What's up doc and end of year projects
The last post here was on Jul 19, 2024, quite a long time... a lot has changed since.
In that time I managed to :
- switch jobs twice (last post was written while I was still working with CKSource, there was a brief stint with CastAI aftewards and I am at Churchdesk now)
- met my daughter
- came to terms with many things in general
Purely coincidentally New Year marks another attempt at adding some regular intervals to my public writing.
Anyway - what happened in the 4rth quarter of 2025? We will hit it off with a shopping list.
Got myself a BAMBU LAB P1S 3d printer and I must say WOW. I've resisted the urge for along time (partly because I was badly burned with a CNC machine I ordered which never arrived) but finally called it in November 2025. I've since printed countless parts, crafting tools (guides, center finders and the like) and toys for my kids.
I am constantly amazed by the collective madness that shines though all the models shared freely on the web. For me it's open source 2.0 it opened the door to random repairs and literally doubled or even tripled the speed at which some household projects are done (mostly thanks to the guides and tools I made for power tools)
It's a great learning tool - we've printed multiple gears, moving objects and even instruments already. My son is very much into that, so he likes it a lot - a cherry on top being we can choose the colors of many of them! Totally recommend on for anyone curious about how stuff works and eager to make some stuff themselves.
One probably obvious confession - part of the things I printed were needed only because I got the printer. Think of the spools for example :)









Anything from cardboard screws, through fidgets, filament holders, whistles and air raid sirens.
My son received a set of Makedo tools. We've had a lot of fun with it already and liked it so much that we've printed around 10 sets & a couple hundred of screws to gift it to our friends and family. This is hands down the best gift I came up with seeing them play together and plan their builds is really fun.




Boss and his makedo creations, works great at last minute costume making
Next on the list is a more powerful power drill and impact driver and again.... WOW. What used to take 3 hours before (using 12V one switched to 18V) now takes maybe 40 minutes. The never ending story of Boboland continues, we will be adding more there soon 😄
We've built his new bed up together (admittedly the old screwdriver was enough for that) and he got a good lesson at how powerful the tools are (well I got one about what he can do on his own as well). We let our kids use any tools we us with supervision, but of course he was exploring boundaries and trying to help. He fastened a 12 mm hole drill to the new LXT and I found it stuck upright in a half drilled hole. The sneaky handyman told me it's a bad tool cause it dropped on his foot, but it's all good.
Bed, pegboard, horizontal ladder and more incoming. We've also repaired our version of Infento Pedalo.





The Boboland in all it's glory, old Makita, work in progress ladder, and Infento Pedalo.
Tech and else
I am working with Node.js again, and JavaScript ecosystem remains beautifully complex to build and work with effectively. This article still holds over 5 years after it was initially written.
We've quite a big setup, with a reasonable amount of proper integration tests. This makes stuff slow - because clean rebuild on Mac M4 24GB takes ~16 seconds with tsc as the compiler.
What I've done to optimize the process and allow me to use my favourite testing tool - Wallaby.js is:
- add path aliases for ts and js - this was surprisingly complex because our tests depended directly on the dist folder
- mapped the files properly so that source maps on test errors take you to the actual source code not the dist folder
- add swc as optional watcher and compiler - so that old workflows still work as before - useful to compare result later
- one wacky thing I needed to do is monkey patch sinon.stub and sinon.spy methods - because of how the compilation happens the results differed for these two methods, nothing a quick require before tests wouldn't fix
All that allowed us to use Wallaby without issues and the compilation time (suboptimal because it still removed the folder and copies all the files that don't go through swc every time you start the watcher / build) went down to <0.8 second levels.
Want to add database cloning for proper test parallelism later - which should make the CI faster - we currently use sequential runner but our runner machines are actually 4 cores.
Added somewhat functional dynamic typing support to the Sequelize we're using. This makes my life easier since I never remember what is on which model (and show me someone who remembers that when you have >100 models.
All in all working on tooling let me dig deep in new codebase and get familiar with the code that existed for at times over 10 years.
Thermal printer project.
I've found https://www.printercow.com/ somewhere and wanted to experiment with that. We already had a spare thermal printer for kids lying around as well which made things easy (that didn't stop me from getting a second, more serious looking one though).
The original idea was to put a little fax machine on my son's workbench - for now it's not there. It's in a functional state, allowing to send photos from webcam short text message or uploading images. I've opened it publicly for a couple of days and received a bunch of funny messages from friends and coworkers. Exposed through both Cloudflare & ngrok worked well.
This was also my first properly vibe coded project from start to finish. I only had to take car of connecting to the printer - since Chat got confused there. Besides that I was merely changing the names of parameters, letting Chat GPT do the rest.
Managed to get both the USB and Bluetooth connection working.
Funny thing - the "serious" thermal printer will actually be used by a little bistro my friends run - they've had some issues with mixing up orders last year (their kitchen is quite far from the register) and this setup will work great for their use case and cost them 15 $ and 2 beers.


Converting simple Infento gokart to a snowmobile.
Not much to say here, unscrewed the front wheels. Locked the side slides, mounted a deeper slide in the center to allow turning. I later on weighted the back of the vehicle, because one kid sitting on the front was sometimes not enough to push the power wheel to the ground. One thing I hope to change if the snow stay around for long enough is get wheels with deeper tread as the current one sometimes spins.


That's roughly it. I will save the reading list for another post!
Member discussion