Lukan Snopovs

About me & Contact information

Journal

• New projects - 2025-12-02T01:23:06.100Z ago

  I've started development on 0bchat - A end to end encrypted chat client using C. With PLib V5.0.1 I've started writing navtui, a navidrome client terminal application, both of these projects are privated at the moment but will go public on their first release. I don't want another incident with unfinished repos ehm motionlayer..

  ---

• ew protonmail - 2025-12-04T00:03:53Z ago

  I've been using proton mail for a few months now, and don't get me wrong I like proton as a company and I understand that they're just trying to make some funding. Personally I'm not a big fan of the corperate presence thats becoming the norm. Now I like to use the aerc email client, small issue with this, proton dosent allow users to use clients other then their offical client.. The only way theyve provided to negate this is to have a proton-mail-plus subscription AND a desktop application (protonmail-bridge-core) which acts as the other end of the encryption that way the client using the proton email can actually read emails.. I have no issue with some features being behind a paywall again proton do need to make some money somehow but this is such a basic feature.. To fix this issue I've created a new email at disroot.org which provides end-to-end symetric encryption using pgp, disroot.org also supports 3rd party clients out of the box. I've set up my accounts so that my proton account forwards emails to my disroot account but in doing so I also noticed that proton does not forward emails sent from its own address so emails sent from your own email to yourself will not be forewarded.. ill also say that there are very few options given to actually change the forwarding settings.. In gmail the options to delete, archive or keep forwarded mail is right there but in proton no such feature exists. Ontop of all of this disroot also has a contact book system that allows you to import your contacts from other services eg google contacts, this is yet another feature that protonmail does not have. This website still shows a google email as a contact email all emails sent to that address will be forwarded to my disroot email. go check out disroot.org, note that they are donation run and therefore have limited time and because of this they close registrations during the weekend. After signing up you have a 48 hour period where a maintainer will have to manually approve your account request. There is no mention of the prior when signing up. EDIT, recently email forwarding was disabled as it is now a premium service!

  ---

• I Use Arch BTW - 2025-12-14T10:48:32Z ago

  The title is a joke cowboy.. calm down. I've been using Archlinux for quite a while now.. 3 years. I recently decided that Archlinux is far too mainstream and so I set out on finding a new OS that could fit my needs as a developer. As a developer I like to have the most up-to-date packages along with quick installation of said packages. When I install an OS I like it to have either a quick and easy setup but a long post-install setup (think BSD) OR a very advanced setup and some other features that make it worth my time. Don't take my last statement as me saying I dislike installing advanced systems, just imagine I want to quickly re-install my development environment onto a new device or from a reinstall, I don't want to have to spend hours compiling the kernal or configuring locales. Now the first OS set I tried was BSD based systems. The few I tried where - OpenBSD - FreeBSD - NetBSD (in order) Now OpenBSD was really nice when installing, just like the other BSD varients it had a very basic installation process using a TUI setup process. I found all of the BSD versions very enjoyable to install. After first installing OpenBSD it prompty took 30 minutes to boot. This was likely due to the fact that OpenBSD is very secure and it performs several memory checks and other system checks on boot, the other more likely reason was that it was the first boot. FreeBSD had about the same installation process give or take, I will say FreeBSD was the only BSD version that I had looked into before. FreeBSD was the next OS I installed, Now I ended up cutting my visit short due to FreeBSD's lack of support for my Realtek ethernet driver, this was because I had downloaded the bootonly option which dosent actually come with full driver support so really this was my fault but don't worry I came back to give it another shot. Now NetBSD was cool and all and it was the first BSD version I actually got to boot, I played around with the default X11 environment before trying to fix the very apparent issue with the lack of graphics drivers which caused X11 to be in 1260x680p which is not ideal on a 1920x1080p monitor.. After some googling I found out that NetBSD dident even have a package for the drivers I needed. I switched back to FreeBSD and figured out how to install graphics drivers. I gotta say FreeBSD was a great system to use. Sure packages where annoying to install but it was a sacrafice I was willing to take. My computer is used for two things and two things only, One being development but two being playing games with my friends. I was a bit disapointed that there is no simple BSD alternative build for either Discord or Steam and the few that where available where buggy and complicated, and I'm far to stupid to even attempt such things. I decided to push the lack of support aside and do some developing just to see if the development experience could make up for it. I like to develop in my own custom environment built offof sucklesses window manager. At the time my configuration used polybar which is a popular lightweight statusbar program, this and a few other dependancies are required for my build to work. I spent an aditional Day installing polybar and the other dependancies and found that the package manager provided was so outdated that I ended up having to manually install from source every time I wanted something up to date. Now again like I said before I have no issue with compiling things myself its no hardship just the time crunch is not acceptable when I'm busy and need a package fast. Now after all of the BSD stuff I started looking into more advanced linux distros. a few that caught my eye where Gentoo and Slackware. First I tried slackware before figuring out the absolutly useless package manager and pre-installed bloat so I immediatly to SalixOS which is built ontop of slackware but also has a more usable backage manager slapt-get. I used slackware for a little until I ran into the same problem as before where packages where outdated and some where just not available. More infuriating Slackware has an online user driven repository of source packages that let you download a packages source code along with a file which can be used to compile the program. This system just reminded me of the arch AUR and PKGBUILDs. Now the issue with this slackware system is that packages are not maintained at all so it was common to find packages that EXISTED but where just uninstallable, and of course this meant installing from source and compiling manually.. nice stuff. Next on the list was Gentoo which was a day long commute to install. Gentoo comes with Systemd which is nice.. uh yeah but guess what the package manager is a compile local package mananger so when installing a package you run a very basic command, yes yes good good, then you wait 20 buisness days for it to download the source and all the dependancies before locally compiling and installing.. holy moly what a waste of time. I'm very sure that it was me that was making these issues so apparent and I'm sure there are millions of alternative routes I could have taken. I've ended up back with arch linux which i guess solved my issue with it being mainstream.. it is mainstream because it is so easy to install now-a-days so every idiot installs it just to say they use linux while running a base kde or gnome setup. I absolultly hate those kinds of people who use something for the main reason of status or reputation.. This was the initial reason I wanted to move over to another OS. I wanted an OS that was not diluted and poisoned by outliers in the community. I guess this week long journey has given me proof that I have an actual reason to be using arch. After all that fuss I think arch is the best distro because it has both up-to-date packages and a comprehensive installation and customization process.

