IRC channel logs

<rekado>“debunk” is a bit of a strong word

<rekado>maybe “dispel” is better here

<rekado>“debunk” to me always has a whiff of dishonesty

<rekado>also: very bold to claim that the GWL has “matured” :D

<civodul>yeah i agree that "dispel" sounds milder and more appropriate

<civodul>GWL received new commits, so it has "matured" :-)

<civodul>this looks like a summary more than perspectives

<civodul>maybe that's okay though?

<rekado>the last sentence veers into perspective territory

<rekado>so I think it’s fine

<civodul>alright :-)

<civodul>it'd be nice to plot the level of expectation of editors over time

<rekado>gotta leave now, will be back in a couple hours

<civodul>np, thanks again!

***civodul is now known as civodul`

***civodul` is now known as civodul

<zimoun>Evening!

<zimoun>civodul: what do you mean by: «zimoun, do you copy me?» http://logs.guix.gnu.org/guix-hpc/2022-02-01.log#141736 ?

<civodul>hey zimoun!

<civodul>i was wondering if you were somehow reading this channel

<civodul>looks like you were not :-)

<zimoun>heh! No yesterday was a long day… as today. :-)

<zimoun>civodul, for relaxing on Friday when listening speakers, Konrad raised an interesting issue about OpenBLAS https://lists.gnu.org/archive/html/guix-devel/2022-02/msg00000.html

<zimoun>which raises this weirdness https://lists.gnu.org/archive/html/guix-devel/2022-02/msg00008.html

<civodul>uh

<civodul>weird indeed

<zimoun>maybe I did a mistake. It is late. :-)

<civodul>yeah, i don't know, but it doesn't sound great

<civodul>in your message you're looking at commits that Konrad didn't mention though, that's confusing

<zimoun>No, the issue of Konrad is one. And when investigating, I found this weirdness.

<zimoun>Well, if I am correct, many packages could be affected. And it is not related to the failure Konrad reported, I guess.

<civodul>ah ok

<civodul>one issue at a time :-)

<civodul>Konrad's commit is from Jan. 2020 and we have some, but not all the substitutes it seems

<zimoun>yeah, that’s why I started another thread. ;-) And commented to the issue in the correct thread, I hope.

<civodul>ah yes, i had overlooked that

<zimoun> https://lists.gnu.org/archive/html/guix-devel/2022-02/msg00011.html

<civodul>ah nice

<civodul> https://issues.guix.gnu.org/51536 is scary

<rekado>yes, I think I reported it a while ago as well

<rekado>it builds more or less dependent on how many features the build node’s CPU supports

<zimoun>Hum, on one machine, my machine displays 0.3.9 and on another machine, I get an error (about (value "Unbound variable: ~S") (value (rust-1.52))) for guix time-machine --commit=4b1538e6ef -- show openblas.

<civodul>"guix time-machine --commit=7357b3d7a52eb5db1674012c50d308d792741c48 -- build openblas" works for me

<civodul>there are substitutes on ci.guix

<civodul>zimoun: did you fiddle with ~/.cache/guix/checkouts/ by hand?

<civodul>perhaps you could try "rm -rf ~/.cache/guix/checkouts", to be sure

<civodul>rekado: oh really?

<zimoun>civodul, about the checkout, I did nothing. It fails for me with --no-grafts --check.

IRC channel logs

2022-02-02.log