IRC channel logs


back to list of logs

<JamesRichardson>Hello guix!
<JamesRichardson>Anyone here use mcron?
<JamesRichardson>Looking for an example that will run a shell script daily and send an email if error occurs.
<marusich_>As I understand it, mcron is supposed to be backwards compatible with vixie cron, so you should be able to specify a job using the usual vixie cron syntax. However, I guess you can also do it in Guile. I haven't used mcron much, but you might find an example in the manual.
<marusich_>sorry I can't help more!
<JamesRichardson>I didn't see an example in the manual.
<marusich_>Maybe you can find an example in the Guix manual?
<marusich_>Another possible place to find an example would be in the "system" tests in Guix. But I don't know if that is an easy place to learn from for the first time.
<JamesRichardson>I can just use vixie syntax for now. Need to learn more scheme ;)
<marusich_>But yeah, if you want to see examples, I'm 99% sure there are some in the "system" tests for Guix.
<JamesRichardson>Had thought about looking in the system tests.
<marusich_>TBH I haven't looked into those much myself; still learning other things
<marusich_>I'm surprised the manual doesn't have examples. Maybe once you figure out a good one, you can submit a patch to add it to the manual!
<JamesRichardson>I will do so.
<marusich_>Oh, check this
<marusich_>That looks like it has some examples but they are using Gexps, which are a Guix thing.
<marusich_>I guess that might not be quite what you were looking for.
<JamesRichardson>Not quite.
<JamesRichardson>I am running (or trying to) mcron from GuixSD, so I guess I could use Gexps.
<marusich>Well, if you get it working, feel free to add your knowledge to the manual if there is anything that needs clarification.
<xd1le>how do I change the default shell in GuixSD?
<xd1le>(say from bash to zsh)
<efraim>JamesRichardson: there's also the mcron manual, I haven't actually written any mcron services but it made it look really easy
<efraim>xd1le: I think there's a default-shell field as part of the users' fields in the OS scm file
<xd1le>efraim: thanks for the pointer, I'll try to find it!
<xd1le>efraim: that was easy, thanks
<xd1le>(for reference:
<xd1le>efraim: hmmm so how would I make a gexp for zsh.. if you know?
<xd1le>I'm new to scheme :/
<efraim>xd1le: haven't looked too much into gexps yet, sorry
<Drakonis>Are there alternate rwp
<Drakonis>Repos yet?
<jmd>A lot of things seem to be "gone" from Hydra.
<xd1le>efraim: no worries
<efraim>hopefully we'll notice it less when core-updates gets merged
<efraim>Interesting idea I don't want to forget about: guix on freebsd doesn't work for a number of reasons that I forget, what about their updated linux-compat layer?
<pksadiq>How can I chroot into a guixSD partition?
<rekado>wow, what’s up with core-updates? ~6200 failures?
***paolo_ is now known as paolo
<rekado>one important failure is tcsh
<Digit>^_^ i reccognise this pattern... install guix on debian(/devuan), then get the itch to have guixsd to ensure saturate exposure to guix as goto packagemanagement system. *tries to resist, to avoid the disruption*
<jmd>I see that the definition of guile has gettext as an input. Why then does (provided? "i18n") return #f ?
<wingo>jmd: probably (provided? 'i18n)
<wingo>instead of a string
<jmd>So then the original question comes back: Why does it not find my translations?
<wingo>ACTION dunno, i've never actually used gettext :)
<rekado>efraim: looks like the dejagnu update you pushed broke gcj on all platforms and thus all Java packages.
<rekado>efraim: could you please take a look at this?
<rekado>we may be able to patch gcj to use “file normalize” instead of “absolute”
<efraim>damn thats no good
<rekado>efraim: the fix seems to be to patch this line: libjava/testsuite/lib/libjava.exp: set d [absolute $objdir]
<rekado>“absolute” to “file normalize”
<efraim>looks like what LFS has
<rekado>sed -i 's/absolute/file normalize/' libjava/testsuite/lib/libjava.exp
<rekado>would you like to patch this?
