Now with added SSL. Although I’d still prefer to dump WordPress and switch to a static site – I just need to summon up the enthusiasm for the migration.
It’s been a pretty uneventful week as far as I can remember. I had a couple of days off work but didn’t really do much with them other than lazing about. I made a bit more progress with the weather station, but most of it seemed to involve yak-shaving around trying to decide the best way to time and/or count the anemometer pulses to give a wind speed that’s real time but not too real time.
We had a bit of snow on Thursday and Friday mornings. It didn’t amount to much, but I was glad not to be out in it. Fortunately the weather was far more pleasant on Saturday (marshalling at parkrun) and Sunday (running the Harwich Half). Especially so for the latter, as there was no bag drop and the start line was a fair walk from the car park, which would have been horrendous if it had been freezing. I ended up with my fastest half marathon time for a few years, but nowhere near a PB (despite what Strava may claim – I assume the one in 2019 which was a couple of minutes quicker must have come up short on the GPS).
I still can’t ride my unicycle, but have graduated from following along the garage wall and using it for balance to pushing off, pedalling a few times (I think my record so far is about 2½ whole revolutions) then falling off. At least the falling bit now generally involves just stepping off and (usually) grabbing it before it hits the ground.
On Friday I dialled into (do we still say that? How many people these days even remember phones with dials? I do actually have a nice black bakelite one plugged in, but it no longer seems to work – not sure whether it’s the phone or my landline, but either way it stops people ringing me, which is nice. Anyway, where was I? Oh yes …) James Shore’s Art of Agile book club, with J. B. Rainsberger and Geepaw Hill, for a discussion on TDD. It was interesting overall, but the main specific tip I picked up was a simple one that seems like it might come in handy. The idea was that when you have a test that occasionally fails because of a race condition, it turns out that it is OK to add a sleep, but you should add it to the production code, not the test. That way you end up with a test that fails reliably, which makes it much easier to debug, and once you’ve fixed the synchronisation problem you obviously remove the sleep.