Tokenize single character symbols

6 files changed, 283 insertions, 327 deletions

diff --git a/Gemfile b/Gemfile
index 235744c..97bdde4 100644
--- a/Gemfile
+++ b/Gemfile
@@ -1,3 +1,6 @@
+# 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/.
 # frozen_string_literal: true
 
 source 'https://rubygems.org'
diff --git a/Rakefile b/Rakefile
index 2bc2683..71b3293 100644
--- a/Rakefile
+++ b/Rakefile
@@ -1,3 +1,6 @@
+# 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/.
 # frozen_string_literal: true
 
 require 'open3'
diff --git a/boot/common-boot.s b/boot/common-boot.s
index 26dad8d..a6fb04e 100644
--- a/boot/common-boot.s
+++ b/boot/common-boot.s
@@ -1,3 +1,7 @@
+# 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/.
+
 .global _is_alpha, _is_digit, _is_alnum, _is_upper, _is_lower
 .global _write_out, _read_file, _write_error, _put_char, _printi
 .global _get, _memcmp, _memchr, _memmem, _memcpy
diff --git a/boot/definitions.inc b/boot/definitions.inc
index 0e2f54e..4d8ab9c 100644
--- a/boot/definitions.inc
+++ b/boot/definitions.inc
@@ -1,3 +1,7 @@
+# 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/.
+
 # The constant should match the index in the keywords array in tokenizer.s.
 
 .equ TOKEN_PROGRAM, 1
@@ -26,3 +30,23 @@
 .equ TOKEN_DEFER, 24
 .equ TOKEN_CASE, 25
 .equ TOKEN_OF, 26
+
+# The constant should match the character index in the byte_keywords string.
+
+.equ TOKEN_AND, 27
+.equ TOKEN_DOT, 28
+.equ TOKEN_COMMA, 29
+.equ TOKEN_COLON, 30
+.equ TOKEN_SEMICOLON, 31
+.equ TOKEN_LEFT_PAREN, 32
+.equ TOKEN_RIGHT_PAREN, 33
+.equ TOKEN_LEFT_BRACKET, 34
+.equ TOKEN_RIGHT_BRACKET, 35
+.equ TOKEN_HAT, 36
+.equ TOKEN_EQUALS, 37
+.equ TOKEN_PLUS, 38
+.equ TOKEN_MINUS, 39
+.equ TOKEN_ASTERISK, 40
+.equ TOKEN_AT, 41
+
+.equ TOKEN_ASSIGN, 42
diff --git a/boot/stage1.s b/boot/stage1.s
index 9ab072d..6761bb2 100644
--- a/boot/stage1.s
+++ b/boot/stage1.s
@@ -1,6 +1,10 @@
+# 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/.
+
 .global _start # Program entry point.
 
-# Global variables or registers.
+# Registers used as global variables:
 # s1 - Contains the current position in the source text.
 # s2 - Label counter.
 
@@ -42,6 +46,10 @@ asm_neg_a0: .ascii "neg a0, a0\n"
 .equ ASM_NEG_A0_SIZE, . - asm_neg_a0
 asm_type: .ascii ".type "
 .equ ASM_TYPE_SIZE, . - asm_type
+asm_type_function: .ascii ", @function\n"
+.equ ASM_TYPE_FUNCTION_SIZE, . - asm_type_function
+asm_type_object: .ascii ", @object\n"
+.equ ASM_TYPE_OBJECT_SIZE, . - asm_type_object
 asm_restore_parameters:
 	.ascii "lw a0, 60(sp)\nlw a1, 56(sp)\nlw a2, 52(sp)\nlw a3, 48(sp)\nlw a4, 44(sp)\nlw a5, 40(sp)\n"
 .equ ASM_RESTORE_PARAMETERS_SIZE, . - asm_restore_parameters
@@ -77,14 +85,6 @@ _compile_import:
 	call _tokenize_next
 	mv s1, a0
 
-	/* DEBUG 
-	lw t0, 0(sp)
-	addi t0, t0, '0'
-	sw t0, 4(sp)
-	addi a0, sp, 4
-	li a1, 1
-	call _write_error*/
-
 	j .Lcompile_import_loop
 
 .Lcompile_import_end:
@@ -104,63 +104,35 @@ _build_binary_expression:
 
 	li a0, 0
 	call _build_expression
-
 	call _skip_spaces
-	call _read_token
-	sw a0, 20(sp)
 
