Alien (BUGFIX 1.7): Routine at 40040

The cat XOR blit for sub-cell offsets k = JonesPos&7 >= 4. Rather than right-shifting each sprite row by k (up to 7 passes), it left-shifts by 8-k (at most 4 passes) with SLA/RL chains and writes the spill byte FIRST — placing the sprite k pixels into the 4-byte output field, the same result as BlitJonesFrame's right-shift path. The number of chain groups to skip is patched into the self-modifying JR at BlitJonesShiftCode2, then 21 rows of 4 bytes are XOR-blitted.

The cat XOR blit for sub-cell offsets k = JonesPos&7 >= 4. Rather than right-shifting each sprite row by k (up to 7 passes), it left-shifts by 8-k (at most 4 passes) with SLA/RL chains and writes the spill byte FIRST — placing the sprite k pixels into the 4-byte output field, the same result as BlitJonesFrame's right-shift path. The number of chain groups to skip is patched into the self-modifying JR at BlitJonesShiftCode2, then 21 rows of 4 bytes are XOR-blitted.
40040	SUB 4	A = k-4: shift groups to SKIP (8-k of 4 run)
40042	LD L,A	patch offset = (k-4) * 7 bytes
40043	RLCA
40044	RLCA
40045	RLCA	A = shift_count * 8
40046	SUB L	A = shift_count * 7 (JR offset into SLA/RL chain)
40047	LD (40067),A	patch BlitJonesShiftCode2 JR offset
40050	LD B,21	21 pixel rows
40052	PUSH BC	save row counter
40053	EX DE,HL	HL = sprite data ptr
40054	LD B,(HL)	load the row's 3 sprite bytes
40055	INC HL	into B, C, E
40056	LD C,(HL)
40057	INC HL
40058	LD E,(HL)
40059	INC HL
40060	PUSH HL	save advanced sprite ptr
40061	LD HL,(39751)	HL = this row's screen address
40064	PUSH HL
40065	XOR A	A = 0: collects bits shifted out of B's top
40066	JR 40066	self-modifying: JR offset patched to skip N SLA/RL groups
40068	SLA E	left-shift the 24-bit row across A:B:C:E;
40070	RL C	shifting left by 8-k places the sprite k
40072	RL B	pixels from byte A's left edge — the same
40074	RLA	result as a right shift by k, but never
40075	SLA E	more than 4 shift-group passes
40077	RL C
40079	RL B
40081	RLA
40082	SLA E
40084	RL C
40086	RL B
40088	RLA
40089	SLA E
40091	RL C
40093	RL B
40095	RLA
40096	XOR (HL)	XOR-blit the 4 result bytes, spill byte first
40097	LD (HL),A
40098	INC HL
40099	LD A,B
40100	XOR (HL)
40101	LD (HL),A
40102	INC HL
40103	LD A,C
40104	XOR (HL)
40105	LD (HL),A
40106	INC HL
40107	LD A,E
40108	XOR (HL)
40109	LD (HL),A
40110	POP HL	back to the row-start screen address
40111	LD A,H	finished pixel row 7 of this char row?
40112	AND 7
40114	CP 7
40116	JR Z,40121	yes: char-row / third stepping
40118	INC H	no: next pixel row is one display bank up
40119	JR 40136
40121	LD A,L	last char row of this screen third?
40122	CP 224
40124	JR NC,40132	yes: hop to the next third
40126	LD BC,63776	no: -1760 = pixel-row 0 of the next char row
40129	ADD HL,BC
40130	JR 40136
40132	LD BC,32	+32 carries into H: row 0 of the next third
40135	ADD HL,BC
40136	LD (39751),HL	store the next row's screen address
40139	POP DE	DE = sprite ptr
40140	POP BC
40141	DJNZ 40052	next of the 21 rows
40143	POP BC	B = caller's iteration count
40144	DEC B	(mirror of BlitJonesFrame's ending)
40145	JP NZ,40007	B was 2: advance pos and redraw
40148	RET	B was 1: single XOR pass only