64tass v1.59 r3120 reference manual

This is the manual for 64tass, the multi pass optimizing macro assembler for the 65xx series of processors. Key features:

Contrary how the length of this document suggests 64tass can be used with just basic 6502 assembly knowledge in simple ways like any other assembler. If some advanced functionality is needed then this document can serve as a reference.

This is a development version. Features or syntax may change as a result of corrections in non-backwards compatible ways in some rare cases. It's difficult to get everything right first time.

Project page: https://sourceforge.net/projects/tass64/

The page hosts the latest and older versions with sources and a bug and a feature request tracker.

Usage tips

64tass is a command line assembler, the source can be written in any text editor. As a minimum the source filename must be given on the command line. The -a command line option is highly recommended if the source is Unicode or ASCII.

64tass -a src.asm

There are also some useful parameters which are described later.

For comfortable compiling I use such Makefiles (for make):

demo.prg: source.asm macros.asm pic.drp music.bin
        64tass -C -a -B -i source.asm -o demo.tmp
        pucrunch -ffast -x 2048 demo.tmp >demo.prg

This way demo.prg is recreated by compiling source.asm whenever source.asm, macros.asm, pic.drp or music.bin had changed.

Of course it's not much harder to create something similar for win32 (make.bat), however this will always compile and compress:

64tass.exe -C -a -B -i source.asm -o demo.tmp
pucrunch.exe -ffast -x 2048 demo.tmp >demo.prg

Here's a slightly more advanced Makefile example with default action as testing in VICE, clean target for removal of temporary files and compressing using an intermediate temporary file:

all: demo.prg
        x64 -autostartprgmode 1 -autostart-warp +truedrive +cart $<

demo.prg: demo.tmp
        pucrunch -ffast -x 2048 $< >$@

demo.tmp: source.asm macros.asm pic.drp music.bin
        64tass -C -a -B -i $< -o $@

.PHONY: all clean
        $(RM) demo.prg demo.tmp

It's useful to add a basic header to your source files like the one below, so that the resulting file is directly runnable without additional compression:

*       = $0801
        .word (+), 2005  ;pointer, line number
        .null $9e, format("%4d", start);will be sys 4096
+       .word 0          ;basic line end

*       = $1000

start   rts

A frequently coming up question is, how to automatically allocate memory, without hacks like *=*+1? Sure there's .byte and friends for variables with initial values but what about zero page, or RAM outside of program area? The solution is to not use an initial value by using ? or not giving a fill byte value to .fill.

*       = $02
p1      .addr ?         ;a zero page pointer
temp    .fill 10        ;a 10 byte temporary area

Space allocated this way is not saved in the output as there's no data to save at those addresses.

What about some code running on zero page for speed? It needs to be relocated, and the length must be known to copy it there. Here's an example:

        ldx #size(zpcode)-1;calculate length
-       lda zpcode,x
        sta wrbyte,x
        dex             ;install to zero page
        bpl -
        jsr wrbyte
;code continues here but is compiled to run from $02
zpcode  .logical $02
wrbyte  sta $ffff       ;quick byte writer at $02
        inc wrbyte+1
        bne +
        inc wrbyte+2
+       rts

The assembler supports lists and tuples, which does not seems interesting at first as it sound like something which is only useful when heavy scripting is involved. But as normal arithmetic operations also apply on all their elements at once, this could spare quite some typing and repetition.

Let's take a simple example of a low/high byte jump table of return addresses, this usually involves some unnecessary copy/pasting to create a pair of tables with constructs like >(label-1).

jumpcmd lda hibytes,x   ; selected routine in X register
        lda lobytes,x   ; push address to stack
        rts             ; jump, rts will increase pc by one!
; Build a list of jump addresses minus 1
_       := (cmd_p, cmd_c, cmd_m, cmd_s, cmd_r, cmd_l, cmd_e)-1
lobytes .byte <_        ; low bytes of jump addresses
hibytes .byte >_        ; high bytes

There are some other tips below in the descriptions.

Expressions and data types

Integer constants

Integer constants can be entered as decimal digits of arbitrary length. An underscore can be used between digits as a separator for better readability of long numbers. The following operations are accepted:

Integer operators and functions
x + yadd x to y2 + 2 is 4
x - ysubtract y from x4 - 1 is 3
x * ymultiply x with y2 * 3 is 6
x / yinteger divide x by y7 / 2 is 3
x % yinteger modulo of x divided by y5 % 2 is 1
x ** yx raised to power of y2 ** 4 is 16
-xnegated value-2 is -2
+xunchanged+2 is 2
~x-x - 1~3 is -4
x | ybitwise or2 | 6 is 6
x ^ ybitwise xor2 ^ 6 is 4
x & ybitwise and2 & 6 is 2
x << ylogical shift left1 << 3 is 8
x >> yarithmetic shift right-8 >> 3 is -1

Integers are automatically promoted to floats as necessary in expressions. Other types can be converted to integer using the integer type int.

