IRC channel logs


back to list of logs

<sideEffffECt_>Hi, I've posted the news about the new Guix release on reddit, so if you have an account, you might wan't to keep an eye on it, since there will be people asking questions about Guix, and I'm sure it will make them happy if you answer those for them
<zacts>lo #guix
<phant0mas>okay I fixed the lihash not found issue , changing some small things and start sending patches
<phant0mas>good night
***tschwing_ is now known as tschwinge
<civodul>Hello Guix!
<phant0mas>civodul: good morning
<phant0mas>I fixed the lihash not found error
<phant0mas>I just had to create a cross_library_path env var
<viric>hi civodul !
<phant0mas>now I just have to make sure it gets built with the i686-pc-gnu gcc gcc
<phant0mas>and I will start sending patches for the glibc
<viric>are we pwed?
<viric>talking about glibc maintainers. :)
<phant0mas>what happened with the nsa has made my stop trusting anything that runs binary blobs on it, even the hardware
<phant0mas>made me*
<viric>well, assange talks about Red Had being pwned
<viric>when the software you run are dozens of milions of lines of source code, it begins to look like a binary blob
<viric>specially in a world where "bugs" can be introduced as if unintentional
<viric>everyone hopes good will in foss
<civodul>viric: i think this blog post is essentially bullshit
<civodul>i haven't found Assange's complete talk
<civodul>but i think it's more interesting to listen to what Appelbaum has to say
<civodul>because he really knows all this very well
<viric>as this modern times, I haven even read it
<viric>only the title.
<civodul>nicely said :-)
<civodul>oops, typo ;-)
<phant0mas>civodul: don't check the glibc work till now patch I sent yesterday
<phant0mas>I am gonna clean the source, break it to smaller patches and I will start sending them for merging with the upstream
<civodul>phant0mas: i didn't look at it yesterday, because i focused on that release thing ;-)
<phant0mas>I know I know :-)
<civodul>but i just wanted something i can build to reproduce the "undefined reference" problem you were having
<civodul>it doesn't matter if it's not "clean" etc.
<civodul>so what would you recommend?
***civodul changes topic to 'GNU Guix --- --- 0.6 is out!'
<phant0mas>the easiest way , just clone my gitorious repo
<civodul>ok, i have it around
<civodul>i'll get back to you when i get there
<phant0mas>but I think it may have been caused from the wrong version of libpthread in the store, because I can't reproduce it now
<civodul>ah, even better then :-)
<Fulax>civodul: the gnu-system image is neat, thanks.
<Fulax>but I didn't figure how to get root to play with dmd
<Fulax>running `deco' as `guest' lead to the issue/patch I reported
<civodul>inspiring talk:
<civodul>(courtesy of ArneBab :-))
<ArneBab>civodul: courtesy of someone on twitte whoose name I forgot ☺
<viric>civodul: do you play twitter?
<civodul>viric: nope i don't
<ArneBab>civodul: if you should ponder using twitter, I’d rather go for gnu social. That can forward to twitter and it is actually nicer for discussions.
<civodul>i play #guile, much better :-)
<civodul>no worries here
<viric>is gnu social a program?
<civodul>it's not completely clear:
<viric>that's why I asked :)
<viric>civodul: trendy. :)
<Digit>afaict, gnu social is kinda riding the coat-tails of the likes of freedombox, and the likes of diaspora... long coat-tails.
<davexunit>gnu social is statusnet.
<davexunit>I don't think statusnet was riding any coat tails. remember before the change? it was pretty popular.
<civodul>viric: arrrgh
<civodul>"riding coat tails"?
<Digit>that maybe comes across more harsh than i intended
<davexunit>gnu social was started, and then it merged with the statusnet project and continued under that name. the statusnet developer moved on to develop and the statusnet project is once again known as gnu social.
<viric>statusnet is a bunch of php files in an apache server
<davexunit>it's run by former FSF campaigns manager Matt Lee
<Digit>thnx davexunit. didnt, know that, n likely wouldnt have ever bothered to learn that. so, yeah, thnx for dispelling the misconceptions and misinfo. :)
<viric>was statusnet a one-man thing?
<ArneBab>viric: see
<davexunit>viric: It was primarily developed by Evan Prodomou
<ArneBab>I run GNU social on my own server. An advantage is that it actually works on commodity servers. It’s not rreally fast, but it works.
