apps/README.md

@@ -74,8 +74,6 @@ There are decimal, hexadecimal and binary literals. A "straight" number is
 parsed as a decimal. Hexadecimal literals must be prefixed with `0x` (`0xf4`).
 Binary must be prefixed with `0b` (`0b01100110`).
 
-A decimal literal cannot start with `0`, with the exception of the `0` literal.
-
 Decimals and hexadecimal are "flexible". Whether they're written in a byte or
 a word, you don't need to prefix them with zeroes. Watch out for overflow,
 however.

apps/at28w/glue.asm

@@ -22,5 +22,6 @@ jp	at28wMain
 .inc "core.asm"
 .inc "lib/util.asm"
 .inc "lib/parse.asm"
+.inc "lib/args.asm"
 .inc "at28w/main.asm"
 USER_RAMSTART:

apps/at28w/main.asm

@@ -11,15 +11,15 @@
 ; *** Code ***
 
 at28wMain:
-	ld	de, AT28W_MAXBYTES
-	ld	a, (hl)
-	or	a
-	jr	z, at28wInner		; no arg
-	call	parseHexadecimal	; --> DE
+	ld	de, .argspecs
+	ld	ix, AT28W_MAXBYTES
+	call	parseArgs
 	jr	z, at28wInner
 	; bad args
 	ld	a, SHELL_ERR_BAD_ARGS
 	ret
+.argspecs:
+	.db	0b111, 0b101, 0
 
 at28wInner:
 	; Reminder: words in parseArgs aren't little endian. High byte is first.

apps/lib/args.asm

@@ -0,0 +1,111 @@
+; *** Requirements ***
+; lib/parse
+;
+; *** Consts ***
+; maximum number of bytes to receive as args in all commands. Determines the
+; size of the args variable.
+.equ	PARSE_ARG_MAXCOUNT	3
+
+; *** Code ***
+
+; Parse arguments at (HL) with specifiers at (DE) into (IX).
+;
+; Args specifiers are a series of flag for each arg:
+; Bit 0 - arg present: if unset, we stop parsing there
+; Bit 1 - is word: this arg is a word rather than a byte. Because our
+;                  destination are bytes anyway, this doesn't change much except
+;                  for whether we expect a space between the hex pairs. If set,
+;                  you still need to have a specifier for the second part of
+;                  the multibyte.
+; Bit 2 - optional: If set and not present during parsing, we don't error out
+;		    and write zero
+;
+; Bit 3 - String argument: If set, this argument is a string. A pointer to the
+;                          read string, null terminated (max 0x20 chars) will
+;                          be placed in the next two bytes. This has to be the
+;                          last argument of the list and it stops parsing.
+; Sets A to nonzero if there was an error during parsing, zero otherwise.
+parseArgs:
+	push	bc
+	push	de
+	push	hl
+	push	ix
+
+	; init the arg value to a default 0
+	xor	a
+	ld	(ix), a
+	ld	(ix+1), a
+	ld	(ix+2), a
+	ld	b, PARSE_ARG_MAXCOUNT
+.loop:
+	ld	a, (hl)
+	; is this the end of the line?
+	or	a		; cp 0
+	jr	z, .endofargs
+
+	; Get the specs
+	ld	a, (de)
+	bit	0, a		; do we have an arg?
+	jr	z, .error	; not set? then we have too many args
+
+	ld	c, a		; save the specs for multibyte check later
+	bit	3, a		; is our arg a string?
+	jr	z, .notAString
+	; our arg is a string. Let's place HL in our next two bytes and call
+	; it a day. Little endian, remember
+	ld	(ix), l
+	ld	(ix+1), h
+	jr	.success	; directly to success: skip endofargs checks
+
+.notAString:
+	call	parseHexPair
+	jr	c, .error
+
+	; we have a good arg and we need to write A in (IX).
+	ld	(ix), a
+
+	; Good! increase counters
+	inc	de
+	inc	ix
+	inc	hl		; get to following char (generally a space)
+
+	; Our arg is parsed, our pointers are increased. Normally, HL should
+	; point to a space *unless* our argspec indicates a multibyte arg.
+	bit	1, c
+	jr	nz, .nospacecheck	; bit set? no space check
+	; do we have a proper space char (or null char)?
+	ld	a, (hl)
+	or	a
+	jr	z, .endofargs
+	cp	' '
+	jr	nz, .error
+	inc	hl
+.nospacecheck:
+	djnz	.loop
+	; If we get here, it means that our next char *has* to be a null char
+	ld	a, (hl)
+	or	a		; cp 0
+	jr	z, .success	; zero? great!
+	jr	.error
+
+.endofargs:
+	; We encountered our null char. Let's verify that we either have no
+	; more args or that they are optional
+	ld	a, (de)
+	or	a
+	jr	z, .success	; no arg? success
+	bit	2, a
+	jr	z, .error	; if unset, arg is not optional. error
+	; success
+.success:
+	xor	a
+	jr	.end
+.error:
+	inc	a
+.end:
+	pop	ix
+	pop	hl
+	pop	de
+	pop	bc
+	ret
+

