comment above it that says how the licenses apply. <roptat>rekado, ok, I'll do that, thanks <ng0>there is no mit license <ng0>there is one, but no one means the mit license when they write mit license <ng0>it helps to compare license texts, they are rather short :) <roptat>so, how should I define the correct license? <wgreenhouse>OOC, does there exist a template for guix/guixsd as an LXC guest? <ng0>o.o something broke in gnome <ng0>closing the gnome filemanager restarts gnome desktop <ng0>repeating this 3 times and you get back to login <ng0>can someone with a current system reproduce this? <ng0>roptat: sorry I can't help. all I can say is, there are websites where licenses are listed and I just compare if I am not sure <rekado>roptat: MIT can mean Expat or X11. Often it’s Expat. X11 has an additional clause. <roptat>and an "and" becomes an "and/or" <roptat>it's the same text, but not exactly the same as in the software <rekado>you could always say ‘non-copyleft’ and link to the file. <rekado>GuixSD releases are in sync with Guix releases (it’s the same software). <jmi2k>Can anyone help me? I'm installing GuixSD, and when I try to install the system (guix system init ...) it fails because it can't find /etc/resolv.conf . I'm using bare-bones.scm <apteryx> ng0: Have you been through the whole manual? I have a feeling this is explained somewhere in there. <jmi2k>catotano: are you talking to me? <apteryx>jmi2k: Can you post the exact error message? <ng0>apteryx: yep, and no it is not <ng0>if you refer to the tor problem <jmi2k>after a bunch of warnings about possibly unbound variables, it says: guix system: error: build failed: while setting up the build environment: getting attributes of path '/etc/resolv.conf': No such file or directory <apteryx>ng0: I was refering to the general use case of adding file configuration using the declarative scheme approach. <jmi2k>I can send you a pastebin with the exact configuration and output if you want <ng0>the unbound variable warnings are important <apteryx>jmi2k: Yes, the full output might prove. Have you tried creating a dummy empty file (touch /etc/resolv.conf) and re-running the guix system init? <jmi2k>Yes, I tried. It failed in a different way. I'm testing on a VM, so I deleted it and started again. I'll put that output too in the pastebin if you want <ng0>I'mgoing through the handbook again, see if I find an example apteryx <apteryx>ng0: I'm reading it as well, but I'm only at chapter 3. I'll take a while. I think most of the in-depth stuff is toward the end of it. <apteryx>jmi2k: I'm not sure if I'll be able to help you, but in all cases, seeing the output might gives us ideas! I myself only tried the lightweight-desktop.scm (or something similar) config, and didn't have that issue. <jmi2k>Give me some minutes, I'm gathering all the output from my VM <apteryx>So maybe comparing these two configs could give hints about something missing? <apteryx>Sorry, just found out that when I use "CTRL-BUTTON1" it kills the current window using RatPoison. Not sure if that's intended or not. <ng0>i think i can't help though.. i've never used uuid in file-systems part <ng0>in practice, X has been superceded by Y" is nice... i haven't updated this system in a while <ng0>looking at the dependencies of sddm and lightdm, i think it would be nice to switch from slim to sddm. sddm is also what replaces kdm. <ng0>lightdm would require gnome-accountservice etc <ng0>but at least it has more deps <jmd>What do you mean by "switch"? <jmd>It would be great to have an alternative to slim. But I don't think we should get rid of it. <apteryx>I see that my guix-profile stored in the system under /var/guix/profiles/per-user/maxim/guix-profile has the permissions: lrwxrwxrwx maxim:users <apteryx>Isn't that too loose? It means any users could mess with