<ArneBab>(on my own cheap server)
<viric>ArneBab: by now I'm in civodul situation. :)
<ArneBab>you can also just use existing servers
<viric>not my style
<ArneBab>viric: the advantage of GNU social over Twitter is that in GNU social many servers can interact (it’s a federated system)
<ArneBab>viric: that’s fully OK ☺
<ArneBab>viric: the advantage of GNU social over Twitter is that in GNU social many servers can interact (it’s a federated system)
<civodul>"federation" is a lie
<ArneBab>(sorry for the duplicate: keyboard fired thrice)
<ArneBab>civodul: why?
<civodul>we know than 90% of users will end up using,, etc. etc.
<civodul>systems must be inherently distributed
<viric>yes, federation isn't anything very good
*ArneBab won’t use - doesn’t work on commodity servers)
<viric>it's something slightly better than centralized
<viric>but that doesn't make it good
<ArneBab>federation is E-Mail
<viric>exactly, email is horrible
<ArneBab>I think that email actually is really good.
<davexunit>I don't know, I think federation is good.
<ArneBab>if we all GnuPG-signed our emails (making E-Mail actually trustable), then E-Mail could be a basis for any kind of service)
<viric>If you pay a "good IP", a good "DNS entry", ... you can achieve to run your own server to participate in the "everyone sending in plaintext" and "all but you having the mailbox in a single company" model.
<civodul>i mean the idea behind federation is certainly good
<viric>And signing every word you send isn't very good either
<civodul>but it's so much easier to use an existing instance of the software, that it doesn't really help, IMO
<ArneBab>when I look at the E-Mail addresses of my offline friends, only few use theh same hosters.
<viric>you can be blamed later.
<ArneBab>viric: sure. If you want non-accountable speech, you need at least OTR or better still: Freenet
<ArneBab>civodul: the trick with federation is that you can pay a local group to host for you
<viric>do you see? email is quite bad
<ArneBab>viric: why? I do not need non-accountable speech in E-Mail
<viric>I have higher expectations
<ArneBab>viric: Add *working*, *interoperable* GnuPG (signing+encryption), and we have a really good system.
<viric>I run my own mailbox
<zerwas>I like Bitmessage as an alternative to email
<viric>but I've to pay a MTA, if I want other MTA to accept my mail
<ArneBab>do you use GnuPG for more than 50% of your emails?
<davexunit>civodul: the mediagoblin lead developer is very aware of that and he's been talking a lot about how to make it easier to run an instance of a web service.
<viric>I don't thing gnupg is comfortable to use.
<ArneBab>viric: that’s because we do *not* have widespread GnuPG. Otherwise I would not get spam, because my client would discard any message without a valid, trusted signature.
<ArneBab>viric: and you’re right in that. The bad usability of encryption and signing is the real problem of E-Mail
<viric>the current state of email is quite sad
<viric>despite any "it could be much better if everyone..."
<ArneBab>viric: except for one part: That you cannot hide the metadata is an actual protocol issue (who sends the E-Mail to whom). Though there are ways around that (anonymizing repeater networks)
<ArneBab>viric: then you do not actually talk about email as protocol?
<viric>email as the current communication system named email
<ArneBab>viric: for GnuPG this is something I’m excited about:
<zerwas>By the way, Bitmessage eliminates the metadata problem
<ArneBab>viric: the webmail interfaces or all the existing email clients?
<ArneBab>zerwas: that’s interesting. Does it require paying for sending emails?
<viric>I mean how people end up communicating with email
<ArneBab>(interesting in a good means)
<zerwas>ArneBab: yes, computation time
<zerwas>ArneBab: the client must do a proof of work when sending a message
<zerwas>It helps to mitigate spam
<viric>that's a jungle where the strongest wins
<zerwas>in terms of spam?
<viric>my quick thought is that it's a too heavy solution for a little win
<ArneBab>viric: well, most people use email in the end
<viric>I don't have a spam filter... I don't receive so much spam
<zerwas>viric: If you really really don't want spam you can increase the demanded difficulty in the settings easily
<ArneBab>viric: on email: ☺ (a bit outdated nowadays, since Facebook has risen to enormeous size, but still quite close)
<ArneBab>zerwas: in freenet we use a web-of-trust: You can be seen if either you solved captchas of the receivers or people the receivers trust marked you as not-spammer
<ArneBab>zerwas: in money normal people will always lose (unequality is much too strong nowadays)
<viric>xk3d? I didn't know the 3d :)
<ArneBab>I didn’t know it either, but google did ☺
<zerwas>ArneBab: I don't think that's the case with Bitmessage
<zerwas>And I don't think Freenet's WoT is very convenient. If it stays this way it will never take off
<ArneBab>zerwas: the WoT in freenet works really well - if you ignore the slow-as-hell implementation.
