Ok apparently plugging my headset into my T495s on Linux causes the kernel to dump core ffs
Back on windows, have rolled back the realtek driver fuckery which was apparently installed by lenovo thinkvantage and it works again.
That was short lived.
We all know you switch gear and opinion faster than most people swap underwear.
The thing slowly undermining you is the fact that you -- in the moment -- declare the change to be final. ("Never again! This is it!")
Pattern observation, however, gives a prognosis that it's highly unlikely that this will be so. You even declared that a week or two back; saying something like "The key is that I keep an ability to switch environments fast while keeping work going".
OTOH, don't stop, it's enjoying to watch.
This is mostly my inner evaluation monologue leaking into my normal life. We all do this sort of stuff regularly. Think of all those times you all bought something on here and it was a lemon. I think my calculated risk vs return rate is quite good. As for why run like this, 25% of my job is sticking my dick in every hole and see which ones bite. Thus avoiding escalating mistakes into ones being made by large development and operations teams. Seems to scale across other areas of life as well (apart from redheads)
Really the point is agility is a survival trait. Agility comes from disparate experience. Disparate experience comes from doing everything all the time.
Yesterday’s Lenovo fun had no impact as I was running Ubuntu off a USB-C external disk so rolling back was pulling the disk out and carrying on. Experiment over. The issue turns out to be the version of Realtek audio drivers that lenovo ship have a compatibility issue with a Dolby component. This pisses about 2000 events a minute into the event log. Thus removing them with driver rollback to the default WHQL MSFT provided drivers solves the problem.
On Linux I learned that the audio issue was not a hardware issue but a software one. The evaluation criteria I have for “works” on a desktop is an spreadsheet which is about 200 odd test cases long. I got to case 81 which was “WebGL works in preferred browser” and was impressed. Hit case 82, test plantronics headset, plugged it in and hard lock, rebooted then did it again with console open instead of X and bam, kernel crash. Evaluation failed.
The only things that have passed these test cases so far are windows and OSX. Unfortunately due to the VPN component I require being deprecated on OSX due to the removal of kernel loadable modules this was added to evaluation and it now fails. So I’m stuck with windows or Linux (if they ever fix all the shitty little bugs)
There is science behind this