my profile. <apteryx>OK at least the parent folder of my profile ("maxim") is 755, so that should protect my guix-profile from being written to. <ng0>i think slim is a terrible default <ng0>it would be good to have *anything* which can use more than en-us <jmd>I don't have any strong opinion on which is the default. <jmd>Actually I didn't know there was a "default". One has to explicitly choose a config at install time. <ng0>slim is the only option at the moment.. if not, it is not documented <efraim>i have switched to lightdm when I left gnome and gdm <ng0>so it's not documented <efraim>slim is default, I think some of the options were parameterized when sddm was added <davexunit>has anyone tried updating our mesa to the 12 series? <rekado>could we also add llvm to the inputs? <ng0>I have a strong opinion on slim.. if it could do more then en-us, we are not providing support for that. and providing support for more keyboard-layouts at login is a basic requirement in my opinion <efraim>isn't it at 12.01 on core-updates? <efraim>turns out vim-full needs python2 for some of the python tests, i'll just skip them by switching python-wrapper to python :) <davexunit>I bought a new AMD GPU that is supposed to have pretty good free drivers provided you are running linux 4.7+ and the latest Mesa <ng0>efraim: really? I removed python2 and it succeeded <davexunit>but I'm not absolutely sure that all of the AMD patches have made it into releases of both projects. just hearsay. <ng0>i have not addressed all the replies on the vim-full thread yet. <ng0>like, should we switch to github releases and just keep up to date with that, so that vim-full can just inherit vim? <ng0>that was suggested as upstream points to github now <efraim>i took out the patches for now, we can figure out the patches separately <ng0>tehy are needed for the testsuite <efraim>two separate issues effectively, one for a new package, one for updating the source <efraim>if upstream wants us to use the github releases to get patched vim then I guess that makes the most sense <ng0>If this wasn't pointed out in the thread, I would've gone with just the patches <ng0>dunno how gentoo does it.. i think with patches instead of github, but i could be wrong <efraim>I don't know what debian does either <efraim>7.4 went over 2000 patches i think <ng0>ok, so gentoo uses github <efraim>with neovim around now I assume they'll go faster <ng0>I think I'll prepare a patch to update vim source to github, and then update vim to at least 8.0.0005 <ng0>afterwards vim-full can be smaller <efraim>vim might need more patching, it still says who compiled it with vim --version <efraim>breaks reproducibility with guix challenge <ng0>i thought the 'who build it' looked like it is something unique to the build environment.. ok, then this needs to be fixed <ng0>but this is a separate issue <jmd>efraim: Somebody ought to submit a bug to the vim developers. <apteryx>10:29 -!- apteryx [~maxim@c-98-248-32-25.hsd1.ca.comcast.net] has joined #guix <apteryx>10:29 -!- Topic set by civodul [~user@unaffiliated/civodul] [Wed Aug 3 07:33:20 2016] <apteryx>10:29 [ abbe ] [ CustosL1men ] [ gry ] [ lewo ] [ orly_owl_ ] [ sprang ] <apteryx>10:29 [ Acou_Bass ] [ cyraxjoe ] [ habs ] [ LnL ] [ otremblay ] [ Steap ] <apteryx>10:29 [ aggelos__ ] [ davexunit ] [ himmAllRight ] [ lxo ] [ pagurus` ] [ steveeJ ] <apteryx>10:29 [ albertoefg ] [ defanor ] [ hodapp ] [ M-yrns ] [ paolo ] [ taylan ] <apteryx>10:29 [ amuck ] [ detrout ] [ ifur ] [ mark_otaris ] [ paroneayea ] [ tg ] <apteryx>10:29 [ amz3 ] [ Dezponia ] [ Igel ] [ mark_weaver ] [ Petter ] [ thomassgn ] <apteryx>10:29 [ amz3` ] [ df_ ] [ ijp ] [ marxistvegan ] [ Phlogistique ] [ tinyOwl ] <apteryx>10:29 [ andschwa ] [ Digitteknohippie] [ jackhill ] [ MatthewAllan93] [ pksadiq`` ] [ Tsutsukakushi] <apteryx>10:29 [ apteryx ] [ divansantana ] [ JamesRichardson] [ MinceR ] [ quasisane ] [ ueno ] <apteryx>10:29 [ ArneBab ] [ domenkozar ] [ jchmrt ] [ mog ] [ quigonjinn ] [ Ulrar ] <apteryx>10:29 [ b4283 ] [ dvn ] [ jluttine ] [ mood ] [ reepca ] [ wgreenhouse ] <apteryx>10:29 [ bavier ] [ efraim ] [ jmd ] [ mordocai ] [ rekado ] [ wingo ] <apteryx>10:29 [ bmpvieira ] [ emmanueloga ] [ jmi2k_ ] [ mt ] [ retroj ] [ Wojciech_K ] <apteryx>10:29 [ boegel ] [ erdic ] [ joachifm ] [ nckx ] [ RichardPaulBck[m] [ Xe ] <hodapp>maybe I can save this by pasting my credit card number by mistake <apteryx>10:29 [ book` ] [ exio4 ] [ jockej ] [ nee` ] [ roptat ] [ ym ] <apteryx>10:29 [ brendyn ] [ fedruantine ] [ joshuaBPMan_ ] [ nekomune ] [ sankey ] [ yrdz ] <hodapp>and drawing attention away from the mis-paste <ng0>can someone quiet apteryx for a short moment? <apteryx>10:29 [ cantstanya ] [ fkz_ ] [ JRHaigh ] [ newdan ] [ sapientech ] [ yrns_ ] <apteryx>10:29 [ catern ] [ fotis2 ] [ jsimonet ] [ ng0 ] [ Satyricon ] [ zacts ] <apteryx>10:29 [ catonano ] [ fps ] [ kelsoo ] [ NhanH ] [ shymega ] [ zmthy_ ] <apteryx>10:29 [ cbaines ] [ fr33domlover ] [ kensington ] [ nichi_ ] [ Sleep_Worker ] [ ZombieChicken] <apteryx>10:29 [ cehteh ] [ Gamayun ] [ kini ] [ OrangeShark ] [ sneek ] [ zw01 ] <apteryx>10:29 [ chipb ] [ gnunet_bot ] [ kmicu ] [ orbea ] [ soloojos ] <apteryx>10:29 [ CompanionCube] [ gnusosa ] [ kragniz ] [ OriansJ` ] [ sourcef ] <apteryx>10:29 [ copumpkin ] [ Gottox ] [ ktosiek ] [ Orin ] [ specing ] <apteryx>10:29 -!- Irssi: #guix: Total of 141 nicks [0 ops, 0 halfops, 0 voices, 141 normal] <apteryx>10:29 -!- Channel #guix created Sun Jan 20 18:45:45 2013 <apteryx>10:29 -!- Irssi: Join to #guix was synced in 1 secs <apteryx>10:29 -!- jmi2k_ [~jmi2k@17.92.217.87.dynamic.jazztel.es] has quit [Client Quit] <apteryx>10:30 < apteryx> Sorry, just found out that when I use "CTRL-BUTTON1" it kills the current window using RatPoison. Not sure if that's intended or not. <apteryx>10:31 -!- jmi2k [~jmi2k@17.92.217.87.dynamic.jazztel.es] has joined #guix <apteryx>10:31 < ng0> i think i can't help though.. i've never used uuid in file-systems part <apteryx>10:34 < catonano> jmi2k: no, I was not talking to you, <apteryx>10:41 -!- alezost [~user@unaffiliated/alezost] has joined #guix <apteryx>11:38 < efraim> isn't it at 12.01 on core-updates? <ng0>yeah can someone kick apteryx for a short moment <apteryx>11:39 < davexunit> efraim: oh cool. that's good to hear. <apteryx>11:39 < efraim> turns out vim-full needs python2 for some of the python tests, i'll just skip them by switching python-wrapper to python :) <apteryx>11:39 < davexunit> I bought a new AMD GPU that is supposed to have pretty good free drivers provided you are running linux 4.7+ and the latest Mesa <apteryx>11:39 < ng0> efraim: really? I removed python2 and it succeeded <apteryx>11:40 < davexunit> but I'm not absolutely sure that all of the AMD patches have made it into releases of both projects. just hearsay. <ng0>this looks like it will go on for some more minutes <apteryx>11:40 < ng0> i have not addressed all the replies on the vim-full thread yet. <apteryx>11:40 < ng0> like, should we switch to github releases and just keep up to date with that, so that vim-full can just inherit vim? <apteryx>11:41 < ng0> that was suggested as upstream points to github now <apteryx>11:41 -!- bill-auger [~quassel@75-138-187-221.dhcp.oxfr.ma.charter.com] has joined #guix <apteryx>11:41 < efraim> i took out the patches for now, we can figure out the patches separately <amz3>apteryx: you are on fire! <apteryx>11:41 < ng0> tehy are needed for the testsuite <apteryx>11:41 < efraim> two separate issues effectively, one for a new package, one for updating the source <apteryx>11:43 < efraim> if upstream wants us to use the github releases to get patched vim then I guess that makes the most sense <apteryx>11:43 < ng0> If this wasn't pointed out in the thread, I would've gone with just the patches <apteryx>11:43 -!