apps/lib/parse.asm

@@ -22,6 +22,42 @@ parseHex:
 	ret
 
 
+; Parses 2 characters of the string pointed to by HL and returns the numerical
+; value in A. If the second character is a "special" character (<0x21) we don't
+; error out: the result will be the one from the first char only.
+; HL is set to point to the last char of the pair.
+;
+; On success, the carry flag is reset. On error, it is set.
+parseHexPair:
+	push	bc
+
+	ld	a, (hl)
+	call	parseHex
+	jr	c, .end		; error? goto end, keeping the C flag on
+	rla \ rla \ rla \ rla	; let's push this in MSB
+	ld	b, a
+	inc	hl
+	ld	a, (hl)
+	cp	0x21
+	jr	c, .single	; special char? single digit
+	call	parseHex
+	jr	c, .end		; error?
+	or	b		; join left-shifted + new. we're done!
+	; C flag was set on parseHex and is necessarily clear at this point
+	jr	.end
+
+.single:
+	; If we have a single digit, our result is already stored in B, but
+	; we have to right-shift it back.
+	ld	a, b
+	and	0xf0
+	rra \ rra \ rra \ rra
+	dec	hl
+
+.end:
+	pop	bc
+	ret
+
 ; Parse the decimal char at A and extract it's 0-9 numerical value. Put the
 ; result in A.
 ;
@@ -104,45 +140,71 @@ parseDecimal:
 	pop	hl	; <-- lvl 1, orig
 	jp	unsetZ
 
-; Parse string at (HL) as a hexadecimal value without the "0x" prefix and
-; return value in DE.
-; HL is advanced to the character following the last successfully read char.
-; Sets Z on success.
+; Parse string at (HL) as a hexadecimal value and return value in DE under the
+; same conditions as parseLiteral.
 parseHexadecimal:
-	ld	a, (hl)
-	call	parseHex
-	jp	c, unsetZ	; we need at least one char
-	push	bc
-	ld	de, 0
-	ld	b, 0
-.loop:
-	; we push to B to verify overflow
-	rl	e \ rl d \ rl b
-	rl	e \ rl d \ rl b
-	rl	e \ rl d \ rl b
-	rl	e \ rl d \ rl b
-	or	e
-	ld	e, a
-	; did we overflow?
-	ld	a, b
-	or	a
-	jr	nz, .end	; overflow, NZ already set
-	; next char
+	call	hasHexPrefix
+	ret	nz
+	push	hl
+	ld	d, 0
+	inc	hl	; get rid of "0x"
 	inc	hl
+	call	strlen
+	cp	3
+	jr	c, .single
+	cp	4
+	jr	c, .doubleShort	; 0x123
+	cp	5
+	jr	c, .double	; 0x1234
+	; too long, error
+	jr	.error
+.double:
+	call	parseHexPair
+	jr	c, .error
+	inc	hl			; now HL is on first char of next pair
+	ld	d, a
+	jr	.single
+.doubleShort:
 	ld	a, (hl)
 	call	parseHex
-	jr	nc, .loop
-	cp	a		; ensure Z
+	jr	c, .error
+	inc	hl			; now HL is on first char of next pair
+	ld	d, a
+.single:
+	call	parseHexPair
+	jr	c, .error
+	ld	e, a
+	cp	a			; ensure Z
+	jr	.end
+.error:
+	call	unsetZ
 .end:
-	pop	bc
+	pop	hl
 	ret
 