Integer division is a floor division (rounding down) so 7 / 4 is 1 and not 1.75. If ceiling division is required (rounding up) that can be done by negating both the divident and the result. Typically it's done like 0 - -5 / 4 which results in 2.

        .byte 23        ; as unsigned
        .char -23       ; as signed

; using negative integers as immediate values
        ldx #-3         ; works as '#-' is signed immediate
num     = -3
        ldx #+num       ; needs explicit '#+' for signed 8 bits

        lda #((bitmap >> 10) & $0f) | ((screen >> 6) & $f0)
        sta $d018

Bit string constants

Bit string constants can be entered in hexadecimal form with a leading dollar sign or in binary with a leading percent sign. An underscore can be used between digits as a separator for better readability of long numbers. The following operations are accepted:

Bit string operators and functions
~xinvert bits~%101 is ~%101
y .. xconcatenate bits$a .. $b is $ab
y x nrepeat%101 x 3 is %101101101
x[n]extract bit(s)$a[1] is %1
x[s]slice bits$1234[4:8] is $3
x | ybitwise or~$2 | $6 is ~$0
x ^ ybitwise xor~$2 ^ $6 is ~$4
x & ybitwise and~$2 & $6 is $4
x << ybitwise shift left$0f << 4 is $0f0
x >> ybitwise shift right~$f4 >> 4 is ~$f

Length of bit string constants are defined in bits and is calculated from the number of bit digits used including leading zeros.

Bit strings are automatically promoted to integer or floating point as necessary in expressions. The higher bits are extended with zeros or ones as needed.

Bit strings support indexing and slicing. This is explained in detail in section Slicing and indexing.

Other types can be converted to bit string using the bit string type bits.

        .byte $33       ; 8 bits in hexadecimal
        .byte %00011111 ; 8 bits in binary
        .text $1234     ; $34, $12 (little endian)

        lda $01
        and #~$07       ; 8 bits even after inversion
        ora #$05
        sta $01

        lda $d015
        and #~%00100000 ;clear a bit
        sta $d015

Floating point constants

Floating point constants have a radix point in them and optionally an exponent. A decimal exponent is e while a binary one is p. An underscore can be used between digits as a separator for better readability. The following operations can be used:

Floating point operators and functions
x + yadd x to y2.2 + 2.2 is 4.4
x - ysubtract y from x4.1 - 1.1 is 3.0
x * ymultiply x with y1.5 * 3 is 4.5
x / yinteger divide x by y7.0 / 2.0 is 3.5
x % yinteger modulo of x divided by y5.0 % 2.0 is 1.0
x ** yx raised to power of y2.0 ** -1 is 0.5
-xnegated value-2.0 is -2.0
+xunchanged+2.0 is 2.0
~xalmost -x~2.1 is almost -2.1
x | ybitwise or2.5 | 6.5 is 6.5
x ^ ybitwise xor2.5 ^ 6.5 is 4.0
x & ybitwise and2.5 & 6.5 is 2.5
x << ylogical shift left1.0 << 3.0 is 8.0
x >> yarithmetic shift right-8.0 >> 4 is -0.5

As usual comparing floating point numbers for (non) equality is a bad idea due to rounding errors.

The only predefined constant is pi.

Floating point numbers are automatically truncated to integer as necessary. Other types can be converted to floating point by using the type float.

Fixed point conversion can be done by using the shift operators. For example an 8.16 fixed point number can be calculated as (3.14 << 16) & $ffffff. The binary operators operate like if the floating point number would be a fixed point one. This is the reason for the strange definition of inversion.

        .byte 3.66e1       ; 36.6, truncated to 36
        .byte $1.8p4       ; 4:4 fixed point number (1.5)
        .sint 12.2p8       ; 8:8 fixed point number (12.2)

Character string constants

Character strings are enclosed in single or double quotes and can hold any Unicode character.

Operations like indexing or slicing are always done on the original representation. The current encoding is only applied when it's used in expressions as numeric constants or in context of text data directives.