-	li t0, '&'
-	sw t0, 16(sp)
 	mv a0, s1
-	lw a1, 20(sp)
-	addi a2, sp, 16
-	call _token_compare
-	beqz a0, .L_build_binary_expression_and
+	addi a1, sp, 16
+	call _tokenize_next
+	lw t0, 16(sp)
 
-	li t0, 0x726f # or
-	sw t0, 16(sp)
-	mv a0, s1
-	lw a1, 20(sp)
-	addi a2, sp, 16
-	call _token_compare
-	beqz a0, .L_build_binary_expression_or
+	li t1, TOKEN_AND
+	beq t0, t1, .L_build_binary_expression_and
 
-	li t0, '='
-	sw t0, 16(sp)
-	mv a0, s1
-	lw a1, 20(sp)
-	addi a2, sp, 16
-	call _token_compare
-	beqz a0, .L_build_binary_expression_equal
+	li t1, TOKEN_OR
+	beq t0, t1, .L_build_binary_expression_or
 
-	li t0, '+'
-	sw t0, 16(sp)
-	mv a0, s1
-	lw a1, 20(sp)
-	addi a2, sp, 16
-	call _token_compare
-	beqz a0, .L_build_binary_expression_plus
+	li t1, TOKEN_PLUS
+	beq t0, t1, .L_build_binary_expression_plus
 
-	li t0, '-'
-	sw t0, 16(sp)
-	mv a0, s1
-	lw a1, 20(sp)
-	addi a2, sp, 16
-	call _token_compare
-	beqz a0, .L_build_binary_expression_minus
+	li t1, TOKEN_EQUALS
+	beq t0, t1, .L_build_binary_expression_equal
 
-	li t0, '*'
-	sw t0, 16(sp)
-	mv a0, s1
-	lw a1, 20(sp)
-	addi a2, sp, 16
-	call _token_compare
-	beqz a0, .L_build_binary_expression_product
+	li t1, TOKEN_ASTERISK
+	beq t0, t1, .L_build_binary_expression_product
+
+	li t1, TOKEN_MINUS
+	beq t0, t1, .L_build_binary_expression_minus
 
 	j .Lbuild_binary_expression_end
 
 .L_build_binary_expression_equal:
-	addi s1, s1, 1 # Skip =.
+	mv s1, a0 # Skip =.
 	li a0, 1
 	call _build_expression
 	la a0, asm_sub_a0_a1
@@ -174,7 +146,12 @@ _build_binary_expression:
 	j .Lbuild_binary_expression_end
 
 .L_build_binary_expression_and:
-	addi s1, s1, 1 # Skip &.
+	/* DEBUG
+	addi a0, s1, 0
+	li a1, 4
+	call _write_error */
+
+	mv s1, a0 # Skip &.
 	li a0, 1
 	call _build_expression
 	la a0, asm_and_a0_a1
@@ -184,7 +161,7 @@ _build_binary_expression:
 	j .Lbuild_binary_expression_end
 
 .L_build_binary_expression_or:
-	addi s1, s1, 2 # Skip or.
+	mv s1, a0 # Skip or.
 	li a0, 1
 	call _build_expression
 	la a0, asm_or_a0_a1
@@ -194,7 +171,7 @@ _build_binary_expression:
 	j .Lbuild_binary_expression_end
 
 .L_build_binary_expression_plus:
-	addi s1, s1, 1 # Skip +.
+	mv s1, a0 # Skip +.
 	li a0, 1
 	call _build_expression
 	la a0, asm_add_a0_a1
@@ -204,7 +181,7 @@ _build_binary_expression:
 	j .Lbuild_binary_expression_end
 
 .L_build_binary_expression_minus:
-	addi s1, s1, 1 # Skip -.
+	mv s1, a0 # Skip -.
 	li a0, 1
 	call _build_expression
 	la a0, asm_sub_a0_a1
@@ -214,7 +191,7 @@ _build_binary_expression:
 	j .Lbuild_binary_expression_end
 
 .L_build_binary_expression_product:
-	addi s1, s1, 1 # Skip *.
+	mv s1, a0 # Skip *.
 	li a0, 1
 	call _build_expression
 	la a0, asm_mul_a0_a1
@@ -937,29 +914,31 @@ _skip_comment:
 
 # Parameters:
 # a0 - Line length.
-.type _compile_assembly, @function
-_compile_assembly:
+.type _compile_procedure_section, @function
+_compile_procedure_section:
 	# Prologue.
 	addi sp, sp, -16
 	sw ra, 12(sp)
 	sw s0, 8(sp)
 	addi s0, sp, 16
 
