IRC channel logs

2022-01-18.log

back to list of logs

<zimoun>hi!
<zimoun>rekado_: where can I find your nice R ’guix.install’ thing?
<civodul>zimoun: hey! it's at https://git.elephly.net/gitweb.cgi?p=software/r-guix-install.git;a=tree
<zimoun>civodul: cool! Thanks! Why not include it in Guix proper?
<civodul>i guess we could include it
<civodul>though the goal was to let users install it via the regular R mechanism, and *then* install the rest with it
<civodul>perhaps that's the reaosn
<civodul>*reason
<zimoun>is it possible to install it via regular CRAN installer?
<zimoun>from my understanding, it is just a simple R function that the user must have in their R workspace.
<civodul>dunno
<civodul>rekado: how's tensorflow 2 going? :-)
<civodul>there are colleagues interested in it
<rekado>no progress, I’m afraid
<rekado>zimoun: it’s just a procedure that you dump in ~/.Rprofile
<rekado>I’d like to make it a CRAN package
<rekado>zimoun: the code is here: https://git.elephly.net/software/r-guix-install.git
<zimoun>rekado: maybe it could go to etc/ in Guix proper, no? At least, a pointer in the manual? WDYT?
<rekado>I don’t know.
<rekado>etc isn’t available when people run ’guix pull’
<rekado>for a pointer in the manual I’d prefer to have a package that we can ask people to install
<rekado>civodul: the disk price suddenly jumped up again
<rekado>it’s now 679.99 vs 614
<rekado>yesterday it was 655
<civodul>rekado: maybe you need to check with private browsing?
<zimoun>rekado: yes, etc/ is not available by pull, but the manual could point to it; waiting a CRAN package for it.
<civodul>something we may want to participate in: https://trex-coe.eu/events/eurohpc-summit-week-2022
<zimoun>rekado: I have not investigated why but r-bigmelon is failing https://ci.guix.gnu.org/build/402864/log/raw
<rekado>zimoun: may need to setenv HOME
<rekado>oh, that’s just a warning
<zimoun>rekado: I just sent #53349 about adding R packages, let me know.. :-)
<zimoun>arf, depressing! Some users just want Conda and other various mess done by hand because… I do not understand why when Guix is working smoothly for their use-case.
<nckx>Familiarity is a feature.
<zimoun>argh, when people are discovering Git submodules… this top-level https://github.com/getzlab/rnaseqc includes this https://github.com/walaj/SeqLib/tree/35c9ec5edb53c4afdbc60c09a5f748571c294c13 which includes BWA. Arf! Time to go home and let all that as an excercise for some users. ;-)
<zimoun>have a nice evening :-)
<zimoun>civodul: I am waiting the answer for a training about FAIR stuff and the dates are the same as the HPC stuff you posted earlier.
<zimoun>nckx: well, I do not think the feature is familiarity but instead kind of laziness. ;-)
<nckx>But… I'm lazy, and I like Guix. What did I do wrong?
<zimoun>hehe! I am lazy too. I think we are the true lazy, thus long-term lazy. And some users I was speaking about are short-term lazy. :-)
<nckx>100% agree with that view.
<nckx>Same thing I think after realising ‘I should have written a script to do this after the 2nd time, not the 50th.’
<zimoun>Yeah, basically, it took me a bit less than 2h, to read their complaint, package all (9 R stuff), and reply them. Done. And because Guix done for (almost) ever, because SWH done for (almost) whatever GitHub or CRAN will go. Their way is quicker now, say 20min, but no guarantee it will work again weeks or months later. Who knows how many time they will spend later? Or they will give up… arf!