Doubling the quotes inside string literals escapes them and results in a single quote.

Character string operators and functions
y .. xconcatenate strings"a" .. "b" is "ab"
y in xis substring of"b" in "abc" is true
a x nrepeat"ab" x 3 is "ababab"
a[i]character from start"abc"[1] is "b"
a[-i]character from end"abc"[-1] is "c"
a[:]no change"abc"[:] is "abc"
a[s:]cut off start"abc"[1:] is "bc"
a[:-s]cut off end"abc"[:-1] is "ab"
a[s]reverse"abc"[::-1] is "cba"

Character strings are converted to integers, byte and bit strings as necessary using the current encoding and escape rules. For example when using a sane encoding "z"-"a" is 25.

Other types can be converted to character strings by using the type str or by using the repr and format functions.

Character strings support indexing and slicing. This is explained in detail in section Slicing and indexing.

mystr   = "oeU"         ; character string constant
        .text 'it''s'   ; it's
        .word "ab"+1    ; conversion result is "bb" usually

        .text "text"[:2]     ; "te"
        .text "text"[2:]     ; "xt"
        .text "text"[:-1]    ; "tex"
        .text "reverse"[::-1]; "esrever"

Byte string constants

Byte strings are like character strings, but hold bytes instead of characters.

Quoted character strings prefixing by b, l, n, p, s, x or z characters can be used to create byte strings. The resulting byte string contains what .text, .shiftl, .null, .ptext and .shift would create. Direct hexadecimal entry can be done using the x prefix and z denotes a z85 encoded byte string. Spaces can be used between pairs of hexadecimal digits as a separator for better readability.

Byte string operators and functions
y .. xconcatenate stringsx"12" .. x"34" is x"1234"
y in xis substring ofx"34" in x"1234" is true
a x nrepeatx"ab" x 3 is x"ababab"
a[i]byte from startx"abcd12"[1] is x"cd"
a[-i]byte from endx"abcd"[-1] is x"cd"
a[:]no changex"abcd"[:] is x"abcd"
a[s:]cut off startx"abcdef"[1:] is x"cdef"
a[:-s]cut off endx"abcdef"[:-1] is x"abcd"
a[s]reversex"abcdef"[::-1] is x"efcdab"

Byte strings support indexing and slicing. This is explained in detail in section Slicing and indexing.

Other types can be converted to byte strings by using the type bytes.

        .enc "screen"   ;use screen encoding
mystr   = b"oeU"        ;convert text to bytes, like .text
        .enc "none"     ;normal encoding

        .text mystr     ;text as originally encoded
        .text s"p1"     ;convert to bytes like .shift
        .text l"p2"     ;convert to bytes like .shiftl
        .text n"p3"     ;convert to bytes like .null
        .text p"p4"     ;convert to bytes like .ptext

Binary data may be embedded in source code by using hexadecimal byte strings. This is more compact than using .byte followed by a lot of numbers. As expected 1 byte becomes 2 characters.

        .text x"fce2"   ;2 bytes: $fc and $e2 (big endian)

If readability is not a concern then the more compact z85 encoding may be used which encodes 4 bytes into 5 characters. Data lengths not a multiple of 4 are handled by omitting leading zeros in the last group.

        .text z"FiUj*2M$hf";8 bytes: 80 40 20 10 08 04 02 01

For data lengths of multiple of 4 bytes any z85 encoder will do. Otherwise the simplest way to encode a binary file into a z85 string is to create a source file which reads it using the line label = binary('filename'). Now if the labels are listed to a file then there will be a z85 encoded definition for this label.

Lists and tuples

Lists and tuples can hold a collection of values. Lists are defined from values separated by comma between square brackets [1, 2, 3], an empty list is []. Tuples are similar but are enclosed in parentheses instead. An empty tuple is (), a single element tuple is (4,) to differentiate from normal numeric expression parentheses. When nested they function similar to an array. Both types are immutable.

List and tuple operators and functions
y .. xconcatenate lists[1] .. [2] is [1, 2]
y in xis member of list2 in [1, 2, 3] is true
a x nrepeat[1, 2] x 2 is [1, 2, 1, 2]
a[i]element from start("1", 2)[1] is 2
a[-i]element from end("1", 2, 3)[-1] is 3
a[:]no change(1, 2, 3)[:] is (1, 2, 3)
a[s:]cut off start(1, 2, 3)[1:] is (2, 3)
a[:-s]cut off end(1, 2.0, 3)[:-1] is (1, 2.0)
a[s]reverse(1, 2, 3)[::-1] is (3, 2, 1)
*aconvert to argumentsformat("%d: %s", *mylist)
... op aleft fold... + (1, 2, 3) is ((1+2)+3)
a op ...right fold(1, 2, 3) - ... is (1-(2-3))

Arithmetic operations are applied on the all elements recursively, therefore [1, 2] + 1 is [2, 3], and abs([1, -1]) is [1, 1].

