IRC channel logs


back to list of logs

<mbakke>I think I'll cancel the remaining staging jobs and start a new evaluation, since the current is lagging more than a week behind master. Hopefully it can be merged in a few days.
<civodul>mbakke: sounds good!
<civodul>i can turn off the queue-runner during that time
<mbakke>civodul: That would be great :-)
<mbakke>I'm getting this error when compiling 'master':
<mbakke>guix/git.scm:115:17: warning: possibly unbound variable `object-lookup-prefix'
<mbakke>Ah, I probably need the new guile-git in my checkout.
<civodul>mbakke: queue-runner stopped!
<civodul>and yes, you can use the newer guile-git, but if you don't, that's ok
<civodul>the code explicitly checks whether that procedure exists
<civodul>ACTION -> zZz
<civodul>good night!
<mbakke>civodul: Excellent, good night!
***Gamayun_ is now known as Gamayun
<jsoo>whew! the docs are not kidding! the install takes a looooong time. it’s averaging <50kib/s. is this normal?
***kmic is now known as kmicu
<efraim>currently aarch64 FTBFS on core-updates, glibc-intermediate fails, i think related to the current binutils
<efraim>Functional language conference in Israel in July
<mbakke>efraim: How does glibc-intermediate fail?
<mbakke>"make" fails to build with glibc 2.27.
<mbakke>Hydra just completed an evaluation, but still says it's "pending":
<bytes83>I sent a small update patch on the guix-patch mailing list. But i never received neither an acknowledgement and the email is not showing up in the tracker also. What can i do ?
<snape>bytes83: it's :-) not guix-patch
<bytes83>Yes. Im sorry for the miscommunication. I sent them to Twice (in a span of 44 mins). But nothing happened.
<mbakke>bytes83: It's probably awaiting moderation.
<rekado>wayland does not build on core-updates / rhel6
<rekado>7 test failures.
<rekado>I’m trying to fix this now.
<efraim>mbakke: probably didn't need the whole log, but here it is:
<rekado>the wayland failures are all related to locking and creating sockets
<bytes83>Oh! I didnt know that the guix-patches list was moderated. (This was my first patch)
<rekado>bytes83: it’s not moderated, but like all GNU mailing lists the very first email is usually delayed.
<rekado>I think that’s called grey-listing, which is meant to prevent spam.
<rekado>with some delay, the message will eventually appear on the list.
<rekado>future messages will not be delayed, if sent from the same address.
<bytes83>Okay!!. Thanks. :-)
<snape>grey-listing works because bots won't try to send the email another time 30min later, wheteras your mail server will.
<rekado>so, this code fails: flock(socket->fd_lock, LOCK_EX | LOCK_NB) < 0
<rekado>any idea why? Could this be a problem only in the build container?
<rekado>(or is this a problem only when building with NFS, which has unusual locking semantics?)
<rekado>the tests pass when I run “make check” manually.
<rekado>I guess for the rhel6 branch I’ll just delete these seven tests and move on.
<vvedantham>Hello I tried installing guix using the install script in and got the following warnings: substitute: guile: warning: failed to install locale substitute: warning: failed to install locale: Invalid argument. Package installation works fine but does anyone know of a fix? I'm using Ubuntu 16.04.
<rekado>vvedantham: the warning is harmless.
<rekado>vvedantham: you can remove the warning by installing the glibc-locales package and setting the GUIX_LOCPATH environment variable to $HOME/.guix-profile/lib/locale
<rekado>this may also have to be set in the environment in which the guix-daemon is running.
<vvedantham>I have created a patch for the CRAN package abbyyR. Is there any way to test this?
<rekado>where did you put the package definition?
<rekado>I would add it to gnu/packages/cran.scm and bind it to a variable “r-abbyyr”
<rekado>you can test it with “./pre-inst-env guix build r-abbyyr”
<vvedantham>Added it to cran.scm, will test it now
<dijong>hi, how can i install two package versions (like gtk+@2 and gtk+@3) at the same time? when i try installing one it upgrades or downgrades the other, and i couldn't find a solution at
<rekado>dijong: you’d have to install them in the same transaction, but even then I think that “guix package -i” prevents that due to the potential for conflicts.
<rekado>“guix environment --ad-hoc” would allow it, though.
<dijong>rekado: thank you, installing them via "guix environment --ad-hoc gtk+@2:out gtk+@3:bin gtk+@3:out" seemed to work -- but only for that shell, how can i do it for my user's global (default) environment?
<rekado>dijong: this is not possible at the moment. It’s a missing feature. See bug #30830.
<mbakke>efraim: The glibc/binutils/aarch64 issue is fixed in the upstream 2.26 stable branch:;a=commit;h=a5db85df69dce7fcf793fefc60c764c684e0c272
<mbakke>And probably also 2.27, which I am still bootstrapping from.
<efraim>mbakke: thanks for looking into it
<efraim>mbakke: build phase succeeded, looks like it worked
<mbakke>efraim: Great. Though we should probably take the rest of that branch if we don't go for 2.27.
<mbakke>Will send in the glibc patch once I've verified that "make" builds, hopefully later today.
<mbakke>Going AFK for a while now.
<efraim>i'm probably going to go to bed soon
<efraim>i'm going to build out to hello overnight to test it, not expecting any problems
<Apteryx>hello! How do I reload my init.scm using herd?
<Apteryx>kill & restart it? :D
<mbakke>Apteryx: that's what I do :P
<wigust->Apteryx: As I rememeber you could reload service objects without killing shepherd.
<Apteryx>I've read that, but could make use of 'reload'. It seems to be wanting pure service definitions in the required file-name argument rather than an init file.
<Apteryx>Not sure though :)
<wigust->Apteryx: Yes, it want's. I think you could put them in a separate file.
<Apteryx>OK! I'll keep that in mind for next time! Thanks :)
<Rukako>civodul: h-hi
<civodul>hey Rukako!
<Rukako>is this a good time or am I late again?
<mbakke>"make" fails to build with glibc 2.27 even with