preproce.inc

一个用纯汇编 写的操作系统 源代码 是用 TASM 编译器写的

; flat assembler core
; Copyright (c) 1999-2006, Tomasz Grysztar.
; All rights reserved.

preprocessor:
	mov	edi,characters
	mov	ecx,100h
	xor	al,al
      make_characters_table:
	stosb
	inc	al
	loop	make_characters_table
	mov	esi,characters+'a'
	mov	edi,characters+'A'
	mov	ecx,26
	rep	movsb
	mov	edi,characters
	mov	esi,symbol_characters+1
	movzx	ecx,byte [esi-1]
	xor	eax,eax

      mark_symbol_characters:
	lodsb
	mov	byte [edi+eax],0
	loop	mark_symbol_characters
	mov	edi,locals_counter
	mov	al,7
	stos	byte [edi]
	movzx	ecx,al
	mov	al,'0'
	rep	stos byte [edi]
	mov	edi,[memory_start]
	mov	[include_paths],edi
	
	mov	esi,include_variable
	call	get_environment_variable

	xor	al,al
	stosb

	mov	[memory_start],edi
	mov	eax,[additional_memory]
	mov	[free_additional_memory],eax
	mov	eax,[additional_memory_end]
	mov	[labels_list],eax
	
	xor	eax,eax
	mov	[display_buffer],eax
	mov	[hash_tree],eax
	mov	[macro_status],al

	mov	esi,[input_file]
	mov	edx,esi
	call	open
	jc	main_file_not_found
	
	mov	edi,[memory_start]
	call	preprocess_file
	
	mov	eax,[error_line]
	mov	[current_line],eax
	cmp	[macro_status],0
	jne	incomplete_macro
	mov	[source_start],edi
	ret

preprocess_file:
	push	[memory_end]
	
	push	esi
	
	mov	al,2		;FILE_END
	xor	edx,edx
	call	lseek
	push	eax
	
	xor	al,al
	xor	edx,edx
	call	lseek
	
	pop	ecx
	
	mov	edx,[memory_end]
	dec	edx
	mov	byte [edx],1Ah
	
	sub	edx,ecx
	jc	out_of_memory
	
	mov	esi,edx
	cmp	edx,edi
	jbe	out_of_memory

	mov	[memory_end],edx
	
	call	read
	call	close

	pop	edx
	xor	ecx,ecx
	mov	ebx,esi
	
      preprocess_source:
	inc	ecx
	mov	[current_line],edi
	mov	eax,edx
	stos	dword [edi]
	mov	eax,ecx
	stos	dword [edi]
	mov	eax,esi
	sub	eax,ebx
	stos	dword [edi]
	xor	eax,eax
	stos	dword [edi]
	
