IRC channel logs


back to list of logs

***kmic is now known as kmicu
***jxself_ is now known as jxself
<civodul>Hello Guix!
*mark_weaver works on more security patches for icecat
<civodul>bootstrap-tarballs not built yet on mips64el-linux
<mark_weaver>civodul: is that of particular importance?
<mark_weaver>there are about 800 builds in the queue for mips. it's not apparent to me how hydra decides what order to build them in. might it be a long while before the bootstrap-tarballs are built?
<civodul>mark_weaver: well we have to make sure that it builds at all
<civodul>or could you build it on another machine?
<civodul>just to make sure
<mark_weaver>I don't have another machine that can do it in a reasonable amount of time.
<mark_weaver>I could ask the mips build slave to do it behind hydra's back, of course.
<civodul>or we speculatively assume that it's gonna be alright?
<civodul>wait, we can do that from hydra itself, so we don't lose the log
<mark_weaver>suppose it was a problem, and they didn't build. how bad would that be?
<mark_weaver>ah, you mean from the shell on hydra?
<mark_weaver>feel free!
*civodul logs in
<civodul>stopping the queue runner in the meantime
<civodul>IMO this is one of the things we shouldn't allow ourselves to break
<mark_weaver>it should be noted that all of the bootstrap tarballs being built by master are currently broken, or at least I suspect they are.
<civodul>because of libgcc_s?
<civodul>oh i thought you fixed it :-)
<civodul>we'll need to check/fix
<mark_weaver>I fixed it in core-updates
<mark_weaver>well, I don't remember clearly. will have to check
<civodul>it's building now
<mark_weaver>I fixed it in 270b501ee2b1b4ea4003a60b5568fc4dcb09d9db
<mark_weaver>which is in core-updates, not master
<civodul>perfect, thanks
<davexunit>good morning, #guix
<davexunit>back on my guix machine
<davexunit>I wonder how guix stacks up with this:
*davexunit is trying to get used to ratpoison
<zacts_pi>davexunit you are using ratpoison/stumpwm again?
<davexunit>I didn't care for stumpwm because common lisp
<zacts_pi>oh yeah
<zacts_pi>hopefully the stumpwm maintainer will adopt a guile version
<davexunit>no way would that happen.
<davexunit>they obviously like common lisp.
<davexunit>guile-wm is what I want to ultimately use
<zacts_pi>Huh, I heard someone saying something like this on #guile
<zacts_pi>oh well
<zacts_pi>yeah improving guile-wm would be cool too
<davexunit>I think I'll be able to do pretty much all of my FSF work from my guix machine today
<bavier``>davexunit: nice to hear that guix is already "in production" ;)
<davexunit>if only I could get the FSF to run a guix server...
<davexunit>but it would be the oddball among the Trisquel machines.
<civodul>silly erc
<civodul>mark_weaver: thanks for all the IceCat fixes!
<davexunit>yay, core-updates is merged!
<rekado_>some of the java tests require running an X server or at least "an Xvfb dummy X server with fvwm2 window manager" --- may I disable these tests or should Java depend on X and a window manager to get past these tests?
<mark_weaver>gtk+-3 depends on xorg-server to do tests, so there's precedent there
<davexunit>rekado_: have you tried using xfvb?
<mark_weaver>though we don't yet have fvwm2
*davexunit wonders how difficult that is to package
<mark_weaver>I think it would be okay to disable the tests that depend on fvwm2 for now, with a FIXME comment mentioning they await the fvwm2 package.
<rekado_>There also seem to be quite a lot of tests that fail for reasons unrelated to the build environment, though the majority passes. E.g. the hotspot VM tests have 3 failures out of more than 150 tests.
<iyzsong>does it really need fvwm to test? well, I could package it next day :)
<rekado_>the jdk has more than 1408 tests of which about a dozen or so fail.
<rekado_>I wonder how far I need to go to fix these tests.
<mark_weaver>well, we need to make sure the test failures don't indicate real problems.
<rekado_>okay, I'll try to get an overview on what the problems are. Luckily there's plenty of test output in the logs (although I fear I might learn more about Java than I hoped).
<rekado_>iyzsong: the tests explicitly mention fvwm2 as the only one of three tested WMs for which a couple of tests pass.
<rekado_>I don't know yet how many really require it.
<mark_weaver>what are the other two tested WMs?
<rekado_>the comments say "mwm fails (key name errors) and twm fails (hangs), but fvwm2 works well"
<rekado_>it looks like there are only two tests that would need it.
