( math32.tal )
( )
( 32-bit integers are represented by two 16-bit integers )
( x** means xhi* xlo* )

%DEBUG { #ff #0e DEO }
%RTN  { JMP2r }
%EMIT { #18 DEO }
%DIGIT { #00 SWP ;digits ADD2 LDA EMIT }
%SPACE { #20 EMIT }
%NEWLINE { #0a EMIT }

%TOR2 { ROT2 ROT2 }
%POP4 { POP2 POP2 }

%X { #0000 #0001 }
%Y { #1234 #ffff }
%Z { #fedc #ba98 }

|0100
    ( run a bunch of test cases )

    #ffff #ffff #2a ;mul16 ;emit-short ;test16 JSR2
    #0001 #ffff #2a ;mul16 ;emit-short ;test16 JSR2
    #ffff #0001 #2a ;mul16 ;emit-short ;test16 JSR2
    #00ff #0001 #2a ;mul16 ;emit-short ;test16 JSR2
    NEWLINE

    #ffff #ffff #ffff #ffff #2a ;mul32 ;emit-long ;test32 JSR2
    #0000 #0001 #ffff #ffff #2a ;mul32 ;emit-long ;test32 JSR2
    #0000 #0002 #0000 #ffff #2a ;mul32 ;emit-long ;test32 JSR2
    #0001 #2345 #0034 #5678 #2a ;mul32 ;emit-long ;test32 JSR2
    NEWLINE

    X Y #2b ;add32 ;emit-long ;test32 JSR2
    Z Y #2b ;add32 ;emit-long ;test32 JSR2
    #fedc #0000 #1234 #0000 #2b ;add32 ;emit-long ;test32 JSR2
    #fedc #0000 #0000 #0000 #2b ;add32 ;emit-long ;test32 JSR2
    #0000 #0000 #0000 #0000 #2b ;add32 ;emit-long ;test32 JSR2
    NEWLINE

    #1234 #1234 #0000 #0001 #2d ;sub32 ;emit-long ;test32 JSR2
    #1234 #1234 #0000 #1234 #2d ;sub32 ;emit-long ;test32 JSR2
    #1234 #1234 #1234 #0001 #2d ;sub32 ;emit-long ;test32 JSR2
    #1234 #1234 #1000 #0000 #2d ;sub32 ;emit-long ;test32 JSR2
    #1234 #1234 #ffff #ffff #2d ;sub32 ;emit-long ;test32 JSR2
    NEWLINE

    X Y #26 ;and32 ;emit-long ;test32 JSR2
    X Y #7c ;or32 ;emit-long ;test32 JSR2
    X Y #5e ;xor32 ;emit-long ;test32 JSR2
    X Y #3d ;eq32 ;emit-byte ;test32 JSR2
    Y Y #3d ;eq32 ;emit-byte ;test32 JSR2
    X Y #7e ;ne32 ;emit-byte ;test32 JSR2
    X X #7e ;ne32 ;emit-byte ;test32 JSR2
BRK

@test16 ( x* y* symbol^ test-addr* emit-addr* -> )
    ,&emitaddr STR2
    ,&testaddr STR2
    ,&testsym STR
    ,&y STR2
    ,&x STR2
    ,&x LDR2 ;emit-short JSR2
    SPACE ,&testsym LDR EMIT SPACE
    ,&y LDR2 ;emit-short JSR2
    SPACE #3d EMIT SPACE
    ,&x LDR2 ,&y LDR2
    ,&testaddr LDR2 JSR2
    ,&emitaddr LDR2 JSR2 NEWLINE
RTN
&testsym $1
&testaddr $2
&emitaddr $2
&x $2 &y $2

@test32 ( x** y** symbol^ test-addr* emit-addr* -> )
    ,&emitaddr STR2
    ,&testaddr STR2
    ,&testsym STR
    ,&ylo STR2 ,&yhi STR2
    ,&xlo STR2 ,&xhi STR2
    ,&xhi LDR2 ,&xlo LDR2 ;emit-long JSR2
    SPACE ,&testsym LDR EMIT SPACE
    ,&yhi LDR2 ,&ylo LDR2 ;emit-long JSR2
    SPACE #3d EMIT SPACE
    ,&xhi LDR2 ,&xlo LDR2
    ,&yhi LDR2 ,&ylo LDR2
    ,&testaddr LDR2 JSR2
    ,&emitaddr LDR2 JSR2 NEWLINE
RTN
&testsym $1
&testaddr $2
&emitaddr $2
&xhi $2 &xlo $2
&yhi $2 &ylo $2

@eq32 ( xhi* xlo* yhi* ylo* -> bool^ )
    ROT2 EQU2 ,&maybe JCN
    POP4 #00 RTN
    &maybe EQU2
RTN

@ne32 ( xhi* xlo* yhi* ylo* -> bool^ )
    ROT2 EQU2 ,&maybe JCN
    POP4 #01 RTN
    &maybe NEQ2
RTN

@and32 ( xhi* xlo* yhi* ylo* -> xhi|yhi* xlo|ylo* )
    ROT2 AND2 TOR2 AND2 SWP2
RTN

@or32 ( xhi* xlo* yhi* ylo* -> xhi|yhi* xlo|ylo* )
    ROT2 ORA2 TOR2 ORA2 SWP2
RTN

@xor32 ( xhi* xlo* yhi* ylo* -> xhi|yhi* xlo|ylo* )
    ROT2 EOR2 TOR2 EOR2 SWP2
RTN