<efraim>i broke it, seems only fair :)
<lfam>I really struck out with Vc, huh? ;)
<lfam>And the vigra change didn't have the desired effect either
<efraim>i tried to fix the i686 build error but it proved harder than expected
<lfam>efraim: For vigra?
<efraim>for vc
<lfam>I expected it to work at least on x86_64 since it had worked on my main machine
<lfam>I guess we can only expect it to work on 32 and 64-bit Intel compatible architectures
<efraim>it seems to optomize for hardware
<lfam>Too bad, that limits Krita to those architectures
<lfam>Although there is work in progress to use ARM's NEON
<efraim>python2-efl on i686 had its on-again off-again test failures with dns lookup
<lfam>Any suggestions on how to figure out why reboot hangs?
<efraim>where does it hang?
<lfam>The final console messages are from ntpd, and then something about the wireless. I'm reconfiguring and rebooting while removing each of the commits that I rebase on top of master to see if one of those is the culprit. Seems likely since I haven't heard of anyone else complaining about reboot failures
<efraim>i sometimes seem to hang on startup or shutdown while it waits for some if-up/if-down on wlan0. I assume its systemd on debian that has the 90 second time-out
<efraim>i'm watching gcj build, just saw it echo timestamp to a file :/
<lfam>Nice catch :)
<lfam>It should use $SOURCE_DATE_EPOCH instead
<efraim>my python-efl test issue should be easy to fix, it discovers all the files and then runs the tests so I can just delete the file
<rekado>found an upstream fix for the tcsh build failure.
<efraim>is it also the latest version?
<rekado>I don’t know. Latest tcsh release seems to be in 2015. The patch is much more recent.
<rekado>trying to build tcsh with this fix. Will push to core-updates if successful.
<lfam>Interesting that so many things depend on tcsh. Is it a common dependency or is there some bottleneck dependency?
<rekado>I’m very surprised about this too
<lfam>I've noticed it before but I didn't investigate because it wasn't a problem then ;)
<lfam>I guess the bottleneck is boost
<rekado>I don’t see boost in the list of packages depending on tcsh
<rekado>(according to guix refresh -l tcsh
<efraim>boost isn't a leaf
<efraim>guix graph -t bag boost | grep tcsh
<lfam>tcsh is a native-input of boost
<efraim>[color = magenta] now I know it must be really important
<lfam>Determined with `grep` ;)
<lfam>So, are we brave enough to graft Bash?
<lfam>That was the wrong link. I meant to paste this one:
<lfam>The patch is now available from the regular place
<lfam>It's that Bash bug that allows code execution if you can set the system's hostname
<htgoebel>ACTION smiles: all python packages build successfully using the refreshed build system
<rekado>htgoebel: woo!
<lfam>htgoebel: That's great :)
<rekado>the patch for tcsh works (after some tweaking), pushing to core-updates
<efraim>still building gcj
<lfam>Can anyone try building imagemagick 6.9.6-1 (the latest release in the series that we package)? There are test failures and I'm wondering if it's something specific to my machine or if the tests just fail everywhere
<jin>lfam: there are 6 FAILS for gucharmap
<lfam>Well, I bit the bullet and made an account on the imagemagick message board so I could report the bug:
<lfam>jin: Thanks for trying it! Do you results match what I got in that link I just posted?
<jin>let me see..
<lfam>I don't see 'gucharmap' in my build's output
<jin>error on copy
<lfam>Looks like a few of the failures are caused by fonts not being available
<lfam>The 'wandtest.tap 1' failure is a total mystery to me
<jin>lfam: the 6 FAILS are the same.
<lfam>jin: Thanks for testing it. It's good to have some confirmation that it's not machine-specific
***Steap_ is now known as Steap