From b7e010e5ac191b1857dc7ea3ba13688c4b341b41 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Juhani=20Krekel=C3=A4?= <juhani@krekelä.fi>
Date: Thu, 24 Feb 2022 19:14:25 +0200
Subject: [PATCH] Check target words list when checking the guess word

---
 Makefile            |  10 +++--
 compress-dict.py    |  16 ++++++-
 compress-targets.py |   8 ++--
 dosdl.asm           | 105 +++++++++++++++++++++++++++++++-------------
 4 files changed, 100 insertions(+), 39 deletions(-)

diff --git a/Makefile b/Makefile
index 4e94363..e018709 100644
--- a/Makefile
+++ b/Makefile
@@ -1,16 +1,20 @@
 NASM ?= nasm
 PYTHON ?= python
 
+# 'murky' is the rarest target word we'll use
+# https://github.com/lynn/hello-wordl/blob/7da40c1f067eb1ec157d4c5b7a9bd8257ed39342/src/Game.tsx#L34
+RAREST_WORD = murky
+
 all: dosdl.com
 
 dosdl.com: dosdl.asm dictionary.inc targets.inc
 	$(NASM) -fbin -o $@ $<
 
-dictionary.inc: dictionary.json compress-dict.py
-	$(PYTHON) compress-dict.py $< $@
+dictionary.inc: dictionary.json targets.json compress-dict.py
+	$(PYTHON) compress-dict.py $< targets.json $(RAREST_WORD) $@
 
 targets.inc: targets.json compress-targets.py
-	$(PYTHON) compress-targets.py $< $@
+	$(PYTHON) compress-targets.py $< $(RAREST_WORD) $@
 
 clean:
 	rm -f *.inc *.com
diff --git a/compress-dict.py b/compress-dict.py
index 1ff9a68..b7e1dbd 100644
--- a/compress-dict.py
+++ b/compress-dict.py
@@ -5,13 +5,27 @@ import sys
 alphabet = 'abcdefghijklmnopqrstuvwxyz'
 
 srcpath = sys.argv[1]
-targetpath = sys.argv[2]
+excludepath = sys.argv[2]
+rarestword = sys.argv[3]
+targetpath = sys.argv[4]
 
 with open(srcpath, 'r') as f:
 	words = json.load(f)
 
+with open(excludepath, 'r') as f:
+	exclude_all = json.load(f)
+
 # We only care about 5-letter words
 words = [word for word in words if len(word) == 5]
+exclude_all = [word for word in exclude_all if len(word) == 5]
+
+exclude = set()
+for word in exclude_all:
+	exclude.add(word)
+	if word == rarestword: break
+
+# Don't include words in the exclude list
+words = [word for word in words if word not in exclude]
 
 # Split dictionary into per-startletter arrays
 arrays = {letter: [] for letter in alphabet}
diff --git a/compress-targets.py b/compress-targets.py
index bc17562..d9f8a45 100644
--- a/compress-targets.py
+++ b/compress-targets.py
@@ -5,17 +5,16 @@ import sys
 alphabet = 'abcdefghijklmnopqrstuvwxyz*'
 
 srcpath = sys.argv[1]
-targetpath = sys.argv[2]
+rarestword = sys.argv[2]
+targetpath = sys.argv[3]
 
 with open(srcpath, 'r') as f:
 	all_words = json.load(f)
 
-# 'murky' is the rarest word we'll use
-# https://github.com/lynn/hello-wordl/blob/7da40c1f067eb1ec157d4c5b7a9bd8257ed39342/src/Game.tsx#L34
 words = []
 for word in all_words:
 	words.append(word)
-	if word == 'murky': break
+	if word == rarestword: break
 
