From 8f0a6ed542a0ffc807eb69b14a578bb4f640432a Mon Sep 17 00:00:00 2001
From: Russell-S-Harper <russell.s.harper@gmail.com>
Date: Fri, 24 Aug 2018 23:40:47 -0400
Subject: [PATCH] Revised MOD to prevent overflow.

---
 common/common.asm | 67 ++++++++++++++++++++++++++++++++++++-----------
 common/common.h   |  4 +--
 common/page6.src  |  8 +++---
 3 files changed, 57 insertions(+), 22 deletions(-)

diff --git a/common/common.asm b/common/common.asm
index ff3e1c6..512aef4 100644
--- a/common/common.asm
+++ b/common/common.asm
@@ -261,11 +261,11 @@ _3	ORA _F
 	RTS
 .)
 
-_DEC	.(		; DEC r			8r		Rr <- dec(Rr)	- convert Rr from hex aabbcc.dd to decimal ######.##
+_DEC	.(		; DEC r			8r		Rr <- dec(Rr)	- convert Rr from hex aabbccdd to decimal ######.###
 	RTS
 .)
 
-_HEX	.(		; HEX r			9r		Rr <- hex(Rr)	- convert Rr from decimal ######.## to hex aabbcc.dd
+_HEX	.(		; HEX r			9r		Rr <- hex(Rr)	- convert Rr from decimal ######.### to hex aabbccdd
 	RTS
 .)
 
@@ -604,7 +604,7 @@ _2	JSR _CPXI0	; copy p to I0
 	STA _I3+2
 	LDA _R0+3,Y
 	STA _I3+3
-	; I4 / I5 will form 64-bit result
+	; I4 / I5 will form 64-bit result for DIV
 	LDX #_I4-_R0
 	JSR _ZERQX
 	LDA _I0+3	; get sign of result
@@ -613,37 +613,37 @@ _2	JSR _CPXI0	; copy p to I0
 	RTS
 .)
 
-_CMPDM	.(		; compare I0/I1 to I2/I3, return result in status
+_CPQXDM	.(		; compare I0/I1 to quadword pointed by X, return result in status
 	LDA _I1+3
-	CMP _I3+3
+	CMP _R1+3,X
 	BCC _1		; definitely less
 	BNE _1		; definitely greater
 	LDA _I1+2
-	CMP _I3+2
+	CMP _R1+2,X
 	BCC _1		; definitely less
 	BNE _1		; definitely greater
 	LDA _I1+1
-	CMP _I3+1
+	CMP _R1+1,X
 	BCC _1		; definitely less
 	BNE _1		; definitely greater
 	LDA _I1
-	CMP _I3
+	CMP _R1,X
 	BCC _1		; definitely less
 	BNE _1		; definitely greater
 	LDA _I0+3
-	CMP _I2+3
+	CMP _R0+3,X
 	BCC _1		; definitely less
 	BNE _1		; definitely greater
 	LDA _I0+2
-	CMP _I2+2
+	CMP _R0+2,X
 	BCC _1		; definitely less
 	BNE _1		; definitely greater
 	LDA _I0+1
-	CMP _I2+1
+	CMP _R0+1,X
 	BCC _1		; definitely less
 	BNE _1		; definitely greater
 	LDA _I0
-	CMP _I2
+	CMP _R0,X
 _1	RTS
 .)
 
@@ -711,7 +711,8 @@ _DIV	.(		; DIV r pq		dr pq		Rr <- Rp / Rq	- division
 	LDX #_I3-_R0	; absolute value of register q saved in I3
 	JSR _ABSX
 	LDY #51		; 51 bits are enough, and ensure alignment
-_1	JSR _CMPDM	; is I0/I1 < I2/I3
+_1	LDX #_I2-_R0	; is I0/I1 < I2/I3
+	JSR _CPQXDM
 	BCC _2		; yes, skip subtraction
 	BEQ _4		; special case when p = q
 	JSR _UPDDM	; I0/I1 -= I2/I3
@@ -733,10 +734,44 @@ _5	LDX #_I4-_R0
 _6	JMP _RETI0X	; pull X, transfer I0 to r register, let it handle the return
 .)
 
-_MOD	.(
-	RTS
+_MOD	.(		; MOD r pq		er pq		Rr <- Rp % Rq	- modulus
+	TXA
+	PHA		; save r register for later
+	JSR _INTDM	; initialize
+	PHA		; save sign of result
+	LDX #_I0-_R0	; absolute value of register p saved in I0
+	JSR _ABSX
+	LDX #_I3-_R0	; absolute value of register q saved in I3
+	JSR _ABSX
+	LDA _I3		; copy |q| in I3 to I4
+	STA _I4
+	LDA _I3+1
+	STA _I4+1
+	LDA _I3+2
+	STA _I4+2
+	LDA _I3+3
+	STA _I4+3
+	LDY #34		; 34 shifts are enough
+_1	LDX #_I4-_R0	; is I0/I1 < I4/I5?
+	JSR _CPQXDM
+	BCC _3		; yes, all done
+	LDX #_I2-_R0	; is I0/I1 < I2/I3
+	JSR _CPQXDM
+	BCC _2		; yes, skip subtraction
+	JSR _UPDDM	; I0/I1 -= I2/I3
+_2	JSR _SHDDM	; I2/I3 /= 2
+	DEY
+	BNE _1
+_3	LDA _F		; clear underflow
+	AND #_F_U^$FF
+	STA _F
+	PLA		; set the sign of the product
+	BEQ _4
+	LDX #_I0-_R0	; negate I0
+	JSR _NEGX
+_4	JMP _RETI0X	; pull X, transfer I0 to r register, let it handle the return
 .)
-	
+
 _ESC	.(		; ESC			00				- escape back into regular assembler
 	PLA		; discard the COMMON _1 return address
 	PLA
diff --git a/common/common.h b/common/common.h
index 893b745..d95cf36 100644
--- a/common/common.h
+++ b/common/common.h
@@ -20,8 +20,8 @@
 ; INR r			5r		Rr <- Rr + 1.0	- increment register
 ; DCR r			6r		Rr <- Rr - 1.0	- decrement register
 ; TST r			7r		F <- Rr <=> 0.0	- test register
-; DEC r			8r		Rr <- dec(Rr)	- convert Rr from hex aabbccdd to decimal #########
-; HEX r			9r		Rr <- hex(Rr)	- convert Rr from decimal ######### to hex aabbccdd
+; DEC r			8r		Rr <- dec(Rr)	- convert Rr from hex aabbccdd to decimal ######.###
+; HEX r			9r		Rr <- hex(Rr)	- convert Rr from decimal ######.### to hex aabbccdd
 ; ADD r pq		ar pq		Rr <- Rp + Rq	- addition
 ; SUB r pq		br pq		Rr <- Rp - Rq	- subtraction
 ; MUL r pq		cr pq		Rr <- Rp * Rq	- multiplication
diff --git a/common/page6.src b/common/page6.src
index f5f71a8..c6235af 100644
--- a/common/page6.src
+++ b/common/page6.src
@@ -5,10 +5,10 @@
 
 HDR(DEMO)
 	CMN
-	SET(R0, 1048575.9995)
-	SET(R1, 3)
-	DIV(R2, R0, R1)
-	MUL(R3, R2, R1)
+	SET(R0, 1048575)
+	SET(R1, 3 * 5 * 5)
+	MOD(R2, R0, R1)
+	;SUB(R3, R0, R2)
 	ESC
 	BRK