@@ -108,7 +108,7 @@ BRK

 		OVR #10 SWP - .Audio0/output DEI #04 SFT < .Screen/color DEO

 		.Screen/y DEI2 #0002 ++ .Screen/y DEO2

 		( incr ) SWP #01 + SWP

-		DUP2 < ,&loop JCN

+		LTHk ,&loop JCN

 	POP2

 BRK

@@ -310,7 +310,7 @@ RTN

 	DUP2 #0018 ++

 	&loop

 		( move ) OVR2 .Screen/y DEO2

-		( draw ) DUPr STHr .Screen/color DEO

+		( draw ) STHrk .Screen/color DEO

 		( incr ) .Screen/addr DEI2 #0008 ++ .Screen/addr DEO2  

 		( incr ) SWP2 #0008 ++ SWP2

 		OVR2 OVR2 << ,&loop JCN

@@ -371,7 +371,7 @@ RTN

 		.Screen/x DEI2 #0001 ++ .Screen/x DEO2

 		( draw ) OVR .Audio0/length DEI2 SWP POP > #02 * #01 + .Screen/color DEO

 		( incr ) SWP #01 + SWP

-		DUP2 < ,&loop JCN

+		LTHk ,&loop JCN

 	POP2

 	( range )

@@ -390,7 +390,7 @@ RTN

 	.wave-view/y1 LDZ2 DUP2 #0020 ++

 	&loop

 		OVR2 .Screen/y DEO2

-		( draw ) DUPr STHr .Screen/color DEO

+		( draw ) STHrk .Screen/color DEO

 		( incr ) SWP2 #0001 ++ SWP2

 		OVR2 OVR2 << ,&loop JCN

 	POP2 POP2

@@ -414,9 +414,9 @@ RTN

 	( draw ) #21 .Screen/color DEO

 	.Screen/x DEI2 #0004 ++ .Screen/x DEO2

 	.Screen/y DEI2 #0008 ++ .Screen/y DEO2

-	;font-hex #00 DUPr STHr #08 * ++ .Screen/addr DEO2

+	;font-hex #00 STHrk #08 * ++ .Screen/addr DEO2

 	( draw ) #21 .Screen/color DEO

-	.Screen/x DEI2 #0004 -- #00 #00 DUPr STHr ;knob-offsetx ++ LDA ++ .Screen/x DEO2

+	.Screen/x DEI2 #0004 -- #00 #00 STHrk ;knob-offsetx ++ LDA ++ .Screen/x DEO2

 	.Screen/y DEI2 #0010 -- #00 #00 STHr ;knob-offsety ++ LDA ++ .Screen/y DEO2

 	;knob-icns #0020 ++ .Screen/addr DEO2

 	( draw ) #25 .Screen/color DEO

@@ -427,13 +427,13 @@ RTN

 	STH SWP 

 	DUP #04 SFT TOS #0008 ** ;font-hex ++ .Screen/addr DEO2

-	( draw ) DUPr STHr .Screen/color DEO

+	( draw ) STHrk .Screen/color DEO

 	#0f AND TOS #0008 ** ;font-hex ++ .Screen/addr DEO2

 	.Screen/x DEI2 #0008 ++ .Screen/x DEO2

-	( draw ) DUPr STHr .Screen/color DEO

+	( draw ) STHrk .Screen/color DEO

 	DUP #04 SFT TOS #0008 ** ;font-hex ++ .Screen/addr DEO2

 	.Screen/x DEI2 #0008 ++ .Screen/x DEO2

-	( draw ) DUPr STHr .Screen/color DEO

+	( draw ) STHrk .Screen/color DEO

 	#0f AND TOS #0008 ** ;font-hex ++ .Screen/addr DEO2

 	.Screen/x DEI2 #0008 ++ .Screen/x DEO2

 	( draw ) STHr .Screen/color DEO

@@ -119,6 +119,8 @@ findopcode(char *s)

 				i |= (1 << 5); /* mode: short */

 			else if(s[3 + m] == 'r')

 				i |= (1 << 6); /* mode: return */

+			else if(s[3 + m] == 'k')

+				i |= (1 << 7); /* mode: keep */

 			else

 				return 0; /* failed to match */

 			m++;

@@ -17,15 +17,15 @@ WITH REGARD TO THIS SOFTWARE.

