IRC channel logs


back to list of logs

<kyamashita>quiliro: Is the package broken or the network transfer?
<quiliro>guix system: error: build failed: some substitutes for the outputs of derivation `/gnu/store/5fh6p3ywchsjp4zphyy6vlqpf80i72lz-e2fsck-static-1.42.13.drv' failed
<quiliro>every time
<quiliro>not with --fallback though
<grillon>Yeah!!!!!! My neovim is fully working. But I used virtualenv and the distro-python to achieve this.
<quiliro>but many package downloading with --fallback
<kyamashita>e2fsck-static works fine when I upgrade. It may be a i686-specific issue if it is not the network.
<quiliro>i have a good network now
<quiliro>so it is not a networking issue
<kyamashita>OriansJ: I have a faint feeling that I've seen this repo before. What type of assembly is this using?
<OriansJ>kyamashita: Knight assembly, it is part of the stage0 bootstrap
<OriansJ>kyamashita: Every piece is written targeting a platform independent VM that produces identical results on all hardware platforms
<OriansJ>it started with just a 280 byte Hex monitor and has grown a text editor, line macro assemblers and other extremely useful bootstrapping programs
<kyamashita>OriansJ: It's cool to see projects trying to address the bootstrapping problem.
<OriansJ>kyamashita: Actually, since the vm implementation is based upon actual running hardware (made out of TTL logic), it will solve it forever once done
<kyamashita>I like the sound of that!
<OriansJ>In the mean time, however the VM is only 41KB when stripped to the core or 98KB with all the bells and wishes on linux
<OriansJ>Everything required in the VM is in exactly 4 files (vm.h vm_minimal.c vm_instructions.c vm_decode.c)
<OriansJ>For a grand total of only 5039 lines of C code
<OriansJ>But for the bells and whistles version using gnu getopt and X11 logic, it is unfortunately bigger :(
<OriansJ>(5171 lines of code inside of the stage0 but still)
<OriansJ>and I just updated the bootstrapping steps to leverage the nicer to work with version
<kyamashita>I'm sure people can eventually code golf the finished implementation.
<OriansJ>kyamashita: possibly and the work this week involves building its first compiler
<OriansJ>also, I'm waiting on Janneke's feedback on how I can improve my lisp to better bootstrap his MES compiler
<kyamashita>Software development is still young, I'm sure we'll make the Maxwell's Equations of Software better over time. Right now, making bootstrappable and reproducible software is difficult enough.
<OriansJ>kyamashita: actually it is trivial to do given the constraints of the VM used in Stage0 because it becomes impossible to add any randomness.
<OriansJ>You have a program rom, a source file and a destination file and every single instruction is perfectly deterministic
<kyamashita>But you started the project with that in mind (and thank goodness you did). Not all projects do that, as many patches in the Guix tree indicate.
<OriansJ>kyamashita: sadly true, I honestly can't imagine why people shove things like time stamps and random garbage into their programs
<OriansJ>Yeah, I get malloc is slightly faster than calloc and when file systems didn't support time stamps your build system might need to shove it into the binaries
<OriansJ>but aside from legacy reasons, could there even be an excuse for that behavior anymore?
<kyamashita>I'm not sure, as I am the evern00b. However, I imagine there are plenty of places to stick metadata these days.
<OriansJ>kyamashita: you are absolutely right but how much metadata does one really require in a binary program?
<kyamashita>Shouldn't be much.
<kyamashita>Maybe projects could supply a "--build-reprodicibly" flag or something like that if they want metadata in the code by default.
<OriansJ>kyamashita: having thought about that for a bit; I don't think there is any valid technical reason, when given a specific compiler, source code and build command that the result shouldn't always be identical. That being said, only defects/variations in the compiler, build command or source code should change the outcome. Since guix allows us to perfectly always use a specific compiler, exact source code and build command, the remaining
<OriansJ>reproducibility bugs have to do with the source code interacting with non-deterministic defects in the tools themselves.
<OriansJ>The places I most find non-deterministic behavior are: floating point, when malloc is used instead of calloc, using randomness from outside sources (say reading from /dev/urandom) or performing fuzzy optimization tricks (the dumbest idea in the history of compilers if you ask me) for a 1-3% performance boost under ideal conditions.
<kyamashita>OriansJ: Interesting that you bring up floating point. I haven't thought about that before.
<OriansJ>kyamashita: haven't you noticed that x86 implicitly converts 32 and 64 bit floats to 80 bits and rounds it back for storage?
<OriansJ>it is a side effect of the design of the x8087 coprocessor
<kyamashita>OriansJ: I read recently about something like that.
<kyamashita>I need to read some CPU and instruction set manuals.
<OriansJ>kyamashita: There are even weirder rounding rules when you get into IEEE floating point support
<efraim>is that why we sometimes have rounding errors in math libraries for non-x86_64 architectures?
<OriansJ>There have been cases where minor variations floating point hardware and ordering make results order of magnitude different.