- prasunanand [~Prasun@103.225.100.51] has quit [Ping timeout: 245 seconds] <apteryx>11:44 < ng0> dunno how gentoo does it.. i think with patches instead of github, but i could be wrong <apteryx>11:44 -!- jmi2k_ [~jmi2k@163.146.18.95.dynamic.jazztel.es] has quit [Quit: Leaving] <apteryx>11:44 < efraim> I don't know what debian does either <ng0>where are the ops when you need them... <apteryx>11:44 < efraim> 7.4 went over 2000 patches i think <apteryx>11:45 < ng0> ok, so gentoo uses github <apteryx>11:45 < efraim> with neovim around now I assume they'll go faster <efraim>this is a really long userstring <ng0>efraim: this must be the most unique fingerprint ever <amz3>so what's up in guix land? <amz3>what are the next dev that must happen? <ng0>apteryx: so what's the real issue with icecat now, apteryx ? <apteryx>Ok, before this paste mega-fail, I just wanted to share that the default useragent string for the IceCat package is showing that we're on the Windows platform. <amz3>I read there is guix ops and another thing I don't recall the name about getting guix on clusters <apteryx>Why would we advertise ourselves as using Windows? Isn't this detrimental to promoting GNU and open standards ;) <ng0>in a world where everyone would use Guix, setting it to guix would not be dangerous <jmd>apteryx: I think it's a privacy issue. <ng0>but setting it to guix would make the users very unique right now <apteryx>apteryx: Well Ubuntu is mainstream enought to hide into the masses without giving points to Windows. <ng0>apteryx: you might want to look into panopticlick <ng0>or what it was called <amz3>well, then why no use ubuntu UA? <ng0>i don't want this changed <ng0>if you want it changed, you can always create a local version <jmd>apteryx: I think the issue is, that *no* information should be divulged to websites. <Petter>There's also a plugin, Random Agent Spoofer. <Petter>You can set it to only use "Linux" browsers. <efraim>is -ffast-math an acceptable compiler setting? for all arches? <apteryx>jmd: Agreed. I guess in the current world leaving the useragent string empty would create issues when accessing some websites? <jmd>Possibly. And I don't think any other browsers have an empty useragent field. So that would defeat the purpose. <Petter>Yes, some sites will say "browser not supported" based on the user-agent. <apteryx>ng0: I guess I'm protected very well from tracking as even the EFF's panopticlick javascript won't be allowed to run by LibreJS :p <ng0>that does only imply you can not run the tests <ng0>if so, yes commiting whois with changes is okay for me <ng0>otherwise i have to getmail and reply <apteryx>Petter: Thanks for the Random Agent Spoofer suggestion. <apteryx>According to panopticlick, the User Agent used (Windows NT6.1 ...) is currently found on 1 out of 16 browsers they have in their DB (161969). <apteryx>Currently the easiest way to fingerprint my browser seems to be the "Hash of canvas fingerprint" which is pretty unique (I'm the only one to have it in all of their database!). This gives 17.31 bits of information according to them. <ng0>yeah those (canvas, webgl) are hashes I want to look at when I package torbrowser at some point <apteryx>canvas size... No idea but probably relates to the size of the tiling frame I'm browsing in? <apteryx>I'll try again when the browser is fullscreen on my 1920x1080 resolution. This should be more common. <ng0>apteryx: there should be stats on screen resolutions etc. <ng0>if you search for them <apteryx>Yeah screen size is always 1920x1080 (24), independent of tiling. <Petter>Wonder if anyone has hardcoded the report of the screen/browser