 /* clang-format off */

 void   push8(Stack *s, Uint8 a) { if (s->ptr == 0xff) { s->error = 2; return; } s->dat[s->ptr++] = a; }

-Uint8  pop8(Stack *s) { if (s->ptr == 0) { s->error = 1; return 0; } return s->dat[--s->ptr]; }

-Uint8  peek8(Stack *s, Uint8 a) { if (s->ptr < a + 1) s->error = 1; return s->dat[s->ptr - a - 1]; }

+Uint8  pop8_keep(Stack *s, Uint8 a) { if (s->ptr < a + 1) s->error = 1; return s->dat[s->ptr - a - 1]; }

+Uint8  pop8_nokeep(Stack *s, Uint8 a) { if (s->ptr == 0) { s->error = 1; return 0; } return s->dat[--s->ptr]; (void) a; }

+static Uint8 (*pop8)(Stack *s, Uint8 a);

 void   mempoke8(Uint8 *m, Uint16 a, Uint8 b) { m[a] = b; }

 Uint8  mempeek8(Uint8 *m, Uint16 a) { return m[a]; }

 void   devpoke8(Device *d, Uint8 a, Uint8 b) { d->dat[a & 0xf] = b; d->talk(d, a & 0x0f, 1); }

 Uint8  devpeek8(Device *d, Uint8 a) { d->talk(d, a & 0x0f, 0); return d->dat[a & 0xf];  }

 void   push16(Stack *s, Uint16 a) { push8(s, a >> 8); push8(s, a); }

-Uint16 pop16(Stack *s) { return pop8(s) + (pop8(s) << 8); }

-Uint16 peek16(Stack *s, Uint8 a) { return peek8(s, a * 2) + (peek8(s, a * 2 + 1) << 8); }

+Uint16 pop16(Stack *s, Uint8 a) { return pop8(s, a * 2) + (pop8(s, a * 2 + 1) << 8); }

 void   mempoke16(Uint8 *m, Uint16 a, Uint16 b) { mempoke8(m, a, b >> 8); mempoke8(m, a + 1, b); }

 Uint16 mempeek16(Uint8 *m, Uint16 a) { return (mempeek8(m, a) << 8) + mempeek8(m, a + 1); }

 void   devpoke16(Device *d, Uint8 a, Uint16 b) { devpoke8(d, a, b >> 8); devpoke8(d, a + 1, b); }

@@ -34,72 +34,72 @@ Uint16 devpeek16(Device *d, Uint16 a) { return (devpeek8(d, a) << 8) + devpeek8(

 void op_brk(Uxn *u) { u->ram.ptr = 0; }

 void op_nop(Uxn *u) { (void)u; }

 void op_lit(Uxn *u) { push8(u->src, mempeek8(u->ram.dat, u->ram.ptr++)); }

-void op_pop(Uxn *u) { pop8(u->src); }

-void op_dup(Uxn *u) { push8(u->src, peek8(u->src, 0)); }

-void op_swp(Uxn *u) { Uint8 b = pop8(u->src), a = pop8(u->src); push8(u->src, b); push8(u->src, a); }

-void op_ovr(Uxn *u) { push8(u->src, peek8(u->src, 1)); }

-void op_rot(Uxn *u) { Uint8 c = pop8(u->src), b = pop8(u->src), a = pop8(u->src); push8(u->src, b); push8(u->src, c); push8(u->src, a); }

+void op_pop(Uxn *u) { pop8(u->src, 0); }

+void op_dup(Uxn *u) { Uint8 a = pop8(u->src, 0); push8(u->src, a); push8(u->src, a); }

+void op_swp(Uxn *u) { Uint8 b = pop8(u->src, 0), a = pop8(u->src, 1); push8(u->src, b); push8(u->src, a); }

+void op_ovr(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b); push8(u->src, a); push8(u->src, b); }

+void op_rot(Uxn *u) { Uint8 c = pop8(u->src, 0), b = pop8(u->src, 1), a = pop8(u->src, 2); push8(u->src, b); push8(u->src, c); push8(u->src, a); }

 /* Logic */

-void op_equ(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); push8(u->src, b == a); }

-void op_neq(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); push8(u->src, b != a); }

-void op_gth(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); push8(u->src, b > a); }

-void op_lth(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); push8(u->src, b < a); }

-void op_jmp(Uxn *u) { Uint8 a = pop8(u->src); u->ram.ptr += (Sint8)a; }

-void op_jnz(Uxn *u) { Uint8 a = pop8(u->src); if (pop8(u->src)) u->ram.ptr += (Sint8)a; }

