Well, my new year started "couci-couça" / "così così" / "so-so" / "so und so".
🤔
Last night, shortly after midnight, I had a glance at my hobby site covid.aequologica.net.
💣
Ach ! Patatras ! Badaboum ! The graph was empty ! Ganz leer !
1/n
Immediately I understood that something with the time/date handling had to be wrong somewhere. A modest personal variation on the infamous "Year 2000 problem", the "year 2021 problem". en.wikipedia.org/wiki/Year_2000…
I was not at home (do not worry, I was safe), so I had to wait until ~2am in the night to be back to my Mac to try to fix the issue.
The hours before had been happily spent in the company of Moët & Chandon 🍾, then a quite nice Crozes-Hermitage 🍷, then back to bubbles ...
... with a Louis Roederer 🍾.
There is little doubt that this kind of company does not promote a healthy debugging.
Nonetheless, I insisted on fixing the problem as soon as possible.
The culprit was found relatively quickly, and, given my mental conditions, I decided not to try to produce a decent fix, but only to comment the offending piece of code.
😎
But after having commented a couple of lines that looked absolutely innocuous, horrible and incomprehensible exceptions started to be raised all over the place.
😱
My drunkenness, though relatively benign, made me behave in the worst of manners : as I was totally unable to figure out what was happening with the code, I felt outrageously humiliated, and wanted to prove the whole world that I was perfectly capable of doing this f...ing fix.
After some time, the duration of which I do not remember well, I had to surrender : some remnants of feeble consciousness helped me to craft a "we are on maintenance" display in place of the void graph, and I went to bed feeling excruciatingly guilty.
I woke up this morning, more or less clear, and I fixed the thing.
Continued: 2 days after …
I woke up this morning, utterly clear; I had a glance at my hobby site covid.aequologica.net. Guess what ?
Not straight 💣, but 🤔.
Though we were already the 3rd of January, the latest data (from Johns Hopkins) was from last year 31th of December.
Did the infamous "year 2021 problem" strike again ?
I consume covid-19 data through a kind-of broker site, that parses the Johns Hopkins University original and ugly CVS to a nicer and more usable JSON.
And …
… 💣.
Line 15 of the broker's script, below a self-accusing "HACK" comment, a blatantly hard-coded "2020" :
😂
I take some pride in having covid.aequologica.net continuing to be close to normal under these adverse conditions.
Even if pushing the 'start date' variable to the extreme left resulted in a graph whose interpretation is better to be left to fervid Philip K. Dick readers.
The good citizen I am duly reported to issue to the broker, and swiftly engineered and deployed to production a counter-hack that allows covid.aequologica.net to report correct 2021 data, even when pushing start date to its minimum.
That's all. For today.
• • •
Missing some Tweet in this thread? You can try to
force a refresh