	push	ebx edx
	
	call	convert_line
	
	call	preprocess_line
	pop	edx ebx

      next_line:
	cmp	byte [esi-1],1Ah
	jne	preprocess_source
	
      file_end:
	nop
	pop	[memory_end]
	clc
	ret

convert_line:
	push	ecx
	test	[macro_status],0Fh
	jz	convert_line_data
	mov	ax,3Bh
	stos	word [edi]
      convert_line_data:
	cmp	edi,[memory_end]
	jae	out_of_memory
	lods	byte [esi]
	cmp	al,20h
	je	convert_line_data
	cmp	al,9
	je	convert_line_data
	mov	ah,al
	mov	ebx,characters
	xlat	byte [ebx]
	or	al,al
	jz	convert_separator
	cmp	ah,27h
	je	convert_string
	cmp	ah,22h
	je	convert_string
	mov	byte [edi],1Ah
	scas	word [edi]
	xchg	al,ah
	stos	byte [edi]
	mov	ebx,characters
	xor	ecx,ecx
      convert_symbol:
	lods	byte [esi]
	stos	byte [edi]
	xlat	byte [ebx]
	or	al,al
	loopnzd convert_symbol
	neg	ecx
	cmp	ecx,255
	ja	name_too_long
	mov	ebx,edi
	sub	ebx,ecx
	mov	byte [ebx-2],cl
      found_separator:
	dec	edi
	mov	ah,[esi-1]
      convert_separator:
	xchg	al,ah
	cmp	al,20h
	jb	control_character
	je	convert_line_data
      symbol_character:
	cmp	al,3Bh
	je	ignore_comment
	cmp	al,5Ch
	je	backslash_character
	stos	byte [edi]
	jmp	convert_line_data
      control_character:
	cmp	al,1Ah
	je	line_end
	cmp	al,0Dh
	je	cr_character
	cmp	al,0Ah
	je	lf_character
	cmp	al,9
	je	convert_line_data
	or	al,al
	jnz	symbol_character
	jmp	line_end
      lf_character:
	lods	byte [esi]
	cmp	al,0Dh
	je	line_end
	dec	esi
	jmp	line_end
      cr_character:
	lods	byte [esi]
	cmp	al,0Ah
	je	line_end
	dec	esi
	jmp	line_end
      convert_string:
	mov	al,22h
	stos	byte [edi]
	scas	dword [edi]
	mov	ebx,edi
      copy_string:
	lods	byte [esi]
	stos	byte [edi]
	cmp	al,0Ah
	je	missing_end_quote
	cmp	al,0Dh
	je	missing_end_quote
	or	al,al
	jz	missing_end_quote
	cmp	al,1Ah
	je	missing_end_quote
	cmp	al,ah
	jne	copy_string
	lods	byte [esi]
	cmp	al,ah
	je	copy_string
	dec	esi
	dec	edi
	mov	eax,edi
	sub	eax,ebx
	mov	[ebx-4],eax
	jmp	convert_line_data
      backslash_character:
	mov	byte [edi],0
	lods	byte [esi]
	cmp	al,20h
	je	concatenate_lines
	cmp	al,9
	je	concatenate_lines
	cmp	al,1Ah
	je	unexpected_end_of_file
	cmp	al,0Ah
	je	concatenate_lf
	cmp	al,0Dh
	je	concatenate_cr
	cmp	al,3Bh
	je	find_concatenated_line
	mov	al,1Ah
	stos	byte [edi]
	mov	ecx,edi
	mov	ax,5C01h
	stos	word [edi]
	dec	esi
      group_backslashes:
	lods	byte [esi]
	cmp	al,5Ch
	jne	backslashed_symbol
	stos	byte [edi]
	inc	byte [ecx]
	jmp	group_backslashes
      backslashed_symbol:
	cmp	al,1Ah
	je	unexpected_end_of_file
	cmp	al,0Ah
	je	extra_characters_on_line
	cmp	al,0Dh
	je	extra_characters_on_line
	cmp	al,20h
	je	extra_characters_on_line
	cmp	al,9
	je	extra_characters_on_line
	cmp	al,22h
	je	extra_characters_on_line
	cmp	al,27h
	je	extra_characters_on_line
	cmp	al,3Bh
	je	extra_characters_on_line
	mov	ah,al
	mov	ebx,characters
	xlat	byte [ebx]
	or	al,al
	jz	backslashed_symbol_character
	mov	al,ah
      convert_backslashed_symbol:
	stos	byte [edi]
	xlat	byte [ebx]
	or	al,al
	jz	found_separator
	inc	byte [ecx]
	jz	name_too_long
	lods	byte [esi]
	jmp	convert_backslashed_symbol
      backslashed_symbol_character:
	mov	al,ah
	stos	byte [edi]
	inc	byte [ecx]
	jmp	convert_line_data
      concatenate_lines:
	lods	byte [esi]
	cmp	al,20h
	je	concatenate_lines
	cmp	al,9
	je	concatenate_lines
	cmp	al,1Ah
	je	unexpected_end_of_file
	cmp	al,0Ah
	je	concatenate_lf
	cmp	al,0Dh
	je	concatenate_cr
	cmp	al,3Bh
	jne	extra_characters_on_line
      find_concatenated_line:
	lods	byte [esi]
	cmp	al,0Ah
	je	concatenate_lf
	cmp	al,0Dh
	je	concatenate_cr
	or	al,al
	jz	concatenate_ok
	cmp	al,1Ah
	jne	find_concatenated_line
	jmp	unexpected_end_of_file
      concatenate_lf:
	lods	byte [esi]
	cmp	al,0Dh
	je	concatenate_ok
	dec	esi
	jmp	concatenate_ok
      concatenate_cr:
	lods	byte [esi]
	cmp	al,0Ah
	je	concatenate_ok
	dec	esi
      concatenate_ok:
	inc	dword [esp]
	jmp	convert_line_data
      ignore_comment:
	lods	byte [esi]
	cmp	al,0Ah
	je	lf_character
	cmp	al,0Dh
	je	cr_character
	or	al,al
	jz	line_end
	cmp	al,1Ah
	jne	ignore_comment
      line_end:
	xor	al,al
	stos	byte [edi]
	pop	ecx
	ret