-void op_jsr(Uxn *u) { Uint8 a = pop8(u->src); push16(u->dst, u->ram.ptr); u->ram.ptr += (Sint8)a; }

-void op_sth(Uxn *u) { Uint8 a = pop8(u->src); push8(u->dst, a); }

+void op_equ(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b == a); }

+void op_neq(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b != a); }

+void op_gth(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b > a); }

+void op_lth(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b < a); }

+void op_jmp(Uxn *u) { Uint8 a = pop8(u->src, 0); u->ram.ptr += (Sint8)a; }

+void op_jnz(Uxn *u) { Uint8 a = pop8(u->src, 0); if (pop8(u->src, 1)) u->ram.ptr += (Sint8)a; }

+void op_jsr(Uxn *u) { Uint8 a = pop8(u->src, 0); push16(u->dst, u->ram.ptr); u->ram.ptr += (Sint8)a; }

+void op_sth(Uxn *u) { Uint8 a = pop8(u->src, 0); push8(u->dst, a); }

 /* Memory */

-void op_pek(Uxn *u) { Uint8 a = pop8(u->src); push8(u->src, mempeek8(u->ram.dat, a)); }

-void op_pok(Uxn *u) { Uint8 a = pop8(u->src); Uint8 b = pop8(u->src); mempoke8(u->ram.dat, a, b); }

-void op_ldr(Uxn *u) { Uint8 a = pop8(u->src); push8(u->src, mempeek8(u->ram.dat, u->ram.ptr + (Sint8)a)); }

-void op_str(Uxn *u) { Uint8 a = pop8(u->src); Uint8 b = pop8(u->src); mempoke8(u->ram.dat, u->ram.ptr + (Sint8)a, b); }

-void op_lda(Uxn *u) { Uint16 a = pop16(u->src); push8(u->src, mempeek8(u->ram.dat, a)); }

-void op_sta(Uxn *u) { Uint16 a = pop16(u->src); Uint8 b = pop8(u->src); mempoke8(u->ram.dat, a, b); }

-void op_dei(Uxn *u) { Uint8 a = pop8(u->src); push8(u->src, devpeek8(&u->dev[a >> 4], a)); }

-void op_deo(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); devpoke8(&u->dev[a >> 4], a, b); }

+void op_pek(Uxn *u) { Uint8 a = pop8(u->src, 0); push8(u->src, mempeek8(u->ram.dat, a)); }

+void op_pok(Uxn *u) { Uint8 a = pop8(u->src, 0); Uint8 b = pop8(u->src, 1); mempoke8(u->ram.dat, a, b); }

+void op_ldr(Uxn *u) { Uint8 a = pop8(u->src, 0); push8(u->src, mempeek8(u->ram.dat, u->ram.ptr + (Sint8)a)); }

+void op_str(Uxn *u) { Uint8 a = pop8(u->src, 0); Uint8 b = pop8(u->src, 1); mempoke8(u->ram.dat, u->ram.ptr + (Sint8)a, b); }

+void op_lda(Uxn *u) { Uint16 a = pop16(u->src, 0); push8(u->src, mempeek8(u->ram.dat, a)); }

+void op_sta(Uxn *u) { Uint16 a = pop16(u->src, 0); Uint8 b = pop8(u->src, 0); mempoke8(u->ram.dat, a, b); }

+void op_dei(Uxn *u) { Uint8 a = pop8(u->src, 0); push8(u->src, devpeek8(&u->dev[a >> 4], a)); }

+void op_deo(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); devpoke8(&u->dev[a >> 4], a, b); }

 /* Arithmetic */

-void op_add(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); push8(u->src, b + a); }

-void op_sub(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); push8(u->src, b - a); }

-void op_mul(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); push8(u->src, b * a); }

-void op_div(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); push8(u->src, b / a); }

-void op_and(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); push8(u->src, b & a); }

-void op_ora(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); push8(u->src, b | a); }

-void op_eor(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); push8(u->src, b ^ a); }

-void op_sft(Uxn *u) { Uint8 a = pop8(u->src), b = pop8(u->src); push8(u->src, b >> (a & 0x07) << ((a & 0x70) >> 4)); }