Arithmetic operations between lists are applied one by one on their elements, so [1, 2] + [3, 4] is [4, 6].

When lists form an array and columns/rows are missing the smaller array is stretched to fill in the gaps if possible, so [[1], [2]] * [3, 4] is [[3, 4], [6, 8]].

Lists and tuples support indexing and slicing. This is explained in detail in section Slicing and indexing.

mylist  = [1, 2, "whatever"]
mytuple = (cmd_e, cmd_g)

mylist  = ("e", cmd_e, "g", cmd_g, "i", cmd_i)
keys    .text mylist[::2]    ; keys ("e", "g", "i")
call_l  .byte <mylist[1::2]-1; routines (<cmd_e-1, <cmd_g-1, <cmd_i-1)
call_h  .byte >mylist[1::2]-1; routines (>cmd_e-1, >cmd_g-1, >cmd_i-1)

Although lists elements of variables can't be changed using indexing (at the moment) the same effect can be achieved by combining slicing and concatenation:

lst     := lst[:2] .. [4] .. lst[3:]; same as lst[2] := 4 would be

Folding is done on pair of elements either forward (left) or reverse (right). The list must contain at least one element. Here are some folding examples:

minimum = size([part1, part2, part3]) <? ...
maximum = size([part1, part2, part3]) >? ...
sum     = size([part1, part2, part3]) + ...
xorall  = list_of_numbers ^ ...
join    = list_of_strings .. ...
allbits = sprites.(left, middle, right).bits | ...
all     = [true, true, true, true] && ...
any     = [false, false, false, true] || ...

The range(start, end, step) built-in function can be used to create lists of integers in a range with a given step value. At least the end must be given, the start defaults to 0 and the step to 1. Sounds not very useful, so here are a few examples:

;Bitmask table, 8 bits from left to right
        .byte %10000000 >> range(8)
;Classic 256 byte single period sinus table with values of 0–255.
        .byte 128 + 127.5 * sin(range(256) * pi / 128)
;Screen row address tables
_       := $400 + range(0, 1000, 40)
scrlo   .byte <_
scrhi   .byte >_


Dictionaries hold key and value pairs. Definition is done by collecting key:value pairs separated by comma between braces {"key":"value", :"default value"}.

Looking up a non-existing key is normally an error unless a default value is given. An empty dictionary is {}. This type is immutable. There are limitations what may be used as a key but the value can be anything.

Dictionary operators and functions
y .. xcombine dictionaries{1:2, 3:4} .. {2:3, 3:1} is {1:2, 2:3, 3:1}
x[i]value lookup{"1":2}["1"] is 2
x.isymbol lookup{.ONE:1, .TWO:2}.ONE is 1
y in xis a key1 in {1:2} is true
; Simple lookup
        .text {1:"one", 2:"two"}[2]; "two"
; 16 element "fader" table 1->15->12->11->0
        .byte {1:15, 15:12, 12:11, :0}[range(16)]
; Symbol accessible values. May be useful as a function return value too.
coords  = {.x: 24, .y: 50}
        ldx #coords.x
        ldy #coords.y


Code holds the result of compilation in binary and other enclosed objects. In an arithmetic operation it's used as the numeric address of the memory where it starts. The compiled content remains static even if later parts of the source overwrite the same memory area.

Indexing and slicing of code to access the compiled content might be implemented differently in future releases. Use this feature at your own risk for now, you might need to update your code later.

Label operators and functions
a.bb member of alabel.locallabel
.b in aif a has symbol b.locallabel in label
a[i]element from startlabel[1]
a[-i]element from endlabel[-1]
a[:]copy as tuplelabel[:]
a[s:]cut off start, as tuplelabel[1:]
a[:-s]cut off end, as tuplelabel[:-1]
a[s]reverse, as tuplelabel[::-1]
mydata  .word 1, 4, 3
mycode  .block
local   lda #0

        ldx #size(mydata) ;6 bytes (3*2)
        ldx #len(mydata)  ;3 elements
        ldx #mycode[0]    ;lda instruction, $a9
        ldx #mydata[1]    ;2nd element, 4
        jmp mycode.local  ;address of local label

Addressing modes

Addressing modes are used for determining addressing modes of instructions.