-; Parse string at (HL) as a binary value (010101) without the "0b" prefix and
-; return value in E. D is always zero.
+; Sets Z if (HL) has a '0x' prefix.
+hasHexPrefix:
+	ld	a, (hl)
+	cp	'0'
+	ret	nz
+	push	hl
+	inc	hl
+	ld	a, (hl)
+	cp	'x'
+	pop	hl
+	ret
+
+; Parse string at (HL) as a binary value (0b010101) and return value in E.
+; D is always zero.
 ; Sets Z on success.
 parseBinaryLiteral:
+	call	hasBinPrefix
+	ret	nz
 	push	bc
 	push	hl
+	ld	d, 0
+	inc	hl	; get rid of "0b"
+	inc	hl
 	call	strlen
 	or	a
 	jr	z, .error	; empty, error
@@ -175,20 +237,17 @@ parseBinaryLiteral:
 	pop	bc
 	ret
 
-; Parses the string at (HL) and returns the 16-bit value in DE. The string
-; can be a decimal literal (1234), a hexadecimal literal (0x1234) or a char
-; literal ('X').
-;
-; As soon as the number doesn't fit 16-bit any more, parsing stops and the
-; number is invalid. If the number is valid, Z is set, otherwise, unset.
-parseLiteral:
-	ld	de, 0		; pre-fill
+; Sets Z if (HL) has a '0b' prefix.
+hasBinPrefix:
 	ld	a, (hl)
-	cp	0x27		; apostrophe
-	jr	z, .char
 	cp	'0'
-	jr	z, .hexOrBin
-	jp	parseDecimal
+	ret	nz
+	push	hl
+	inc	hl
+	ld	a, (hl)
+	cp	'b'
+	pop	hl
+	ret
 
 ; Parse string at (HL) and, if it is a char literal, sets Z and return
 ; corresponding value in E. D is always zero.
@@ -196,45 +255,43 @@ parseLiteral:
 ; A valid char literal starts with ', ends with ' and has one character in the
 ; middle. No escape sequence are accepted, but ''' will return the apostrophe
 ; character.
-.char:
+parseCharLiteral:
+	ld	a, 0x27		; apostrophe (') char
+	cp	(hl)
+	ret	nz
+
 	push	hl
 	inc	hl
 	inc	hl
 	cp	(hl)
-	jr	nz, .charEnd	; not ending with an apostrophe
+	jr	nz, .end	; not ending with an apostrophe
 	inc	hl
 	ld	a, (hl)
 	or	a		; cp 0
-	jr	nz, .charEnd	; string has to end there
+	jr	nz, .end	; string has to end there
 	; Valid char, good
+	ld	d, a		; A is zero, take advantage of that
 	dec	hl
 	dec	hl
-	ld	e, (hl)
+	ld	a, (hl)
+	ld	e, a
 	cp	a		; ensure Z
-.charEnd:
+.end:
 	pop	hl
 	ret
 
