14 Mar 2023 |
| butterchicken joined the room. | 23:00:39 |
17 Mar 2023 |
| gr_h_m joined the room. | 08:52:30 |
20 Mar 2023 |
| TyberiusPrime (smilodon inopinatus) changed their display name from TyberiusPrime (fiber at home when) to TyberiusPrime (limited to 25 messages per 3 hours. Expect further significant reductions). | 09:51:21 |
21 Mar 2023 |
Carl Thomé | Any feedback/tips on this setup?
https://github.com/carlthome/ml-runtimes
I'm playing around with the idea of replacing devcontainers at work but felt jupyterWith was a little too brittle in my previous tinkering. | 23:16:36 |
22 Mar 2023 |
| valconius joined the room. | 00:00:19 |
| mjlbach changed their display name from mjlbach to mjlbach (Old). | 05:14:34 |
| mjlbach changed their display name from mjlbach (Old) to mjlbach. | 05:51:06 |
23 Mar 2023 |
necrophcodr | I've been trying to work on making developer shells for R projects, but I'm having some trouble with renv and openssl (specifically, but probably not exclusively). Is there a good way to setup developer shells where renv works with openssl, or perhaps even a good way to parse renv.lock files into NixOS R derivations? | 14:55:51 |
jbedo | Renv and nix have overlapping functionality, was there something you specifically needed renv for? | 21:03:13 |
jbedo | You might also be interested in https://github.com/TyberiusPrime/nixR if the framework in nixpkgs isn't sufficient for your use case | 21:15:20 |
necrophcodr | In reply to @jb:vk3.wtf Renv and nix have overlapping functionality, was there something you specifically needed renv for? Yes, R development on Windows. We do not work in a homogeneous environment | 22:08:44 |
necrophcodr | If everyone used NixOS I wouldn't even be asking either, we'd just use Nix flakes for the entire thing and have it be reproducible that way. Unfortunately that is not an option here with too much overlap between not only departments but also organisations. | 22:10:31 |
necrophcodr | Technically I suppose it could be possible to generate an renv.lock file from the nix development environment too. If that's a viable route, I am definitely not opposed to going the opposite direction. | 22:12:01 |
necrophcodr | In reply to @necrophcodr:matrix.org Technically I suppose it could be possible to generate an renv.lock file from the nix development environment too. If that's a viable route, I am definitely not opposed to going the opposite direction. (only I have no real idea how to do that) | 22:12:21 |
jbedo | you couldn't generate the lockfile entirely from an expression, as rPackages uses sh256 hashes and cran et al use md5 | 22:35:59 |
jbedo | actually maybe you can build an R environment with nix (that contains renv) and use snapshot | 22:57:36 |
jbedo | snapshot(type='all') | 22:57:48 |
jbedo | i think that'll pick up all the R packages | 22:57:53 |
necrophcodr | I've tried that and it didn't really work, but maybe I've been doing that wrong. I'll give it another go. I'm facing an issue with renv as well where it seems that running R inside an area where renv is already initialized seems to ignore all R packages added with rWrapper.override | 22:58:37 |
necrophcodr | Heck I might also end up rolling my own solution that'll generate nix expressions for this instead, or generate an renv.lock from a nix expression. | 23:00:15 |
necrophcodr | I'd like to avoid doing that as it isn't as neat, but if R -e "renv::snapshot(type='all')" doesn't work, I might just end up going the long way around | 23:00:53 |
jbedo | renv might be changing the library paths to only point at its project local dirs, you could read the R_LIBS_SITE environment variable and add them to your library paths after loading | 23:03:21 |
necrophcodr | I'll make a note of that when I give this a shot tomorrow morning for sure. Thanks for all the tips so far either way! | 23:03:59 |
jbedo | no worries :) | 23:04:25 |
24 Mar 2023 |
| Hilmar (he/him) left the room. | 22:07:09 |
28 Mar 2023 |
| Robert Michael joined the room. | 08:06:19 |
1 Apr 2023 |
| Nano joined the room. | 13:15:49 |
| nanoni17728 joined the room. | 13:36:25 |
4 Apr 2023 |
| Kai joined the room. | 06:13:41 |
5 Apr 2023 |
| Redstone joined the room. | 05:49:39 |