+void op_add(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b + a); }

+void op_sub(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b - a); }

+void op_mul(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b * a); }

+void op_div(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b / a); }

+void op_and(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b & a); }

+void op_ora(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b | a); }

+void op_eor(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b ^ a); }

+void op_sft(Uxn *u) { Uint8 a = pop8(u->src, 0), b = pop8(u->src, 1); push8(u->src, b >> (a & 0x07) << ((a & 0x70) >> 4)); }

 /* Stack */

 void op_lit16(Uxn *u) { push16(u->src, mempeek16(u->ram.dat, u->ram.ptr++)); u->ram.ptr++; }

-void op_pop16(Uxn *u) { pop16(u->src); }

-void op_dup16(Uxn *u) { push16(u->src, peek16(u->src, 0)); }

-void op_swp16(Uxn *u) { Uint16 b = pop16(u->src), a = pop16(u->src); push16(u->src, b); push16(u->src, a); }

-void op_ovr16(Uxn *u) { push16(u->src, peek16(u->src, 1)); }

-void op_rot16(Uxn *u) { Uint16 c = pop16(u->src), b = pop16(u->src), a = pop16(u->src); push16(u->src, b); push16(u->src, c); push16(u->src, a); }

+void op_pop16(Uxn *u) { pop16(u->src, 0); }

+void op_dup16(Uxn *u) { Uint16 a = pop16(u->src, 0); push16(u->src, a); push16(u->src, a); }

+void op_swp16(Uxn *u) { Uint16 b = pop16(u->src, 0), a = pop16(u->src, 1); push16(u->src, b); push16(u->src, a); }

+void op_ovr16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push16(u->src, b); push16(u->src, a); push16(u->src, b); }

+void op_rot16(Uxn *u) { Uint16 c = pop16(u->src, 0), b = pop16(u->src, 1), a = pop16(u->src, 2); push16(u->src, b); push16(u->src, c); push16(u->src, a); }

 /* Logic(16-bits) */

-void op_equ16(Uxn *u) { Uint16 a = pop16(u->src), b = pop16(u->src); push8(u->src, b == a); }

-void op_neq16(Uxn *u) { Uint16 a = pop16(u->src), b = pop16(u->src); push8(u->src, b != a); }

-void op_gth16(Uxn *u) { Uint16 a = pop16(u->src), b = pop16(u->src); push8(u->src, b > a); }

-void op_lth16(Uxn *u) { Uint16 a = pop16(u->src), b = pop16(u->src); push8(u->src, b < a); }

-void op_jmp16(Uxn *u) { u->ram.ptr = pop16(u->src); }

-void op_jnz16(Uxn *u) { Uint16 a = pop16(u->src); if (pop8(u->src)) u->ram.ptr = a; }

-void op_jsr16(Uxn *u) { push16(u->dst, u->ram.ptr); u->ram.ptr = pop16(u->src); }

-void op_sth16(Uxn *u) { Uint16 a = pop16(u->src); push16(u->dst, a); }

+void op_equ16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push8(u->src, b == a); }

+void op_neq16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push8(u->src, b != a); }

+void op_gth16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push8(u->src, b > a); }

+void op_lth16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push8(u->src, b < a); }

+void op_jmp16(Uxn *u) { u->ram.ptr = pop16(u->src, 0); }

+void op_jnz16(Uxn *u) { Uint16 a = pop16(u->src, 0); if (pop8(u->src, 0)) u->ram.ptr = a; }

+void op_jsr16(Uxn *u) { push16(u->dst, u->ram.ptr); u->ram.ptr = pop16(u->src, 0); }

+void op_sth16(Uxn *u) { Uint16 a = pop16(u->src, 0); push16(u->dst, a); }

 /* Memory(16-bits) */

-void op_pek16(Uxn *u) { Uint8 a = pop8(u->src); push16(u->src, mempeek16(u->ram.dat, a)); }

-void op_pok16(Uxn *u) { Uint8 a = pop8(u->src); Uint16 b = pop16(u->src); mempoke16(u->ram.dat, a, b); }

-void op_ldr16(Uxn *u) { Uint8 a = pop8(u->src); push16(u->src, mempeek16(u->ram.dat, u->ram.ptr + (Sint8)a)); }

-void op_str16(Uxn *u) { Uint8 a = pop8(u->src); Uint16 b = pop16(u->src); mempoke16(u->ram.dat, u->ram.ptr + (Sint8)a, b); }

-void op_lda16(Uxn *u) { Uint16 a = pop16(u->src); push16(u->src, mempeek16(u->ram.dat, a)); }

