IRC channel logs


back to list of logs

<rekado>guix-past fails to build for me: (exception unbound-variable (value #f) (value "Unbound variable: ~S") (value (rgbds)) (value #f))
<rekado>oh, I just need a more recent Guix
<rekado>odd, though, that this happens with “guix pull”…
<rekado>no, my bad: I used “time-machine --commit” but still built extra channels, which obviously don’t work with that old commit of the “guix” channel.
<civodul>rekado: i think guix-past is tested against 1.1.0
<civodul>+ master
<civodul>the problem you had was with a commit older than 1.1.0?
<rekado>I’m trying a commit to install R 3.6.0
<rekado>that’s pretty far back
<civodul>ah ok
<civodul>perhaps you'll use an inferior
<rekado>no, I’ll just provide a new channels file that only uses the “guix” channel at that commit. That’s all I really need here :)
<efraim>rekado: I have ...good news?
<efraim>UTHSC needs a newer tensorflow, which means I have to get something working
<efraim>do you have any notes from bootstrapping bazel? Even from their "bootstrap" 300 MB zip file?
<civodul>efraim: i guess a {non,half}-bootstrapped bazel could go to guix-science
<civodul>though... it would be best if we had a plan to bootstrap it
<civodul>with rekado's notes, we could get in touch with Bazel folks we met before at the R-B summit and see if we can get them on-board
<efraim>I'm hoping I can get something building and then rip out the bundled parts as possible
<civodul>yeah it can be an incremental process i guess
<rekado>efraim: I have no notes about bazel
<efraim>alright, thanks anyway
<rekado>I just remember that there are about 40 bundled Jars.
<rekado>some of them are the full JDK
<civodul>of course this does not mean you should feel discouraged, efraim
<efraim>of course :)
<civodul>40 jars aren't enough to empty the ocean