size. <ng0>torbrowser (the one you can build on gentoo) defaults to 1000x600x24 btw <Petter>Does it report this value after a resize as well? <ng0>if you find a solution, talk to mozilla <Petter>function getScreenWidth() { return 1000; } <apteryx>Petter: I think the resolution detected is supposed to be your monitor resolution rather than related to your browser window size. <ng0>no, it's the size of your browser window <apteryx>ng0: window.screen.availHeight / availWidth says otherwise. <apteryx>Unless IceCat is confused by ratpoison which is possible of course :) <ng0>idk, and I don't care anymore right now. ***orly_owl_ is now known as orly_owl
<apteryx>Thanks for pointing the privacy issues. I can now understand that using the Windows user agent is better for privacy reason. <apteryx>Simply switching the Windows reference to GNU/Linux in the string made it unique. <ng0>50G should be enough for a VM where you also build things in and create a full system, right? i have no reference as I use way more here <roptat>how do you transform "1.16.0" into "1_16_0" in scheme? <ng0>i think gnurl package has or had this at some point <ng0>i don't know if it's the best solution <paroneay`>roptat: is it literally "replace every . with _"? <paroneay`>roptat: (string-map (lambda (c) (if (eq? c #\\.) #\\_ c)) "21.24.23") => "21_24_23" <roptat>(string-join (string-split "21.24.23" #\\.) "_")) <ng0>what currently is being used for the gnurl example I gave <paroneayea>roptat: it's probably a bit faster because it only walks over the list once, but it doesn't matter much probably <ng0>hm.. I need to update gnurl with the next curl release, I'll keep that in mind, updating the version string <marusich> I want to wrap text using a backslash followed by a newline as described in '(guile) String Syntax'. I'm using ParEdit to write my scheme file in Emacs. However, inside of double quotes, when I type \\ followed by enter, I get "\\^M". Why is this and how can I make it work like I wanted? <marusich>One workaround is to temporarily switch to text-mode and then back to scheme-mode, but there must be a more ergonomic way. <rekado>marusich: I cannot reproduce this. I’m using Emacs 25.1 with paredit. I don’t get ‘\\^M’. The line is broken instead. <ng0>efraim: gentoo even separates vim and gvim. I never used gvim that's because I forgot this <marusich>apteryx, that seems to work as I expected. <marusich>My enter key is bound to RET, if that makes any difference. <ng0>so I think vim, vim-full, and gvim (based on vim-full) should be a thing we could do because vim-full could be wanted to run on a system without X <marusich>"RET (translated from <return>) runs the command newline, which is an <marusich>interactive compiled Lisp function in `simple.el'." <marusich>So, inspection of ^M via C-x = reveals that this character is RET (codepoint 13, #o15, #xd)... I am not sure why that got inserted. <marusich>Are you using version 24 of paredit, rekado? <apteryx>marusich: It's probably a feature of paredit. "\\" trigger an escape sequence. <marusich>Oh, I think I see what's happening. C-h <return> says "RET (translated from <return>) runs the command newline", so probably somehow my return key is set up to behave as if I have pressed the RET key (maybe keybinds? I don't know much about those in Emacs). Therefore, in normal circumstances, pressing the return key will make it seem like I pressed RET, which inserts a newline by calling the 'newline' function. But when I type a backspace in <marusich>ParEdit, an interactive prompt comes up saying "Character to escape: ", so if I hit enter, I am probably literally telling ParEdit to escape RET, rather than a newline. <marusich>By "C-h <return>" I meant "C-h k <return>" <marusich>apteryx, I didn't know I could insert a newline directly with C-j; thanks for that info