Status Report - Week 35 2026

Real life

While kids are only going back to school next week, Nathalie had two groups of students come every day for "warm-up sessions", hence a clear smell of "we're back to it" at home. I did help her but only a bit, because (1) she did not really need me, and (2) I promised myself to try to be nice and patient with her students this year, and I didn't want to start eating into my (limited) reserve of patience :-)

Anton and I had an interesting day when he asked me to check why he could not activate anti-cheating features in one of his video games: we had to update the BIOS and I somehow managed to destroy Windows' boot record (?) without even actually initiating the BIOS update. Fortunately I was able to fix things without a reinstall, update the BIOS and get the anti-cheating features activated. This caught my curiosity and I read a bit about those. They look pretty advanced - as do the cheating techniques. I don't understand why people would cheat at a video game, but the "need" is clearly big given the R&D spent for and against it...

Apart from that, I spent a large chunk of my Friday evening and Saturday watching the UTMB live stream, which reminded me that mountains are really beautiful <3

In terms of reading, I continued The best and the brightest, and Maths MP-MP* - Tout-en-un.

I did run a fair bit - 26.10 km, and had a very light week a GeoGuessr, with only 1 win out of 3 duels.

Open source

I did a bit of work on the SourceHut, particularly to understand why build artifacts were not saved for failed builds in Production in spite of the patch I'd sent for this. So I had to spend the time to set up an actual builds worker locally using a VM, which eventually showed that the code was working!?! Then the only reason would be that the code was not deployed to Production... so I asked and it was indeed the case!

Doing this work was not lost though: I now know how to set up an actual builds worker, and I added a log stating what version a worker is running; both should seriously reduce investigation times if/when a similar weirdness occurs in the future.

I also submitted a fix for an event visibility issue in the project hub, as well as a clean-up I identified when investigating another issue.