| 2 Sep 2024 |
 emily | my understanding is that there is heavy pressure to Bazel all the things, though I'm sure it's not completely universal | 20:35:15 |
| emily | I have definitely heard "it's not relevant for Google because they just use their own build system" before, at least | 20:35:40 |
 trofi | Google is big :) | 20:36:59 |
| trofi | It's hard to get uniformity in a huge company. And one has to co-exist with an external world. Thus there are places where openssl has to be used as-is. | 20:38:01 |
| emily | right | 20:39:27 |
| emily | I thought there was policy around BoringSSL specifically but I'm just an outsider who occasionally hears whispers | 20:39:56 |
 Alyssa Ross | if only there was a better OpenSSL build system tho | 20:40:00 |
| emily | I'm sure there's very few properties that hold across the entirety of google3 | 20:40:05 |
| Alyssa Ross | I hate having to keep adding entries to the mapping of platforms to OpenSSL config names | 20:40:18 |
| emily | we could just ship BoringSSL! | 20:40:26 |
| emily | (probably we couldn't) | 20:40:29 |
| Alyssa Ross | almost certainly not lol | 20:40:45 |
| emily | there should be branding flags so that if you're not inside Google it calls itself ExcitingSSL | 20:41:08 |
| trofi | Gah, I stand corrected. Both boringssl and openssl proper(openssl_vendor) use local implementation of the build system and don't seem to rely on the original one any more. sudo is a bit closer to the original (just as an example of a package). | 20:53:08 |
| 4 Sep 2024 |
|  SomeoneSerge (utc+3) changed their display name from SomeoneSerge (UTC+3) to SomeoneSerge (nix.camp). | 21:48:44 |
| 6 Sep 2024 |
| emily | an awful lot of packages seem to manually set a bunch of --<foo>dirs that are already set by stdenv. what's up wi that? | 12:37:42 |
| emily | * an awful lot of packages seem to manually set a bunch of --<foo>dirs that are already set by stdenv. what's up with that? | 12:37:45 |
| emily | e.g. jq:
configureFlags = [
"--bindir=\${bin}/bin"
"--sbindir=\${bin}/bin"
"--datadir=\${doc}/share"
"--mandir=\${man}/share/man"
| 12:37:56 |
| emily | lvm2:
] ++ lib.optionals (!enableCmdlib) [
"--bindir=${placeholder "bin"}/bin"
"--sbindir=${placeholder "bin"}/bin"
"--libdir=${placeholder "lib"}/lib"
"--with-libexecdir=${placeholder "lib"}/libexec"
| 12:38:26 |
|  K900 joined the room. | 12:38:35 |
| K900 | Uhh | 12:38:45 |
 Artturin | Blame and see if those lines are from tens of years ago | 12:38:46 |
| K900 | Cool more rooms I never knew existed | 12:38:53 |
| emily | I'm setting --sbindir to /bin globally (a bunch of packages have to work around it anyway), but I'm scared of breaking things | 12:38:56 |
| emily | In reply to @k900:0upti.me
Cool more rooms I never knew existed don't worry this one is kind of dead | 12:39:05 |
| K900 | Well now it's not | 12:39:15 |
| Alyssa Ross | Possibly because the default outputs don't make sense? | 12:39:27 |
| emily | https://github.com/NixOS/nixpkgs/commit/9f8390701363a767cb409a2931402a4effb972c7 🤔 | 12:39:55 |
| emily | surely automatically passing the multiple output parameters dates back to older than 2020? | 12:40:12 |
| Alyssa Ross | e.g. jq sets datadir to $doc/share, when I can imagine normally we'd want "data" installed elsewhere | 12:40:17 |