-	sw a0, 4(sp) # a0 - Line length.
+.Lcompile_procedure_section_loop:
+	call _skip_spaces
+	call _skip_comment
+	call _skip_spaces
 
-	# Write the source to the standard output.
 	mv a0, s1
-	lw a1, 4(sp)
-	call _write_out
-
-	lw t0, 4(sp)
-	add s1, s1, t0
+	addi a1, sp, 0
+	call _tokenize_next
+	li t0, TOKEN_PROC
+	lw t1, 0(sp)
+	bne t0, t1, .Lcompile_procedure_section_end
 
-	li a0, '\n'
-	call _put_char
+	call _compile_procedure
 
-	addi s1, s1, 1 # Skip the new line.
+	j .Lcompile_procedure_section_loop
 
+.Lcompile_procedure_section_end:
 	# Epilogue.
 	lw ra, 12(sp)
 	lw s0, 8(sp)
@@ -1038,15 +1017,19 @@ _compile_constant:
 	sw s0, 8(sp)
 	addi s0, sp, 16
 
-	call _read_token
+	mv a0, s1
+	addi a1, sp, 0
+	call _tokenize_next
 
-	mv a1, a0 # The identifier length from _read_token should be in a1.
-	mv a0, s1 # Save the identifier pointer before advancing it.
-	add s1, s1, a1
+	sub a1, a0, s1 # The identifier end from _tokenize_next should be in a0.
+	mv a0, s1
+	add s1, s1, a1 # Save the identifier pointer before advancing it.
 	call _write_out
 
-	call _skip_spaces
-	addi s1, s1, 2 # Skip the assignment sign.
+	mv a0, s1
+	addi a1, sp, 0
+	call _tokenize_next
+	mv s1, a0 # Skip the assignment sign.
 
 	# : .long
 	li t0, 0x20676e6f # ong_
@@ -1154,42 +1137,10 @@ _compile_variable:
 	lw a1, 24(sp)
 	call _write_out
 
-	li t0, 0x0a74 # t\n
-	sw t0, 12(sp)
-	li t0, 0x63656a62 # bjec
-	sw t0, 8(sp)
-	li t0, 0x6f40202c # , @o
-	sw t0, 4(sp)
-	addi a0, sp, 4
-	li a1, 10
-	call _write_out
-
-	# .size identifier, size
-	li t0, 0x2065 # e_
-	sw t0, 12(sp)
-	li t0, 0x7a69732e # .siz
-	sw t0, 8(sp)
-	addi a0, sp, 8
-	li a1, 6
-	call _write_out
-
-	lw a0, 28(sp)
-	lw a1, 24(sp)
-	call _write_out
-
-	li t0, 0x202c # ,_
-	sw t0, 12(sp)
-	addi a0, sp, 12
-	li a1, 2
-	call _write_out
-
-	lw a0, 20(sp)
-	lw a1, 16(sp)
+	la a0, asm_type_object
+	li a1, ASM_TYPE_OBJECT_SIZE
 	call _write_out
 
-	li a0, '\n'
-	call _put_char
-
 	# identifier: .zero size
 	lw a0, 28(sp)
 	lw a1, 24(sp)
@@ -1239,14 +1190,8 @@ _compile_procedure:
 	lw a1, 16(sp)
 	call _write_out
 
-	li t0, 0x0a6e6f69 # ion\n
-	sw t0, 12(sp)
-	li t0, 0x74636e75 # unct
-	sw t0, 8(sp)
-	li t0, 0x6640202c # , @f
-	sw t0, 4(sp)
-	addi a0, sp, 4
-	li a1, 12
+	la a0, asm_type_function
+	li a1, ASM_TYPE_FUNCTION_SIZE
 	call _write_out
 
 	lw a0, 20(sp)
@@ -1356,7 +1301,7 @@ _compile_procedure:
 	beqz a0, .Lcompile_procedure_end
 
 	lw a0, 12(sp)
-	call _compile_line
+	call _compile_statement
 	j .Lcompile_procedure_body
 
