devlog 2 - subtraction and integers

in this devlog, i added subtraction and the ability to add integers of variable length. it may not sound like much, but i’m trying to provide a good base for the rest of the interpreter. i have my obsidian notes on lexical analysis pasted below, but i’m not sure how well it will format. for the next devlog, i plan to implement operator precedence and multiplication + division, and also learn about writing tests in go (since writing tests is probably a smart thing to do)

notes

Lexical Analysis is a stage in interpretation that turns a string of text into analytical tokens, typically for a later stage like the parser.

What is a token?

A token is a type and a value. For example, an integer 8 would be converted into a token of type: Integer and value: 8.

How-to Lexical Analysis

The lexical analyzer, or lexer, breaks down a string of characters into tokens. It goes character-by-character, and has no context for the surrounding characters. It then passes the list of tokens to the interpreter to evaluate.

Lexemes

A lexeme is a collection of characters forming a token. For example, this can include simple 1 character tokens (operators, parenthesis, etc.), but also include integers, floats, etc.

A relationship between token types and lexemes can be seen in the table below:

first devlog!
i have a pretty simple first devlog. the two main things i completed are: adding two numbers and creating a project README, license, etc.
i spent a lot of time researching interpreters and learning certain terminology that is specific to them. for example, the lexer is the part of the interpreter that takes in a string and outputs a series of tokens that can be interpreted later.
i also did a lot of research into different licenses (copyleft vs. permissive), and i landed on using LGPLv3 for my project. in the future, i will be posting my Obsidian notes into my devlogs to show more of my learning.
i attached a little video demonstrating the (limited) capabilities of my calculator. i have big plans for the future however!

the crash at the end was intentional, don’t worry.

adventures in pong
the entity–component–system i designed was pretty bad. it wasn’t awful for a first attempt, but i want to redesign it to something more akin to Flecs. the ecs quickly became unruly for even simple entity relationships, and it became too much to wrangle even with a simple pong game. before tackling larger projects, i definitely want to redesign the ecs to be much more ergonomic.

future game projects will occur on other projects on hack club (i think). i still have major plans for this engine (physics, networking, misc. debug stuff), but i need to take care of some primary issues with it first. i honestly have no clue what i was thinking when i was writing the ecs.

ecs and other misc stuff
it’s been six days since my last devlog and, since i’ve been programming for ~1 hr a day, i’ve made some progress on the engine. firstly, i have a prototype ecs working! it’s not fast, and is not “technically” an ecs, but it behaves similarly to one. i plan to rewrite it one day to be more adherent to data-oriented design, but it will work for now. i also wrapped up the opengl backend and cleaned up some stuff.

the first game i plan to make with this engine is a pong clone. i don’t want to do anything crazy, and instead focus on simple stuff. i will make progressively more advanced clones of simple games until my engine is ready to tackle an original game (current idea is a coop wizarding game)

note: i recently discovered that glfwSwapInterval(1) (which is used to enable vsync) is broken on apple silicon macs! i truthfully don’t know what i’m going to do to fix it, and i think i will just deal with it until it becomes more unbearable than it already is. it’s stuttering like crazy, but there’s not much i can do.

unfortunately, i gave up on bgfx. it was a neat idea, but sort of overkill for the simple 2d engine i’m making, and it also had poor shader documentation. i switched over to opengl and it has been great. i got a textured quad rendering in ~2 hours, and it’s been much easier than bgfx. i just need to abstract some rendering stuff now, and then i just have to make an ecs and then i’m done!

this one took a while. i’m not completely sure if the time is correct, since i spent a lot of time reading documentation and looking at my code. some stuff that was added to the engine:

• generic graphic backend capability (however, bgfx should be the only one i need)
• robust resource loading
• image loading with stbi
• window resizing
  i also made several misc changes to the engine, but it’s been many hours of programming and i forget most of it. however, i got a cube rendering! very exciting stuff.

ref-counted resource management
i realized that, before rendering, i should make some sort of asset management. i went sort of overkill and created a ref-counted cached resource management system, but it’s actually pretty simple. there’s a ResourceManager, which is exposed to the api. ResourceManager can register ResourceLoaders (load and sizeOf), and then, using ResourceManager.load(path), users can load a resource after registering a loader for that type of resource. i only made TextResourceLoader so far, but i plan to implement binary, images, and possibly compiled shaders (but idk yet). i also still need to make it async, but kotlin has pretty good async capabilities, so i’m not too worried.

got bgfx working after an hour of debugging. macos really hates rendering on background threads. next: rendering structure and draw commands.

so far I have general engine structure down. however, i’m having trouble with bgfx on mac.