| 9 Mar 2024 |
 hexa | not with --no-preserve=mode 🙂 | 19:40:56 |
| hexa | u+w recursively is probably the best way | 19:41:04 |
 @kyle:iteratee.net | I used rsync with the chmod instead of copy to solve exactly this problem. | 19:41:56 |
| @kyle:iteratee.net | * I used rsync with the chmod flag instead of copy to solve exactly this problem. | 19:42:09 |
 Lily Foster | rsync -rlE would probably do it too | 19:42:50 |
| Lily Foster | i think | 19:42:53 |
| Lily Foster | idk rsync scares me | 19:42:56 |
| hexa | I think rsync might be overkill 😄 | 19:43:04 |
| @kyle:iteratee.net | But it's a one-liner. | 19:43:26 |
| Lily Foster | nah, we need to make as many packages have rsync as a build-time reverse dep as possible (/s) | 19:43:34 |
| Lily Foster | * nah, we need to make as many packages have rsync as a build-time reverse deps as possible (/s) | 19:44:20 |
| Lily Foster | In reply to @lily:lily.flowers
nah, we need to make as many packages have rsync as a build-time reverse deps as possible (/s) (for context, iirc the last time an rsync bump erroneously went to master instead of staging was because the PR author didn't expect it to have thousands of reverse deps) | 19:45:02 |
| @kyle:iteratee.net | Once you give up on a one-liner, cp and then chmod has the nice property of being simple and obvious (to a reader) | 19:45:04 |
| hexa | In reply to @kyle:iteratee.net
But it's a one-liner. yeah, but it increases the closure 🙂 | 19:50:08 |
| hexa | In reply to @kyle:iteratee.net
But it's a one-liner. * yeah, but it increases the build closure 🙂 | 19:50:15 |
| Lily Foster | rsync in linux stdenv when | 19:50:36 |
 Qyriad | can't install do this? | 19:50:58 |
| Qyriad | I can never remember exactly what intersection of functionality install has, but it can copy and set permissions at the same time at least | 19:51:22 |
| Lily Foster | In reply to @qyriad:katesiria.org
I can never remember exactly what intersection of functionality install has, but it can copy and set permissions at the same time at least it needs to be able to ignore some perm bits but keep the execute bits the same (rather than setting a static mode). i forget if it supports that, but if so that would definitely be the best idea | 19:52:18 |
| Qyriad | I don't see it in the man page 😔 | 19:54:26 |
| @kyle:iteratee.net | I think you can use find to locate things by their permissions. You could use that along with `--no-preserve=mode` to add back +x bits. | 19:58:04 |
| @kyle:iteratee.net | (find things in source with +x bits and then arrange to chmod the same thing in destination) | 19:59:13 |
| Qyriad | we're getting back to "might as well just use rsync" territory though | 19:59:58 |
| Qyriad | or like, a python script | 20:00:03 |
| 10 Mar 2024 |
|  @kstiehl:matrix.numericas.de left the room. | 08:59:57 |
| 11 Mar 2024 |
 szucsitg | In reply to @adis:blad.is
https://github.com/NixOS/nixpkgs/pull/289231 Is there any reason behind the missing linkNodeModulesHook in this PR comparing to your source repo? | 20:41:26 |
| 14 Mar 2024 |
|  @federicodschonborn:matrix.org left the room. | 02:04:35 |
|  countoren changed their display name from oren to countoren. | 14:09:42 |
 @janne.hess:helsinki-systems.de | Is there a way to get from a set of .tgz files to a proper npm cache? | 15:13:41 |
| @janne.hess:helsinki-systems.de | I tried just mapping them over fetchurl and then npm cache add but that breaks prefetch-npm-deps for some reason (Error: invalid value: integer 1710428006885, expected u8 at line 1 column 223) | 15:14:15 |