-void op_sta16(Uxn *u) { Uint16 a = pop16(u->src); Uint16 b = pop16(u->src); mempoke16(u->ram.dat, a, b); }

-void op_dei16(Uxn *u) { Uint8 a = pop8(u->src); push16(u->src, devpeek16(&u->dev[a >> 4], a)); }

-void op_deo16(Uxn *u) { Uint8 a = pop8(u->src); Uint16 b = pop16(u->src); devpoke16(&u->dev[a >> 4], a, b); }

+void op_pek16(Uxn *u) { Uint8 a = pop8(u->src, 0); push16(u->src, mempeek16(u->ram.dat, a)); }

+void op_pok16(Uxn *u) { Uint8 a = pop8(u->src, 0); Uint16 b = pop16(u->src, 0); mempoke16(u->ram.dat, a, b); }

+void op_ldr16(Uxn *u) { Uint8 a = pop8(u->src, 0); push16(u->src, mempeek16(u->ram.dat, u->ram.ptr + (Sint8)a)); }

+void op_str16(Uxn *u) { Uint8 a = pop8(u->src, 0); Uint16 b = pop16(u->src, 0); mempoke16(u->ram.dat, u->ram.ptr + (Sint8)a, b); }

+void op_lda16(Uxn *u) { Uint16 a = pop16(u->src, 0); push16(u->src, mempeek16(u->ram.dat, a)); }

+void op_sta16(Uxn *u) { Uint16 a = pop16(u->src, 0); Uint16 b = pop16(u->src, 1); mempoke16(u->ram.dat, a, b); }

+void op_dei16(Uxn *u) { Uint8 a = pop8(u->src, 0); push16(u->src, devpeek16(&u->dev[a >> 4], a)); }

+void op_deo16(Uxn *u) { Uint8 a = pop8(u->src, 0); Uint16 b = pop16(u->src, 0); devpoke16(&u->dev[a >> 4], a, b); }

 /* Arithmetic(16-bits) */

-void op_add16(Uxn *u) { Uint16 a = pop16(u->src), b = pop16(u->src); push16(u->src, b + a); }

-void op_sub16(Uxn *u) { Uint16 a = pop16(u->src), b = pop16(u->src); push16(u->src, b - a); }

-void op_mul16(Uxn *u) { Uint16 a = pop16(u->src), b = pop16(u->src); push16(u->src, b * a); }

-void op_div16(Uxn *u) { Uint16 a = pop16(u->src), b = pop16(u->src); push16(u->src, b / a); }

-void op_and16(Uxn *u) { Uint16 a = pop16(u->src), b = pop16(u->src); push16(u->src, b & a); }

-void op_ora16(Uxn *u) { Uint16 a = pop16(u->src), b = pop16(u->src); push16(u->src, b | a); }

-void op_eor16(Uxn *u) { Uint16 a = pop16(u->src), b = pop16(u->src); push16(u->src, b ^ a); }

-void op_sft16(Uxn *u) { Uint16 a = pop16(u->src), b = pop16(u->src); push16(u->src, b >> (a & 0x0007) << ((a & 0x0070) >> 4)); }

+void op_add16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push16(u->src, b + a); }

+void op_sub16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push16(u->src, b - a); }

+void op_mul16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push16(u->src, b * a); }

+void op_div16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push16(u->src, b / a); }

+void op_and16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push16(u->src, b & a); }

+void op_ora16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push16(u->src, b | a); }

+void op_eor16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push16(u->src, b ^ a); }

+void op_sft16(Uxn *u) { Uint16 a = pop16(u->src, 0), b = pop16(u->src, 1); push16(u->src, b >> (a & 0x0007) << ((a & 0x0070) >> 4)); }

 void (*ops[])(Uxn *u) = {

 	op_brk, op_lit, op_nop, op_pop, op_dup, op_swp, op_ovr, op_rot,

@@ -128,9 +128,10 @@ haltuxn(Uxn *u, char *name, int id)

 void

 opcuxn(Uxn *u, Uint8 instr)

 {

-	Uint8 op = instr & 0x3f, freturn = instr & 0x40;

+	Uint8 op = instr & 0x3f, freturn = instr & 0x40, fkeep = instr & 0x80;

 	u->src = freturn ? &u->rst : &u->wst;

 	u->dst = freturn ? &u->wst : &u->rst;

+	pop8 = fkeep ? pop8_keep : pop8_nokeep;

 	(*ops[op])(u);

 }