lsmtree VS db_tutorial

For learning, I built components of the database engine like an SQL parser, a file-based B+ tree, a log-structured merge-tree. And I planned to merge them into the simple database engine. But I paused the project due to the changes in life obstacles.

I am trying to recreate work from https://cstack.github.io/db_tutorial/ Basically it is recreating very simple SQLite, loading data from a file at the start of the program, and storing data in memory when the program runs then saving data to a file when we close the program. I believe that my issues are connected to working in the Windows instead of the Linux. Right now I am at part 5 when we start saving our data to the file when we close our program.

Stoped open source contributions recently, will be working on Build your Own Text Editor and Let's Build a Simple Database probably after landing an internship or a job

Last I heard, micro controllers still extensively use C and C-like sibling languages (like The Arduino Language). Most system applications use UI kits like Qt or GTK, which are also C libraries. Most of the Linux kernel consists of C code. Both SQL and NoSQL database backend are written mostly in C, with C++ coming in second, and only a few written in Rust, Zig or Golang (if you're interested in creating your own simple database, this is a great tutorial). TensorFlow and Pytorch, the machine learning libraries used for creating ANN-based architectures, uses libtensorflow and libtorch behind the scene that uses a mix of C and C++ libraries, then there's also GPU stuff like CUDA and RoCM that uses a mix of Fortran, C and Julia. Vulkan is also written in C and is supported in C++. Almost all performant libraries in Ruby, Python, Rust or Julia uses C bindings.

Have you thought about writing a DB from scratch in C? You'll learn lots of cool stuff about memory and OS abstractions: cstack.github.io/db_tutorial/

There's an online tutorial started a while back: https://cstack.github.io/db_tutorial