lower_case:
	mov	edi,converted
	mov	ebx,characters
      convert_case:
	lods	byte [esi]
	xlat	byte [ebx]
	stos	byte [edi]
	loop	convert_case
      case_ok:
	ret

get_directive:
	push	edi
	mov	edx,esi
	mov	ebp,ecx
	call	lower_case
	pop	edi
      scan_directives:
	mov	esi,converted
	movzx	eax,byte [edi]
	or	al,al
	jz	no_directive
	mov	ecx,ebp
	inc	edi
	mov	ebx,edi
	add	ebx,eax
	mov	ah,[esi]
	cmp	ah,[edi]
	jb	no_directive
	ja	next_directive
	cmp	cl,al
	jne	next_directive
	repe	cmps byte [esi],[edi]
	jb	no_directive
	je	directive_ok
      next_directive:
	mov	edi,ebx
	add	edi,2
	jmp	scan_directives
      no_directive:
	mov	esi,edx
	mov	ecx,ebp
	stc
	ret
      directive_ok:
	lea	esi,[edx+ebp]
	movzx	eax,word [ebx]
	add	eax,preprocessor
	clc
	ret


preprocess_line:
	mov	eax,esp
	sub	eax,100h
	jc	stack_overflow
	cmp	eax,[stack_limit]
	jb	stack_overflow
	push	ecx esi
      preprocess_current_line:
	mov	esi,[current_line]
	add	esi,16
	cmp	word [esi],3Bh
	jne	line_start_ok
	add	esi,2
      line_start_ok:
	test	[macro_status],0F0h
	jnz	macro_preprocessing
	cmp	byte [esi],1Ah
	jne	not_fix_constant
	movzx	edx,byte [esi+1]
	lea	edx,[esi+2+edx]
	cmp	word [edx],031Ah
	jne	not_fix_constant
	mov	ebx,characters
	movzx	eax,byte [edx+2]
	xlat	byte [ebx]
	ror	eax,8
	mov	al,[edx+3]
	xlat	byte [ebx]
	ror	eax,8
	mov	al,[edx+4]
	xlat	byte [ebx]
	ror	eax,16
	cmp	eax,'fix'
	je	define_fix_constant
      not_fix_constant:
	call	process_fix_constants
	jmp	initial_preprocessing_ok
      macro_preprocessing:
	call	process_macro_operators
      initial_preprocessing_ok:
	mov	esi,[current_line]
	add	esi,16
	mov	al,[macro_status]
	test	al,2
	jnz	skip_macro_block
	test	al,1
	jnz	find_macro_block
      preprocess_instruction:
	mov	[current_offset],esi
	lods	byte [esi]
	movzx	ecx,byte [esi]
	inc	esi
	cmp	al,1Ah
	jne	not_preprocessor_symbol
	cmp	cl,3
	jb	not_preprocessor_directive
	push	edi
	mov	edi,preprocessor_directives
	call	get_directive
	pop	edi
	jc	not_preprocessor_directive
	mov	byte [edx-2],3Bh
	jmp	near eax
      not_preprocessor_directive:
	xor	ch,ch
	call	get_preprocessor_symbol
	jc	not_macro
	mov	byte [ebx-2],3Bh
	mov	[struc_name],0
	jmp	use_macro
      not_macro:
	mov	[struc_name],esi
	add	esi,ecx
	lods	byte [esi]
	cmp	al,':'
	je	preprocess_label
	cmp	al,1Ah
	jne	not_preprocessor_symbol
	lods	byte [esi]
	cmp	al,3
	jne	not_symbolic_constant
	mov	ebx,characters