 # We only care about 5-letter words
 words = [word for word in words if len(word) == 5]
@@ -29,6 +28,7 @@ for word in words:
 	array.append(packed)
 
 with open(targetpath, 'w') as f:
+	f.write(f'targets_len equ {len(array)}\n')
 	f.write('targets:\n')
 	for packed in array:
 		f.write(f'\tdb {", ".join(str(byte) for byte in packed)}\n')
diff --git a/dosdl.asm b/dosdl.asm
index b197e40..af58eb4 100644
--- a/dosdl.asm
+++ b/dosdl.asm
@@ -7,42 +7,16 @@ section .code
 
 select_target:
 	; TODO: Implement target word selection
-	mov si, 1
+	; https://github.com/lynn/hello-wordl/blob/7da40c1f067eb1ec157d4c5b7a9bd8257ed39342/src/util.ts#L14
+	mov si, 1397
 
-extract_target:
 	shl si, 1
 	shl si, 1
 	add si, targets
 
 	mov di, target
 
-	; Target word is stored packed 5 bits per letter into 32 bits
-	; 2221_1111 4333_3322 5555_4444 0000_0005
-	; Load 4333_3322_2221_1111 into dx and extracts first three letters
-	mov dx, [si]
-	add si, 2
-	mov cl, 5
-%rep 3
-	mov al, dl
-	and al, 0x1f
-	add al, 'a'
-	stosb
-	shr dx, cl
-%endrep
-
-	; Load 0000_0005_5555_4444 into ax, shift over by one, combine into dx
-	; which has 0000_0000_0000_0004, and extract last two letters
-	mov ax, [si]
-	add si, 2
-	shl ax, 1
-	or dx, ax
-%rep 2
-	mov al, dl
-	and al, 0x1f
-	add al, 'a'
-	stosb
-	shr dx, cl
-%endrep
+	call unpack_target
 
 read_guess:
 	; Number of thus far guessed letters stored in bx
@@ -170,7 +144,30 @@ check_dictionary:
 		loop .dictionary_loop
 
 check_targets:
-	; TODO: Check whether word is in the list of target words
+	mov cx, targets_len
+	mov si, targets
+
+	.targets_loop:
+		mov di, dictionary_entry
+		call unpack_target
+
+		xor bp, bp
+		.compare_loop:
+			cmp bp, 5
+			je .match
+
+			mov al, [di + bp]
+			cmp al, [guess + bp]
+			jne .not_match
+
+			inc bp
+			jmp .compare_loop
+
+		.match:
+		jmp word_found
+
+		.not_match:
+		loop .targets_loop
 
 word_not_found:
 	mov ah, 9
@@ -385,6 +382,52 @@ newline:
 	pop ax
 	ret
 
+; in:
+;  si = packed target word
+;  di = 5 byte buffer for unpacking the target
+; out:
+;  si = first byte after the packed target word
+unpack_target:
+	push ax
+	push cx
+	push dx
+	push di
+
+	; Target word is stored packed 5 bits per letter into 32 bits
+	; 2221_1111 4333_3322 5555_4444 0000_0005
+	; Load 4333_3322_2221_1111 into dx and extracts first three letters
+	mov dx, [si]
+	add si, 2
+	mov cl, 5
+%rep 3
+	mov al, dl
+	and al, 0x1f
+	add al, 'a'
+	stosb
+	shr dx, cl
+%endrep
+
+	; Load 0000_0005_5555_4444 into ax, shift over by one, combine into dx
+	; which has 0000_0000_0000_0004, and extract last two letters
+	mov ax, [si]
+	add si, 2
+	shl ax, 1
+	or dx, ax
+%rep 2
+	mov al, dl
+	and al, 0x1f
+	add al, 'a'
+	stosb
+	shr dx, cl
+%endrep
+
+	pop di
+	pop dx
+	pop cx
+	pop ax
+
+	ret
+
 section .data
 target times 5 db 0
 guess times 5 db 0
@@ -394,7 +437,7 @@ wrong_places times 5 db 0
 
 guesses db 0
 
-dictionary_entry times 4 db 0
+dictionary_entry times 5 db 0
 
 section .rodata