For indexing there must be no white space between the comma and the register letter, otherwise the indexing operator is not recognized. On the other hand put a space between the comma and a single letter symbol in a list to avoid it being recognized as an operator.

Addressing mode operators
#+signed immediate
#-signed immediate
( )indirect
[ ]long indirect
,bdata bank indexed
,ddirect page indexed
,kprogram bank indexed
,rdata stack pointer indexed
,sstack pointer indexed
,xx register indexed
,yy register indexed
,zz register indexed

Parentheses are used for indirection and square brackets for long indirection. These operations are only available after instructions and functions to not interfere with their normal use in expressions.

Several addressing mode operators can be combined together. Currently the complexity is limited to 4 operators. This is enough to describe all addressing modes of the supported CPUs.

Valid addressing mode operator combinations
#immediatelda #$12
#+signed immediatelda #+127
#-signed immediatelda #-128
#addr,#addrmovemvp #5,#6
addrdirect or relativelda $12 lda $1234 bne $1234
bit,addrdirect page bitrmb 5,$12
bit,addr,addrdirect page bit relative jumpbbs 5,$12,$1234
(addr)indirectlda ($12) jmp ($1234)
(addr),yindirect y indexedlda ($12),y
(addr),zindirect z indexedlda ($12),z
(addr,x)x indexed indirectlda ($12,x) jmp ($1234,x)
[addr]long indirectlda [$12] jmp [$1234]
[addr],ylong indirect y indexedlda [$12],y
#addr,bdata bank indexedlda #0,b
#addr,b,xdata bank x indexedlda #0,b,x
#addr,b,ydata bank y indexedlda #0,b,y
#addr,ddirect page indexedlda #0,d
#addr,d,xdirect page x indexedlda #0,d,x
#addr,d,ydirect page y indexedldx #0,d,y
(#addr,d)direct page indirectlda (#$12,d)
(#addr,d,x)direct page x indexed indirectlda (#$12,d,x)
(#addr,d),ydirect page indirect y indexedlda (#$12,d),y
(#addr,d),zdirect page indirect z indexedlda (#$12,d),z
[#addr,d]direct page long indirectlda [#$12,d]
[#addr,d],ydirect page long indirect y indexedlda [#$12,d],y
#addr,kprogram bank indexedjsr #0,k
(#addr,k,x)program bank x indexed indirectjmp (#$1234,k,x)
#addr,rdata stack indexedlda #1,r
(#addr,r),ydata stack indexed indirect y indexedlda (#$12,r),y
#addr,sstack indexedlda #1,s
(#addr,s),ystack indexed indirect y indexedlda (#$12,s),y
addr,xx indexedlda $12,x
addr,yy indexedlda $12,y

Direct page, data bank, program bank indexed and long addressing modes of instructions are intelligently chosen based on the instruction type, the address ranges set up by .dpage, .databank and the current program counter address. Therefore the ,d, ,b and ,k indexing is only used in very special cases.

The immediate direct page indexed #0,d addressing mode is usable for direct page access. The 8 bit constant is a direct offset from the start of actual direct page. Alternatively it may be written as 0,d.

The immediate data bank indexed #0,b addressing mode is usable for data bank access. The 16 bit constant is a direct offset from the start of actual data bank. Alternatively it may be written as 0,b.

The immediate program bank indexed #0,k addressing mode is usable for program bank jumps, branches and calls. The 16 bit constant is a direct offset from the start of actual program bank. Alternatively it may be written as 0,k.

The immediate stack indexed #0,s and data stack indexed #0,r accept 8 bit constants as an offset from the start of (data) stack. These are sometimes written without the immediate notation, but this makes it more clear what's going on. For the same reason the move instructions are written with an immediate addressing mode #0,#0 as well.

The immediate (#) addressing mode expects unsigned values of byte or word size. Therefore it only accepts constants of 1 byte or in range 0–255 or 2 bytes or in range 0–65535.

The signed immediate (#+ and #-) addressing mode is to allow signed numbers to be used as immediate constants. It accepts a single byte or an integer in range −128–127, or two bytes or an integer of −32768–32767.

The use of signed immediate (like #-3) is seamless, but it needs to be explicitly written out for variables or expressions (#+variable). In case the unsigned variant is needed but the expression starts with a negation then it needs to be put into parentheses (#(-variable)) or else it'll change the address mode to signed.