 .Lcompile_procedure_end:
@@ -1577,7 +1522,7 @@ _compile_if:
 
 	call _read_line
 	li a1, 1
-	call _compile_line
+	call _compile_statement
 
 	j .Lcompile_if_loop
 
@@ -1614,8 +1559,8 @@ _compile_if:
 #
 # Returns 1 in a0 if the parsed line contained a text section element such a
 # procedure or the program entry point. Otherwise sets a0 to 0.
-.type _compile_line, @function
-_compile_line:
+.type _compile_statement, @function
+_compile_statement:
 	# Prologue.
 	addi sp, sp, -32
 	sw ra, 28(sp)
@@ -1626,45 +1571,17 @@ _compile_line:
 	sw a0, 20(sp)
 	sw a1, 16(sp)
 
-	beqz a0, .Lcompile_line_empty # Skip an empty line.
-
-	lbu t0, (s1)
-	li t1, '('
-	beq t0, t1, .Lcompile_line_comment
-
-	li t0, 0x636f7270 # proc
-	sw t0, 12(sp)
-	mv a0, s1
-	addi a1, sp, 12
-	li a2, 4
-	call _memcmp
-	beqz a0, .Lcompile_line_procedure
-
-	li t0, 0x69676562 # begi
-	sw t0, 12(sp)
-	mv a0, s1
-	addi a1, sp, 12
-	li a2, 4
-	call _memcmp
-	beqz a0, .Lcompile_line_begin
-
-	li t0, 0x2e646e65 # end.
-	sw t0, 12(sp)
-	mv a0, s1
-	addi a1, sp, 12
-	li a2, 4
-	call _memcmp
-	beqz a0, .Lcompile_line_exit
+	call _skip_comment
 
 	mv a0, s1
 	lw a1, 20(sp)
 	call _is_local_identifier
-	bnez a0, .Lcompile_line_identifier
+	bnez a0, .Lcompile_statement_identifier
 
 	mv a0, s1
 	li a1, 2
 	call _is_register_identifier
-	bnez a0, .Lcompile_line_identifier
+	bnez a0, .Lcompile_statement_identifier
 
 	li t0, 0x6f746f67 # goto
 	sw t0, 12(sp)
@@ -1672,7 +1589,7 @@ _compile_line:
 	addi a1, sp, 12
 	li a2, 4
 	call _memcmp
-	beqz a0, .Lcompile_line_goto
+	beqz a0, .Lcompile_statement_goto
 
 	li t0, 0x75746572 # retu
 	sw t0, 12(sp)
@@ -1680,7 +1597,7 @@ _compile_line:
 	addi a1, sp, 12
 	li a2, 4
 	call _memcmp
-	beqz a0, .Lcompile_line_return
+	beqz a0, .Lcompile_statement_return
 
 	li t0, 0x6669 # if
 	sw t0, 12(sp)
@@ -1688,77 +1605,42 @@ _compile_line:
 	addi a1, sp, 12
 	li a2, 2
 	call _memcmp
-	beqz a0, .Lcompile_line_if
+	beqz a0, .Lcompile_statement_if
 
 	lbu t0, (s1)
 	li t1, '.'
-	beq t0, t1, .Lcompile_line_label
+	beq t0, t1, .Lcompile_statement_label
 	li t1, '_'
-	beq t0, t1, .Lcompile_line_identifier
+	beq t0, t1, .Lcompile_statement_identifier
 
-	j .Lcompile_line_unchanged # Else.
+	j .Lcompile_statement_empty # Else.
 
-.Lcompile_line_if:
+.Lcompile_statement_if:
 	call _compile_if
-	j .Lcompile_line_section
+	j .Lcompile_statement_end
 
-.Lcompile_line_label:
+.Lcompile_statement_label:
 	lw a0, 20(sp)
 	call _compile_label
-	j .Lcompile_line_section
+	j .Lcompile_statement_end
 
-.Lcompile_line_return:
+.Lcompile_statement_return:
 	call _compile_return
-	j .Lcompile_line_section
+	j .Lcompile_statement_end
 
-.Lcompile_line_goto:
+.Lcompile_statement_goto:
 	call _compile_goto
-	j .Lcompile_line_section
+	j .Lcompile_statement_end
 
-.Lcompile_line_identifier:
+.Lcompile_statement_identifier:
 	call _compile_identifier
-	j .Lcompile_line_section
+	j .Lcompile_statement_end
 