  ---

• RE: I Use Arch BTW - 2026-03-06T07:52:00.961Z ago

  I Installed CRUX linux a day before Tony Banters (@tonybtw) made a video on it. I Used FreeBSD for a long while and I was interested in CRUX mainly for its port-style package management system. I installed CRUX and off the bat the installation was pretty straight forward. The CRUX 3.8 Handbook was very in-depth about how to install and setup CRUX, the only thing I added that wasn't in the handbook was full disk encryption which worked seamlessly. CRUX has a script styled init similar to FreeBSD and so this system seemed to be the highest contender for a Linux distrobution that felt like FreeBSD. After installing CRUX linux it took about a week of bugfixing for things like steam and my graphics card to work but after that the experience has been perfect, the package system is amazing and minimal and the init system is dead-simple, zero systemd OR openrc bloat..

  ---

• Projects update - 2026-03-06T08:20:43.811Z ago

  Plib has gotten to the point where I see no new features other than minor QOL macros, this is good, check it out in the projects section of this site, Plate, my template generator has also come a long way, not anywhere near perfection like Plib but its getting there! This site you're viewing was generated statically using Plate! Check it out in the projects above too!

  ---

• Is there no such thing as a cheap split-layout keyboard? - 2026-03-06T08:22:45.366Z ago

  The Corne V4 series split keyboards are amazing and I would get one if they weren't 500 of my currency...

  ---

• Thinkpad 10 working!! - 2026-03-06T08:25:36.792Z ago

  I've had this old Thinkpad 10 Tablet (Gen 1) which has been sitting rather sadly on my bench for a few months now right next to my Thinkpad T560.. Today a new charger finally arrived and I was able to start debugging the several battery issues I ran into.. Luckily most of the issues where resolved by simply re-seating the battery header cables, the only other functionality affected was the physical docking keyboard mount that will light up but not work.. I intend on using the Thinkpad 10 for my everyday carry, y'know dumbphones EDC style..

  ---

• [New project] BLUI - 2026-03-12T08:21:45.998Z ago

  BLUI is a basic api for creating Terminal User Interfaces (TUI). BLUI is incredibly small and robust, BLUI is cross platform and works on Windows, Linux and MacOS. BLUI is compatable with both C and C++. HOW IT WORKS: Inside the header file it has one global variable called _BLUI, this is where all the data for EVERYTHING is held. You can use a function to append a new "element" to the main app structure. Each "element" can have one of two types, it can either be a box or a text "element", text elements can be aligned to the left right or center whereas all "elements" have properties for x and y position on the screen. BLUI doesn't have any input management yet, this is on my to-do list! Feel free to send in a PR or issue!

