What programming/technical projects have you been working on?

I've been playing around with arduino lately. I bought a big bag of random components off of aliexpress to play around with. For my first real-use project (as in, not just messing around for fun), I decided to create a little notification gadget that would blink and play a tune when I received a notification on my computer. I've got the arduino side done-ish so if you send it a "p" over serial, it'll trigger the notification code. I'm mostly interested in notifications for Slack and Gmail, but I haven't done much research into how to have those notifications trigger the device's.

In typical ADHD-me fashion, I got distracted by some rabbit trail in the middle of the project. The tutorial I was using to understand how to interface with the speaker included some code for a tune player. You fed it an array of notes and durations and it would play the tune back accordingly. So then of course I just had to find some familiar songs, which ultimately led to this repo of Linux beep commands for a few songs. Manually converting the /beep commands to the Note objects I'd created was slow, so I wrote a JS script to use regex to extract the command arguments and dump them out in the console as an array of Note objects. Then I'd just copy/paste those into my C++ code and I was good to go aside from some minor tweaks to the playback speed.

Forgive the code, because even this simple bit of code is the most complex C++ I've ever written. The jump from single-file to multi-file was a HUGE leap for me. I'm JS and PHP guy...I like my languages high level. So even getting to this point was a huge effort on my part. On the upside, all this typed code is going to make my transition to Typescript in the next few months a little easier to wrap my head around.

https://github.com/jd13313/nano-notifier/tree/master/src

I'm open to feedback on the code though. C++ is completely new to me, so even the most basic of suggestions would be extremely helpful.

I've been enjoying messing around with an Inkplate 10. It's an ESP32 programmable e-ink screen that can be used as a photo gallery, weather station or anything you can code up in Micropython or Arduino.

I use it as a weather report and a slideshow of kid photos.

I've been working on rescaffold, which is a tool to unpack reusable parts of codebases (scaffolds) into multiple projects as needed. It's roughly similar to yo/create-react-app and the like, but it can update a codebase to use updated versions of the scaffolds in it!

It's functional enough to actually be used, even though it can't git pull yet, and doesn't have sophisticated conflict management. But it can unpack a scaffold into an existing project, with template replacement for configurable things like project names, and it can also update existing files (and won't overwrite files you've changed yourself). It will also remove files that have been removed in the updated version scaffold they came from.

For example, you can make a small scaffold that adds source for an executable to a Go project. It might just contain one file: pkgs/go/cmd/<command name>/main.go. Adding this scaffold will ask you what the project name should be. If you say "foo", it'll create pkgs/go/cmd/foo/main.go, creating any directories that don't exist yet, and leaving everything else alone. Then, if the scaffold source is updated, you can update the file with the new contents, but only if you haven't modified the file yourself. You can use many small scaffolds this way, or few huge ones, and it should just all work together!

My current collection of projects are all similar- web themed.

On top of steadily expanding upon my template git repository (which can be specialised for a number of different project types through the use of a python script [py, c++ executable, c++ lib, http]), I've been working on a number of common or shared tools and code between my personal landing page, a portfolio which I'm procrastinating on populating with text (I am hopeless at the whole self-advertising thing), and a simple mongodb based image sharing site. (Deep breath).

I've been moving away from PHP in favour of very light-weight client-sided page rendering. I'm not a fan of the idea of larger bloated JavaScript frameworks so everything is vanilla, optionally minified. All API calls are passed through an NGINX reverse proxy to a very simple low-level socket based python script I've dubbed, imaginatively, echopy. This is what interfaces with mongo.

You'd have thought that after years of approaching development in this way I'd get a little less stubborn- a little more willing to utilise existing, and I'm sure very powerful tools. Maybe not. At least I'm not producing any of this stuff commercially, eh?

Otherwise I've been working on expanding further on my personal Discord bot. Who doesn't have one at this point? The latest changes include a refactor which improves the behaviour of many of the modules while the bot resides in multiple guilds at once, as well as a module for calculating, given a number of entry and exit dates, the number of Shengen Zone days a non-EU citizen has left during a rolling period (thanks, Brexit).

I've been hacking on a few "transformers" python functions ranging from image resizing to notes transcription. It's been a good exercise in understanding scale and getting ready for a media portfolio project I'm building out.