-.Lcompile_line_exit:
-	call _compile_exit
-	j .Lcompile_line_section
-
-.Lcompile_line_begin:
-	lw a1, 16(sp)
-	bnez a1, .Lcompile_line_compile_entry
-	call _compile_text_section
-.Lcompile_line_compile_entry:
-	call _compile_entry_point
-	li a0, 1
-	j .Lcompile_line_end
-
-.Lcompile_line_procedure:
-	lw a1, 16(sp)
-	bnez a1, .Lcompile_line_compile_procedure
-	call _compile_text_section
-.Lcompile_line_compile_procedure:
-	call _compile_procedure
-	li a0, 1
-	j .Lcompile_line_end
-
-.Lcompile_line_comment:
-	lw a0, 20(sp)
-	call _skip_comment
-	j .Lcompile_line_section
-
-.Lcompile_line_empty:
+.Lcompile_statement_empty:
 	addi s1, s1, 1
-	j .Lcompile_line_section
+	j .Lcompile_statement_end
 
-.Lcompile_line_unchanged:
-	lw a0, 20(sp)
-	call _compile_assembly
-	j .Lcompile_line_section
-
-.Lcompile_line_section:
-	mv a0, zero
-
-.Lcompile_line_end:
+.Lcompile_statement_end:
 	sw a0, 12(sp)
 	call _skip_spaces
 	call _skip_comment
@@ -1804,20 +1686,25 @@ _compile_entry_point:
 
 	addi s1, s1, 6 # Skip begin\n.
 
-	# Epilogue.
-	lw ra, 4(sp)
-	lw s0, 0(sp)
-	addi sp, sp, 8
-	ret
+	# Generate the body of the procedure.
+.Lcompile_entry_point_body:
+	call _skip_spaces
+	call _read_line
+	sw a0, 12(sp)
+	li t0, 0x2e646e65 # end
+	sw t0, 8(sp)
+	mv a0, s1
+	addi a1, sp, 8
+	li a2, 4
+	call _memcmp
 
-.type _compile_exit, @function
-_compile_exit:
-	# Prologue.
-	addi sp, sp, -8
-	sw ra, 4(sp)
-	sw s0, 0(sp)
-	addi s0, sp, 8
+	beqz a0, .Lcompile_entry_point_end
+
+	lw a0, 12(sp)
+	call _compile_statement
+	j .Lcompile_entry_point_body
 
+.Lcompile_entry_point_end:
 	la a0, asm_exit
 	li a1, ASM_EXIT_SIZE
 	call _write_out
@@ -1857,30 +1744,13 @@ _compile:
 	sw s0, 8(sp)
 	addi s0, sp, 16
 
-	sw zero, 4(sp) # Whether the text section header was already emitted.
-
 	call _compile_module_declaration
 	call _compile_import
 	call _compile_constant_section
 	call _compile_variable_section
-
-.Lcompile_do:
-	lbu t0, (s1) # t0 = Current character.
-	beqz t0, .Lcompile_end # Exit the loop on the NUL character.
-
-	call _skip_spaces
-	call _read_line
-	lw a1, 4(sp)
-	call _compile_line
-
-	beqz a0, .Lcompile_do
-	# Update whether the text section header was already emitted.
-	lw t0, 4(sp)
-	or t0, t0, a0
-	sw t0, 4(sp)
-
-	j .Lcompile_do
-.Lcompile_end:
+	call _compile_text_section
+	call _compile_procedure_section
+	call _compile_entry_point
 
 	# Epilogue.
 	lw ra, 12(sp)
@@ -1888,22 +1758,6 @@ _compile:
 	addi sp, sp, 16
 	ret
 
-.type _main, @function
-_main:
-	# Prologue.
-	addi sp, sp, -8
-	sw ra, 4(sp)
-	sw s0, 0(sp)
-	addi s0, sp, 8
-
-	li s2, 1
-
-	# Epilogue.
-	lw ra, 4(sp)
-	lw s0, 0(sp)
-	addi sp, sp, 8
-	ret
-
 # Entry point.
 .type _start, @function
 _start:
@@ -1912,8 +1766,7 @@ _start:
 	li a1, SOURCE_BUFFER_SIZE # Buffer size.
 	call _read_file
 
-	mv a0, s1
-	call _main
+	li s2, 1
 	call _compile
 
 	# Call exit.
diff --git a/boot/tokenizer.s b/boot/tokenizer.s
index 4315f66..67b2602 100644
--- a/boot/tokenizer.s
+++ b/boot/tokenizer.s
@@ -1,4 +1,10 @@
-.global _tokenize_next, classification, transitions, keywords
+# 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/.
+
+.global _tokenize_next, classification, transitions, keywords, byte_keywords
+
+.include "boot/definitions.inc"
 
