| 27 Jul 2025 |
 K900 | And it just dies there | 05:53:45 |
 Charles | the only thing i can think of is that exec doesn't create a new process, so maybe somehow the shell process has some kind of state somewhere that the exec'd thing ends up inheriting that causes it to explode, and i have no idea how to go about testing/falsifying that idea | 05:55:41 |
| K900 | Oh god | 05:56:34 |
| K900 | I think I figured it out | 05:56:38 |
| Charles | what did you see, denvercoder9 | 05:57:23 |
| K900 | Or maybe not | 06:00:09 |
| K900 | It's like it's getting killed at a random point | 06:00:17 |
| K900 | I wonder if this is some python subprocess.popen nonsense | 06:00:36 |
| K900 | Because Legendary is just doing subprocess.popen | 06:00:45 |
| Charles | not a consistent point? | 06:00:49 |
| K900 | Nope | 06:04:38 |
| K900 | OK so I think what MIGHT be happening is that something is not detaching somewhere | 06:05:19 |
| K900 | So when the Python parent dies, we die with it | 06:05:26 |
| K900 | WAIT IS THIS BUBBLEWRAP | 06:05:29 |
| K900 | Motherfucker it is | 06:07:08 |
| K900 | We default to --die-with-parent | 06:07:15 |
| K900 | And Legendary doesn't wait for the process | 06:07:29 |
| K900 | So we just die | 06:07:30 |
| K900 | Immediately | 06:07:32 |
| K900 | @aidalgol gottem | 06:08:31 |
 aidalgol | yay! | 06:10:06 |
| aidalgol | thank you! | 06:10:08 |
| aidalgol | So where does the change need to be made? In the nix derivation that lives in the umu repo? | 06:10:38 |
| K900 | I'm not sure if die-with-parent by default even makes sense for fhsenvs tbh | 06:10:42 |
| K900 | I'm doing one final test | 06:10:58 |
| K900 | And I can push to your PR I guess | 06:11:04 |
| K900 | Basically all I did is add nixpkgs umu to heroic PATH | 06:11:18 |
| K900 | And then change it to not die with parent | 06:11:24 |
|  @audiotrope:matrix.org left the room. | 06:11:25 |
| K900 | 
Download image.png | 06:12:02 |