So I got to the point where I felt my livestream translation project was usable enough to be open sourced. I think I'll still try to improve it, but even as it stands now it might be useful to some people. I called it smatter

I think @wababa said they're be interested in seeing it, so I'm tagging them here.

I'm hoping to make it possible to host a local "restream" next, with subtitles included. Ideally this makes it possible to watch it on a mobile device or TV.

I've been working on my website. Basically it's just one project with a thousand little sub-projects that I can jump between.

Recently, I spent a bunch of time implementing a "forum" like feature that will probably never be used. It has some of the basic features implemented, like sign ups, login/logout, posting, and commenting, but a lot of other basic features aren't there yet. Things like notifications, replying to comments, post tagging, or even comment/post deletion don't exist at the moment.

I plan on implementing all of that down the line, but got busy with the recent holiday and work beginning to pick up.

My implementation has all just been plain PHP and a MySQL database for the backend, and a bit of Bootstrap theming for the frontend. Quick, simple, and easy to implement. The only issue is that Bootstrap is a much bigger library than I really want to deal with. My goal is to make the site as light and easy to load as possible, and Bootstrap/jQuery bumps up the minimum size quite a bit. Down the road I might redesign the look to be more in line with the simplicity I am hoping for.

I do not do my job in my free time. I have far too many other interests to indulge.

If you want to, more power to you. It's not a negative as I have done this before, but I don't think engineers spending all their free time engineering should not be normalized.

Of course, my hobbies include woodworking which is basically me getting to do that parts of making stuff that I don't normally get to do.

I'm a data engineer that does the majority of my work with the AWS platform. I'm trying to build some projects that would look attractive to game companies in Europe. I enjoy making these kind of 'one-button' cloudformation builds so what I think i'm going to do is something that depolys a game server and sets up some sort of kinesis data stream and storage. I have other demos of my work, but just trying to set up something specific to gaming. I'm confident in my ability, just not exactly sure what would look good to them.

In addition, i'm having fun fiddling with game mechanics in Unreal. I've always had this deep love of the real nitty gritty aspects of game mechanics, down to the atomic parts of how your character even walks - the choice to make them snappy or have rotation time. How they stop or ramp up basic movements. I love this sort of planning of how to give a character weight. The more i learn, the more I can see how it takes very large teams to make a game as all these little things can be investigated and ruminated upon in great depth.

Currently writing a script to grab all the namespace quotas from our clusters so we know how much CPU/memory are allocated to each app. It's also an easy way to see which namespaces have no quotas. Currently works on one cluster. I now have to update the inventory file and let it run through all the non-production clusters.

I've been learning Swift and building an RSS reader that is a Catalyst app (meaning, basically the same codebase compiles to iOS, iPadOS and MacOS). It is quite fun to work on.

I've been amazed at how easy some things are, like animations and navigation handling. Then baffled by hard some things are like capturing if the escape key was pressed.

Anyway, it's the first app I've built entirely by myself and I'm excited to eventually release it.

I've been porting postmarketOS to an old Huawei phone i had laying around, pmOS for who doesn't know is linux for phones, completely replacing android. It's actually very fun and relatively easy, i got as far to get wifi working, i haven't yet "uploaded" my work to pmOS since i couldn't get git to work properly with gitlab

Does anyone have some good boilerplate for reading + buffering ffmpeg rawinput frames into Python?

I wrote a short script to "play a video" on my RGB keyboard, by mapping the coordinates of all my keys to areas of a video frame, and sampling the average color for each key area.

Next, I hooked my system audio up to ffmpeg, and set up a filter to have ffmpeg convert my audio into a video of the waveform, so that I can have my keyboard bumping along to whatever beats are coming from the speakers/headphones.

It works... for about 15 second. But ffmpeg appears to crash and burn after that, feeding python the same stalled frame on repeat. I believe I've run into this problem working on more serious projects in the past, and IIRC, it's because I'm not processing frames fast enough and ffmpeg's buffers fill up.

What I did in the past was set up a worker thread to stage frames into a short FIFO queue, and just drop frames after that queue was filled... but the rest of this script took ~2 hrs to put together and I'm not eager to dive into the quagmire that is python concurrency hell. So - anyone got some precomposed snippets or package recs to smooth over the frame processing?

