IRC channel logs

2014-01-24.log

back to list of logs

<civodul>Hello Guix!
<jmd>Hi ludo
<jmd>After much procrastin^H^H^H^H^H^H^H deliberation I have decided this year not to partake in the annual pilgrimage to Brussels.
<civodul>jmd: oh, too bad
<zacts>hi guix
<zacts>what are some high priority newbie projects I can work on?
<jmd>zacts: Packaging
<civodul>zacts: indeed, and using it, reporting problems, adding any packages that you need and are missing
<civodul>plenty of things to do :-)
<civodul>there's also a 'TODO' file, but that's perhaps more advanced and assuming knowledge of Scheme
<jmd>civodul: Is there a list of already packaged software?
<viric>jmd: I think that's a good task for a newbie. Know where to find the packaged software. :)
<civodul>indeed :-)
<civodul>jmd, zerwas: there's this list http://www.gnu.org/software/guix/package-list.html
<civodul>updated daily
<jmd>civodul: BTW, regarding the discussion about blurbs. It would be good if they were synced to directory.fsf.org
<civodul>this has been discussed with Karl, but it's pretty hard to export from there
<civodul>it nearly impossible to do non-interactively
<jmd>Yeah I thought that might be an issue.
<civodul>it's in 'TODO' though, with some details
<civodul>BTW, on listing available packages: http://www.gnu.org/software/guix/manual/guix.html#GNU-Distribution
<zacts>I am learning scheme, but I'm by no means an expert yet
<zacts>I'm still reading Scheme and the Art of Programming
<jmd>A gentle introduction which can give satisfying results is to create a guix package. You can copy an existing one and start from there.
<zacts>jmd: ok thanks! :-)
<zacts>will gnunet be used for DNS by default on the guix OS?
<zacts>s/OS/distro/
<civodul>zacts: GNUnet will be used for GNS
<civodul>i guess
<civodul>and for other things too, hopefully, like distributing binaries
<civodul>but we're not there yet
<jmd>I wish I understood the "big picture" with gnunet.
<viric>it's tricky to understand gnunet
<zacts>jmd: I know nothing about gnunet, I've just discovered it recently
<viric>GNS is a decentralised naming system
<viric>not in the lines of namecoin, but giving the user full awareness of the decentralisation
<viric>so you have your network of friends defining name-address relations, and you ask that way.
<viric>website.bob.ludo.zacts.gns or something like that.
<jmd>viric I understand that much. But GNS is just one aspect of gnunet
<viric>sure. gnunet also provides point-to-point secure transport (mesh), a DHT for shared content (kind of p2p file sharing) with optional anonymisation, ...
<viric>there is a vpn thing, but I don't understand it yet. For what I understand, it's motly to get access to "the internet" through gnunet, with some gateways in there.
<civodul>speaking of GNUnet: hi sriharsha :-)
<sriharsha>Hi civodul
<sriharsha>let me go through the logs
<civodul>viric: it's more like setting up an open VPN, unlike GVPE which allows you to setup a closed VPN
<civodul>heh :-)
<viric>civodul: yes, but there is not any internal IP address system
<viric>I think that there is some part touching iptables, to map some ipv4 addresses to the proper target, adding and removing dnat
<civodul>ah yes
<sriharsha>viric, there should also be support for tun interfaces
<viric>tricky
<civodul>well it sets up a tun/tap interface with "local" IP addresses
<civodul>right
<sriharsha>yes
<viric>but there is not an internal ip namespace
<viric>ip subnet
<viric>that's my main difficulty with it :)
<sriharsha>i never tried it though
<sriharsha>but grothoff on #gnunet once demonstrated running a website over VPN with GNS name resolution
<viric>it's like extending ipv4 address meaning, while keeping ipv4 addresses.
<viric>by tricks of iptables.
<sriharsha>better ask him on #gnunet
<viric>for me it's very hard to take that as something reliable
<viric>sure
<viric>I asked that to him more than once I think. But I still can't take it as something reliable, although he explained. :)
<sriharsha>ok
<viric>I still think it'd be easier to take some ipv6 subnet, as many vpn systems do.
<viric>but so I guess I don't understand it well :)
<sriharsha>gstreamer-1.0.10 testcase fails on my computer
<civodul>sriharsha: how?
<civodul>oh, i see your message
<sriharsha>civodul, I get this when the testcase fails
<sriharsha>Unexpected critical/warning: Source ID 1 was not found when
<sriharsha>attempting to remove it 90%: Checks: 10, Failures: 1, Errors: 0
<sriharsha>gstcheck.c:72:F:stresstest:test_watch_with_custom_context:0:
<sriharsha>Unexpected critical/warning: Source ID 1 was not found when
<sriharsha>attempting to remove it FAIL: gst/gstbus
<sriharsha>ah, ok
<civodul>that's x86_64, right?
<sriharsha>yes
<sriharsha>is it known to fail?
<civodul>sriharsha: i don't think so, but i just reproduced it locally
<jmd>It would be nice if guix download had some confidence indicators
<jmd>Err.. Is perl an implicit dependency somehow??
<atheia>G'day all.
<jmd>Is there any preference between a patch file, and use of (substitute ..)
<jmd>?
<civodul>jmd: oooh, Octave!
<civodul>congrats on that one :-)
<jmd>np.
<jmd>regarding your feedback on my git comments:
<jmd>Some weeks ago I remember you telling me NOT to put the rationale for a change in a commit message (which I thought was odd). Now you are telling me that I should. Am I missing something?
<civodul>no no, i'm telling you to put it in a comment, not in the commit log
<civodul>i'm pretty consistent about it ;-)
<civodul>it's just so that future readers know why things are this way
<civodul>like "foo.h includes bar.h so it need to be propagated"
<civodul>that's enough
<civodul>that's the whole thing about packaging: it often takes a lot of time, but the end result is just 10-20 lines
<jmd>Ahh ok. So you're happy with the commit message I wrote?
<civodul>so my opinion is that knoweledge acquired during that time must be shared in comments
<civodul>yes, perfect :-)
<jmd>ok. I certainly agree with your other statements.
<sriharsha>which license from the ones in licenses.scm is best suited for liberal MIT license?
<zerwas>sriharsha: I *think* that's "x11"
<zerwas>sriharsha: see http://lists.gnu.org/archive/html/guix-devel/2014-01/msg00018.html
<sriharsha>ok, thanks