<ArneBab>zerwas: For the interface I might mentor a GSoC project this year (better webinterface)
<viric>I quite like the freenet web interface
<ArneBab>viric: it hes a huge number of rough edges ☹
<zerwas>viric: The current interface does not have good usability though
<zerwas>It's easy to use for you and me, but it's too complex for the average joe (I've seen users not understanding it more than once)
<ArneBab>viric: the hardest part of my work as mentor will likely be to keep the project in scope: Ensure that the work actually helps the users
<ArneBab>(and does not get stuck in half-finished state)
<viric>it sounds hard, yes
<civodul>davexunit: sure
<civodul>davexunit: personally i'm interested in supporting projects that address the root causes of the problems
<civodul>that includes GNUnet, Tahoe-LAFS, Tor, and a few others
<zacts>what is Tahoe-LAFS?
<civodul>i remember discussing that with Christopher on the GNU mailing list ;-)
<zacts>Gnunet is neat, I want to eventually contribute
<civodul>and i should mention Twister, of course
<zacts>do any of you use mosh?
<zacts>I'm wondering if that would be a worthwhile project to contribute to also..
<zacts>oh and I need to check if you've ported mosh to guix
<davexunit> ?
<davexunit>civodul: I think mediagoblin has plans to use tahoe lafs
<davexunit>civodul: so bascially, in order to achieve real decentralization, things should be peer-to-peer?
<civodul>davexunit: yes, i think so
<civodul>great if they use Tahoe-LAFS
<davexunit>I'm excited for the p2p features planned for guix. :)
<ArneBab>zacts: Tahoe-LAFS is something like Freenet, but instead of censorship resistant publication they want to provide decentral storage.
<ArneBab>zacts: the core difference is that when Tahoe-LAFS is full, you cannot store more, while when Freenet is full, it starts to overwrite the least recently accessed data.
<ArneBab>zacts: if you have detailed questions on freenet, you can ask them in #freenet ☺
<zacts>davexunit: indeed
<zacts>davexunit: oh cool p2p features?
<civodul>davexunit: though we're not there yet either ;-)
<zacts>davexunit: we should port mosh to guix.
<davexunit>zacts: probably wouldn't be very hard.
<zacts>davexunit: maybe I'll try it this evening, but if you get to it before me, then less work for me. =)
<davexunit>I won't be able to get to it.
<davexunit>going to be busy with other things the next few days.
<zacts>ok, I'll mark it on my list of things to remember todo today
<viric>mosh was easy to package in nix
<civodul>let's "guix import /.../nixpkgs mosh" :-)
<zacts>civodul: can you do that?
<zacts>it would be nice, as I can focus my time on other things
<civodul>i'll see what i can do
<civodul>interesting thoughts on marketing:
<zacts>cool civodul
<Fulax>civodul: thanks for your comments about my tentative patch. However, it seems there are some encoding issues in some sources files, see the copyright notice here:
<civodul>thanks for the patch, Fulax
<civodul>yeah i noticed it's latin-1
<civodul>should recode them
<Fulax>actually, half files are utf-8
<civodul>oh, i only half-borked it then :-)
<civodul>damn, it seems we/i kinda fail to spread the word about the new release
<davexunit>civodul: I posted to HN but got nothing
<davexunit>and posted to twitter where wingo retweeted me.
<davexunit>I guess we need to find better ways to get attention for guix.
<civodul>yes, definitely
<civodul>i'm bad at that
<civodul>i wonder how to attract people to contribute to this part of the job
<Digit>distros that use guix by default... that'd get some attention
<atheia`>I don't know — I think the next big splash would be able to install guix straight on machines.
<davexunit>Digit: well guix will be the distro.