 .section .rodata
 
@@ -8,7 +14,7 @@
 #
 # Classification:
 #
-.equ CLASS_INVALID, 0x0
+.equ CLASS_INVALID, 0x00
 .equ CLASS_DIGIT, 0x01
 .equ CLASS_CHARACTER, 0x02
 .equ CLASS_SPACE, 0x03
@@ -25,9 +31,11 @@
 .equ CLASS_EOF, 0x0e
 .equ CLASS_DOT, 0x0f
 .equ CLASS_MINUS, 0x10
-.equ CLASS_DOUBLE_QUOTE, 0x11
+.equ CLASS_QUOTE, 0x11
+.equ CLASS_GREATER, 0x12
+.equ CLASS_LESS, 0x13
 
-.equ CLASS_COUNT, 18
+.equ CLASS_COUNT, 20
 
 .type classification, @object
 .size classification, 128
@@ -66,12 +74,12 @@ classification:
 	.byte CLASS_INVALID # 1F US
 	.byte CLASS_SPACE # 20 Space
 	.byte CLASS_SINGLE # 21 !
-	.byte CLASS_DOUBLE_QUOTE # 22 "
+	.byte CLASS_QUOTE # 22 "
 	.byte 0x00 # 23 #
 	.byte 0x00 # 24 $
 	.byte CLASS_SINGLE # 25 %
 	.byte CLASS_SINGLE # 26 &
-	.byte 0x00 # 27 '
+	.byte CLASS_QUOTE # 27 '
 	.byte CLASS_LEFT_PAREN # 28 (
 	.byte CLASS_RIGHT_PAREN # 29 )
 	.byte CLASS_ASTERISK # 2A *
@@ -92,9 +100,9 @@ classification:
 	.byte CLASS_DIGIT # 39 9
 	.byte CLASS_COLON # 3A :
 	.byte CLASS_SINGLE # 3B ;
-	.byte 0x00 # 3C <
+	.byte CLASS_LESS # 3C <
 	.byte CLASS_EQUALS # 3D =
-	.byte 0x00 # 3E >
+	.byte CLASS_GREATER # 3E >
 	.byte 0x00 # 3F ?
 	.byte CLASS_SINGLE # 40 @
 	.byte CLASS_CHARACTER # 41 A
@@ -220,7 +228,10 @@ keywords:
 	.ascii "case"
 	.word 2
 	.ascii "of"
-.size keywords, . - keywords
+
+.type byte_keywords, @object
+byte_keywords: .ascii "&.,:;()[]^=+-*@"
+.equ BYTE_KEYWORDS_SIZE, . - byte_keywords
 
 .section .data
 
@@ -240,78 +251,66 @@ keywords:
 #   handles each action.
 #
 .type transitions, @object
-.size transitions, 17 * CLASS_COUNT # state count * CLASS_COUNT
+.size transitions, 14 * CLASS_COUNT # state count * CLASS_COUNT
 transitions:
 	#     Invalid Digit   Alpha   Space   :       =       (       )     
 	#     *       _       Single  Hex     0       x       NUL     .
-	#     -       "
-	.word 0x00ff, 0x0103, 0x0102, 0x0300, 0x0101, 0x0105, 0x0106, 0x0107
-	.word 0x0108, 0x0102, 0x010b, 0x0102, 0x010c, 0x0102, 0x00ff, 0x010e # 00 Start
-	.word 0x010f, 0x0110
+	#     -       " or '  >       <
+	.word 0x00ff, 0x0103, 0x0102, 0x0300, 0x0101, 0x06ff, 0x0106, 0x06ff
+	.word 0x06ff, 0x0102, 0x06ff, 0x0102, 0x010c, 0x0102, 0x00ff, 0x0108
+	.word 0x0105, 0x0110, 0x0104, 0x0107 # 0x00 Start
 
-	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x0104, 0x02ff, 0x02ff
-	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff # 01 Colon
-	.word 0x02ff, 0x02ff
+	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x07ff, 0x02ff, 0x02ff
+	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
+	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff # 0x01 Colon
 
 	.word 0x05ff, 0x0102, 0x0102, 0x05ff, 0x05ff, 0x05ff, 0x05ff, 0x05ff
-	.word 0x05ff, 0x0102, 0x05ff, 0x0102, 0x0102, 0x0102, 0x05ff, 0x05ff # 02 Identifier
-	.word 0x05ff, 0x05ff
+	.word 0x05ff, 0x0102, 0x05ff, 0x0102, 0x0102, 0x0102, 0x05ff, 0x05ff
+	.word 0x05ff, 0x05ff, 0x05ff, 0x05ff # 0x02 Identifier
 
