| 15 Sep 2025 |
 Lun | lol their CI blows up when you PR from a fork | 02:23:11 |
 emily | contributing to these sorts of things is always kind of miserable | 02:24:33 |
| emily | Lun: we should probably send the ROCm update to staging, perhaps? to batch up those huge rebuilds | 02:34:31 |
| Lun | I asked hexa if ck rebuilds hitting master was ok and got a ≈don't do it too often in the infra channel fairly recently | 02:35:32 |
| Lun | I don't mind either way | 02:35:36 |
| emily | fair enough | 02:35:49 |
| emily | you can probably drop the Boost stuff from https://github.com/NixOS/nixpkgs/pull/427944/commits/d844a2f11124be2a736e9ac3eb6fcefefd8012bd after my PR, but I guess we'd have to split it for ordering, so meh | 02:36:06 |
| Lun | Want to send the propagatedBuildInputs change as a standalone PR? | 02:36:49 |
| Lun | it probably won't fix the ones where I had to -I bodge it, those are due to the messed up compiler situation for rocm so it can't even fix all the boost lines :/ | 02:37:19 |
| Lun | looking forward to trying the config file thing you suggested | 02:37:33 |
| emily | I don't mind just cleaning it up later | 02:37:38 |
| emily | especially if we staging | 02:37:44 |
| Lun | I'll target staging then. about halfway done on a review run | 02:38:35 |
| emily | fwiw
diff --git a/pkgs/stdenv/adapters.nix b/pkgs/stdenv/adapters.nix
index 255ff6a688..636fc298e8 100644
--- a/pkgs/stdenv/adapters.nix
+++ b/pkgs/stdenv/adapters.nix
@@ -95,12 +95,12 @@
if (args.__structuredAttrs or false) || (args ? env.NIX_CFLAGS_LINK) then
{
env = (args.env or { }) // {
- NIX_CFLAGS_LINK = toString (args.env.NIX_CFLAGS_LINK or "") + " -static";
+ NIX_CFLAGS_LINK = toString (args.env.NIX_CFLAGS_LINK or "") + " -static-pie";
};
}
else
{
- NIX_CFLAGS_LINK = toString (args.NIX_CFLAGS_LINK or "") + " -static";
+ NIX_CFLAGS_LINK = toString (args.NIX_CFLAGS_LINK or "") + " -static-pie";
}
)
// lib.optionalAttrs (!(args.dontAddStaticConfigureFlags or false)) {
(completely untested)
| 04:05:50 |
 K900 | Akonadi depends on mariadb-server | 05:37:26 |
| K900 | It grabs whatever version we have aliased to the non-versioned one though | 05:37:58 |
| Lun | emily: there's a issue with llama-cpp in the ROCm bump now that looks to be from a llama-cpp bump this week that included a PR that broke selecting what instructions to use on some ISAs. I'm happy to send it to staging and fix it in post if that seems reasonable, probably a simple fix to some GFX… conditionals I can PR (and upstream) tomorrow and it'd be nice to unblock your msgpack. failing that I can poke it tomorrow/. | 06:17:47 |
| K900 | @emily how's the cmake 4 stuff going? | 07:24:59 |
| K900 | We might have 25.05 built VERY quickly somehow | 07:25:05 |
 nim65s | As a warmup exercice I started by the things I maintain, and I already have some fixes at https://github.com/nim65s/nixpkgs/tree/cmake4-integration | 07:49:27 |
| nim65s | Feel free to grab whatever you neeed | 07:49:39 |
| nim65s | Hum, forgot tu push | 07:50:17 |
| K900 | Hmm | 07:51:08 |
| K900 | I actually like the replace-fail thing | 07:51:14 |
| K900 | @emily thoughts? | 07:51:19 |
| K900 | This basically gets us the same thing as the patches but without the extra fetch | 07:51:33 |
| nim65s | Done | 07:52:21 |
| nim65s | Something is wrong with triton-llvm though | 07:52:38 |
|  xanderio joined the room. | 11:39:19 |
| emily | can we do a cycle without please | 12:04:54 |