  ---

• Update on thinkpad + working T560 - 2026-03-19T01:53:35.460Z ago

  The Thinkpad 10 (gen 1) has been put to rest, sadly it was not a viable machine for daily work, the only USB slot did not seem to function nor did the external keyboard, this means there was no way to write using an external keyboard or even flash a new operating system. I had another thinkpad though, a T560 which I received with a few missing parts. It had no internet card, ram or SSD, the monsters even took the motherboard connector for the SSD! Thankfully I had 16GB (2x8GB) DDR3 that I could use for it, I also had a spare SATA M.2 drive I could use for storage drive. I ended up purchasing a T560 Motherboard header to SATA connector and a NVME to SATA converter so I could use the M.2 in place of the SATA SSD it had originally. While I waited for the parts I used a USB to boot an OS off of. The OS(s) I used for this was TinyCore linux and then AntiX Linux, both of which are made for ancient hardware. Anyway, I finally got the new parts for the T560 in the mail yesterday and I put it all together into a fully working T560! I can say it works perfectly now.

  ---

• New projects update - 2026-03-19T21:05:43.067Z ago

  Plib 7.x.x is in development, new features like combined shorthand and a better interface system will be included in the new version, required and extra features will be separated into a plib7.h and plib7e.h (e for extra).

  ---

• new DWM updates! - 2026-03-26T05:53:21.442Z ago

  Hey, this is very personal but my custom build of Sucklesses Dynamic Window Manager [DWM] has gotten some great new updates! Firstly I've fixed the long time issue where ST (suckless terminal) dosn't show up on the titlebar correctly.. this was fixed by patching ST with the title parsing fix patch. I've now completly removed the use of lemonbar and now I'm running a full native DWMBar system! I had to patch status2d with lots of custom code to allow for positioning text in the center and right side of the bottom bar! this allowed me to port my old lemonbar-bar style over to dwmbar, this wasn't too hard as my bar was designed to be modular.. anyway. check it out it you want too..

  ---

• RE: Is there no such thing as a cheap split-layout keyboard? - 2026-04-03T05:53:48.367Z ago

  Hi! Guess what I found.. I found a cheap Corne V4.1 on ebay for under 70 dollars! (61.20 to be exact). This was a great find as the pcbs alone cost more then that. This particular Corne has both QMK and VIAL support which is great, it has both RGB and because its the V4.1 it has no screens which I did not like on the Corne V3's.. Its the wired version which I also prefer. The entire build costs around $100 (my local currency) but I'm able to sell the board for over double that as there is a very small and concentrated market for Cornes in my country.. Hopefully the Corne lives up to its reputation, I've heard they are great for forcing yourself to learn 10-finger-typing and they reduce wrist-strain because of their staggered design. I've wanted a Corne keyboard for so long and I'm quite happy this showed up! I'll post another comment about how my typing style has changed after using it. Side-note, I find it quite funny to think about how people use my computer. This is because I use DWM which already is quite keyboard driven, I've made majority of my computer operable through keybinds alone. I switched to DVORAK layout a while ago and that will put a lot of people off this setup and now lastly I'm moving over from a 61 key keyboard (custom RK61) to a 46 key Corne V4.1..

  ---

• Plate TODOs - 2026-04-06T12:29:34.919Z ago

  Hello! If you're reading this then im sure yourve seen the mess above in the projects section, this is due to PLATE bugs that I haven't solved yet. I'm busy with other commitments (school) so I will not be able to attend to these issues, I'm sorry for the inconvenience!

