What programming/technical projects have you been working on?

I could list a couple of half-baked going nowhere projects, but let me do something else. For the last two-and-a-half years, I have been working alone on a backend in a remote job. No-one reviewing my work, no learning from the work of others and also not shipping anything because the priorities keep changing. I am starting to feel burned out.

To remedy this, I have arranged a break for six weeks to two months with the specific goal of participating in projects with others. I have an eye on a few open-source projects that I perhaps could join for the time. But if anyone here needs the help of a seasoned Go developer, or knows of a project that could use the help: I am all ears.

For a while, I have had the wish to start with Rust, so that would be cool too. But so far, I have not gotten beyond the hello world example, so I would probably not be of much use there. A project that combines the two would be golden though.

Have been learning C for about 6 months now, maybe 5-6 hours a week when I can fit it in, mostly just working through exercises in a textbook. It's fun, but it's a lot of stuff to internalise. Haven't even gotten to pointers yet. The end goal is to make some kind of GUI music notation application, probably switching over to C++ eventually but you never know! Going to be a long road before I can reach that kind of programming expertise though.

I got my audio player that is using a Raspberry Pi and a clickwheel to control Spotify working on the software side, which is really exciting. Currently I am working on cleaning up and documenting the code, to prepare for adding in new features.

On the hardware side of things I have everything now, but the battery controller appears to be defective so that will need to be returned. It is a bit disappointing that it is defective, as I was about to create a demo video to share of the project in its current state before I put everything into an enclosure. I do have all the hardware at the moment before I return the battery controller, so I am going to take measurements and figure out 3D modelling software to 3D print an enclosure for the project

I removed all the key caps and scissor switches on my Freewrite Traveler, cleaned them all and the keyboard chassis, then put them back on. I wasn't great at it. I broke about 10 scissor switches. Thankfully, the company is sending replacements for a total of $5. Now I'll just need to be more careful with my inspiration Cabernet.

I'm even more proud to say that I just finished cleaning and leveling the keys and "hammers" in a Korg Concert 5000 keyboard. This involved:
-disassembling the piano to remove the keyboard
-removing each of the 88 flat springs, 88 hammers, and 88 key covers
-wiping down the keyboard chassis, lubricating all contact points with dielectric grease
-washing the key covers with detergent
-wiping down and relubricating the hammers (these were so grungy)
-lubricating contact points on the key covers
-reassemble each key complex on the keyboard
-reassemble the piano
-(along the way I also had to super glue one of the plastic hooks back on a key cover. These make sure the key cover can't be pulled upward)

It works! I just have to get the felt strip reglued tomorrow, and it will be as good as new!

Well, sort of. I bought this thing at a steep discount from a local working musician as a gift for my spouse 13 years ago. It is battered to hell on the outside, but now at least the inside isn't full of bar smoke grease crap from 100s of gigs. But the keys were so sticky after just a few months of use (and some misuse by small children along the way) that we could never stand to play it. Super great mindfulness activity, ending in a feeling of having rejuvenated something.

Honestly for the past few weeks I have been ignoring my project. It’s actually got my anxiety acting up. It’s at the point where I am having a hard time fitting it all in my mind, which is a problem because the point I am trying to work on right now is where all the functionality I have made so far is supposed to fit together.

I've been working on a project to use SDL GPU to render a many-entity simulation, like simulating a million worker ant colony. I started the project because the idea of making the simulation itself sounded fun, but I've ended up down a tangential rabbit hole of playing with the graphics stuff. Trying out different shader languages and compilers, trying different approaches to multi threading, hooking shader complication into my CMake build process, and so on. I'm having a surprising amount of fun diving into the weeds of SDL's Vulkan implementation. This has been a problem for making progress because I'm too distracted by using the project as a bit of a Slang/Vulkan/SDL playground.

Edit: I should've originally added what my favorite multi-threading synchronization structure has been for multi-threading GPU commands from the CPU. This favorite structure has basically been just no synchronization at all. Instead, I'm embedding all the data into a fixed set of GPU buffers, and drawing whatever their latest state is at any given time. So I can do things like change the number of entities being rendered without synchronizing the main rendering thread and the main simulation thread because the simulation thread embeds all that data into the buffer it is writing to. The main rendering thread just makes an identical indirect draw call every frame without needing to know what's in the buffer.

A while back I posted about rmk, a rust keyboard firmware based on embassy. I made a configuration for my ferris sweep, but couldnt get the communication between the 2 halves working without lowering the baud rate by 10x. Turns out, this is a limitation of the RP2040, where its internal pull up resistors are somewhat weak, meaning that if you have a longer TRRS cable between your split halves, or even a badly routed trace, it can degrade comms.