-.hexOrBin:
-	inc	hl
-	ld	a, (hl)
-	inc	hl		; already place it for hex or bin
-	cp	'x'
-	jr	z, .hex
-	cp	'b'
-	jr	z, .bin
-	; special case: single '0'. set Z if we hit have null terminating.
-	or	a
-.hexOrBinEnd:
-	dec	hl \ dec hl	; replace HL
-	ret			; Z already set
-
-.hex:
-	push	hl
+; Parses the string at (HL) and returns the 16-bit value in DE. The string
+; can be a decimal literal (1234), a hexadecimal literal (0x1234) or a char
+; literal ('X').
+;
+; As soon as the number doesn't fit 16-bit any more, parsing stops and the
+; number is invalid. If the number is valid, Z is set, otherwise, unset.
+parseLiteral:
+	call	parseCharLiteral
+	ret	z
 	call	parseHexadecimal
-	pop	hl
-	jr	.hexOrBinEnd
-
-.bin:
+	ret	z
 	call	parseBinaryLiteral
-	jr	.hexOrBinEnd
+	ret	z
+	jp	parseDecimal
+

apps/lib/util.asm

@@ -5,31 +5,6 @@ isWS:
 	cp	0x09
 	ret
 
-; Advance HL to next WS.
-; Set Z if WS found, unset if end-of-string.
-toWS:
-	ld	a, (hl)
-	call	isWS
-	ret	z
-	or	a
-	jp	z, unsetZ
-	inc	hl
-	jr	toWS
-
-; Consume following whitespaces in HL until a non-WS is hit.
-; Set Z if non-WS found, unset if end-of-string.
-rdWS:
-	ld	a, (hl)
-	call	isWS
-	jr	nz, .ok
-	or	a
-	jp	z, unsetZ
-	inc	hl
-	jr	rdWS
-.ok:
-	cp	a	; ensure Z
-	ret
-
 ; Copy string from (HL) in (DE), that is, copy bytes until a null char is
 ; encountered. The null char is also copied.
 ; HL and DE point to the char right after the null char.

apps/zasm/glue.asm

@@ -68,6 +68,7 @@ jp	zasmMain
 .inc "lib/util.asm"
 .inc "lib/ari.asm"
 .inc "lib/parse.asm"
+.inc "lib/args.asm"
 .inc "zasm/util.asm"
 .equ	IO_RAMSTART	USER_RAMSTART
 .inc "zasm/io.asm"

apps/zasm/gluea.asm

@@ -25,6 +25,7 @@ jp	zasmMain
 .inc "lib/util.asm"
 .inc "lib/ari.asm"
 .inc "lib/parse.asm"
+.inc "lib/args.asm"
 .inc "zasm/util.asm"
 .equ	IO_RAMSTART	USER_RAMSTART
 .inc "zasm/io.asm"

apps/zasm/main.asm

@@ -23,34 +23,34 @@
 ; HL is set to the last lineno to be read.
 ; Sets Z on success, unset on error. On error, A contains an error code (ERR_*)
 zasmMain:
-	; Parse args in (HL)
-	; blkdev in
-	call	parseHexadecimal	; --> DE
-	jr	nz, .badargs
-	ld	a, e
+	; Parse args. HL points to string already
+	; We don't allocate memory just to hold this. Because this happens
+	; before initialization, we don't really care where those args are
+	; parsed. That's why we borrow zasm's RAMSTART for a little while.
+	ld	de, .argspecs
+	ld	ix, ZASM_RAMSTART
+	call	parseArgs
+	jr	z, .goodargs
+	; bad args
+	ld	hl, 0
+	ld	de, 0
+	ld	a, SHELL_ERR_BAD_ARGS
+	ret
+
+.goodargs:
+	; HL now points to parsed args
+	; Init I/O
+	ld	a, (ZASM_RAMSTART)	; blkdev in ID
 	ld	de, IO_IN_BLK
 	call	blkSel
-
-	; blkdev in
-	call	rdWS
-	jr	nz, .badargs
-	call	parseHexadecimal	; --> DE
-	jr	nz, .badargs
-	ld	a, e
+	ld	a, (ZASM_RAMSTART+1)	; blkdev out ID
 	ld	de, IO_OUT_BLK
 	call	blkSel
 