Script for reference

A few podcasts have too many ads (e.g. Smartless, Fly on the Wall, etc) -- and they also happen to have youtube videos for their episodes. I hate ads and I like yt-dlp + sponsorblock.

So I set up a cron job to pull the latest episodes from a list of playlists. It names them YYYY-MM-DD Podcast Name - Episode Title.m4a and then I have this poorly modified PHP script to generate the RSS feed that I use with PocketCasts.

Its been pretty great, to be honest. Having all of these playlists in one feed is pretty good, too.

I wouldn't mind fixing up the PHP, but I don't know PHP very well -- especially not enough to write it from scratch.

I've got a fork of a docker image/container for vsftp cloned to my github that I need to finish working on. It allows for persistent user storage despite container rebuilds/updates, which was something I haven't readily found in any other vsftp image.

I have a duplex document scanner on my LAN that I use to scan documents in to paperless-ngx via ftp. Currently I achieve that by running vsftpd on the host, but I'd prefer to move it to a container.

I have finally gotten to play around with my tiny VPS I have been paying for over a year. My motivation was mostly to circumvent the very draconian corporate firewall, which does deep packet inspection, and basically everything is blocked besides ports 80/443, and even most sites using those.

So I installed Apache Guacamole on my VPS with an RDP server on the same instance, and use it to control a graphical linux interface plainly from the browser. (Funnily enough it only works with http://IP, but not with any other combination of (http, https)://(IP, domain) due to the firewall. But it does work!)

I also bought a domain and set up https and an nginx reverse proxy, with NixOS of course. The resulting config is very elegant! Also enabled zram and normal swap, because the default 2GB RAM is very tiny for even something as lean as xfce.

Will probably follow up with moving the server to an ARM-based instance, that way I can get a better server for the same money at my chosen cloud provider.

For the last month in my freetime at work, I've been building an Excel sheet and vsb script that will create and update the locations of all my cable heads on my data center computer room. I'd love to use literally anything other than vbs, but because I'm not a dev, I don't have access to any of the tools I'd like to use.

My reason behind this project is that the panel schedules have not been maintained in over 20 years and tons of racks have been added and removed without things being updated. When I joined 9 months ago, I was horrified when I was handed a paper binder and told that was our panel schedules. I've already managed to transcribe what is accurate from those 100 panels, but the big task is still ahead, verifying what is where for every circuit in the room.

I just want my visualizer working before I start tackling that problem, but it's being a major pain because vbs.

I am completely new to programming but I’ve been toying around with twine for a bit. I’ve been trying to create a branching narrative with a lot of different flags but it’s been very slow going. It seems every time I learn something new I restart the project completely with the new techniques in mind.

My "on again, off again" project... I got distracted by buying a house. :)

I have a room full of comic books. They are 99% alphabetized and inventoried in both a flat and relational database.

What I want to do is an Alexa style query asking for a specific thing and having a series of smart lights light up the correct box(es) Wheel of Fortune style.

So far, using a Raspberry Pi, I got the database query to work. I can input a query with text, and output one or multiple box numbers.

What I wasn't able to crack was the voice input or smart light output. Then I bought a house and moved and haven't been able to get back to it. :(

I've been writting a bot for Runescapes NXT client. The game is old as hell and bots have plauged it since its inception but it was more about the process of actually figuring out how to do it than it was about botting in the game for me.

Its written in C#, it hooks into the games loaded memory and reads out the state of the objects, NPC's, inventory items, etc, etc. Then sends WinAPI messages to the client to interact. This means the whole thing can run in the background on your desktop and theoretically support multiple clients at the same time though it doesn't at the moment.

It was a bit of a challenge at first but once I laid down the foundations it was a pretty simple process to get working bots going. It has a fighter bot, an abyss crafting bot, a bunch of smaller bots its even able to do pathing (provide a list of world coords and it will walk them).

My biggest problem I currently have with it is the pointer paths keep breaking with updates and its painful to keep rebuilding them. So I'm currently working on automating that whole process. My current implementation is not as fast as Chest Engines but its getting there.

Edit: some spelling