Parse charmap.asm for dump_text.py and dump_names.py

4 files changed, 104 insertions, 84 deletions

diff --git a/charmap.asm b/charmap.asm
index ad6de2f..ef56ab7 100644
--- a/charmap.asm
+++ b/charmap.asm
@@ -2,8 +2,10 @@
 
 ; Control characters (see home/text.asm)
 
+	charmap "<NULL>",    $00
+
 	charmap "<PLAY_G>",  $14 ; "<PLAYER>くん" or "<PLAYER>ちゃん"; same as "<PLAYER>" in English
-	charmap "<JP_18>",   $18 ; "ノ゛"? (ungrammatical)
+
 	charmap "<NI>",      $1d ; "に　"
 	charmap "<TTE>",     $1e ; "って"
 	charmap "<WO>",      $1f ; "を　"
@@ -44,15 +46,47 @@
 	charmap "<ROCKET>",  $5e ; "ROCKET"
 	charmap "<DEXEND>",  $5f
 
-; Actual characters (from other graphics files)
+; Other characters
+
+	charmap "■", $60
+	charmap "▲", $61
+	charmap "☎", $62
+
+	charmap "D", $63
+	charmap "E", $64
+	charmap "F", $65
+	charmap "G", $66
+	charmap "H", $67
+	charmap "I", $68
+	charmap "V", $69
+	charmap "S", $6a
+	charmap "L", $6b
+	charmap "M", $6c
+
+	charmap "：", $6d
+
+	charmap "「", $70
+	charmap "」", $71
+	charmap "『", $72
+	charmap "』", $73
+	charmap "・", $74
+	charmap "…", $75
 
-	; needed for _LoadFontsExtra1 (see engine/load_font.asm)
-	charmap "■",         $60 ; gfx/font/black.2bpp
-	charmap "▲",         $61 ; gfx/font/up_arrow.png
-	charmap "☎",         $62 ; gfx/font/phone_icon.2bpp
+	charmap "┌", $79
+	charmap "─", $7a
+	charmap "┐", $7b
+	charmap "│", $7c
+	charmap "└", $7d
+	charmap "┘", $7e
+	charmap "　", $7f
 
 ; Japanese kana
 
+	charmap "イ゛", $01
+	charmap "ヴ",  $02
+	charmap "エ゛", $03
+	charmap "オ゛", $04
+
 	charmap "ガ", $05
 	charmap "ギ", $06
 	charmap "グ", $07
@@ -69,11 +103,17 @@
 	charmap "デ", $12
 	charmap "ド", $13
 
+	charmap "ネ゛", $17
+	charmap "ノ゛", $18
+
 	charmap "バ", $19
 	charmap "ビ", $1a
 	charmap "ブ", $1b
 	charmap "ボ", $1c
 
+	charmap "ィ゛", $20
+	charmap "あ゛", $21
+
 	charmap "が", $26
 	charmap "ぎ", $27
 	charmap "ぐ", $28
@@ -106,25 +146,14 @@
 	charmap "ぺ", $47
 	charmap "ぽ", $48
 
-	charmap "「", $70
-	charmap "」", $71
-	charmap "『", $72
-	charmap "』", $73
-	charmap "·", $74
-	charmap "⋯", $75
+	charmap "も゜", $4d
 
+	charmap "ぃ", $6e
+	charmap "ぅ", $6f
 	charmap "ぁ", $76
 	charmap "ぇ", $77
 	charmap "ぉ", $78
 
-	charmap "┌", $79
-	charmap "─", $7a
-	charmap "┐", $7b
-	charmap "│", $7c
-	charmap "└", $7d
-	charmap "┘", $7e
-	charmap "　", $7f
-
 	charmap "ア", $80
 	charmap "イ", $81
 	charmap "ウ", $82
diff --git a/tools/dump_names.py b/tools/dump_names.py
index c7ef02d..a80677e 100644
--- a/tools/dump_names.py
+++ b/tools/dump_names.py
@@ -1,18 +1,7 @@
 #!/usr/bin/env python
 
 import sys, os, io
-
-def parse_int(s):
-	s = s.strip()
-	if s.startswith('$'):
-		return int(s[1:], 16)
-	if s.startswith('%'):
-		return int(s[1:], 2)
-	return int(s)
-
-def parse_string(s):
-	# assumes strings are literal, no STRCAT() etc
-	return s.strip('" ')
+from read_charmap import get_project_dir, read_charmap
 
 def calc_bank(p):
 	return p // 0x4000
@@ -26,39 +15,10 @@ def get_sym_loc(p):
 	b, a = calc_bank(p), calc_address(p)
 	return '%02x:%04x' % (b, a)
 
