uxn-utils/cli/lz/ulzenc.tal

( uxncli ulzenc.rom example.txt a.ulz )

|10 @Console &vector $2 &read $1 &pad $4 &type $1 &write $1 &error $1
|a0 @File &vector $2 &success $2 &stat $2 &delete $1 &append $1 &name $2 &length $2 &read $2 &write $2
|000

	@pad $2
	@src $30
	@dst $30
	@match-len $2
	@output-ptr $2
	@dict-best $2
	@combine $2
	@dict-data $2
	@dict-len $2

|100

@ready-src ( -> )
	;meta #06 DEO2
	.Console/type DEI #03 AND ?{ ;dict/usage <perr>/
		#010f DEO }
	;&await .Console/vector DEO2
	BRK
	&await ( -> )
	.Console/read DEI .src skey ?ready-dst
	BRK

@ready-dst ( -> )
	;&await .Console/vector DEO2
	BRK
	&await ( -> )
	.Console/read DEI .dst skey ?on-ready
	BRK

@meta $1
	( name ) "Ulzenc 0a
	( desc ) "ULZ 20 "Encoder 0a
	( auth ) "By 20 "Devine 20 "Lu 20 "Linvega 0a
	( date ) "5 20 "Jun 20 "2024 $1
	( exts ) 00

@on-ready ( -> )
	( | load raw )
	;src .File/name DEO2
	#4000 .File/length DEO2
	;raw .File/read DEO2
	;raw .File/success DEI2 uxn_lz_compress
	( | write )
	;dst .File/name DEO2
	.output-ptr LDZ2 ;compressed SUB2 .File/length DEO2
	;compressed .File/write DEO2
	( | summary )
	;dict/decompressed <pstr>
	;src <pstr>/
	;dict/spacer <pstr>
	;dst <pstr>/
	[ LIT2 "( 18 ] DEO
	.output-ptr LDZ2 ;compressed SUB2 <pdec>
	;dict/bytes <pstr>
	[ LIT2 ") 18 ] DEO
	#0a18 DEO
	( halt ) #800f DEO
	BRK

@<append-byte> ( byte -- )
	.output-ptr LDZ2 INC2k .output-ptr STZ2
	STA
	JMP2r

@<append-short> ( short* -- )
	.output-ptr LDZ2 INC2k INC2 .output-ptr STZ2
	STA2
	JMP2r

@uxn_lz_compress ( input* length* -- )
	( | fill variables )
	;compressed .output-ptr STZ2
	ADD2k NIP2 SWP2
	&w ( end* start* -- )
	EQU2k ?&end
	( | get available dictionary size )
	DUP2 ;raw SUB2 #0100 LTH2k ?{ SWP2 }
	POP2 .dict-len STZ2
	( | size of the string to search for )
	SUB2k #3fff #0004 ADD2 LTH2k ?{ SWP2 }
	POP2 ,&string-len STR2
	( | itterate through the dictionary )
	#0000 .match-len STZ2
	DUP2 .dict-len LDZ2 SUB2 .dict-data STZ2
	&for1 ( for ; dict_len; dict++, dict_len-- )
	.dict-len LDZ2 #0000 EQU2 ?&end-for1
	( Find common prefix length with the string ) #0000
	&for2 ( for i = 0;; i++ )
	( | If we reach the end of the string, it's the best possible match. )
	DUP2 [ LIT2 &string-len $2 ] NEQ2 ?{
		DUP2 .match-len STZ2
		.dict-data LDZ2 .dict-best STZ2
		POP2 !&done-search }
	( | in[i] != dict[i % dict_len] break; )
	
	( a ) ADD2k LDA STH
	( b ) DUP2 .dict-len LDZ2 DIV2k MUL2 SUB2 .dict-data LDZ2 ADD2
	( res ) LDA STHr NEQ ?&end-for2
	INC2 ORAk ?&for2
	&end-for2 ( i > match_len )
	DUP2 .match-len LDZ2 LTH2 ?{
		DUP2 .match-len STZ2
		.dict-data LDZ2 .dict-best STZ2 }
	POP2 .dict-data LDZ2 INC2 .dict-data STZ2
	.dict-len LDZ2 #0001 SUB2 .dict-len STZ2
	!&for1
	&end-for1
	&done-search ( -- )
	( CPY ) .match-len LDZ2 #0003 GTH2 ?op-cpy
	( LIT ) !op-lit
	&end POP2 POP2 JMP2r

(
@|opcodes )

@op-cpy ( in* -- )
	( | More numeric range: treat 0 as 4, 1 as 5, etc. )
	.match-len LDZ2 #0004 SUB2
	( | CPY2 )
	DUP2 #003f GTH2 ?&cpy2
	( *output_ptr++ = match_ctl
	| 0x80; ) NIP #80 ORA <append-byte>
	!&cpy-resume
	&cpy2 ( match-ctl* -- )
	SWP #c0 ORA <append-short>
	&cpy-resume ( -- )
	( | *output_ptr++ = in - dict_best - 1; )
	DUP2 .dict-best LDZ2 SUB2 #0001 SUB2 NIP <append-byte>
	( | in += match_len; Advance input by size of the match )
	.match-len LDZ2 ADD2
	( | Disable combining previous literal, if any )
	#0000 .combine STZ2
	!uxn_lz_compress/w

@op-lit ( in* -- )
	.combine LDZ2 ORA ?&combine
	( start a new literal )
	( Store this address, and later use it to increment the literal size. )
	( | combine = output_ptr++; )
	.output-ptr LDZ2 INC2k .output-ptr STZ2
	.combine STZ2
	( | *combine = 0; )
	#00 .combine LDZ2 LDA2 STA
	( | *output_ptr++ = *in++; )
	LDAk <append-byte>
	INC2 !uxn_lz_compress/w
	&combine ( -- )
	( | if ++*combine == 127 )
	.combine LDZ2 LDA INC DUP #7f NEQ ?{ POP #00 }
	.combine LDZ2 STA
	LDAk <append-byte>
	INC2 !uxn_lz_compress/w

(
@|stdlib )

@skey ( key buf -- proc )
	OVR #21 LTH ?{
		#00 SWP sput #00 JMP2r }
	POP2 #01 JMP2r

@scap ( str* -- end* )
	INC2 & LDAk ?scap
	JMP2r

@sput ( chr str* -- )
	scap/ INC2k #00 ROT ROT STA
	STA
	JMP2r

@<pstr> ( str* -- )
	LDAk #18 DEO
	INC2 & LDAk ?<pstr>
	POP2 JMP2r

@<perr> ( str* -- )
	LDAk #19 DEO
	INC2 & LDAk ?<perr>
	POP2 JMP2r

@<pdec> ( short* -- )
	#2710 [ LIT2r 00fb ]
	&>w ( -- )
		DIV2k #000a DIV2k MUL2 SUB2 SWPr EQUk OVR STHkr EQU AND ?{
			DUP [ LIT "0 ] ADD #19 DEO
			INCr }
		POP2 #000a DIV2 SWPr INCr STHkr ?&>w
	POP2r POP2 POP2 JMP2r

(
@|memory )

@dict &usage "usage: 20 "ulzenc.rom 20 "a.ulz 20 "b.bin 0a $1
	&decompressed "Compressed 20 $1
	&spacer 20 "-> 20 $1
	&bytes 20 "bytes $1

@raw $8000

@compressed