  ---

• Corne V4.1 updates - 2026-05-05T00:08:36.004Z ago

  Order list: - Corne V4.1 PCB & 3D printed case (navy blue) - Gateron White linear switches - ORings - Low profile semi-transparent keycaps Total cost: under $100 NZD! Modifications: - I used the orings on the keycaps ontop of the gateron white switches and holy moly it sounds very nice. Better then my other keyboard (Gamakay Phoenix linear switches) - The keyboard came in navy blue, sadly there where no alternative colors so this is what I get, the keyboard also came with a weird material finish on it which i do not like. To amend this, I sanded the keyboard until smooth, sadly I ended up rounding off the staggered design of the case at the back but I dont really care too much. After sanding I painted the keyboard a gloss black. - The keyboard only had QMK firmware and I dont like the QMK firmware UI so I went ahead and flashed the VIA/VIAL firmware onto the Corne. Thats all on the modifications so far, Im planning on coin+tape modding it if i need too. I also intend on adding magnetic pucks to allow for it to be used on a stand Issues: - Lack of number row, I knew that this keyboard would be a challenge when I was researching it, The lack of a number row has been quite the challenge and I ended up adding a new layer (super cool feature) that lets me access all of the number-related functions. - Non-staggered keyboard rows. On a regular keyboard the keys are staggered outwards from bottom to top. The 'a' key is 0.5u to the left of the 'q' key and the 'z' key is nearly 1u to the left of the 'q' key. This has been quite disorientating as Im used to the bottom row being ~1 key to the left whereas on the corne there is no staggering so all the 'q', 'a' and 'z' keys are in the exact same column. I've been struggling with re-learning the position of keys such as 'b' and 'v'. - The color of the case was horrible, who wants a dark navy keyboard.. That was one of the first things I wanted to do (paint the case). - New keybinds: The corne keyboard uses layered binds as a compromise to only having 46 keys. A layer is another level of keybinds that can be toggled on and off in a few different ways, the main way I've used layers is provide a way to use the number keys. I have a layer that has the top row being the number keys the home row being the corrosponding symbols for each key and then finally on the bottom row ive put macro binds that press Mod+n which on my desktop changes the workspace. Ive put a bind on the bottom modrow (bottom three keys) that toggles the layer to the number layer WHEN HELD so in order to press a key you have to press the LT1 (change to layer one) and then while holding it you have access to the described number, symbol and macro rows. - Firmware issues, for some reason when removing the keyboard while vial is running (the firmware tool for the corne) the binds disappear, this is mostly a me issue. Other: - This keyboard has a bit of a learning curve but I think that its put me a step in the right direction towards losing some of my bad typing habits such as ten-finger-typing and correct finger placement for keys like 'b' and 'y'. - I specifically bought the wired version of this keyboard because the wireless version has issues with the firmware, I'm also lazy and I don't want to have to charge the battery all the time. - Adding onto the prior list item; The RGB on this keyboard is amazing, I didn't buy it for the lighting but I'm truly glad it came with it. The lights are very bright and thanks to the VIAL firmware its incredibly customisable and you can even set binds through the firmware that lets you change the lighting mode through keybinds. All in all the keyboard has a large learning curve but its very much worth it.

  ---

• Updates on projects - 2026-05-05T00:11:20.775Z ago

  Due to large school projects I've been unable to work on any personal projects, I do intend on updating my website using a PHP backend. maybe a new markdown system for the posts on this page. I'm not sure! I'll update PLATE in a few weeks. It needs a new rebuild.. Sorry for the state of this page, I had to roll it back to fit the capabilities of Plate.

  ---

• Update on Corne V4.1 - 2026-06-07T23:01:41.331Z ago

  It came with VIAL out of the box, I've created a new repo "crkbd_qmk" that has a new and improved QMK implementation of my setup. I however have had a small incident with a hardware issue. The corne keyboard does not handle hotplugging well. Whist I was flashing QMK firmware, I must have accidentally pulled out the connection TRS/TRRS cable while the board was still powered, this has lead to irreversable MCU RP2040 damage. This cannot be fixed by me so I've bought another Corne V4.1 PCB.. This issue is not a undefined behavior equals broken board issue, what I mean by this is that; if you treat the keyboard well and you're careful around known issues (removing connection cable) then the board will last a long time..

  ---

• I made a IRC-like server - 2026-06-11T22:52:11.988Z ago

  Over the last two days I have been writing a custom relay-like server for fun. My server is multi-threaded and allows for multiple clients to connect to the server. My server has just two endpoints, one for posting a message and another for streaming messages sent to the server. This server is slightly different to an IRC server, my server uses a set-size backlog that holds onto a certain amount of the last sent messages, this means that when a client connects to the server stream, the last 10 messages can be displayed. IRC does not do this and usually has a temporary queue for managing multiple messages sent at the same time. As long as messages sent to the server abide by the size limit, the server does not change or modify messages at all, this was intentional as its up the CLIENT to handle how connections occur. You can view the files for this project under TRCS, above, in the projects section of this page. Included in the repo are a few scripts that let you test the server in plaintext, read the README file for information about this. --- Along with writing the entire server (its written in ANSI C99), I've started writing a E2E client using the Fernet cryptography library from python. I've started experimenting with Cython which allows trans-interpretation from Python to C. So with the Fernet code (client/fernet.pyx) being only a few lines, I've decided to try and make a fully working client in under 100 LOC (lines of code). Thank you for reading, check out the project if youre interested, email me at the contact options above if you have any questions.

