| 21 Oct 2025 |
 Eelco | so it has to touch at least 1 byte every 4096 bytes | 21:07:52 |
 Sergei Zimmerman (xokdvium) | Yeah that was the overhead. A loop over all the pages | 21:08:28 |
| Sergei Zimmerman (xokdvium) | 1.23 │ lea -0x10000(%rsp),%r11
0.23 │ 15: sub $0x1000,%rsp
1.01 │ orq $0x0,(%rsp)
59.12 │ cmp %r11,%rsp
0.27 │ ↑ jne 15
| 21:08:42 |
| Eelco | right, that's to avoid a segfault if you have guard pages enabled (which I think is the default) | 21:09:13 |
| Eelco | I would expect the overhead for that loop to be pretty trivial though | 21:09:31 |
| Eelco | in the case where the pages are present | 21:09:52 |
| 22 Oct 2025 |
|  0xcafca changed their profile picture. | 10:21:53 |
| 0xcafca changed their profile picture. | 10:23:31 |
 tomberek | @niksnut:matrix.org: builtins.fetchTree cannot take advantage of the "__final" optimization. This means usages of flake-compat will re-fetch inputs unnecessarily. Is there a way to expose `prim_fetchFinalTree`. This can create a large performance regression. | 15:11:06 |
| Eelco | I think we should allow fetchTree { final = true; ... } | 15:11:43 |
 Robert Hensing (roberth) | I get what it does but I never felt like I had a complete understanding somehow. If we were wrong about final we could always design something better without the pressure and call it fetchSource :) | 15:14:03 |
| Eelco | final just means it won't add more attributes | 15:17:39 |
| tomberek | @roberthensing:matrix.org: is the concern that it would be abused ir ossify some behavior? | 15:26:18 |
| Robert Hensing (roberth) | I guess I just expected it to be prettier | 15:26:44 |
| Robert Hensing (roberth) | sometimes things just aren't, and that's ok | 15:27:35 |
| tomberek | I suspect this can get better with lazy paths/trees, but that seems to be further in the future. | 15:28:00 |
| Robert Hensing (roberth) | the basic question seems to be: do we want to trust the lock file, and I think usually the answer is yes, final = true; | 15:29:24 |
| Robert Hensing (roberth) | if you can't trust your lock file, you're either editing it by hand, which you shouldn't do, or letting people you don't trust update, which you know, you'd have bigger problems | 15:29:52 |
| Robert Hensing (roberth) | I feel like true should probably be the default in a future version of this primop if we have one | 15:30:55 |
| Robert Hensing (roberth) | anyway, making final part of the public interface seems fine to me | 15:31:52 |
| Robert Hensing (roberth) | iirc a lazier fetchTree would be a mitigation for the performance loss but not a complete fix | 15:32:17 |
| tomberek | Would it make sense for the final to be default if fetchTree has been given enough attributes? (If provided with all the info.) | 15:43:55 |
| 23 Oct 2025 |
 lovesegfault | First batch of small PRs to enable multipart uploads:
cc. John Ericson Sergei Zimmerman (xokdvium)
| 02:29:28 |
| lovesegfault | These are all tiny so it should be easy to review and merge. If you want to see the big picture, look at https://github.com/NixOS/nix/pull/14330 | 02:30:43 |
 fzakaria | I finished the Kaitai spec | 04:21:54 |
| fzakaria | https://github.com/fzakaria/nix-nar-kaitai-spec | 04:21:55 |
| fzakaria | it has test suite also | 04:21:59 |
| lovesegfault | y'all why is Headers not a map | 06:28:04 |
| lovesegfault | seems strange to use a vec of pair and force me to scan | 06:28:21 |
 Mic92 | I did some more work over the last weekend on the installer: https://github.com/NixOS/experimental-nix-installer/pull/51 i hopefully get ci green again this weekend and than i can do some more tests. | 09:56:46 |