-	; .org high byte
-	ld	e, 0			; in case we .skipOrgSet
-	call	rdWS
-	jr	nz, .skipOrgSet		; no org argument
-	call	parseHexadecimal	; --> DE
-	jr	nz, .badargs
-
-.skipOrgSet:
-	; Init .org with value of E
+	; Init .org
+	; This is the 3rd argument, optional, will be zero if not given.
 	; Save in "@" too
-	ld	a, e
+	ld	a, (ZASM_RAMSTART+2)
 	ld	(ZASM_ORG+1), a		; high byte of .org
 	ld	(DIREC_LASTVAL+1), a
 	xor	a
@@ -82,11 +82,8 @@ zasmMain:
 .end:
 	jp	ioLineNo		; --> HL, --> DE, returns
 
-.badargs:
-	; bad args
-	ld	a, SHELL_ERR_BAD_ARGS
-	ret
-
+.argspecs:
+	.db	0b001, 0b001, 0b101
 .sFirstPass:
 	.db	"First pass", 0
 .sSecondPass:

recipes/sms/romasm/glue.asm

@@ -16,6 +16,7 @@
 	jp	upcase
 	jp	findchar
 	jp	parseHex
+	jp	parseHexPair
 	jp	blkSel
 	jp	blkSet
 	jp	fsFindFN
@@ -144,9 +145,9 @@ basFindCmdExtra:
 	jp	basPgmHook
 .mycmds:
 	.db "ed", 0
-	.dw 0x1f00
+	.dw 0x1e00
 	.db "zasm", 0
-	.dw 0x2400
+	.dw 0x2300
 	.db 0xff
 
 f0GetB:
@@ -165,13 +166,13 @@ f1PutB:
 	ld	ix, FS_HANDLES+FS_HANDLE_SIZE
 	jp	fsPutB
 
-; last time I checked, PC at this point was 0x1e92. Let's give us a nice margin
+; last time I checked, PC at this point was 0x1df8. Let's give us a nice margin
 ; for the start of ed.
-.fill 0x1f00-$
+.fill 0x1e00-$
 .bin "ed.bin"
 
-; Last check: 0x23b0
-.fill 0x2400-$
+; Last check: 0x22dd
+.fill 0x2300-$
 .bin "zasm.bin"
 
 .fill 0x7ff0-$

recipes/sms/romasm/user.h

@@ -14,6 +14,7 @@
 .equ	upcase			@+3
 .equ	findchar		@+3
 .equ	parseHex		@+3
+.equ	parseHexPair	@+3
 .equ	blkSel			@+3
 .equ	blkSet			@+3
 .equ	fsFindFN		@+3

tools/tests/unit/test_lib_parse.asm

@@ -1,85 +0,0 @@
-jp	test
-
-.inc "core.asm"
-.inc "lib/util.asm"
-.inc "lib/parse.asm"
-
-zasmGetPC:
-	ret
-
-testNum:	.db 1
-
-test:
-	ld	sp, 0xffff
-
-	call	testParseHex
-	call	testParseHexadecimal
-
-	; success
-	xor	a
-	halt
-
-testParseHex:
-	ld	a, '8'
-	call	parseHex
-	jp	c, fail
-	cp	8
-	jp	nz, fail
-	call	nexttest
-
-	ld	a, 'e'
-	call	parseHex
-	jp	c, fail
-	cp	0xe
-	jp	nz, fail
-	call	nexttest
-
-	ld	a, 'x'
-	call	parseHex
-	jp	nc, fail
-	call	nexttest
-	ret
-
-testParseHexadecimal:
-	ld	hl, .s99
-	call	parseHexadecimal
-	jp	nz, fail
-	ld	a, e
-	cp	0x99
-	jp	nz, fail
-	call	nexttest
-
-	ld	hl, .saB
-	call	parseHexadecimal
-	jp	nz, fail
-	ld	a, e
-	cp	0xab
-	jp	nz, fail
-	call	nexttest
-
-	; The string "Foo" will not cause a failure. We will parse up to "o"
-	; and then stop.
-	ld	hl, .sFoo
-	call	parseHexadecimal
-	jp	nz, fail
-	ld	a, e
-	cp	0xf
-	call	nexttest
-	ret
-
-.sFoo:		.db "Foo", 0
-.saB:		.db "aB", 0
-.s99:		.db "99", 0
-
-nexttest:
-	ld	a, (testNum)
-	inc	a
-	ld	(testNum), a
-	ret
-
-fail:
-	ld	a, (testNum)
-	halt
-
-; used as RAM
-sandbox:

tools/tests/unit/test_parse.asm

@@ -0,0 +1,136 @@
+jp	test
+
+.inc "core.asm"
+.inc "lib/util.asm"
+.inc "lib/parse.asm"
+.inc "lib/args.asm"
+
+zasmGetPC:
+	ret
+
+testNum:	.db 1
+
+test:
+	ld	hl, 0xffff
+	ld	sp, hl
+
+	call	testParseHex
+	call	testParseHexPair
+	call	testParseArgs
+
+	; success
+	xor	a
+	halt
+
+testParseHex:
+	ld	a, '8'
+	call	parseHex
+	jp	c, fail
+	cp	8
+	jp	nz, fail
+	call	nexttest
+
+	ld	a, 'e'
+	call	parseHex
+	jp	c, fail
+	cp	0xe
+	jp	nz, fail
+	call	nexttest
+
+	ld	a, 'x'
+	call	parseHex
+	jp	nc, fail
+	call	nexttest
+	ret
+
+testParseHexPair:
+	ld	hl, .s99
+	call	parseHexPair
+	jp	c, fail
+	cp	0x99
+	jp	nz, fail
+	call	nexttest
+
+	ld	hl, .saB
+	call	parseHexPair
+	jp	c, fail
+	cp	0xab
+	jp	nz, fail
+	call	nexttest
+
+	ld	hl, .sFoo
+	call	parseHexPair
+	jp	nc, fail
+	call	nexttest
+	ret
+
+.sFoo:		.db "Foo", 0
+.saB:		.db "aB", 0
+.s99:		.db "99", 0
+
+
+testParseArgs:
+	ld	hl, .t1+6
+	ld	de, .t1
+	ld	iy, .t1+3
+	call	.testargs
+
+	ld	hl, .t2+6
+	ld	de, .t2
+	ld	iy, .t2+3
+	call	.testargs
+
+	ld	hl, .t3+6
+	ld	de, .t3
+	ld	iy, .t3+3
+	call	.testargs
+	ret
+
+; HL and DE must be set, and IY must point to expected results in IX
+.testargs:
+	ld	ix, sandbox
+	call	parseArgs
+	jp	nz, fail
+	ld	a, (ix)
+	cp	(iy)
+	jp	nz, fail
+	ld	a, (ix+1)
+	cp	(iy+1)
+	jp	nz, fail
+	ld	a, (ix+2)
+	cp	(iy+2)
+	jp	nz, fail
+	jp	nexttest
+
+; Test data format: 3 bytes specs, 3 bytes expected (IX), then the arg string.
+
+; Empty args with empty specs
+.t1:
+	.db 0b0000, 0b0000, 0b0000
+	.db 0, 0, 0
+	.db 0
+
+; One arg, one byte spec
+.t2:
+	.db 0b0001, 0b0000, 0b0000
+	.db 0xe4, 0, 0
+	.db "e4", 0
+
+; 3 args, 3 bytes spec
+.t3:
+	.db 0b0001, 0b0001, 0b0001
+	.db 0xe4, 0xab, 0x99
+	.db "e4 ab 99", 0
+
+nexttest:
+	ld	a, (testNum)
+	inc	a
+	ld	(testNum), a
+	ret
+
+fail:
+	ld	a, (testNum)
+	halt
+
+; used as RAM
+sandbox: