18 Oct 2024 |
Alyssa Ross | maybe the .so has some other C++ symbols that will be more identifiable | 19:55:19 |
Pol | mmh OK | 19:58:40 |
Pol | I will try to disable the test, build the package and run some experimentations. | 20:00:37 |
Pol | If you have suggestions, feel free to share :) | 20:00:57 |
Pol | Download image.png | 20:04:56 |
Pol | Looks like no symbols are missing when using ldd . | 20:05:16 |
Alyssa Ross | Does ldd show missing symbols? | 20:06:13 |
Pol | Download image.png | 20:08:44 |
Pol | Nope... | 20:08:55 |
Alyssa Ross | no, I mean, I'm not sure showing missing symbols is functionality ldd has | 20:11:13 |
Alyssa Ross | ldd prints linked libraries | 20:11:22 |
Pol | Oh. My mistake then. | 20:12:23 |
| Ihar Hrachyshka joined the room. | 21:24:32 |
20 Oct 2024 |
| @dan:djc.id.au left the room. | 06:34:05 |
Theo Paris | What would be the best way to package upstream mkbootimg from mainline AOSP outside of nixpkgs? The upstream repository doesn't contain a setup.py or pyproject.toml, nor does it have an init.py structure, so I can't use buildPythonApplication. The only way I know how to do this is via mkDerivation and manually copying *.py to $out/bin, however that doesn't account for the gki directory | 06:47:07 |
Theo Paris | * What would be the best way to package upstream mkbootimg from mainline AOSP outside of nixpkgs? The upstream repository doesn't contain a setup.py or pyproject.toml, nor does it have an __init__.py structure, so I can't use buildPythonApplication. The only way I know how to do this is via mkDerivation and manually copying *.py to $out/bin, however that doesn't account for the gki directory | 06:47:25 |
Theo Paris | * What would be the best way to package upstream mkbootimg from mainline AOSP outside of nixpkgs? The upstream repository doesn't contain a setup.py or pyproject.toml, nor does it have an __init__.py structure, so I can't use buildPythonApplication. The only way I know how to do this is via mkDerivation and manually copying *.py to $out/bin, however that doesn't account for the gki directory that gets imported | 06:47:38 |
emily | it's already in android-tools | 06:53:19 |
Theo Paris | I know | 06:56:14 |
Theo Paris | I tried writing a setup.py in postUnpack but the problem is the python files aren't in a subdirectory for a module, they are in the root of the repo | 06:57:39 |
Theo Paris | * I tried writing a setup.py in postUnpack but the problem is the python files aren't in a subdirectory for a module, they are in the root of the repo. I guess I could mv them | 06:57:52 |
emily | if you want it as a separate package for closure size reasons, make it a separate output of android-tools ? | 06:58:12 |
emily | ah, you said "outside of Nixpkgs", I thought you said "inside" | 06:58:28 |
emily | well, I don't know what the motivation is, but can't you just reuse the existing build system that android-tools uses? | 06:58:39 |
emily | you can perhaps convince it to only build mkbootimg | 06:58:46 |
kirillrdy | is leaking of propagated build inputs for python applications a known issue ? https://github.com/NixOS/nixpkgs/pull/300186#discussion_r1805586570 cc hexa (PEP725 when) natsukium ? | 09:32:22 |
kirillrdy | there are a few examples of derivations deleting rm $out/nix-support/propagated-build-inputs | 09:35:37 |
hexa (PEP725 when) | Yss | 11:18:19 |
hexa (PEP725 when) | * | 11:19:20 |
21 Oct 2024 |
| (artur 'manuel) changed their display name from (lambda (f l) (format nil "~a ~a")) "Artur" "Manuel" to (artur 'manuel). | 20:03:29 |