# This Source Code Form is subject to the terms of the Mozilla Public License,
# v. 2.0. If a copy of the MPL was not distributed with this file, You can
# obtain one at https://mozilla.org/MPL/2.0/.

.equ SOURCE_BUFFER_SIZE, 81920

.equ SYS_READ, 63
.equ SYS_WRITE, 64
.equ SYS_EXIT, 93
.equ SYS_MMAP2, 222
.equ STDIN, 0
.equ STDOUT, 1
.equ STDERR, 2

.section .rodata

.type keyword_equ, @object
keyword_equ: .ascii ".equ"
.equ KEYWORD_EQU_SIZE, 4

.type keyword_section, @object
keyword_section: .ascii ".section"
.equ KEYWORD_SECTION_SIZE, 8

.type keyword_type, @object
keyword_type: .ascii ".type"
.equ KEYWORD_TYPE_SIZE, 5

.type keyword_type_object, @object
keyword_type_object: .ascii "object"
.equ KEYWORD_TYPE_OBJECT_SIZE, 6

.type keyword_type_function, @object
keyword_type_function: .ascii "function"
.equ KEYWORD_TYPE_FUNCTION_SIZE, 8

.type keyword_ret, @object
keyword_ret: .ascii "ret"
.equ KEYWORD_RET_SIZE, 3

.type keyword_global, @object
keyword_global: .ascii ".globl"
.equ KEYWORD_GLOBAL_SIZE, 6

.type keyword_proc, @object
keyword_proc: .ascii "proc "
.equ KEYWORD_PROC_SIZE, 5

.type keyword_end, @object
keyword_end: .ascii "end"
.equ KEYWORD_END_SIZE, 3

.type keyword_begin, @object
keyword_begin: .ascii "begin"
.equ KEYWORD_BEGIN_SIZE, 5

.type keyword_var, @object
keyword_var: .ascii "var"
.equ KEYWORD_VAR_SIZE, 3

.type asm_prologue, @object
asm_prologue: .string "\taddi sp, sp, -32\n\tsw ra, 28(sp)\n\tsw s0, 24(sp)\n\taddi s0, sp, 32\n"

.type asm_epilogue, @object
asm_epilogue: .string "\tlw ra, 28(sp)\n\tlw s0, 24(sp)\n\taddi sp, sp, 32\n\tret\n"

.type asm_type_directive, @object
asm_type_directive: .string ".type "

.type asm_type_function, @object
asm_type_function: .string ", @function\n"

.type asm_colon, @object
asm_colon: .string ":\n"

.type asm_call, @object
asm_call: .string "\tcall "

.section .bss

.type source_code, @object
source_code: .zero SOURCE_BUFFER_SIZE

.type source_code_position, @object
source_code_position: .word 0

.section .text

# Reads standard input into a buffer.
# a0 - Buffer pointer.
# a1 - Buffer size.
#
# Returns the amount of bytes written in a0.
.type _read_file, @function
_read_file:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	mv a2, a1
	mv a1, a0
	li a0, STDIN
	li a7, SYS_READ
	ecall

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

# Writes a character from a0 into the standard output.
.type _write_c, @function
_write_c:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	sb a0, 20(sp)
	li a0, STDOUT
	addi a1, sp, 20
	li a2, 1
	li a7, SYS_WRITE
	ecall

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	add sp, sp, 32
	ret

# Write null terminated string.
#
# Parameters:
# a0 - String.
.type _write_z, @function
_write_z:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	sw a0, 20(sp)

.write_z_loop:
	# Check for 0 character.
	lb a0, (a0)
	beqz a0, .write_z_end

	# Print a character.
	li a0, STDOUT
	lw a1, 20(sp)
	li a2, 1
	li a7, SYS_WRITE
	ecall

	# Advance the input string by one byte.
	lw a0, 20(sp)
	addi a0, a0, 1
	sw a0, 20(sp)

	j .write_z_loop

.write_z_end:
	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

# Detects if a0 is an uppercase character. Sets a0 to 1 if so, otherwise to 0.
.type _is_upper, @function
_is_upper:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	li t0, 'A' - 1
	sltu t1, t0, a0 # t1 = a0 >= 'A'

	sltiu t2, a0, 'Z' + 1 # t2 = a0 <= 'Z'
	and a0, t1, t2 # t1 = a0 >= 'A' & a0 <= 'Z'

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

# Detects if a0 is an lowercase character. Sets a0 to 1 if so, otherwise to 0.
.type _is_lower, @function
_is_lower:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	li t0, 'a' - 1
	sltu t2, t0, a0 # t2 = a0 >= 'a'

	sltiu t3, a0, 'z' + 1 # t3 = a0 <= 'z'
	and a0, t2, t3 # t2 = a0 >= 'a' & a0 <= 'z'