Turns out, QMK (the more popular C-based firmware for keyboards), also had to fight this problem, and came up with a solution to it, essentially by making the receiving side weakly pull the line up to simulate a really strong pull-up resistor. I implemented this solution in rust, and built a breadboard based 'keyboard' with 2 keys and 2 pi picos to test it without risking my expensive keyboard controllers. (Even though QMK uses this solution, I was still wary about making 2 pins pull a line in opposite directions, its the sort of thing that burns out pins.)

Anyway, the tests went perfectly, and my pull request was just merged!

I have been working on making a handwired mechanical keyboard, specifically the ScottoErgo design by the youtuber JoeScotto. It's one of those small ones so it only has 36 switches. Though I modified his design to use a USB C connector instead of the aviator connector he used.

The past couple days I've been playing around with cookies and I made a little cookie testing site: https://unli.xyz/cookietest.html

And I was wondering why the cookieDisplay wasn't showing up in Firefox but not in Chromium and it was because of my I don't care about cookies addon which was hiding that div. So that gave me a silent chuckle and only took about 30 seconds to figure out (I renamed cookieDisplay to c: that's good enough for me).

This next part is more vexing... I'm trying to use the site to check that cookies are maintained across page loads:

python -m library.lb text --selenium --cookies-from-browser firefox https://unli.xyz/cookietest.html?cookieValue=t https://unli.xyz/cookietest.html?cookieValue=test2 https://unli.xyz/cookietest.html?cookieValue -vv # --manual

https://github.com/chapmanjacobd/library/commit/d9d2488081f6844bcbf7d565d6ad40a871213985

But the site is behaving differently inside of selenium and outside--it's almost like it can't read the cookie once I set it after the first page load. It's probably something simple but the solution is eluding me for the past 45 mins... I'll take a short break.

I'm not currently working on it, but planning on soon developing an Obsidian plugin to allow me to recursively copy all headlines and sub-headlines from a file or a folder (all files within that folder). Simple little project but would be incredibly useful for me as I document anything that I'm learning on Obsidian. So like if I encounter some question / pattern and other stuff I make sure to note it down so next time I can quickly look it up. I also sometimes include the material itself (like from a book).

So it would allow me to quickly copy all headlines from a folder and feed it into an LLM to help me search through what I'm looking for and where it could be for cases when ordinary search doesn't work that well (like if I don't recall precisely how the thing I'm searching for was worded) I'm talking about hundreds or thousands of headlines /sub-headlines here over lots of files here, by the way.

I haven't found anything on the plugin store that would do that.

Been knee deep in a refactor for a bit that's actually gone quite well in the end. I have datasets that have various rules applied to them. I need to remove records that fail some rules ( but still output them at the end so i can investigate why), and then after those I need to do a cleanup pass to adjust some records.

Reduced the logic of it all to something easier dealing with ints for feedback but it's probably one of the things i'm happier with. Lots of useage of the things that make F# great, and should be very reusable in my workflows.

type Data =
    {   GoodRecords: int array
        ErrorRecords: int array
        Modifier: int }

module Remove = 
    let removeTemplate logic record = if logic then Error record else Ok record
    let notEven record = removeTemplate (record % 2 <> 0) record
    let is0 record = removeTemplate (record = 0) record

    let rules =
        [|  notEven
            is0 |]

module Clean = 
    let cleanNames record : Result<int,int> =
        Ok record

    let doStuff record : Result<int,int>=
        Ok record

    let applyModifier modifier record : Result<int,int> =
        let newRecord = modifier + record
        Ok newRecord

    let rules modifier =
        [|  cleanNames
            doStuff
            applyModifier modifier |]

let applyRules rules record =
    rules
    |> Array.fold (fun record rule -> Result.bind rule record) record

let unwrapResult r = Result.defaultWith id r
let allRules modifier =
    Clean.rules modifier
    |> Array.append Remove.rules

let Apply data =
    let good, errors =
        data.GoodRecords
        |> Array.map(fun record -> applyRules (allRules data.Modifier) (Ok record))
        |> Array.partition Result.isOk

    { data with 
        GoodRecords = good |> Array.map unwrapResult
        ErrorRecords = errors |> Array.map unwrapResult }

let testData =
    {   GoodRecords = [| 1; 2; 3; 4|]
        ErrorRecords = [||]
        Modifier = 1}

Apply testData