%COMPLEMENT32 { SWP2 #ffff EOR2 SWP2 #ffff EOR2 }

@complement32 ( x** -> ~x** )
    COMPLEMENT32
RTN

@negate32 ( x** -> -x** )
    COMPLEMENT32
    INC2 ( ~xhi -xlo )
    DUP2 #0000 NEQ2 ( ~xhi -xlo non-zero? )
    ,&done JCN ( xlo non-zero => don't inc hi )
    SWP2 INC2 SWP2 ( -xhi -xlo )
    &done
RTN

@left-by-16 ( xhi* xlo* -> xlo* 0000 )
    SWP2 POP2 #0000
RTN

@add32 ( xhi* xlo* yhi* ylo* -> zhi* zlo* )
    ,&y2 STR2 ,&y0 STR2 ( save ylo, yhi )
    ,&x2 STR2 ,&x0 STR2 ( save xlo, xhi )
    #0000 #0000 ,&z0 STR2 ,&z2 STR2 ( reset zhi, zlo )

    ( x3 + y3 => z2z3 )
    #00 ,&x3 LDR #00 ,&y3 LDR ADD2 ,&z2 STR2

    ( x2 + y2 + z2 => z1z2 )
    #00 ,&x2 LDR ,&z1 LDR2 ADD2 ,&z1 STR2
    #00 ,&y2 LDR ,&z1 LDR2 ADD2 ,&z1 STR2

    ( x1 + y1 + z1 => z0z1 )
    #00 ,&x1 LDR ,&z0 LDR2 ADD2 ,&z0 STR2
    #00 ,&y1 LDR ,&z0 LDR2 ADD2 ,&z0 STR2

    ( x0 + y0 + z0 => z0 )
    ,&x0 LDR ,&z0 LDR ADD ,&z0 STR
    ,&y0 LDR ,&z0 LDR ADD ,&z0 STR

    ( load zhi,zlo )
    ,&z0 LDR2 ,&z2 LDR2
RTN
( registers for add32 )
[ &x0 $1 &x1 $1 &x2 $1 &x3 $1 ]
[ &y0 $1 &y1 $1 &y2 $1 &y3 $1 ]
[ &z0 $1 &z1 $1 &z2 $2 ]

@sub32 ( x** y** -> z** )
    ;negate32 JSR2 ;add32 JSR2
RTN

@mul16 ( x* y* -> z** )
    ,&y1 STR ,&y0 STR ( save ylo, yhi )
    ,&x1 STR ,&x0 STR ( save xlo, xhi )
    #0000 #00 ,&z0 STR2 ,&z2 STR ( reset z0,z1,z2 )
    #0000 #00 ,&a0 STR2 ,&a2 STR ( reset a0,a1,a2 )

    ( x1 * y1 => z1z2 )
    #00 ,&x1 LDR #00 ,&y1 LDR MUL2 ,&z1 STR2

    ( x0 * y1 => z0z1 )
    #00 ,&x0 LDR #00 ,&y1 LDR MUL2 ,&z0 LDR2 ADD2 ,&z0 STR2

    ( x1 * y0 => a1a2 )
    #00 ,&x1 LDR #00 ,&y0 LDR MUL2 ,&a1 STR2

    ( x0 * y0 => a0a1 )
    #00 ,&x0 LDR #00 ,&y0 LDR MUL2 ,&a0 LDR2 ADD2 ,&a0 STR2

    ( add z and a<<8 )
    #00 ,&z0 LDR2 ,&z2 LDR
    ,&a0 LDR2 ,&a2 LDR #00
    ;add32 JSR2
RTN
[ &x0 $1 &x1 $1 ]
[ &y0 $1 &y1 $1 ]
[ &z0 $1 &z1 $1 &z2 $1 ]
[ &a0 $1 &a1 $1 &a2 $1 ]

@mul32 ( x** y** -> z** )
    ,&y1 STR2 ,&y0 STR2 ( save ylo, yhi )
    ,&x1 STR2 ,&x0 STR2 ( save xlo, xhi )
    ,&y1 LDR2 ,&x1 LDR2 ;mul16 JSR2 ( [x1*y1] )
    ,&z1 STR2 ,&z0 STR2 ( sum = x1*y1, save zlo, zhi )

    ,&y1 LDR2 ,&x0 LDR2 MUL2 ( [x0*y1]<<16 )
    ,&y0 LDR2 ,&x1 LDR2 MUL2 ( [x1*y0]<<16 )
    ( [x0*y0]<<32 will completely overflow )
    ADD2 ,&z0 LDR2 ADD2 ( sum += x0*y1<<16 + x1*y0<<16 )
    ,&z1 LDR2
RTN
[ &x0 $2 &x1 $2 ]
[ &y0 $2 &y1 $2 ]
[ &z0 $2 &z1 $2 ]

@emit-long ( hi* lo* -> )
    SWP2 ( lo* hi* )
    ;emit-short JSR2
    ;emit-short JSR2
RTN

%EMIT-BYTE { DUP #04 SFT DIGIT #0f AND DIGIT }

@emit-short ( x* -> )
    SWP ( lo^ hi^ )
    EMIT-BYTE EMIT-BYTE
RTN

@emit-byte ( x^ -> )
    EMIT-BYTE
RTN

@digits
    30 31 32 33 34 35 36 37
    38 39 61 62 63 64 65 66