-def strip_comment(s):
-	# assumes ";" is not in the charmap
-	return s.split(';')[0].rstrip()
-
-def get_project_dir():
-	script_path = os.path.realpath(__file__)
-	script_dir = os.path.dirname(script_path)
-	project_dir = os.path.join(script_dir, '..')
-	return os.path.normpath(project_dir)
-
-def get_charmap_path():
-	project_dir = get_project_dir()
-	return os.path.join(project_dir, 'charmap.asm')
-
 def get_baserom_path():
 	project_dir = get_project_dir()
 	return os.path.join(project_dir, 'baserom.gb')
 
-def read_charmap():
-	charmap_path = get_charmap_path()
-	charmap = {}
-	with io.open(charmap_path, 'r', encoding='utf-8') as f:
-		lines = f.readlines()
-	for line in lines:
-		line = strip_comment(line).lstrip()
-		if not line.startswith('charmap '):
-			continue
-		char, value = line[len('charmap '):].rsplit(',', 1)
-		char = parse_string(char)
-		value = parse_int(value)
-		charmap[value] = char
-	return charmap
-
 def dump_strings(data):
 	charmap = read_charmap()
 	ss = []
diff --git a/tools/dump_text.py b/tools/dump_text.py
index 8f7c07d..3242c31 100644
--- a/tools/dump_text.py
+++ b/tools/dump_text.py
@@ -1,26 +1,12 @@
 #!/usr/bin/python3
 
 from sys import argv, stdout
+from read_charmap import read_charmap
 
 
-char_table = [
-	"?", "イ゛", "ヴ", "エ゛", "オ゛", "ガ", "ギ", "グ", "ゲ", "ゴ", "ザ", "ジ", "ズ", "ゼ", "ゾ", "ダ",
-	"ヂ", "ヅ", "デ", "ド", "<PLAY_G>", "<0x15>", "<0x16>", "ネ゛", "<JP_18>", "バ", "ビ", "ブ", "ボ", "<NI>", "<TTE>", "<WO>",
-	"ィ゛", "あ゛", "<TA!>", "<KOUGEKI>", "<WA>", "<NO>", "が", "ぎ", "ぐ", "げ", "ご", "ざ", "じ", "ず", "ぜ", "ぞ",
-	"だ", "ぢ", "づ", "で", "ど", "<ROUTE>", "<WATASHI>", "<KOKO_WA>", "<RED>", "<GREEN>", "ば", "び", "ぶ", "べ", "ぼ", "<ENEMY>",
-	"パ", "ピ", "プ", "ポ", "ぱ", "ぴ", "ぷ", "ぺ", "ぽ", "<MOM>", "<GA>", "<_CONT>", "<SCROLL>", "も゜", "<NEXT>", "<LINE>",
-	"@", "<PARA>", "<PLAYER>", "<RIVAL>", "#", "<CONT>", "<……>", "<DONE>", "<PROMPT>", "<TARGET>", "<USER>", "<PC>", "<TM>", "<TRAINER>", "<ROCKET>", "<DEXEND>",
-	"■", "▲", "☎", "D", "E", "F", "G", "H", "I", "V", "S", "L", "M", ":", "ぃ", "ぅ",
-	"「", "」", "『", "』", "・", "<・・・>", "ぁ", "ぇ", "ぉ", "┌", "─", "┐", "│", "└", "┘", " ",
-	"ア", "イ", "ウ", "エ", "オ", "カ", "キ", "ク", "ケ", "コ", "サ", "シ", "ス", "セ", "ソ", "タ",
-	"チ", "ツ", "テ", "ト", "ナ", "ニ", "ヌ", "ネ", "ノ", "ハ", "ヒ", "フ", "ホ", "マ", "ミ", "ム",
-	"メ", "モ", "ヤ", "ユ", "ヨ", "ラ", "ル", "レ", "ロ", "ワ", "ヲ", "ン", "ッ", "ャ", "ュ", "ョ",
-	"ィ", "あ", "い", "う", "え", "お", "か", "き", "く", "け", "こ", "さ", "し", "す", "せ", "そ",
-	"た", "ち", "つ", "て", "と", "な", "に", "ぬ", "ね", "の", "は", "ひ", "ふ", "へ", "ほ", "ま",
-	"み", "む", "め", "も", "や", "ゆ", "よ", "ら", "り", "る", "れ", "ろ", "わ", "を", "ん", "っ",
-	"ゃ", "ゅ", "ょ", "ー", "゜", "゛", "？", "！", "。", "ァ", "ゥ", "ェ", "▷", "▶", "▼", "♂",
-	"円", "×", ".", "/", "ォ", "♀", "０", "１", "２", "３", "４", "５", "６", "７", "８", "９"
-]
+bank_size = 0x4000
+
+charmap = read_charmap()
 
