IRC channel logs

2014-04-08.log

back to list of logs

<civodul>phant0mas: that would be very surprising; if you have a test case, i'm happy to test tomorrow
<civodul>Steap: Hydra will restart in one hour
<civodul>good night/day! :-)
<Sumi>Hello , I have a basic knowledge of C and linux programming , I want to contribute to your project , how can I ,
<civodul>Hello Guix!
<phant0mas>good morning guix
<civodul>hey phant0mas
<phant0mas>did you check the email?
<phant0mas>I am going to find a proffessor to ask him about something and I will be right back
<civodul>ok, haven't checked email yet
<civodul>but i saw youpi this morning and he told me "wtf is going on with git?" ;-)
<phant0mas>hahaha
*civodul upgrades OpenSSL: http://www.openssl.org/news/secadv_20140407.txt
<civodul> http://s3.jspenguin.org/ssltest.py
***tschwing_ is now known as tschwinge
<mark_weaver>we need a better way to patch security problems in core libraries :-(
<mark_weaver>doing it the normal way, it will probably take me at at least a day or two to get my YeeLoong secure again.
*mark_weaver considers mutating his openssl-1.0.1f store
<bavier>mark_weaver: it's the rebuild times that are a problem?
<mark_weaver>yes
*mark_weaver goes offline awaiting the rebuild. I might be offline a long while...
<civodul>hey sriharsha
<sriharsha>Hi
<civodul>sriharsha: are you going to upgrade GNUnet in Guix?
<sriharsha>I shall give a try, if all the tests work out fine, I shall post a patch; if not, I only have time to fix the tests next week.
<civodul>ok
<civodul>davexunit: man-db is ready to push, no?
<davexunit>civodul: seems to be. I'll push once I'm home from work.
<davexunit>in about 2 hours.
<civodul>cool!
<civodul>i think i'll do the release tomorrow morning (UTC)
<davexunit>awesome.
<davexunit>that will be great.
<davexunit>I was hoping to have notmuch packaged, too, but man-db proved difficult.
<civodul>yeah, good job on that one
<civodul>i looked at it once, and postponed forever :-)
<davexunit>as always, mark_weaver helped me immensely. :)
<davexunit>and you too, civodul. so thanks!
<civodul>me? :-)
<davexunit>my development environment is much improved now that I can test with only guix packages in my PATH and such.
<civodul>ah, that's good
<civodul>eating one's own dog food is a great experience
<davexunit>yeah, it helps a lot. it's also taught me a lot about the utility programs that I take for granted
<davexunit>"what do you mean 'less' cannot be found?"
<civodul>heh :-)
<phant0mas>civodul: firstly I did what you told me , and added an #error message after the else, and it seems it never uses the empty definitions of define hook
<phant0mas>secondly when libpthread asks for -lihash does it mean libihash? Because I have already created a hurd-minimal package and added it as a propagated input
<civodul>phant0mas: could you send me the patch and command line for the first issue, so i can look tomorrow?
<civodul>for the second issue, yes, it wants libihash
<civodul>we need to see why it isn't found
<phant0mas>my hurd minimal package builds libihash only
<phant0mas>maybe it wants more libs to be built
<phant0mas>or it justs looks in the wrong place
<phant0mas>I am checking it out
<phant0mas>LS
<phant0mas>wrong buffer :P
<phant0mas>civodul: for the first issue it never entered the else part because it didn't fail while building
<phant0mas>for what should I look in libc.so to confirm that?
<civodul>phant0mas: could you send me the patch and command line for the first issue, so i can look tomorrow? :-)
<phant0mas>you mean the patch for the glibc? of course
<civodul>just enough to i can reproduce the "undefined reference" issue
<civodul>s/to/so/
<phant0mas>of cource
<phant0mas>btw about the git issue I had , I totally forgot that store doesn't change packages unless the hash changes
<civodul>yeah, that happens ;-)
<bavier>do the #:modules and #:imported-modules arguments only work in the gnu-build-system?
<civodul>they work in some others
<civodul>which one do you have in mind?
<bavier>cmake-build-system
<bavier>I was looking back at the lapack package, trying to fix some rpath troubles in the dynamic libs
<bavier>but I got a "no code for module (guix build rpath)" error after adding that module
<civodul>looking at build-system/cmake.scm, cmake-build-system supports both
<civodul>aah
<civodul>yeah, you need to add it both to imported-modules and modules
<bavier>btw, what is the difference between the two?
<civodul>an example is cdparanoia
<civodul>some modules are part of Guile, while others aren't
<civodul>#:modules tells which modules should be used in the expression's environment
<civodul>whereas #:imported-modules tells which modules should be imported into the build chroot
<bavier>I see, ok
<bavier>great, now it's working. thanks.
<civodul>cool
<civodul>do you know if Octave builds now?
<civodul>hydra has trouble handling it because of the TeXLive dependency (takes forever to transfer among build machines)
<bavier>the texlive build was flagged as failing on hydra, so the octave build didn't go through
<civodul>yeah
<bavier>it built for me
<bavier>I need to check that octave doesn't suffer from the rpaths in the lapack library
<civodul>ok
<bavier>so, for augment-rpath, patchelf needs to be listed as an input?
<bavier>or as a native-input?
<civodul>native-input, yes