	movzx	eax,byte [esi]
	xlat	byte [ebx]
	ror	eax,8
	mov	al,[esi+1]
	xlat	byte [ebx]
	ror	eax,8
	mov	al,[esi+2]
	xlat	byte [ebx]
	ror	eax,16
	cmp	eax,'equ'
	je	define_equ_constant
	mov	al,3
      not_symbolic_constant:
	mov	ch,1
	mov	cl,al
	call	get_preprocessor_symbol
	jc	not_preprocessor_symbol
	push	edx esi
	mov	esi,[struc_name]
	mov	[struc_label],esi
	sub	[struc_label],2
	mov	cl,[esi-1]
	mov	ch,10b
	call	get_preprocessor_symbol
	jc	struc_name_ok
	mov	ecx,[edx+12]
	add	ecx,3
	lea	ebx,[edi+ecx]
	mov	ecx,edi
	sub	ecx,[struc_label]
	lea	esi,[edi-1]
	lea	edi,[ebx-1]
	std
	rep	movs byte [edi],[esi]
	cld
	mov	edi,[struc_label]
	mov	esi,[edx+8]
	mov	ecx,[edx+12]
	add	[struc_name],ecx
	add	[struc_name],3
	call	move_data
	mov	al,3Ah
	stos	byte [edi]
	mov	ax,3Bh
	stos	word [edi]
	mov	edi,ebx
	pop	esi
	add	esi,[edx+12]
	add	esi,3
	pop	edx
	jmp	use_macro
      struc_name_ok:
	mov	edx,[struc_name]
	movzx	eax,byte [edx-1]
	add	edx,eax
	mov	al,3Ah
	mov	[edx],al
	inc	al
	xchg	al,[edx+1]
	dec	al
	mov	[edx+2],al
	pop	esi edx
	jmp	use_macro
      preprocess_label:
	dec	esi
	sub	esi,ecx
	lea	ebp,[esi-2]
	mov	ch,10b
	call	get_preprocessor_symbol
	jnc	symbolic_constant_in_label
	lea	esi,[esi+ecx+1]
	jmp	preprocess_instruction
      symbolic_constant_in_label:
	mov	ebx,[edx+8]
	mov	ecx,[edx+12]
	add	ecx,ebx
      check_for_broken_label:
	cmp	ebx,ecx
	je	label_broken
	cmp	byte [ebx],1Ah
	jne	label_broken
	movzx	eax,byte [ebx+1]
	lea	ebx,[ebx+2+eax]
	cmp	ebx,ecx
	je	label_constant_ok
	cmp	byte [ebx],':'
	jne	label_broken
	inc	ebx
	jmp	check_for_broken_label
      label_broken:
	push	line_preprocessed
	jmp	replace_symbolic_constant
      label_constant_ok:
	mov	ecx,edi
	sub	ecx,esi
	mov	edi,[edx+12]
	add	edi,ebp
	push	edi
	lea	eax,[edi+ecx]
	push	eax
	cmp	esi,edi
	je	replace_label
	jb	move_rest_of_line_up
	rep	movs byte [edi],[esi]
	jmp	replace_label
      move_rest_of_line_up:
	lea	esi,[esi+ecx-1]
	lea	edi,[edi+ecx-1]
	std
	rep	movs byte [edi],[esi]
	cld
      replace_label:
	mov	ecx,[edx+12]
	mov	edi,[esp+4]
	sub	edi,ecx
	mov	esi,[edx+8]
	rep	movs byte [edi],[esi]
	pop	edi esi
	inc	esi
	jmp	preprocess_instruction
      not_preprocessor_symbol:
	mov	esi,[current_offset]
	call	process_equ_constants
      line_preprocessed:
	pop	esi ecx
	ret

get_preprocessor_symbol:
	push	ebp edi esi
	mov	ebp,ecx
	shl	ebp,22
	movzx	ecx,cl
	mov	ebx,hash_tree
	mov	edi,10
      follow_hashes_roots: