| 6 Aug 2022 |
 @v0id:nltrix.net | hexa: fyi | 13:40:49 |
| @v0id:nltrix.net | * on line 565 there are definitions for thirdparty | 13:42:46 |
| 7 Aug 2022 |
 K900 | Posted this in #bootspec:nixos.org but also going to post here: my refind/sbctl/secure boot thing is actually at a point where you might want to run it: https://gitlab.com/K900/bootis | 08:04:09 |
| K900 | If you want to give it a try, please do and report back | 08:04:20 |
| 8 Aug 2022 |
|  ar left the room. | 07:23:51 |
| 9 Aug 2022 |
 Winter (she/her) | So the zlib issue is perplexing me:
On Darwin, I can't get the test that fails to to fail at all, while on Linux I can only get it to segfault in the build sandbox -- it fails without segfaulting otherwise.
😅 not sure what to think of this | 15:09:28 |
 Mindavi | Probably the behaviour of curl is different when sandboxed (the way nix does it on linux) | 15:24:51 |
| Winter (she/her) | In reply to @rick:matrix.ciphernetics.nl
Probably the behaviour of curl is different when sandboxed (the way nix does it on linux) I don't see why curl would do that tbh | 15:30:23 |
| Mindavi | I remember something about dns resolving, but I also don't know much details | 15:50:43 |
 Wanja Hentze | getaddrinfo and friends are disgustingly environment-dependent | 17:21:50 |
| Wanja Hentze | https://tailscale.com/blog/sisyphean-dns-client-linux/image4.png | 17:21:55 |
| Wanja Hentze | this is the simple version, with no nsswitch.conf in the mix | 17:22:07 |
| Wanja Hentze | so I can totally see it doing something different depending on what files it does/doesn't see in the sandbox | 17:22:26 |
| Wanja Hentze | all kinds of random things start happening once you sandbox programs away from their "conventional" environment. I have observed toybox' grep starting up more than 10x faster than normally when there are no locale databases present | 17:36:51 |
| Wanja Hentze | * all kinds of random things start happening once you sandbox programs away from their "conventional" environment. I have observed toybox' grep starting up more than 10x faster than normally when there are no locale databases present (it still works exactly as it should, of course, because locales are a misfeature nobody has ever missed, ever) | 17:45:49 |
| Winter (she/her) | In reply to @whentze:matrix.org
all kinds of random things start happening once you sandbox programs away from their "conventional" environment. I have observed toybox' grep starting up more than 10x faster than normally when there are no locale databases present (it still works exactly as it should, of course, because locales are a misfeature nobody has ever missed, ever) The thing is, the curl dev was able to repro the segfault without a sandbox, while I can only repro it within a sandbox. Just seems so weird... | 18:14:46 |
 hexa | I think a few people (ajs, ris, me) have been playing fallback maintainer for packages that should have maintainers, but surprisingly don't 😄 | 20:00:17 |
 @andreas.schraegle:helsinki-systems.de | so, anyone want to sign up for maintaining openssl? | 20:00:27 |
| hexa | or curl | 20:00:34 |
| hexa | or microcodeIntel | 20:00:41 |
| @andreas.schraegle:helsinki-systems.de | I already have nss, maybe I should just take up all the tls libraries | 20:00:44 |
| hexa | don't forget botan2 then | 20:00:54 |
| Winter (she/her) | In reply to @hexa:lossy.network
or curl tempted to take this tbh | 20:01:04 |
| hexa | mbedtls anyone? | 20:01:19 |
| * K900 In reply to @hexa:lossy.network
mbedtls anyone? runs away screaming | 20:01:29 |
| Winter (she/her) | mbedtls isn't hat bad last i checked? | 20:01:37 |
| Winter (she/her) | * mbedtls isn't that bad last i checked? | 20:01:39 |
| Winter (she/her) | what's bad about it K900 | 20:01:48 |
| @andreas.schraegle:helsinki-systems.de | In reply to @hexa:lossy.network
mbedtls anyone? polar…ssl? | 20:01:55 |
| hexa | there's also lots of packages maintained by eelco, but eelcon hasn't really been contributing to nixpkgs in that sense in a while | 20:02:18 |
