[00:01] <MTecknology> For all this abstraction, it feels a lot like spaghetti code mixed with hard-coded logic; like the identity view implementing a password check field and a realname and username editor, but then pulling the password editor from somewhere else, even though one would presume these objects will always either be in this view (or at least the code
[00:01] <MTecknology> presumes this).
[00:05] <MTecknology> I'm struggling to find any way to disable that disk partitioning prompt. I could drop it from client.py, but that file is already loaded in order to run my early-commands, so it's too late to simply drop those lines from my version; and it's from snap, so modifying it on the iso is going to be exceptionally non-trivial.
[00:13] <MTecknology> dbungert: Without actually digging deeper, when I look at update_for_state() in installprogress.py, I get a feeling that simply removing one of the cases would cause an automatic failure due to an unhandled non-exception, which seems like the sorta logic that is going to cause lots of other headaches from surprise cross-project dependencies.
[00:14] <MTecknology> in other words -> https://dpaste.com/A5AHY46GX.txt
[00:51] <MTecknology> Ah ... I guess I can't actually make changes to this anyway because it's running inside of a read-only file system.
[01:15] <MTecknology> How the heck do I get nvidia to stop trying to install it's own mok cert??
[02:41] <MTecknology> Found it! Ubuntu's shim-signed uses hard-coded paths, completely disregarding the standard created by the dkms package. (seems like a bug)
[10:38] <ice9> what is the correct way to redirect specific log to a file in rsyslog? I tried ":msg, contain, ...." but it's still going to the kernel log.