 	.word 0x02ff, 0x0103, 0x00ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
-	.word 0x02ff, 0x02ff, 0x02ff, 0x00ff, 0x0103, 0x02ff, 0x02ff, 0x02ff # 03 Integer
-	.word 0x02ff, 0x02ff
+	.word 0x02ff, 0x02ff, 0x02ff, 0x00ff, 0x0103, 0x02ff, 0x02ff, 0x02ff
+	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff # 0x03 Integer
 
+	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x04ff, 0x02ff, 0x02ff
 	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
-	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff # 04 Assign
-	.word 0x02ff, 0x02ff
+	.word 0x02ff, 0x02ff, 0x04ff, 0x02ff # 0x04 Greater
 
-	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
-	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff # 05 Eauals
-	.word 0x02ff, 0x02ff
+	.word 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff
+	.word 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff
+	.word 0x06ff, 0x06ff, 0x04ff, 0x06ff # 0x05 Minus
 
 	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
-	.word 0x0109, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff # 06 Left paren
-	.word 0x02ff, 0x02ff
+	.word 0x0109, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
+	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff # 0x06 Left paren
 
 	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
-	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff # 07 Right paren
-	.word 0x02ff, 0x02ff
-
 	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
-	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff # 08 Asterisk
-	.word 0x02ff, 0x02ff
+	.word 0x02ff, 0x02ff, 0x02ff, 0x04ff # 0x07 Less
+
+	.word 0x02ff, 0x0102, 0x0102, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
+	.word 0x02ff, 0x0102, 0x02ff, 0x0102, 0x0102, 0x0102, 0x02ff, 0x02ff
+	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff # 0x08 Dot
 
 	.word 0x0109, 0x0109, 0x0109, 0x0109, 0x0109, 0x0109, 0x0109, 0x0109
-	.word 0x010a, 0x0109, 0x0109, 0x0109, 0x0109, 0x0109, 0x00ff, 0x0109 # 09 Comment
-	.word 0x0109, 0x0109
+	.word 0x010a, 0x0109, 0x0109, 0x0109, 0x0109, 0x0109, 0x00ff, 0x0109
+	.word 0x0109, 0x0109, 0x0109, 0x0109 # 0x09 Comment
 
 	.word 0x00ff, 0x0109, 0x0109, 0x0109, 0x0109, 0x0109, 0x0109, 0x04ff
-	.word 0x010a, 0x0109, 0x0109, 0x0109, 0x0109, 0x0109, 0x00ff, 0x0109 # 0a Closing comment
-	.word 0x0109, 0x0109
+	.word 0x010a, 0x0109, 0x0109, 0x0109, 0x0109, 0x0109, 0x00ff, 0x0109
+	.word 0x0109, 0x0109, 0x0109, 0x0109 # 0x0a Closing comment
 
-	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
-	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff # 0b Single character token
-	.word 0x02ff, 0x02ff
+	.word 0x00ff, 0x010b, 0x010b, 0x010b, 0x010b, 0x010b, 0x010b, 0x0110
+	.word 0x010b, 0x010b, 0x010b, 0x010b, 0x010b, 0x010b, 0x010b, 0x0110
+	.word 0x010b, 0x04ff, 0x010b, 0x010b # 0x0b String
 
 	.word 0x02ff, 0x00ff, 0x00ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
-	.word 0x02ff, 0x02ff, 0x02ff, 0x00ff, 0x00ff, 0x010d, 0x02ff, 0x02ff # 0c Zero
-	.word 0x02ff, 0x02ff
+	.word 0x02ff, 0x02ff, 0x02ff, 0x00ff, 0x00ff, 0x010d, 0x02ff, 0x02ff
+	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff # 0x0c Zero
 
 	.word 0x02ff, 0x010d, 0x00ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
-	.word 0x02ff, 0x02ff, 0x02ff, 0x010d, 0x010d, 0x00ff, 0x2ff, 0x02ff # 0d Hexadecimal
-	.word 0x00ff, 0x02ff
-
-	.word 0x02ff, 0x0102, 0x0102, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
-	.word 0x02ff, 0x0102, 0x02ff, 0x0102, 0x0102, 0x0102, 0x02ff, 0x02ff # 0e Dot
-	.word 0x02ff, 0x02ff
-
-	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff # 0f Minus
-	.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
-	.word 0x02ff, 0x02ff
-
-	.word 0x00ff, 0x0110, 0x0110, 0x0110, 0x0110, 0x0110, 0x0110, 0x0110 # 10 Starting string.
-	.word 0x0110, 0x0110, 0x0110, 0x0110, 0x0110, 0x0110, 0x0110, 0x0110
-	.word 0x0110, 0x04ff
+	.word 0x02ff, 0x02ff, 0x02ff, 0x010d, 0x010d, 0x00ff, 0x2ff, 0x02ff
+	.word 0x00ff, 0x02ff, 0x02ff, 0x02ff # 0x0d Hexadecimal
 
