Replace _read_token with the lexer

1 files changed, 72 insertions, 170 deletions

diff --git a/boot/stage1.s b/boot/stage1.s
index b32815d..525da11 100644
--- a/boot/stage1.s
+++ b/boot/stage1.s
@@ -409,11 +409,15 @@ _build_expression:
 	mv a0, s1
 	addi a1, sp, 24
 	call _tokenize_next
+	sw a0, 20(sp)
 
-	call _skip_spaces
-	call _read_token
-	sw s1, 32(sp)
-	sw a0, 28(sp)
+	/* DEBUG
+	lw a0, 32(sp)
+	lw a1, 28(sp)
+	call _write_error
+	lw a0, 28(sp)
+	li a1, 8
+	call _write_error */
 
 	lw a0, 24(sp)
 
@@ -423,14 +427,15 @@ _build_expression:
 	li t0, TOKEN_AT
 	beq a0, t0, .Lbuild_expression_address
 
-	lbu a0, (s1)
-	call _is_digit
-	bnez a0, .Lbuild_expression_literal
+	li t0, TOKEN_INTEGER
+	beq a0, t0, .Lbuild_expression_literal
 
-	lbu a0, (s1)
+	lw a0, 32(sp)
+	lbu a0, (a0)
 	li t0, '_'
 	beq a0, t0, .Lbuild_expression_call
 
+	lw s1, 32(sp)
 	lw a0, 28(sp)
 	lw a1, 36(sp)
 	call _compile_identifier_expression
@@ -438,7 +443,7 @@ _build_expression:
 	j .Lbuild_expression_advance
 
 .Lbuild_expression_negate:
-	addi s1, s1, 1 # Skip the -.
+	lw s1, 20(sp) # Skip the -.
 	mv a0, zero
 	call _build_expression
 
@@ -446,7 +451,7 @@ _build_expression:
 	li a1, ASM_NEG_A0_SIZE
 	call _write_out
 
-	j .Lbuild_expression_advance
+	j .Lbuild_expression_end
 
 .Lbuild_expression_address:
 	lw t1, 36(sp)
@@ -463,10 +468,10 @@ _build_expression:
 	li a1, 13
 	call _write_out
 
-	addi s1, s1, 1 # Skip @.
-	call _read_token
-	sw s1, 32(sp)
-	sw a0, 28(sp)
+	lw a0, 20(sp) # Skip @.
+	addi a1, sp, 24
+	call _tokenize_next
+	mv s1, a0
 
 	lw a0, 32(sp)
 	lw a1, 28(sp)
@@ -477,13 +482,16 @@ _build_expression:
 	li a0, '\n'
 	call _put_char
 
-	j .Lbuild_expression_advance
+	j .Lbuild_expression_end
 
 .Lbuild_expression_call:
+	lw a0, 20(sp)
+	addi a1, sp, 8
+	call _tokenize_next
+	mv s1, a0
+
 	lw a0, 32(sp)
 	lw a1, 28(sp)
-	add s1, s1, a1
-	addi s1, s1, 1
 	call _compile_call
 
 	j .Lbuild_expression_end
@@ -509,8 +517,7 @@ _build_expression:
 	j .Lbuild_expression_advance
 
 .Lbuild_expression_advance:
-	lw a0, 28(sp)
-	add s1, s1, a0
+	lw s1, 20(sp)
 
 .Lbuild_expression_end:
 	# Epilogue.
@@ -747,102 +754,6 @@ _compile_call:
 	addi sp, sp, 32
 	ret
 
-# Reads a token and returns its length in a0.
-# _read_token doesn't change s1, it finds the length of the token s1 is pointing to.
-.type _read_token, @function
-_read_token:
-	# Prologue.
-	addi sp, sp, -16
-	sw ra, 12(sp)
-	sw s0, 8(sp)
-	addi s0, sp, 16
-
-	lbu t0, (s1) # t0 = Current character.
-	sw zero, 4(sp)
-
-	li t1, '.'
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, ','
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, ':'
-	beq t0, t1, .Ltoken_character_colon
-
-	li t1, ';'
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, '('
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, ')'
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, '['
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, ']'
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, '^'
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, '&'
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, '='
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, '+'
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, '-'
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, '*'
-	beq t0, t1, .Ltoken_character_single
-
-	li t1, '@'
-	beq t0, t1, .Ltoken_character_single
-# Expect an identifier or a number.
-.Ltoken_character_loop_do:
-	lw t6, 4(sp)
-	add t1, s1, t6
-	lbu a0, (t1) # a0 = Current character.
-
-	call _is_alnum
-
-	beqz a0, .Ltoken_character_end
-	lw t6, 4(sp)
-	addi t6, t6, 1
-	sw t6, 4(sp)
-	j .Ltoken_character_loop_do
-
-.Ltoken_character_single:
-	lw t6, 4(sp)
-	addi t6, t6, 1
-	sw t6, 4(sp)
-	j .Ltoken_character_end
-
-.Ltoken_character_colon:
-	lbu t0, 1(s1) # t0 = The character after the colon.
-	lw t6, 4(sp)
-	addi t6, t6, 1
-	sw t6, 4(sp)
-
-	li t1, '='
-	beq t0, t1, .Ltoken_character_single
-	j .Ltoken_character_end
-
-.Ltoken_character_end:
-	lw a0, 4(sp)
-
-	# Epilogue.
-	lw ra, 12(sp)
-	lw s0, 8(sp)
-	addi sp, sp, 16
-	ret
-
 # Skips the spaces till the next non space character.
 .type _skip_spaces, @function
 _skip_spaces:
@@ -911,15 +822,14 @@ _skip_comment:
 	addi sp, sp, 16
 	ret
 
