Status Report - Week 42 2026

Real life

This was the last week before the kids' (and ours ;-)) two-week break, and I mostly spent time working on maths, and helped with the organization of Anton's school "course solidaire" where students would run in teams for N minutes for a charity.

My role was to count the number of laps "my" team made, a pretty easy task though it was a bit more funny when I had two teams to track. It was interesting to see that the notion that there's no winner except the charity does not register with most kids: they really want to perform better than the others :-) After this, they also asked parents to volunteer to run as well for the cause, and I happily obliged.

I also took Anton twice to play basketball, which was a very good reminder that I am not good at it :-)

On the reading front, I continue A certain idea of France and in parallel read the excellent and super scary If Anyone Builds It, Everyone Dies that I found in a blog post by Timothy Gowers. They make multiple very good points, but the one I think is the very important is that they claim (and I agree) that there's no reason why a super intelligence would NOT strive to destroy humanity at some point. Technically we're not yet f..ed but I also don't see humanity do any of the things needed so that it stays like this...

I finally finished reading Maths MP-MP* - Tout-en-un, and instead of taking a break, started reading Mathématiques pour l'agrégation - Algèbre et géométrie right away.

Finally, I had a good week at running, with 38.08 km (my last real one before the Den Haag Marathon on November 1st), and a light but good one at GeoGuessr, with 5 wins out of the 9 duels I played, enough to stay in Master III.

Open source

I did not do any work on SourceHut except pinging a few of my outstanding patches, and continuing to monitor the success rate of sr.ht-integration-testing's daily tester.

After a pretty long stretch of successes, I had one failure that I don't really understand, but showed that logging could be improved, which I did here.

I don't have much to show in terms of my work on GCC, which does not mean I did not spend lots of time on it (this are the joys of GCC development :-)): I tried to debug PR c++/127505 and while I'm not there yet, it got me to find issues in the functions you can use to show the C++ parser's state, and I sent a patch to fix them.