  ---

• TRCS in-depth breakdown - 2026-06-11T23:47:38.144Z ago

  TRCS is a mostly-single-file 400 line program written in ANSI C99. Connection loop: TRCS sets up the socket connection before starting a basic client-accept loop, this loop will basically just await a new client connection. If you're unaware, when I refer to a "client" I refer to any incoming connection to my server, this could refer to a person connecting to the server on their browser or a person pinging the server locally, just to name some examples. What are headers? Headers are used as per-request configurations, you can use `curl -i ...` to print the HTTP headers when connecting to an address. The server returns certain headers to indicate return statuses too. For example, if you connect to my server it might return "200 OK" which signifies a successful connection, other more known http codes are 404 (not found) and 500 (internal error). Thread stage1: When a client connects to the server a new thread is created which is given the unique client_fd of the connecting client. When connecting to the server, clients send a bunch of information including the http endpoint requested. Here is an example of what the server receives in HTTP headers if you where to connect to "localhost:6060/hello/world": GET /hello/world HTTP/1.1 Host: localhost:6060 User-Agent: curl/8.20.0 Accept: */* Back to the client thread, the threads one job is to filter through the clients headers and fetch the endpoint, in this case the endpoint is "/hello/world". this endpoint is then passed off to the parsing function "thread_handle_path()", this function splits the endpoint into seperate parts. Lets assume that we are iterating through the parts of the endpoint, "hello" being the first part and "world" being the second part. My server expects the following endpoint format, the first part should be either "sock" to indicate the client wants to start a stream, or "post" to indicate the client is trying to post a message. The second part must be the ID of the server, when stating the server you are required to provide an ID as the last argument, this ID most be given in the second part of the request, this is used asa simple form of authentication, it is not very secure but it works for now, later on I'll write a rolling key system that would work in its place. Now for the third part, if the client is trying to post a new message an additional third part must be given which should contain the URL-safe text content that will be added as a new message on the server. When the thread_handle_path() function is called, it will first claim its thread, this refers to the system in place that limits the amount of total clients/threads that can connect to the server at one time. This works by storing an array of available threads with each value being a dummy value of -1, (this value is referenced as THREAD_EMPTY in the code). When a thread wants to claim a place in the thread array, it will search for the first index in the array that is equal to THREAD_EMPTY, this index is then set to a value that is not THREAD_EMPTY to show that that item in the array is allocated to a thread. The messaging system / posts: To send a message the client must connect to /post with the appropriate part format, when this is authenticated (through thread_handle_path()) the given message will be appended to a char* array called "backlog" this backlog holds a set amount of messages, by default this is ten. The backlogs job is to work as a sort of temporary queue to hold messages if multiple messages where sent at the same time. If instead I just relayed incoming messages directly to the client threads, because the threads only check for new messages every so often (and not all the time which would be impossible for computational and resource limits) it means that if messages where sent within the same interval there is a large chance that one of these messages will be overwritten in memory before the client threads are able to relay it back to the client. This backlog solves this issue by holding the last 10 messages which means that if threads check for new messages every one second, a maximum of 10 messages can be sent and handled correctly at one time. You might be wondering "What about if there is more then 10 messages sent at one time?" then messages will overflow and yes messages will be lost, to combat this indefinitely I added a MAX_THREAD value which limits the amount of clients that can connect to the server at one time. If this value is equal to ten it means that if every user that could possibly be connected sent a message at the exact same time, then the server would be able to handle this fine. Streaming: TRCS uses a basic SSE (Server-Send-Events) system for maintaining a constant connection from a client to the server. This is very easily done using the pre-defined HTTP headers from before, you just need to include the following options for it to start as a SSE: Content-type: text/event-stream Cache-Control: no-cache, no-store, must-revalidate Connection: keep-alive Back to the thread claiming system, the reason the thread needs to claim a thread is not just for thread limiting but also for message syncing, the thread will update its index integer with the total amount of messages that thread has sent to the client. In the TRCS config file you can define how many times a thread checks for new messages a second, the default is once every second. When a thread checks for new messages it simply removes the threads total message count by a server-stored, if the resulting value is not equal to 0 it means that the thread is behind, when this occurs the thread will then loop through the backlog and print out the messages that it missed. When a client exits the thread detects this (when the thread sends data to the streaming client and it fails it means the client has closed the connection) the thread sets its index in the threads array back to THREAD_EMPTY so that another thread can open up and use it. One more thing to note is that if the user is POSTING a message, the thread will not check to see if there is an empty thread, this will only occur on stream connections. Again, like the last post, if you have any questions, email me using the contact information above. This project is still a work in progress so if you experience any issues please create an issue on github or again, email me. Thank you for reading!

  ---