initial parsing
This commit is contained in:
parent
39e8d6ae96
commit
46053090f4
SSH key fingerprint:
SHA256:bB1qbbqmDXZNT0KKD5c2Dfjg53JGhj7B3CFcLIzSqq8
|
|
@ -1,7 +1,7 @@
|
|||
# Makefile: Compile and link main.asm using nasm and mold, intermediate files in target/
|
||||
|
||||
TARGET_DIR := target
|
||||
SRC := src/lib.asm src/int_to_str.asm src/vec.asm src/tokeniser.asm src/file.asm src/alloc.asm
|
||||
SRC := src/lib.asm src/int_to_str.asm src/vec.asm src/tokeniser.asm src/file.asm src/alloc.asm src/ast.asm
|
||||
OBJ := $(patsubst src/%.asm,$(TARGET_DIR)/%.o,$(SRC))
|
||||
|
||||
BIN_SRC := src/main.asm src/panic.asm
|
||||
|
|
|
|||
439
lang/src/ast.asm
Normal file
439
lang/src/ast.asm
Normal file
|
|
@ -0,0 +1,439 @@
|
|||
default rel
|
||||
|
||||
%include "src/tokeniser.inc"
|
||||
|
||||
section .rdata
|
||||
AST_FUNCTION equ 1
|
||||
AST_BLOCK equ 2
|
||||
AST_VARIABLE equ 3
|
||||
AST_NUMBER equ 4
|
||||
AST_BINARY_OP equ 5
|
||||
AST_RETURN_STATEMENT equ 6
|
||||
|
||||
TYPE_VOID equ 1
|
||||
TYPE_BOOL equ 2
|
||||
TYPE_I32 equ 3
|
||||
TYPE_U32 equ 4
|
||||
TYPE_STR equ 5
|
||||
|
||||
section .text
|
||||
extern vec_init_with
|
||||
extern vec_push
|
||||
extern vec_get
|
||||
extern panic
|
||||
extern memcpy
|
||||
extern vec_binary_search_by
|
||||
extern vec_insert
|
||||
|
||||
extern bump_alloc
|
||||
|
||||
extern tokeniser_init
|
||||
extern find_lexeme
|
||||
extern peek_lexeme
|
||||
extern expect_token
|
||||
extern unwrap_token
|
||||
extern peek_expect_token
|
||||
|
||||
extern str_to_int
|
||||
|
||||
;; start very simple, with only functions and addition
|
||||
;; struct Ast {
|
||||
;; nodes: Vec<AstNode>,
|
||||
;; }
|
||||
;;
|
||||
;; struct AstNode {
|
||||
;; kind: u8,
|
||||
;; data: *const (),
|
||||
;; }
|
||||
;;
|
||||
;; struct Argument {
|
||||
;; name: *const u8,
|
||||
;; name_len: usize,
|
||||
;; arg_type: Type,
|
||||
;; }
|
||||
;;
|
||||
;; struct Type {
|
||||
;; kind: u8,
|
||||
;; }
|
||||
|
||||
;; rdi: *mut Ast
|
||||
;; fn parse_func(ast: *mut Ast) -> u64
|
||||
parse_func:
|
||||
push rbp
|
||||
mov rbp, rsp
|
||||
push rdi
|
||||
sub rsp, 48
|
||||
; name: *const u8 [0..8]
|
||||
; name_len: usize [8..16]
|
||||
; args_ptr: *const Arg [16..24]
|
||||
; args_len: usize [24..32]
|
||||
; return_type: Type [32..40]
|
||||
; body: u64 [40..48]
|
||||
|
||||
mov dil, TOKEN_FN
|
||||
call unwrap_token
|
||||
mov dil, TOKEN_IDENT
|
||||
call unwrap_token
|
||||
mov [rsp], rax ; function name
|
||||
mov [rsp + 8], rdx ; function name length
|
||||
mov dil, TOKEN_LPARENS
|
||||
call unwrap_token
|
||||
mov dil, TOKEN_RPARENS
|
||||
call expect_token
|
||||
test rax, rax
|
||||
je .args
|
||||
|
||||
.after_args:
|
||||
mov dil, TOKEN_ARROW
|
||||
call unwrap_token
|
||||
mov rdi, [rsp + 48] ; Ast
|
||||
call parse_type
|
||||
mov [rsp + 32], rax ; return type
|
||||
mov dil, TOKEN_LBRACE
|
||||
call peek_expect_token
|
||||
test rax, rax
|
||||
je panic
|
||||
mov rdi, [rsp + 48] ; Ast
