langatator/README.md

233 lines
4.8 KiB
Markdown

# Langatator
A very basic interpreted programming language.
## Background
The goal of this project is to create a simple implementation of a BASIC-like language just to learn a few things along the way and practice my C programming skills.
No need to do complex things, just to create a simple interpreted language that can be used to do some arithmetics
and create for example a number guessing game.
I didn't really study how others languages works beforehand, I'm just guessing how I'm implementing things so that I can make mistakes to learn from.
## Progress
- 2022-04-29 Implementation of a basic evaluator engine to evaluate arithmetic expressions and call functions
- 2022-05-15 Clear progress being able to use simple while and if though thre are many glitches
ToDo List:
- [X] pow operator
- [X] binary operators
- [X] implement basic math functions
- [X] implement random_int(min, max)
- [X] implement print_number(message)
- [X] add unit tests
- [X] allow to set variables
- [X] read line comments
- [X] add modulus operator '%'
- [ ] add support for multiple characters operators
- [ ] add inclusive operators like '!=', '>=', '<='
- [ ] add support for priority operators
- [X] add input_number() std function
- [ ] add type() std function
- [ ] add ceil() and floor() std functions
- [X] base of the CLI
- [ ] evaluate expression from stdin
- [X] read a file
- [X] if statements
- [X] while statements (with break and continue)
- [ ] repeat statement
- [ ] add functions support
- [ ] add config header file
- [ ] ability to modify keywords and customize the lang
- [ ] add static string support (just for ui)
- [ ] add print_string function
- [ ] add basic number list support
- [ ] add fully features strings support
- [ ] add [classic problem solving](https://rosettacode.org) with code examples
- [ ] add Web Assembly support and publish a demo website
## Installation
You will need to compile the code from source.
- Clone this repository
- Then compile (eg. with `make`)
I use GNU Make with GCC, but I'm sure you can use any C compilers though you may need to edit some part of the code to cope with other compilers (eg. binary constants).
## The language
You would be able to use the lang directly via CLI, via a REPL or by writing in a file (file ext `.ltor`).
One instruction set per line.
### Comments
Can only use single line comments with `#`
```
# this is a comment
```
### Data types
To begin with and order to simplify things we would only have numbers as datatypes.
When a variable is declared it can be 32 bit integer or 32 bit float.
The language may support strings literals in the future.
### Expression evaluation
function calls: func(arg_a, arg_b)
operators: +, *, /, ^, %, =, <, >, &, |, !
### Set a variable
```
set {VARNAME} to {EXPRESSION}
```
### Evaluate an expression without using the result
```
print_number(42)
```
### function definition
```
function {NAME} do
...
end
```
### Conditional structure
```
if {EXPRESSION} then
...
end
```
### Conditional loop
```
while {EXPRESSION} do
...
end
```
### Unconditional loop
```
repeat {INT EXPRESSION} do
...
end
```
```
repeat i from {INT EXPRESSION} to {INT EXPRESSION} do
...
end
```
### std functions
```
abs(nb)
sqrt,sin,cos,exp,ln,log etc.
print_number(nb)
input_number()
ceil(nb)
floor(nb)
random_int(min, max)
random_float(min, max)
type(var) -> return the type of a var as int
is_int(var)
is_float(var)
print_string(str)
print_newline()
print_ascii(nb)
```
# Evaluator (draft)
EDIT: that's not actually quite how I implemented the evaluator.
Map
componentList:
bytes steam
first byte is a uint8 representing the type of the component
then depending on the type of the component there is 0 or N bytes
components types:
- name: integer
size: 4 bytes (for int)
- name: open parenthesis
size: 0 bytes
- name: close parenthesis
size: 0 bytes
- name: float
size: 32 bytes (IEEE 754)
- name: operator
size: 1 byte (255 operators are possible)
- name: function designator
desc: id of a function
size: 2 bytes (65536 functions are possibles)
Example map of `9+(5*(8+6))`
lexed stream
-TYPE:NUMERAL
VALUE:9
-TYPE:OPERATOR
VALUE:PLUS
-TYPE:OPEN_PARENTHESIS
VALUE:NULL
-TYPE:NUMERAL
VALUE:5
-TYPE:OPERATOR
VALUE:TIMES
-TYPE:OPEN_PARENTHESIS
VALUE:NULL
-TYPE:NUMERAL
VALUE:8
-TYPE:OPERATOR
VALUE:PLUS
-TYPE:NUMERAL
VALUE:6
-TYPE:CLOSE_P
VALUE:NULL
-TYPE:CLOSE_P
VALUE:NULL
scan the whole lexed stream
So there will be a kind of time line with reference, a list of reference
in the begining we allocate two lists, one for the component type, one for the component values
## Dynamics lists
we allocate N bytes (uint_8 list[2048])
methods:
list_length()
list_assign_int(int index, int value)
list_assign_float(int index, float value)
list_assign_uint_8(int index, uint_8 value)
list_get_int(int index)
etc...