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

# Detects if the passed character is a 7-bit alpha character or an underscore.
#
# Paramters:
# a0 - Tested character.
#
# Sets a0 to 1 if the character is an alpha character or underscore, sets it to 0 otherwise.
.type _is_alpha, @function
_is_alpha:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	sw a0, 20(sp)

	call _is_upper
	sw a0, 16(sp)

	lw a0, 20(sp)
	call _is_lower

	lw t0, 20(sp)
	xori t1, t0, '_'
	seqz t1, t1

	lw t0, 16(sp)
	or a0, a0, t0
	or a0, a0, t1

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

# Detects whether the passed character is a digit
# (a value between 0 and 9).
#
# Parameters:
# a0 - Exemined value.
#
# Sets a0 to 1 if it is a digit, to 0 otherwise.
.type _is_digit, @function
_is_digit:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	li t0, '0' - 1
	sltu t1, t0, a0 # t1 = a0 >= '0'

	sltiu t2, a0, '9' + 1 # t2 = a0 <= '9'

	and a0, t1, t2

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

# Reads the next token.
#
# Returns token length in a0.
.type _read_token, @function
_read_token:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	la t0, source_code_position # Token pointer.
	lw t0, (t0)
	sw t0, 20(sp) # Current token position.
	sw zero, 16(sp) # Token length.

.read_token_loop:
	lb t0, (t0) # Current character.

	# First we try to read a derictive.
	# A derictive can contain a dot and characters.
	li t1, '.'
	beq t0, t1, .read_token_next

	lw a0, 20(sp)
	lb a0, (a0)
	call _is_alpha
	bnez a0, .read_token_next

	lw a0, 20(sp)
	lb a0, (a0)
	call _is_digit
	bnez a0, .read_token_next

	j .read_token_end

.read_token_next:
	# Advance the source code position and token length.
	lw t0, 16(sp)
	addi t0, t0, 1
	sw t0, 16(sp)

	lw t0, 20(sp)
	addi t0, t0, 1
	sw t0, 20(sp)

	j .read_token_loop

.read_token_end:
	lw a0, 16(sp)

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

# a0 - First pointer.
# a1 - Second pointer.
# a2 - The length to compare.
#
# Returns 0 in a0 if memory regions are equal.
.type _memcmp, @function
_memcmp:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	mv t0, a0
	li a0, 0

.Lmemcmp_loop:
	beqz a2, .Lmemcmp_end

	lbu t1, (t0)
	lbu t2, (a1)
	sub a0, t1, t2

	bnez a0, .Lmemcmp_end

	addi t0, t0, 1
	addi a1, a1, 1
	addi a2, a2, -1

	j .Lmemcmp_loop

.Lmemcmp_end:
	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

# Advances the token stream by a0 bytes.
.type _advance_token, @function
_advance_token:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	# Skip the .equ directive.
	la t0, source_code_position
	lw t1, (t0)
	add t1, t1, a0
	sw t1, (t0)

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

.type _compile_section, @function
_compile_section:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	# Print the .section directive and a space after it.
	li a0, STDOUT
	la a1, source_code_position
	lw a1, (a1)
	li a2, KEYWORD_SECTION_SIZE + 1
	li a7, SYS_WRITE
	ecall

	# Skip the .equ directive.
	li a0, KEYWORD_SECTION_SIZE + 1
	call _advance_token

	# Read the section name.
	call _read_token
	sw a0, 16(sp)

	# Print the section name and newline.
	li a0, STDOUT
	la a1, source_code_position
	lw a1, (a1)
	lw a2, 16(sp)
	addi a2, a2, 1
	li a7, SYS_WRITE
	ecall

	# Skip the section name.
	lw a0, 16(sp)
	addi a0, a0, 1
	call _advance_token

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

# Prints and skips a line.
.type _skip_comment, @function
_skip_comment:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	la t0, source_code_position
	lw t1, (t0)

.skip_comment_loop:
	# Check for newline character.
	lb t2, (t1)
	li t3, '\n'
	beq t2, t3, .skip_comment_end

	# Advance the input string by one byte.
	addi t1, t1, 1
	sw t1, (t0)

	j .skip_comment_loop

.skip_comment_end:
	# Skip the newline.
	addi t1, t1, 1
	sw t1, (t0)

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

# Prints and skips a line.
.type _compile_line, @function
_compile_line:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	la a0, source_code_position
	lw a1, (a0)

.compile_line_loop:
	# Check for newline character.
	lb t0, (a1)
	li t1, '\n'
	beq t0, t1, .compile_line_end

	# Print a character.
	li a0, STDOUT
	li a2, 1
	li a7, SYS_WRITE
	ecall

	# Advance the input string by one byte.
	la a0, source_code_position
	lw a1, (a0)
	addi a1, a1, 1
	sw a1, (a0)

	j .compile_line_loop

.compile_line_end:
	# Print and skip the newline.
	li a0, STDOUT
	li a2, 1
	li a7, SYS_WRITE
	ecall

	la a0, source_code_position
	lw a1, (a0)
	addi a1, a1, 1
	sw a1, (a0)