|
||||
call parse_block
|
||||
mov [rsp + 40], rax ; body
|
||||
.epilogue:
|
||||
mov rdi, 48
|
||||
call bump_alloc
|
||||
mov rsi, rsp
|
||||
mov rdi, rax
|
||||
mov rdx, 48
|
||||
call memcpy
|
||||
mov byte [rsp], AST_FUNCTION ; kind
|
||||
mov [rsp + 8], rdi ; data
|
||||
mov rdi, [rsp + 48] ; Ast
|
||||
lea rsi, [rsp]
|
||||
call vec_push
|
||||
mov rax, [rsp + 48] ; Ast
|
||||
mov rax, [rdi + 8] ; return Ast.nodes.len()
|
||||
add rsp, 48
|
||||
pop rdi
|
||||
pop rbp
|
||||
ret
|
||||
|
||||
.args:
|
||||
mov rdi, [rsp + 48] ; Ast
|
||||
call parse_args
|
||||
mov [rsp + 16], rax ; args_ptr
|
||||
mov [rsp + 24], rdx ; args_len
|
||||
jmp .after_args
|
||||
.panic:
|
||||
call panic
|
||||
|
||||
;; rdi: *mut Ast
|
||||
;; fn parse_args(ast: *mut Ast) -> (*const Argument, usize)
|
||||
parse_args:
|
||||
push rbp
|
||||
mov rbp, rsp
|
||||
push rdi
|
||||
sub rsp, 64
|
||||
|
||||
lea rdi, [rsp + 24] ; vec
|
||||
mov rsi, 24 ; size of Argument
|
||||
mov rdx, 0 ; drop = None
|
||||
mov rcx, 16 ; capacity
|
||||
call vec_init_with
|
||||
|
||||
.loop:
|
||||
mov dil, TOKEN_RPARENS
|
||||
call expect_token
|
||||
test rax, rax
|
||||
jz .done_args
|
||||
mov dil, TOKEN_IDENT
|
||||
call unwrap_token
|
||||
mov [rsp], rax ; arg name
|
||||
mov [rsp + 8], rdx ; arg name length
|
||||
mov dil, TOKEN_COLON
|
||||
call unwrap_token
|
||||
mov rdi, [rsp + 64] ; Ast
|
||||
call parse_type
|
||||
mov [rsp + 16], rax ; arg type
|
||||
|
||||
lea rdi, [rsp + 24] ; vec
|
||||
lea rsi, [rsp] ; arg
|
||||
call vec_push
|
||||
|
||||
mov dil, TOKEN_COMMA
|
||||
call expect_token
|
||||
test rax, rax
|
||||
jz .end_loop
|
||||
jmp .loop
|
||||
.end_loop:
|
||||
mov dil, TOKEN_RPARENS
|
||||
call unwrap_token
|
||||
.done_args:
|
||||
mov rax, [rsp + 24] ; args_ptr
|
||||
mov rdx, [rsp + 32] ; args_len
|
||||
add rsp, 64
|
||||
pop rdi
|
||||
pop rbp
|
||||
ret
|
||||
|
||||
;; rdi: lexeme ptr
|
||||
;; rsi: lexeme len
|
||||
parse_number:
|
||||
push rbp
|
||||
mov rbp, rsp
|
||||
push rbx
|
||||
sub rsp, 16
|
||||
mov [rsp], rdi ; lexeme ptr
|
||||
mov [rsp + 8], rsi ; lexeme len
|
||||
|
||||
cmp rsi, 2
|
||||
jbe .dec_radix
|
||||
|
||||
mov al, byte [rdi]
|
||||
mov bl, byte [rdi + 1]
|
||||
cmp bl, 'x'
|
||||
jne .dec_radix
|
||||
cmp al, '0'
|
||||
je .hex_radix
|
||||
cmp al, 'o'
|
||||
je .oct_radix
|
||||
cmp al, 'b'
|
||||
je .bin_radix
|
||||
jmp .panic ; invalid radix prefix
|
||||
.hex_radix:
|
||||
mov rax, 16
|
||||
jmp .radix_set
|
||||
.oct_radix:
|
||||
mov rax, 8
|
||||
jmp .radix_set
|
||||
.bin_radix:
|
||||
mov rax, 2
|
||||
jmp .radix_set
|
||||
.dec_radix:
|
||||
mov rax, 10
|
||||
jmp .parse
|
||||
.radix_set:
|
||||
add qword [rsp], 2
|
||||
sub qword [rsp + 8], 2
|
||||
.parse:
|
||||
mov rdi, [rsp] ; lexeme ptr
|
||||
mov rsi, [rsp + 8] ; lexeme len
|
||||
mov rdx, rax ; radix
|
||||
call str_to_int
|
||||
add rsp, 16
|
||||
pop rbx
|
||||
pop rbp
|
||||
ret
|
||||
.panic:
|
||||
call panic
|
||||
|
||||
;; rdi: *mut Ast
|
||||
;; fn parse_expr(ast: *mut Ast) -> u64
|
||||
parse_primary_expr:
|
||||
push rbp
|
||||
mov rbp, rsp
|
||||
sub rsp, 8
|
||||
mov [rsp], rdi ; Ast
|
||||
|
||||
mov dil, TOKEN_NUMBER
|
||||
call expect_token
|
||||
test rax, rax
|
||||
jz .number
|
||||
jmp .panic
|
||||
.number:
|
||||
mov rdi, rax ; lexeme ptr
|
||||
mov rsi, rdx ; lexeme len
|
||||
call parse_number
|
||||
mov rdi, [rsp]
|
||||
push rdi
|
||||
mov byte [rsp], AST_NUMBER ; kind
|
||||
mov [rsp + 8], rax ; data
|
||||
lea rsi, [rsp]
|
||||
call vec_push
|
||||
pop rdi
|
||||
mov rax, [rdi + 8] ; return Ast.nodes.len()
|
||||
add rsp, 8
|
||||
pop rbp
|
||||
ret
|
||||
.panic:
|
||||
call panic
|
||||
|
||||
|
||||
;; rdi: *mut Ast
|
||||
;; sil: precedence
|
||||
;; fn parse_expr(ast: *mut Ast) -> u64
|
||||
parse_binary_expr:
|
||||
push rbp
|
||||
mov rbp, rsp
|
||||
sub rsp, 64
|
||||
; lexeme: Lexeme [32..56]
|
||||
; right: u64 [24..32]
|
||||
; precedence: u8 [17..18]
|
||||
; operator: u8 [16..17]
|
||||
; left: u64 [8..16]
|
||||
; rdi: *mut Ast [0..8]
|
||||
mov [rsp], rdi ; Ast
|
||||
mov byte [rsp + 17], sil ; upper_precedence
|
||||
mov byte [rsp + 16], 0
|
||||
|
||||
call parse_primary_expr
|
||||
mov [rsp + 8], rax ; left
|
||||
|
||||
.loop:
|
||||
lea rdi, [rsp + 32] ; lexeme
|
||||
call peek_lexeme
|
||||
mov rax, [rsp + 32]
|
||||
mov dil, [rsp + 17]
|
||||
cmp al, dil ; our_precedence <= upper_precedence
|
||||
jle .done ; also covers some non-binary operator tokens
|
||||
cmp al, TOKEN_PLUS
|
||||
je .plus
|
||||
jmp .done
|
||||
|
||||
.plus:
|
||||
mov dil, TOKEN_PLUS
|
||||
call unwrap_token
|
||||
mov byte [rsp + 16], TOKEN_PLUS
|
||||
jmp .right
|
||||
|
||||
.right:
|
||||
mov rdi, [rsp] ; Ast
|
||||
mov sil, [rsp + 16]
|
||||
call parse_binary_expr
|
||||
mov [rsp + 24], rax ; right
|
||||
|
||||
mov byte [rsp + 32], AST_BINARY_OP ; kind
|
||||
lea rax, [rsp + 8]
|
||||
mov [rsp + 40], rax ; data ptr
|
||||
mov rdi, [rsp] ; Ast
|
||||
lea rsi, [rsp + 32] ; AstNode
|
||||
call vec_push
|
||||
mov rdi, [rsp] ; Ast
|
||||
mov rax, [rdi + 8] ; Ast.nodes.len()
|
||||
mov [rsp + 8], rax ; left
|
||||
|
||||
.done:
|
||||
mov rax, [rsp + 8] ; left
|
||||
add rsp, 56
|
||||
pop rbp
|
||||
ret
|
||||
|
||||
|
||||
;; rdi: *mut Ast
|
||||
;; fn parse_expr(ast: *mut Ast) -> u64
|
||||
parse_expr:
|
||||
push rbp
|
||||
mov rbp, rsp
|
||||
sub rsp, 8
|
||||
mov [rsp], rdi ; Ast
|
||||
|
||||
;; rdi: *mut Ast
|
||||
;; fn parse_statement(ast: *mut Ast) -> u64
|
||||
parse_statement:
|
||||
push rbp
|
||||
mov rbp, rsp
|
||||
sub rsp, 32
|
||||
mov [rsp + 24], rdi ; Ast
|
||||
|
||||
mov dil, TOKEN_RETURN
|
||||
call expect_token
|
||||
test rax, rax
|
||||
jz .return
|
||||
jmp .panic
|
||||
|
||||
.return:
|
||||
call parse_expr
|
||||
mov [rsp + 16], rax ; expression
|
||||
mov byte [rsp], AST_RETURN_STATEMENT ; kind
|
||||
lea rax, [rsp + 16] ; data ptr
|
||||
mov [rsp + 8], rax ; data
|
||||
mov rdi, [rsp + 24] ; Ast