<OriansJ>efraim: It could definitely be the cause, as floating point is the best performance way to deal with large/tiny numbers
<efraim> i kept on running to similar bugs on aarch64 for gsl
<OriansJ>efraim: Oh yeah, that is a floating point calculation size assumption bug
<kyamashita>Darn. Just took a look at some physics, and it looks like infinite-precision real numbers are not physically attainable.
<OriansJ>kyamashita: There are few things in this universe that require more than 42 digits of precision and all of them are programs
<OriansJ>But honestly, if you are trying to optimize for performance your tests need to be better informed of the reality that not all hardware is x86
<OriansJ>And if you are a compiler writer and you put floats in your code, you should be beaten
<kyamashita>cd gcc-6.3.0/; grep -R float
<kyamashita>I can't find where these timestamps are coming from!!! (hunting through a haskell package-to-be)
<OriansJ>kyamashita: Where are the timestamps being written?
<OriansJ>If it is in the ELF header, check the compiler
<kyamashita>Directly into the binary, but it's not in the ELF header.
<kyamashita>They all seem to be after subdirectory names from the source tree though.
<kyamashita>Excepting the "--version" flag related code that I've already corrected.
<OriansJ>kyamashita: have you tried stubbing functions to find which of them it could be in?
<kyamashita>Stubbing functions? Like replacing a function with a "dummy" version?
<kyamashita>Whatever the meaning, not yet.
<OriansJ>kyamashita: It is my last ditch solution to finding non-deterministic code. Works every time but is a massive pain for big programs
<kyamashita>Looks like about 3MB of source code.
<OriansJ>Basically find the main and start removing trees until it builds reproducibly, then work your way down the tree and repeat until you find the function where it exists
<OriansJ>s/trees/function calls/g
<kyamashita>It's worth a shot.
<kyamashita>Oh wow. I think I missed an obvious aberration. I thought it was dead code, but apparently not
<kyamashita>Time for another reproducibility check. Hopefully the rest of the huge differences between the check builds were downstream from this one.
<OriansJ>kyamashita: If the code is dead code, it should be deleted code. If we ever need it back, it can be reverted or studied and reintroduced.
<kyamashita>The program is Darcs, and the suspicious code is something like "#define darcsBug(__FILE__,__LINE__,__TIME__,__DATE__)". The file was small and unassuming, but I'm going to...
<kyamashita>Definitely not dead code; the build just failed. -_-'
<brendyn>Is there a way to make this inherit less redundant
<brendyn>Since there are only 3 symbols that are different
<efraim>not that I know of for trivial-build-system
<brendyn>if i use a let statement outside package, will inherit copy that over, then i can use another let to overshadow it?
<brendyn>i suppose it doesn't even matter if it isn't copied over. but i havent seen a let used like that before
<kyamashita>brendyn: I see (substitute-keyword-arguments (package-arguments wpa-supplicant-minimal) used in some packages. Maybe you can use that?
<efraim>we normally end up with just define for that case, but I think it's fine with that level of substitution
<kyamashita>Well, that comes from admin.scm specifically, but that pattern appears.
<brendyn>efraim: how does one use define. i tried it like this but it says style is undefined
<brendyn>it seems making it a procedure is a mistake?
<brendyn>maybe i need quote and unquote
***jonsger1 is now known as jonsger
<ZombieChicken>Anyone here have experience using menu-entries in grub-configuration? I seem to be having some problems getting it to see another distro's kernel and initramfs
<ZombieChicken>Specifically, I'm having a problem with it not accepting it as a list, and I'm not sure where I'm screwing up at
***jonsger1 is now known as jonsger
<ZombieChicken>nvm. Found out the problem
<marusich>Even though my LANG environment variable is set properly to en_US.utf8, Python seems to think everything is ASCII. :(
<marusich>When I try to call locale.setlocale(locale.LC_ALL, '') it complains, "unsupported locale setting" even though supposedly that is supposed to work:
<marusich>I have no idea why, but the result is that programs like youtube-dl won't save files with non-ASCII characters, which is super dumb because my locale is en_US.utf8 and thus it can support all Unicode characters!
<marusich>Does anyone know why something like this might happen? It's very frustrating.
<roptat>hi, one package I try to get is licensed under this:
<roptat>CDDL is a free software license ( but lists only version 1.0, while the software I have is under CDDL 1.1
<roptat>also 1.0 is not compatible with the gpl, but the software still has a double license
<marusich>Regarding my earlier statement, it seems this might be some kind of regression in youtube-dl.
<marusich>I still don't quite understand how Python is getting its locale information, but it seems an older version of youtube-dl does not exhibit the problem I'm experiencing, so it must be a regression.
<ng0>is there an easier way than the hashsmokedlong-url i get for the tarball here, to use in (pypi-uri) maybe?
<ng0>I finally picked up the kallithea branch again, and I'm fixing up the test suite now
<ng0>okay. I can just use it like this... pypi-uri already has a version to use
<OriansJ>sneek: later tell janneke: You might find this interesting
<sneek>Got it.
<OriansJ>sneek: botsnack!