<davexunit>I think the text used in this retro GNU badge would make good MOTD text:
<yourbeau>is this distro ready to use ?
<davexunit>yourbeau: we are in alpha currently
<davexunit>I'm typing this from a machine running the guix distro
<davexunit>we have the things that most hackers need to do their work, but we don't yet have things like the GNOME desktop environment.
<rekado_>these tests are so frustrating. There's one, for example, that tries to find some substring in the output of javap. The output is plain text and the content matches but the number of spaces does not match (2 spaces expected but 4 printed).
<rekado_>I don't know why this happens and if upstream simply has a somewhat disfunctional test suite.
<rekado_>according to the stated intent of the test, this should pass.
<bavier``>oh, didn't realise guix's mplayer already had libdvdcss available.
<yourbeau>which package manager is used APT or RPM ?
<davexunit>yourbeau: hehe
<yourbeau>what ?
<davexunit>yourbeau: guix *is* a package manager
<DusXMT>yourbeau: guix is used :)
<yourbeau>where am I gonna get the packages from ?
<davexunit>from guix! :)
<yourbeau>is it gnu/linux ?
<davexunit>when you get the guix source, you get the package recipes.
<davexunit>yourbeau: gnu/linux-libre, yes.
<yourbeau>or with hurd ?
<DusXMT>Either someone writes a package definition, or you do
<davexunit>hurd is a WIP
<yourbeau>WIP ?
<DusXMT>Both systems are planned to be supported
<yourbeau>work in process ?
<DusXMT>But Hurd support isn't a main priority
<taylanub>work in progress, yeah
<DusXMT>yourbeau: the distribution is actually called guixotic (or GSD, if that got past, seemed to have a lot of supporters)
<yourbeau>DusXMT, I see, which DE is used ? it looks ugly
<yourbeau>like BeOS
<DusXMT>None by default, but we have Window maker, ratpoison, openbox, and I think xfce?
<taylanub>don't think there's an official DE of choice
<yourbeau>pictures here are ugly
<DusXMT>What you saw was probably WindowMaker, and that's supposed to look like NeXTSTEP :)
<taylanub>those use GNUstep/WindowMaker
<DusXMT>Doesn't :)
<yourbeau>why are you making it look ugly ?
<DusXMT>Because that's how WindowMaker looks
<DusXMT>It's the project's goal to be similar to NeXTSTEP, look on their homepage
<bavier``>yourbeau: you're welcome to help on the aesthetics ;)
<yourbeau>DusXMT, why do we need a new package manager based distro ?
<DusXMT>yourbeau: When you install Guixotic, there's no DE by default, so I don't get why you ask `Why are you making it ugly', since WindowMaker is only one of several choices available, and there's no default
<DusXMT>yourbeau: because guix is innovative
<bavier``>yourbeau: please have a look at the manual and some of the talk videos
<yourbeau>any live CD ?
<yourbeau>I want to give it a host
<DusXMT>There is the installer USB image, which is a live system
<davexunit>yourbeau: installation images are available on
<yourbeau>ok I try
<yourbeau>does it support cups ?
<davexunit>be advised: there is *no* graphical environment here.
<yourbeau>and hplip ?
<davexunit>yourbeau: someone wrote a CUPS package, but it's not done yet.
<rekado_>yourbeau: CUPS is almost there. I still need to push it to the repository.
<rekado_>but this means that it's not included in the live system, obviously.
<davexunit>and by someone I mean rekado_ :P
*rekado_ has had enough of Java for this week and leaves.
*rekado_ waves
<yourbeau>it is not ready for general use
<yourbeau>its only for hackers right now
<yourbeau>its hackable
<yourbeau>but not hacked enough
<DusXMT>THat's why it's called alpha :)
<yourbeau>would RMS switch to GUIX one its ready ?
<davexunit>that's up to RMS. :)
<yourbeau>RMS sucks @ gui
<yourbeau>I saw it in 31c3
<DusXMT>We all use our computers differently, and someone who doesn't use GUIs often, is going to have troubles with them sometimes, it's no biggie
<bavier``>gui's are tools, just like any other
<yourbeau>I did not realize
<taylanub>yourbeau: was I too honest in #emacs? :D
<taylanub>it's really great if you're a programmer and have time to learn it. I heard of non-programmers learning it too though.
<_`_>Is guile-emacs still alive?
<_`_>meh I should look for the answer myself
<taylanub>_`_: bipt works on it on every GSoC, basically ... other contributors absolutely welcome I'd say :)
*taylanub AFK
<yourbeau>taylanub, yep