| 20 Jun 2026 |
 hexa | Redacted or Malformed Event | 09:30:32 |
 pveierland | Oh. Chunky đź‘Ť | 09:30:35 |
 flokli | It's not a fair comparison, we didn't pipe all store paths to it, we only lazily ingest what's requested. | 09:31:01 |
| flokli | And there's still a bit of fine tuning on the parameters and on-disk format to do. | 09:31:21 |
| flokli | But what's in right now is a good dataset for real-world usage. | 09:31:44 |
| pveierland | Will experiment a bit đź‘Ť Provenance is important since the store for each artifact is tracked during indexing. The ranged read research is for client side file system | 09:44:17 |
| pveierland | * | 09:44:44 |
|  Nikita Mikhailov joined the room. | 12:15:06 |
| 21 Jun 2026 |
 llakala | https://github.com/NixOS/nix/blob/f8bb823a23bf6d62f4c8feb792a77702d7a49fe1/src/libexpr/primops.cc#L3276 | 12:59:08 |
| llakala | might it be more performant to loop twice through the attrs - once calculating the number of attrs that actually exist, and the second time actually copying their values? | 12:59:45 |
| llakala | i say this because this removeAttrs has consistently shown up in my nixpkgs profiling | 13:02:25 |
| llakala | there's a lot of names to remove
removedOrReplacedAttrNames = [
"checkInputs"
"installCheckInputs"
"nativeCheckInputs"
"nativeInstallCheckInputs"
"__contentAddressed"
"__darwinAllowLocalNetworking"
"__impureHostDeps"
"__propagatedImpureHostDeps"
"sandboxProfile"
"propagatedSandboxProfile"
"disallowedReferences"
"disallowedRequisites"
"allowedReferences"
"allowedRequisites"
"allowedImpureDLLs"
];
| 13:02:50 |
| llakala | * there's a lot of names to remove on every mkDerivation
removedOrReplacedAttrNames = [
"checkInputs"
"installCheckInputs"
"nativeCheckInputs"
"nativeInstallCheckInputs"
"__contentAddressed"
"__darwinAllowLocalNetworking"
"__impureHostDeps"
"__propagatedImpureHostDeps"
"sandboxProfile"
"propagatedSandboxProfile"
"disallowedReferences"
"disallowedRequisites"
"allowedReferences"
"allowedRequisites"
"allowedImpureDLLs"
];
| 13:02:57 |
 Sergei Zimmerman (xokdvium) | We can do a better filterAttrs, but I don’t think those showed any improvements | 14:21:00 |
| 22 Jun 2026 |
|  fsnkty joined the room. | 09:25:06 |
| fsnkty set a profile picture. | 09:26:03 |
| 23 Jun 2026 |
|  debrazen joined the room. | 10:50:02 |
|  Leo J. Wagner joined the room. | 14:41:44 |
|  citerism joined the room. | 18:14:20 |
|  @knowledgeless:unredacted.org left the room. | 19:45:05 |
|  isabel changed their profile picture. | 19:54:41 |
| 25 Jun 2026 |
 connor (burnt/out) (UTC-8) | Domen Kožar Mic92 re: https://github.com/NixOS/nix/pull/16067#issuecomment-4802639722, I’d love to know more! Is there anywhere you’re working on this? | 19:37:32 |
 Mic92 | patchelf | 19:38:26 |
| Mic92 | check the pr of domen | 19:38:37 |
| Sergei Zimmerman (xokdvium) | IIUC https://github.com/NixOS/patchelf/pull/647 isn't going to address the core issue with slow startups and -z now / RELRO | 19:48:30 |
| Sergei Zimmerman (xokdvium) | I.e. the DT_NEEDED resolution is going to result in less stat storms (which is good), but resolving dynamic relocations isn't going to be any faster | 19:48:58 |
| connor (burnt/out) (UTC-8) | llakala do you happen to have a utility or script (or maybe there’s something already in tree that I’m missing) that lets you produce the diff in stats (like in https://github.com/NixOS/nix/pull/15885)? | 19:53:56 |
| llakala | yep! i put it up in a flake of mine | 20:15:07 |
| llakala | currently undocumented but you can see the source here | 20:16:16 |
| llakala | if you really want to just run it, you can feel free to copy the jq script yourself and run it with jq -s -f script.jq before.json after.json (where stats are generated with NIX_SHOW_STATS) | 20:17:25 |