 if len(argv) != 4:
 	print(f"Usage: {argv[0]} path/to/ROM.gb start_offset end_offset\noffsets are in the form bank:address (both hex), and end_offset is *not* included.")
@@ -31,9 +17,9 @@ try:
 	start_bank,start_addr = [ int(s, 16) for s in argv[2].split(':') ]
 	end_bank,  end_addr   = [ int(s, 16) for s in argv[3].split(':') ]
 	if start_bank != 0:
-		start_addr += (start_bank - 1) * 0x4000
+		start_addr += (start_bank - 1) * bank_size
 	if end_bank != 0:
-		end_addr   += (end_bank   - 1) * 0x4000
+		end_addr   += (end_bank   - 1) * bank_size
 except Error:
 	print("Please specify valid offsets (bank:address, both hex)")
 	exit(1)
@@ -41,10 +27,12 @@ except Error:
 
 with open(argv[1], "rb") as f:
 	f.seek(start_addr)
-	
+
 	string = ""
 	while start_addr < end_addr:
-		string += char_table[ int.from_bytes(f.read(1), "little") ]
+		b = f.read(1)
+		v = int.from_bytes(b, "little")
+		string += char_table[ v ]
 		start_addr += 1
-	
+
 	stdout.buffer.write( f"db \"{string}\"\n".encode('utf-8') )
diff --git a/tools/read_charmap.py b/tools/read_charmap.py
new file mode 100644
index 0000000..9f396b4
--- /dev/null
+++ b/tools/read_charmap.py
@@ -0,0 +1,43 @@
+import os, io
+
+def parse_int(s):
+	# assumes integers are literal; no +-*/, etc
+	s = s.strip()
+	if s.startswith('$'):
+		return int(s[1:], 16)
+	if s.startswith('%'):
+		return int(s[1:], 2)
+	return int(s)
+
+def parse_string(s):
+	# assumes strings are literal; no STRCAT() etc
+	return s.strip('" ')
+
+def strip_comment(s):
+	# assumes ";" is not in the charmap
+	return s.split(';')[0].rstrip()
+
+def get_project_dir():
+	script_path = os.path.realpath(__file__)
+	script_dir = os.path.dirname(script_path)
+	project_dir = os.path.join(script_dir, '..')
+	return os.path.normpath(project_dir)
+
+def get_charmap_path():
+	project_dir = get_project_dir()
+	return os.path.join(project_dir, 'charmap.asm')
+
+def read_charmap():
+	charmap_path = get_charmap_path()
+	charmap = {}
+	with io.open(charmap_path, 'r', encoding='utf-8') as f:
+		lines = f.readlines()
+	for line in lines:
+		line = strip_comment(line).lstrip()
+		if not line.startswith('charmap '):
+			continue
+		char, value = line[len('charmap '):].rsplit(',', 1)
+		char = parse_string(char)
+		value = parse_int(value)
+		charmap[value] = char
+	return charmap