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

.type _compile_object, @function
_compile_object:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	call _compile_line

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

.type _compile_function_statements, @function
_compile_function_statements:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

.compile_function_statements_loop:
	la t0, source_code_position
	lw t1, (t0)
	addi t1, t1, 1 # Skip the tab.

	mv a0, t1
	la a1, keyword_ret
	li a2, KEYWORD_RET_SIZE
	call _memcmp

	beqz a0, .compile_function_statements_end

	call _compile_line
	j .compile_function_statements_loop

.compile_function_statements_end:
	call _compile_line

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

.type _compile_call, @function
_compile_call:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	call _read_token
	sw a0, 20(sp)

	la a0, asm_call
	call _write_z

	li a0, STDOUT
	la a1, source_code_position
	lw a1, (a1)
	lw a2, 20(sp)
	li a7, SYS_WRITE
	ecall

	# Skip parens, semicolon and newline.
	lw a0, 20(sp)
	addi a0, a0, 4
	call _advance_token

	li a0, '\n'
	call _write_c

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

.type _compile_statement, @function
_compile_statement:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	# This is a call if the statement starts with an underscore.
	la t0, source_code_position
	lw t0, (t0)
	# First character after alignment tab.
	addi t0, t0, 1
	lb t0, (t0)
	
	li t1, '_'
	beq t0, t1, .compile_statement_call

	call _compile_line
	j .compile_statement_end

.compile_statement_call:
	li a0, 1
	call _advance_token
	call _compile_call

	j .compile_statement_end

.compile_statement_end:
	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

.type _compile_procedure_body, @function
_compile_procedure_body:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

.compile_procedure_body_loop:
	la a0, source_code_position
	lw a0, (a0)
	la a1, keyword_end
	li a2, KEYWORD_END_SIZE
	call _memcmp

	beqz a0, .compile_procedure_body_epilogue

	call _compile_statement
	j .compile_procedure_body_loop

.compile_procedure_body_epilogue:
	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

.type _compile_procedure, @function
_compile_procedure:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	# Skip "proc ".
	li a0, KEYWORD_PROC_SIZE
	call _advance_token

	call _read_token
	sw a0, 20(sp) # Save the procedure name length.

	# Write .type _procedure_name, @function.
	la a0, asm_type_directive
	call _write_z

	li a0, STDOUT
	la a1, source_code_position
	lw a1, (a1)
	lw a2, 20(sp)
	li a7, SYS_WRITE
	ecall

	la a0, asm_type_function
	call _write_z

	# Write procedure label, _procedure_name:
	li a0, STDOUT
	la a1, source_code_position
	lw a1, (a1)
	lw a2, 20(sp)
	li a7, SYS_WRITE
	ecall

	la a0, asm_colon
	call _write_z

	# Skip the function name and trailing parens, semicolon, "begin" and newline.
	lw a0, 20(sp)
	addi a0, a0, KEYWORD_BEGIN_SIZE + 1 + 4
	call _advance_token

	la a0, asm_prologue
	call _write_z

	call _compile_procedure_body

	# Write the epilogue.
	la a0, asm_epilogue
	call _write_z

	li a0, KEYWORD_END_SIZE + 2
	call _advance_token

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

.type _compile_function, @function
_compile_function:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	# Write the function header.
	call _compile_line
	call _compile_function_statements

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

.type _compile_type, @function
_compile_type:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	# Print the .type directive and a space after it.
	li a0, STDOUT
	la a1, source_code_position
	lw a1, (a1)
	li a2, KEYWORD_TYPE_SIZE + 1
	li a7, SYS_WRITE
	ecall

	# Skip the .type directive.
	li a0, KEYWORD_TYPE_SIZE + 1
	call _advance_token

	# Read and print the symbol name.
	call _read_token
	sw a0, 20(sp)

	# Print the symbol name, comma, space and @.
	li a0, STDOUT
	la a1, source_code_position
	lw a1, (a1)
	lw a2, 20(sp)
	addi a2, a2, 3
	li a7, SYS_WRITE
	ecall

	# Skip the constant name, comma, space and @.
	lw a0, 20(sp)
	addi a0, a0, 3
	call _advance_token

	# Read the symbol type.
	call _read_token
	sw a0, 16(sp)
	la t0, source_code_position
	lw t0, (t0)
	sw t0, 12(sp)

	# Print the symbol type and newline.
	li a0, STDOUT
	la a1, source_code_position
	lw a1, (a1)
	lw a2, 16(sp)
	addi a2, a2, 1
	li a7, SYS_WRITE
	ecall

	lw a0, 16(sp)
	addi a0, a0, 1
	call _advance_token

	lw a0, 12(sp)
	la a1, keyword_type_object
	li a2, KEYWORD_TYPE_OBJECT_SIZE
	call _memcmp

	beqz a0, .compile_type_object

	lw a0, 12(sp)
	la a1, keyword_type_function
	li a2, KEYWORD_TYPE_FUNCTION_SIZE
	call _memcmp

	beqz a0, .compile_type_function

	j .compile_type_end

.compile_type_object:
	call _compile_object

	j .compile_type_end

.compile_type_function:
	call _compile_function

	j .compile_type_end

.compile_type_end:
	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

.type _compile_equ, @function
_compile_equ:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	# Print the .equ directive and a space after it.
	li a0, STDOUT
	la a1, source_code_position
	lw a1, (a1)
	li a2, KEYWORD_EQU_SIZE + 1
	li a7, SYS_WRITE
	ecall

	# Skip the .equ directive.
	li a0, KEYWORD_EQU_SIZE + 1
	call _advance_token

	# Read and print the constant name.
	call _read_token
	sw a0, 20(sp)

	# Print the constant name, comma and space.
	li a0, STDOUT
	la a1, source_code_position
	lw a1, (a1)
	lw a2, 20(sp)
	addi a2, a2, 2
	li a7, SYS_WRITE
	ecall

	# Skip the constant name, comma and the space after it.
	lw a0, 20(sp)
	addi a0, a0, 2
	call _advance_token

	# Read the constant value.
	call _read_token
	sw a0, 16(sp)

	# Print the constant value and newline.
	li a0, STDOUT
	la a1, source_code_position
	lw a1, (a1)
	lw a2, 16(sp)
	addi a2, a2, 1
	li a7, SYS_WRITE
	ecall

	lw a2, 16(sp)
	addi a2, a2, 1
	call _advance_token

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

.type _skip_newlines, @function
_skip_newlines:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	# Skip newlines.
	la t0, source_code_position
	lw t1, (t0)

.skip_newlines_loop:
	lb t2, (t1)
	li t3, '\n'
	bne t2, t3, .skip_newlines_end
	beqz t2, .skip_newlines_end

	addi t1, t1, 1
	sw t1, (t0)

	j .skip_newlines_loop

.skip_newlines_end:
	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

# Process the source code and print the generated code.
.type _compile, @function
_compile:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

.compile_loop:
	call _skip_newlines

	la t0, source_code_position
	lw t0, (t0)
	lb t0, (t0)
	beqz t0, .compile_end
	li t1, '#'
	beq t0, t1, .compile_comment

	la a0, source_code_position
	lw a0, (a0)
	la a1, keyword_equ
	li a2, KEYWORD_EQU_SIZE
	call _memcmp

	beqz a0, .compile_equ

	la a0, source_code_position
	lw a0, (a0)
	la a1, keyword_section
	li a2, KEYWORD_SECTION_SIZE
	call _memcmp

	beqz a0, .compile_section

	la a0, source_code_position
	lw a0, (a0)
	la a1, keyword_type
	li a2, KEYWORD_TYPE_SIZE
	call _memcmp

	beqz a0, .compile_type

	la a0, source_code_position
	lw a0, (a0)
	la a1, keyword_proc
	li a2, KEYWORD_PROC_SIZE
	call _memcmp

	beqz a0, .compile_procedure

	la a0, source_code_position
	lw a0, (a0)
	la a1, keyword_global
	li a2, KEYWORD_GLOBAL_SIZE
	call _memcmp

	beqz a0, .compile_global

	j .compile_end # Not a known token, exit.

.compile_equ:
	call _compile_equ

	j .compile_loop

.compile_section:
	call _compile_section

	j .compile_loop

.compile_type:
	call _compile_type

	j .compile_loop

.compile_global:
	call _compile_line

	j .compile_loop

.compile_comment:
	call _skip_comment

	j .compile_loop

.compile_procedure:
	call _compile_procedure

	j .compile_loop

.compile_end:
	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret

# Entry point.
.globl _start
.type _start, @function
_start:
	# Prologue.
	addi sp, sp, -32
	sw ra, 28(sp)
	sw s0, 24(sp)
	addi s0, sp, 32

	# Read the source from the standard input.
	la a0, source_code
	li a1, SOURCE_BUFFER_SIZE # Buffer size.
	call _read_file

	# Save the pointer to the beginning of the source code in a global variable.
	la t0, source_code
	la t1, source_code_position
	sw t0, (t1)

	call _compile

	# Call exit.
	li a0, 0 # Use 0 return code.
	li a7, SYS_EXIT
	ecall

	# Epilogue.
	lw ra, 28(sp)
	lw s0, 24(sp)
	addi sp, sp, 32
	ret