IRC channel logs

2014-07-02.log

back to list of logs

<DusXMT>Yas, that was the problem. After rebuilding guild and properly testing it beforehand, guix seems to work just fine
<DusXMT>s/guild/guile/
<bavier>DusXMT: I'm glad things worked out
<bavier>let us know if you have any other questions
***jmdx is now known as jmd
<jmd>How can I discover what a built package references (if anything).
<civodul>Hello Guix!
<jmd>Hi
***civodul changes topic to 'GNU Guix --- http://gnu.org/s/guix/ --- 0.7 is in the works: https://lists.gnu.org/archive/html/guix-devel/2014-05/msg00110.html --- testing needed: http://lists.gnu.org/archive/html/guix-devel/2014-06/msg00154.html'
<davexunit>civodul: hey. I haven't gotten the chance to test that patch for the install image on my desktop computer yet. I hope I'll have some time tonight but I'm getting ready to go on vacation so it be a few days before I can test.
<civodul>davexunit: np, and thanks for testing
<civodul>it's on real hardware that we uncover real problems :-)
<davexunit>I was pleasantly surprised to see a guile backtrace in my kernel panic :)
<civodul>it's only pleasant the first time you see it i'm afraid ;-)
<davexunit>it's cool that the initrd is scheme, though.
<davexunit>civodul: so, it looks I've been given the opportunity to give a brief presentation about Guix next week at a hackathon.
<davexunit> http://www.open-bio.org/wiki/Codefest_2014
<civodul>davexunit: that's cool!
<civodul>the event is very much "open source"
<davexunit>yeah
<civodul>you'll have to mention user freedom ;-)
<davexunit>yes, I will.
<davexunit>it's only a short talk, which is good for me because I've never done this before.
<davexunit>but I need to figure out how to present Guix to capture their attention.
<civodul>yeah
<civodul>i'd be good to ask Pjotr what he thinks makes Guix interesting for Cloudbiolinux
<civodul>*it'd
<davexunit>yeah, I've been in contact with him.
<civodul>ok
<davexunit>I think it's that these developers could benefit greatly from reproducible software installations for dealing with large amounts of data and servers.
<civodul>yeah, i was thinking that reproducibility must be valuable in experimental setups
<civodul>that and control over the packaging details, customizability
<davexunit>when I have a few slides prepared, would you mind reading them?
<civodul>sure, np
<civodul>you can reuse stuff at http://git.savannah.gnu.org/cgit/guix/maintenance.git if you want
<davexunit>yeah, I was thinking of borrowing some of that.
<jmd>What's the next thing that needs packaging?
<davexunit>I tried packaging glib-networking last night, but make install tried to copy files into glib's store directory. Not sure the best way to deal with it. :(
<jmd>davexunit: I had that issue with gnumeric
<jmd>I redirected it to a subdir in its own build
<jmd>I wonder how we should educate upstreams not to do such things?
<davexunit>jmd: thanks, I'll check gnumeric to see how it was done.
<jmd>Why does http://hydra.gnu.org/jobset/gnu/master say "with errors" ?
<jmd>844 jobs failing sounds serious.
<bavier>civodul: you beat me to some of those other glib modifications
<civodul>hey bavier
<civodul>bavier: oh, sorry then
<bavier>np, I still wanted to check that everything builds with the upgraded glib
<jmd>There is a conflict between glib.h and libjpeg.h
<bavier>jmd: oh?
<jmd>glib.h contains #define FALSE (0)
<jmd>and jpeglib.h contains enum {FALSE = 0, TRUE = 1};
<bavier>nasty
<jmd>How can I debug a something in a build ?
<bavier>jmd: you can use the --keep-failed option of 'guix build'
<jmd>Yes. And then??
<bavier>from there, you can source the environment_variables file to "recreate" the environment that the source was being built in
<jmd>But that environment builds without debug symbols
<bavier>you should be able to add additional compiler flags when building there
<jmd>how can I tell gdb the correct path to the source?
<bavier>I'm not so skilled with gdb, but I think you can use the -d flag
<jmd>What is the format of the pattern in subsitute* ?
<jmd>Why does (substitute* "tools/img_cmp.c"
<jmd> (("g_object_unref (buffer.);") ""))
<jmd>
<jmd>do absolutely nothing?
<civodul>hello!
<civodul>jmd: because of (buffer)
<civodul>where parentheses are interpreted as subexpression grouping
<civodul>that must be changed to: "g_object_unref \\\\(buffer.\\\\);"
<civodul>(assuming the dot is really needed)
<jmd>Thanks
<jmd>Why doesn't it work like sed?
<civodul>jmd: it uses so-called extended regexps by default, unlike sed
<civodul>but i think it's a good thing
<jmd>Oh the nice thing about standards is that there are so many to choose from!
<civodul>:-)
<jmd>Bed time
<sirius>Is there a way to prevent guix from deleting the build directories in /tmp when a build failed?
<bavier>sirius: Use the --keep-failed options
<bavier>*option
<sirius>Thanks for the fast answer :)
<civodul>good night/day!