|
||||
mov rsi, rsp ; AstNode
|
||||
call vec_push
|
||||
mov rdi, [rsp + 24] ; Ast
|
||||
mov rax, [rdi + 8] ; Ast.nodes.len()
|
||||
mov [rsp], rax
|
||||
|
||||
mov dil, TOKEN_SEMI
|
||||
call unwrap_token
|
||||
mov rax, [rsp] ; expression
|
||||
add rsp, 32
|
||||
pop rbp
|
||||
ret
|
||||
.panic:
|
||||
call panic
|
||||
|
||||
;; rdi: *mut Ast
|
||||
;; fn parse_block(ast: *mut Ast) -> u64
|
||||
parse_block:
|
||||
push rbp
|
||||
mov rbp, rsp
|
||||
push rdi
|
||||
|
||||
; struct Block {
|
||||
; statements: &[u64],
|
||||
; }
|
||||
|
||||
sub rsp, 56
|
||||
; statements: Vec<Statement> [0..40]
|
||||
; statement: u64 [40..48]
|
||||
|
||||
mov dil, TOKEN_LBRACE
|
||||
call unwrap_token
|
||||
.loop:
|
||||
mov rdi, [rsp + 16] ; Ast
|
||||
call parse_statement
|
||||
test rax, rax
|
||||
je .done
|
||||
lea rdi, [rsp + 16] ; vec
|
||||
mov [rsp + 8], rax ; statement
|
||||
lea rsi, [rsp + 8]
|
||||
call vec_push
|
||||
jmp .loop
|
||||
.done:
|
||||
mov rdi, [rsp + 56] ; Ast
|
||||
lea rsi, [rsp + 16] ; statements vec-slice
|
||||
call vec_push
|
||||
mov rdi, [rsp + 56] ; Ast
|
||||
mov rax, [rdi + 8] ; Ast.nodes.len()
|
||||
add rsp, 56
|
||||
pop rdi
|
||||
pop rbp
|
||||
ret
|
||||
|
||||
;; rdi: *mut Ast
|
||||
;; fn parse_type(ast: *mut Ast) -> Type
|
||||
parse_type:
|
||||
push rbp
|
||||
mov rbp, rsp
|
||||
push rdi
|
||||
|
||||
sub rsp, 24
|
||||
mov rdi, rsp
|
||||
call find_lexeme ; TODO: use peek here to allow failing gracefully
|
||||
mov rax, [rsp] ; token kind
|
||||
cmp al, TOKEN_I32
|
||||
je .i32_type
|
||||
cmp al, TOKEN_U32
|
||||
je .u32_type
|
||||
cmp al, TOKEN_VOID
|
||||
je .void_type
|
||||
cmp al, TOKEN_BOOL
|
||||
je .bool_type
|
||||
jmp .panic
|
||||
.i32_type:
|
||||
mov rax, TYPE_I32
|
||||
jmp .epilogue
|
||||
.u32_type:
|
||||
mov rax, TYPE_U32
|
||||
jmp .epilogue
|
||||
.void_type:
|
||||
mov rax, TYPE_VOID
|
||||
jmp .epilogue
|
||||
.bool_type:
|
||||
mov rax, TYPE_BOOL
|
||||
jmp .epilogue
|
||||
.epilogue:
|
||||
add rsp, 24
|
||||
pop rdi
|
||||
pop rbp
|
||||
ret
|
||||
.panic:
|
||||
call panic
|
||||
35
lang/tests/ast.rs
Normal file
35
lang/tests/ast.rs
Normal file
|
|
@ -0,0 +1,35 @@
|
|||
#[path = "shared/shared.rs"]
|
||||
mod util;
|
||||
|
||||
#[repr(C)]
|
||||
struct Ast {
|
||||
nodes: util::BlobVec,
|
||||
}
|
||||
|
||||
#[repr(C)]
|
||||
struct Argument {
|
||||
name: *const u8,
|
||||
name_len: usize,
|
||||
arg_type: u8,
|
||||
}
|
||||
|
||||
#[repr(C)]
|
||||
struct AstNode {
|
||||
node_type: u8,
|
||||
data: usize,
|
||||
}
|
||||
|
||||
unsafe extern "C" {
|
||||
unsafe fn bump_init();
|
||||
|
||||
unsafe fn parse_func(ast: *mut Ast) -> u64;
|
||||
unsafe fn parse_args(ast: *mut Ast) -> (*const Argument, usize);
|
||||
unsafe fn tokeniser_init(path: *const i8) -> ();
|
||||
}
|
||||
|
||||
fn main() {
|
||||
unsafe {
|
||||
bump_init();
|
||||
}
|
||||
println!("Bump allocator initialized.");
|
||||
}
|
||||
Loading…
Reference in a new issue