<ofosos1>hi ugix
<ofosos1>hi guix :)
<amz3`>hey i just discovered
<amz3`>it seems related to what guix pack does or something like ostree not sure
<amz3`>except guix pack doesn't handle sandboxing
<amz3`>as in containers
<amz3`>ofosos1: o/
<ofosos1>i got a problem updating recently:
<ofosos1>this problem occurs when i use the mirrors as well as when i directly use hydra. something broken in source code pro?
<grillon>hi there
<amz3`>ofosos1: doesn't it mean the mirror is corrupted?
<ofosos1>i think so, but it's hydra *and* the mirror
<roptat>other packages are also corrupted in the same way
<grillon>I have started to read guix manual. seems a very nice package manager. I have spend a day to make neovim pass the checkhealth command but it s worthwhile
<efraim>ofosos1: did you try adding --fallback?
<ofosos1>efraim: no, not yet
<ofosos1>efraim: it appears to do it, i'm going to have some ice cream :)
<ng0>does someone use gajim here? would you know if there is something not obvious to keep in mind when you install the extensions systemwide with a package manager? is the very obvious thing I fail to see that gajim needs to know about the paths?
<ng0>i started with a "catch them all" package for guix as there's a reported upstream bug for the single extension tarballs
<ng0>because they are not getting picked up I think there mustr be some way to communicate their store path to the parent package, gajim
<grillon>is there some people using guixsd?
<catonano_>grillon: several !
***kelsoo1 is now known as kelsoo
<ng0>uh.. funny. the maintainer of cairo-5c just told me its testsuite requires the bitstream vera sans font. iirc we don't have that, right?
<alezost>ng0: there is "font-bitstream-vera" package
<ng0>oh. okay, I'll try with this
<ng0>okay, now the improvement is that this font does not exactly match the one the testsuite wants
<rekado_>ng0: I use gajim occasionally. I don’t know about where it looks for plugins, but the source code should give you some hints.
<ng0>it's python based, and I've just seen the main application has a wrap phase
<ng0>the issue I talked about:
<ng0>so at least one other distribution is interested in it
<rekado_>ng0: we can use a git repository, though.
<ng0>that's what I do
<ng0>it would just be nicer to not have all addons but a selective choice
<rekado_>this can be done with “inherit”. You can create one package for one plugin and then inherit from it for all other plugins.
<ng0>yeah I don't know if it's worth the lines of code to remove all but the (package-name) folder everytime
<ng0>it's like 30 or 40 plugins
<ng0>all i wanted was omemo
<Azna>i'm trying to install guixsd with 2 luks encrypted partitions and unencrypted boot using this config
<Azna>i have to manually add 'insmod luks' and 'cryptomount -a' to grub.cfg, and the system hangs while booting
<Azna>uncommenting the second '(dependencies mapped-devices)' line, gives me the same error as in
<Azna>any ideas?
<Azna>the file system labels used are checked in /dev/disks/by-label to be correct, as well as the uuids
<ng0>i think there's a bug I opened on this
<ng0>could be this:
<ng0>if you can deal with one partition, and mount the other one later I'd suggest this as a work around
<Azna>ng0: could this be done with a service? or do you mean manually?
<ng0>eventually someone will fix the bug, but for me it's not grave enough as I can boot from one drive and mount when I need others
<ng0>it's still annoying
<Azna>ng0: i'll try that, thanks
***mo is now known as Guest57900
<catonano>I try to clone this repo and git replies that such repo does not exists
<catonano_>does it exist or doesn' t it ?
<piotr>Hi there!
<piotr>I'm trying to boot from the USB stick. I've followed instructions in the manual. Everything goes smoothly and at the end I get a partition named "gnu-disk-image" with flag "boot". Sounds good. When I mount it it's a classical GNU / Linux root however when I reboot the key isn't recognised by my laptop so I'm unable to install it :'( Any ideas how I could solve that?
<Gamayun>Could just be a funky bios if everything else looks good. I have a netbook that will only recognise images on about half my usb sticks.. :/
<catonano>it's an g repo
<sapientech>trying to use offload facility, got to `guix offload test ~/.machines.scm` but getting: guix offload: error: failed to load SSH private key from '/home/sapientech/.ssh/id_ed25519': Unable to import a key from the file
<sapientech>this is the ssh key i use to connect to the build machine
<sapientech>the info on shows all these commands are with root... thats confusing since the machines.scm specifies the user
<sapientech>and i only have an ssh key for the user. does this mean i have to make an ssh key for my root account and connect that way?
<piotr>Well, I tried with Ubuntu on the same stick and it worked like a charm :/
<piotr>How could I get a iso file from the archive provided by guix?
<pmikkelsen>hi guix
<pmikkelsen>do you know why emacs never can find my fonts after i run guix gc? I'm on guixsd by the way
<piotr>Where do you set Emacs font?
<pmikkelsen>in my .emacs.d/init.el
<sapientech>update: is it because this key is passwd protected?
<piotr>Perhaps gc triggers a hook?
<kyamashita>pmikkelsen: Usually I upgrade my packages after garbage collecting. That tends to resolve the issue for me.
<kyamashita>It'd be better if this didn't happen, though.
<pmikkelsen>kyamashita: i'll try that, thanks.