-# Parameters:
-# a0 - Line length.
+# Walks through the procedure definitions.
 .type _compile_procedure_section, @function
 _compile_procedure_section:
 	# Prologue.
-	addi sp, sp, -24
-	sw ra, 20(sp)
-	sw s0, 16(sp)
-	addi s0, sp, 24
+	addi sp, sp, -32
+	sw ra, 28(sp)
+	sw s0, 24(sp)
+	addi s0, sp, 32
 
 .Lcompile_procedure_section_loop:
 	call _skip_spaces
@@ -938,9 +848,9 @@ _compile_procedure_section:
 
 .Lcompile_procedure_section_end:
 	# Epilogue.
-	lw ra, 20(sp)
-	lw s0, 16(sp)
-	addi sp, sp, 24
+	lw ra, 28(sp)
+	lw s0, 24(sp)
+	addi sp, sp, 32
 	ret
 
 .type _compile_module_declaration, @function
@@ -970,10 +880,10 @@ _compile_module_declaration:
 .type _compile_constant_section, @function
 _compile_constant_section:
 	# Prologue.
-	addi sp, sp, -24
-	sw ra, 20(sp)
-	sw s0, 16(sp)
-	addi s0, sp, 24
+	addi sp, sp, -32
+	sw ra, 28(sp)
+	sw s0, 24(sp)
+	addi s0, sp, 32
 
 	mv a0, s1
 	addi a1, sp, 4
@@ -988,19 +898,24 @@ _compile_constant_section:
 	call _write_out
 
 .Lcompile_constant_section_item:
-	call _skip_spaces
-	lbu a0, (s1)
-	call _is_upper
-	beqz a0, .Lcompile_constant_section_end
+	mv a0, s1
+	addi a1, sp, 12
+	call _tokenize_next
+
+	lw t0, 12(sp)
+	li t1, TOKEN_IDENTIFIER
+
+	bne t0, t1, .Lcompile_constant_section_end
+	lw s1, 20(sp)
 
 	call _compile_constant
 	j .Lcompile_constant_section_item
 
 .Lcompile_constant_section_end:
 	# Epilogue.
-	lw ra, 20(sp)
-	lw s0, 16(sp)
-	addi sp, sp, 24
+	lw ra, 28(sp)
+	lw s0, 24(sp)
+	addi sp, sp, 32
 	ret
 
 .type _compile_constant, @function
@@ -1042,7 +957,6 @@ _compile_constant:
 
 	li a0, '\n'
 	call _put_char
-	call _skip_spaces
 
 	# Epilogue.
 	lw ra, 28(sp)
@@ -1053,10 +967,10 @@ _compile_constant:
 .type _compile_variable_section, @function
 _compile_variable_section:
 	# Prologue.
-	addi sp, sp, -24
-	sw ra, 20(sp)
-	sw s0, 16(sp)
-	addi s0, sp, 24
+	addi sp, sp, -32
+	sw ra, 28(sp)
+	sw s0, 24(sp)
+	addi s0, sp, 32
 
 	mv a0, s1
 	addi a1, sp, 4
@@ -1071,19 +985,23 @@ _compile_variable_section:
 	call _write_out
 
 .Lcompile_variable_section_item:
-	call _skip_spaces
-	lbu a0, (s1)
-	call _is_lower
-	beqz a0, .Lcompile_variable_section_end
+	mv a0, s1
+	addi a1, sp, 12
+	call _tokenize_next
+	lw t0, 12(sp)
+	li t1, TOKEN_IDENTIFIER
+
+	bne t0, t1, .Lcompile_variable_section_end
+	lw s1, 20(sp) # Advance to the beginning of the variable name.
 
 	call _compile_variable
 	j .Lcompile_variable_section_item
 
 .Lcompile_variable_section_end:
 	# Epilogue.
-	lw ra, 20(sp)
-	lw s0, 16(sp)
-	addi sp, sp, 24
+	lw ra, 28(sp)
+	lw s0, 24(sp)
+	addi sp, sp, 32
 	ret
 
 # Compile a global variable.
@@ -1111,18 +1029,6 @@ _compile_variable:
 	call _tokenize_next # Skip the type.
 	mv s1, a0
 
-	/* DEBUG
-	lw a0, 24(sp)
-	add a0, a0, '0'
-	sw a0, 24(sp)
-	addi a0, sp, 24
-	li a1, 1
-	call _write_error
-	lw a0, 28(sp)
-	li a1, 8
-	call _write_error
-	*/
-
 	# .type identifier, @object
 	la a0, asm_type
 	li a1, ASM_TYPE_SIZE
@@ -1478,10 +1384,6 @@ _compile_statement:
 	j .Lcompile_statement_end
 
 .Lcompile_statement_end:
-	sw a0, 12(sp)
-	call _skip_comment
-	lw a0, 12(sp)
-
 	# Epilogue.
 	lw ra, 28(sp)
 	lw s0, 24(sp)
@@ -1492,19 +1394,19 @@ _compile_statement:
 .type _compile_text_section, @function
 _compile_text_section:
 	# Prologue.
-	addi sp, sp, -8
-	sw ra, 4(sp)
-	sw s0, 0(sp)
-	addi s0, sp, 8
+	addi sp, sp, -16
+	sw ra, 12(sp)
+	sw s0, 8(sp)
+	addi s0, sp, 16
 
 	la a0, section_text
 	li a1, SECTION_TEXT_SIZE
 	call _write_out
 
 	# Epilogue.
-	lw ra, 4(sp)
-	lw s0, 0(sp)
-	addi sp, sp, 8
+	lw ra, 12(sp)
+	lw s0, 8(sp)
+	addi sp, sp, 16
 	ret
 
 .type _compile_entry_point, @function