 .section .text
 
@@ -406,6 +405,57 @@ _classify_identifier:
 	addi sp, sp, 16
 	ret
 
+# Takes a symbol and determines its type.
+#
+# Parameters:
+# a0 - Token character.
+#
+# Sets a0 to the appropriate token type.
+.type _classify_single, @function
+_classify_single:
+	# Prologue.
+	addi sp, sp, -16
+	sw ra, 12(sp)
+	sw s0, 8(sp)
+	addi s0, sp, 16
+
+	mv a1, a0
+	li a2, BYTE_KEYWORDS_SIZE
+	la a0, byte_keywords
+	call _memchr
+
+	la a1, byte_keywords
+	sub a0, a0, a1
+	addi a0, a0, 27
+
+	# Epilogue.
+	lw ra, 12(sp)
+	lw s0, 8(sp)
+	addi sp, sp, 16
+	ret
+
+# Classified a symbol containing multiple characters (probably 2).
+#
+# Parameters:
+# a0 - Token length.
+# a1 - Token pointer.
+#
+# Sets a0 to the appropriate token type.
+.type _classify_composite, @function
+_classify_composite:
+	lbu t0, 0(a1)
+	li t1, ':'
+	beq t0, t1, .Lclassify_composite_assign
+
+	j .Lclassify_composite_end
+
+.Lclassify_composite_assign:
+	li a0, TOKEN_ASSIGN
+	j .Lclassify_composite_end
+
+.Lclassify_composite_end:
+	ret
+
 # Initializes the classification table.
 #
 # Paramaters:
@@ -453,12 +503,18 @@ _tokenize_next:
 	li t0, 0x03 # Skip action.
 	beq t1, t0, .Ltokenize_next_skip
 
-	li t0, 0x04 # Comment action.
+	li t0, 0x04 # Delimited string action.
 	beq t1, t0, .Ltokenize_next_comment
 
 	li t0, 0x05 # Finalize identifier.
 	beq t1, t0, .Ltokenize_next_identifier
 
+	li t0, 0x06 # Single character symbol action.
+	beq t1, t0, .Ltokenize_next_single
+
+	li t0, 0x07 # An action for symbols containing multiple characters.
+	beq t1, t0, .Ltokenize_next_composite
+
 	j .Ltokenize_next_reject
 
 .Ltokenize_next_reject:
@@ -481,24 +537,17 @@ _tokenize_next:
 
 .Ltokenize_next_print:
 	/* DEBUG
-	lw a0, 4(sp)
-	mv a1, s1
-	sub a1, a1, a0
-	call _write_error
-	DEBUG */
+	addi a0, a0, 21
+	sw a0, 0(sp)
+	addi a0, sp, 0
+	li a1, 1
+	call _write_error */
 
 	j .Ltokenize_next_end
 
 .Ltokenize_next_comment:
 	addi s1, s1, 1
 
-	/* DEBUG
-	lw a0, 4(sp)
-	mv a1, s1
-	sub a1, a1, a0
-	call _write_error
-	DEBUG */
-
 	j .Ltokenize_next_end
 
 .Ltokenize_next_identifier:
@@ -512,6 +561,26 @@ _tokenize_next:
 
 	j .Ltokenize_next_end
 
+.Ltokenize_next_single:
+	lw a0, 4(sp)
+	addi s1, a0, 1
+	lbu a0, (a0)
+	call _classify_single
+	lw a1, 0(sp)
+	sw a0, (a1)
+
+	j .Ltokenize_next_end
+
+.Ltokenize_next_composite:
+	addi s1, s1, 1
+	lw a1, 4(sp)
+	sub a0, s1, a1
+	call _classify_composite
+	lw a1, 0(sp)
+	sw a0, (a1)
+
+	j .Ltokenize_next_end
+
 .Ltokenize_next_end:
 	mv a0, s1 # Return the advanced text pointer.