<atheia`>My thoughts exactly Digit
<civodul>yeah, we really need to boot now
<Digit>kit em out with dev tools, n less windows-refugee conveniences.... that'd attract the developer userbase. :)
<atheia`>oh yeah… I thought digit meant what I said ;-)
<davexunit>I don't know so much about what it takes to get an OS to boot. what are we missing?
<Digit>i thought guix was a packagemanager++. now it /is/ the distro? o_O
<civodul>dmd robustness, then things like support for suid binaries, using udev, etc.
<civodul>Digit: actually the distro is simply called "GNU"
*Digit clearly hasnt been paying attention
<Digit>i call gnu an os, of which, there are many distributions
<davexunit>this will be the official GNU OS.
<Digit>but i do like the idea i heard oft espoused that guix become like "the" gnu package manager.
<zacts>civodul: guix, as a distro, will be called GNU?
<zacts>not gnu guix?
<civodul>well, strictly speaking: "GNU Guix is the package manager of the GNU system"
<civodul>or "the GNU system uses GNU Guix as its package manager"
<civodul>see? :-)
<zacts>ah cool, so we will have our first gnu system
<zacts>especially using linux-libre
<zacts>gnu linux-libre
<civodul>that's the idea
<Digit>or hurd. mmmm.
<Digit>some day...
<zacts>Digit: there is a project to use the FreeBSD kernel and parabola
<zacts>that would be an interesting project for guix eventually gnu/kNuBSD-Fire
<Digit>oh cool. glad parabola are getting in on that sort of thing. :)
<phant0mas>Digit: hurd soon ;-)
<zacts>phant0mas: oh yeah you've been doing lots of work on that right?
<Digit>i'm fascinated by the possibilities of bedrocklinux, and have this notion that perhaps i could get guix into bedrock core, to help manage all the client distros.
<zacts>Digit: bedrock is an interesting concept
<phant0mas>zacts: later today after I finish today's lessons I will start sending the glibc/hurd patches
<Digit>i hope its possible for anyone to take guix and install it in *whatever GNU+'Linux|Hurd|*BSD*'
<zacts>Digit: that would be cool!
<zacts>I'm also interested in minix3 for educational reasons, it would be cool to put gnu on top of it also, once it matures a bit
<zacts>there are no threads / job control / etc..
<davexunit>Digit: it's easy enough to use guix to supplement the host system's software.
<Digit>minix has been frustratingly ever pushed farther back on priorities. been wanting to delve into it for a long time... pesky illusion of necessity of productivity keeps getting in the way though.
<zacts>Digit: see my blog
<Digit>thnx davexunit. that gives lift to my dreams. :)
<zacts>there is a tutorial on how to import commands into minix3, we are making progress
<zacts>and there is a release 3.3.0 that will happen soon. (with tetris / bsdgames / tmux / etc..)
<zacts>I'm also working on porting schemes to minix3
<Digit>nooooooo! stop! XD i cant get sucked into minix now. i'm only half way through setting up my first client os of my latest bedrock install. cant go running off into that intensely enticing geekfest.
<zacts>I eventually would <3 to work on a HURD like system and apply what I learn to that.
<zacts>but minix3 is a really great educational OS, and I love the book that goes with it
<zacts>the OS core is so tiny the entire src fits in the back of the book
<zacts>but guix will be my first choice for a gnu/minix distro
<zacts>x15 / propel looks cool too, but braunr has been too busy to work on it.
*Digit dreams of a minix/busybox/ii/tinywm/nanoX/{insert other essentials in such clear brief code that one can have an entire OS and have read every line of its code... preferably only over a weekend, if even that.}
<zacts>heh I tried to make a perl cli interface to ii once
<zacts>Digit: minix3 has been ported to the beagle bone black
<Digit>s/busybox/lazybox/ and s/nanoX/femtoX/ (for imaginary versions that are even lighter on the LOC)
<zacts>I would wait for the 3.3.0 release and then try it out.
<Digit>yup, ii is a great little itch. :) i made some little wrapper scripts for it myself. intend to flesh it out further in time.
<zacts>chibi scheme works on minix3 when statically linked
<phant0mas>in order to finish building glibc, libpthread must be built,which needs a basic hurd lib,which in turn needs glibc
<phant0mas>I had some tnt somewhere around
<phant0mas>I will have to build an intermediate glibc , build libihash,and rebuild glibc with libpthread
<zacts>phant0mas: nice! =)
<zacts>I think it's really cool you are working on guix + hurd