Normally addressing mode operators are used in expressions right after instructions. They can also be used for defining stack variable symbols when using a 65816, or to force a specific addressing mode.

param   = #1,s            ;define a stack variable
const   = #1              ;immediate constant
        lda #0,b          ;always "absolute" lda $0000
        lda param         ;results in lda #$01,s
        lda param+1       ;results in lda #$02,s
        lda (param),y     ;results in lda (#$01,s),y
        ldx const         ;results in ldx #$01
        lda #-2           ;negative constant, $fe

Uninitialized memory

There's a special value for uninitialized memory, it's represented by a question mark. Whenever it's used to generate data it creates a hole where the previous content of memory is visible.

Uninitialized memory holes without previous content are not saved unless it's really necessary for the output format, in that case it's replaced with zeros.

It's not just data generation statements (e.g. .byte) that can create uninitialized memory, but .fill, .align or address manipulation as well.

*       = $200          ;bytes as necessary
        .word ?         ;2 bytes
        .fill 10        ;10 bytes
        .align 64       ;bytes as necessary


There are two predefined boolean constant variables, true and false.

Booleans are created by comparison operators (<, <=, !=, ==, >=, >), logical operators (&&, ||, ^^, !), the membership operator (in) and the all and any functions.

Normally in numeric expressions true is 1 and false is 0, unless the -Wstrict-bool command line option was used.

Other types can be converted to boolean by using the type bool.

Boolean values of various types
bitsAt least one non-zero bit
boolWhen true
bytesAt least one non-zero byte
codeAddress is non-zero
floatNot 0.0
intNot zero
strAt least one non-zero byte after translation


The various types mentioned earlier have predefined names. These can used for conversions or type checks.

Built-in type names
addressAddress type
bitsBit string type
boolBoolean type
bytesByte string type
codeCode type
dictDictionary type
floatFloating point type
gapUninitialized memory type
intInteger type
listList type
strCharacter string type
tupleTuple type
typeType type

Bit and byte string conversions can take a second parameter to specify and exact size. Values which can fit in shorter space will be padded but longer ones give an error.

bits(<expression>[, <bit count>])
Convert to the specific number of bits. If the number of bits is negative then it's a signed.
bytes(<expression>[, <byte count>])
Convert to the specific number of bytes. If the number of bits is negative then it's a signed.
        .cerror type(var) != str, "Not a string!"
        .text str(year)   ; convert to string


Symbols are used to reference objects. Regularly named, anonymous and local symbols are supported. These can be constant or re-definable.

Scopes are where symbols are stored and looked up. The global scope is always defined and it can contain any number of nested scopes.

Symbols must be uniquely named in a scope, therefore in big programs it's hard to come up with useful and easy to type names. That's why local and anonymous symbols exists. And grouping certain related symbols into a scope makes sense sometimes too.

Scopes are usually created by .proc and .block directives, but there are a few other ways. Symbols in a scope can be accessed by using the dot operator, which is applied between the name of the scope and the symbol (e.g. myconsts.math.pi).

Regular symbols

Regular symbol names are starting with a letter and containing letters, numbers and underscores. Unicode letters are allowed if the -a command line option was used. There's no restriction on the length of symbol names.

Care must be taken to not use duplicate names in the same scope when the symbol is used as a constant as there can be only one definition for them.

Duplicate names in parent scopes are not a problem and this gives the ability to override names defined in lower scopes. However this can just as well lead to mistakes if a lower scoped symbol with the same name was meant so there's a -Wshadow command line option to warn if such ambiguity exists.

Case sensitivity can be enabled with the -C command line option, otherwise all symbols are matched case insensitive.

For case insensitive matching it's possible to check for consistent symbol name use with the -Wcase-symbol command line option.

A regular symbol is looked up first in the current scope, then in lower scopes until the global scope is reached.

f       .block
g        .block
n        nop            ;jump here

        jsr f.g.n       ;reference from a scope
f.x     = 3             ;create x in scope f with value 3

Local symbols

Local symbols have their own scope between two regularly named code symbols and are assigned to the code symbol above them.

Therefore they're easy to reuse without explicit scope declaration directives.

Not all regularly named symbols can be scope boundaries just plain code symbol ones without anything or an opcode after them (no macros!). Symbols defined as procedures, blocks, macros, functions, structures and unions are ignored. Also symbols defined by .var, := or = don't apply, and there are a few more exceptions, so stick to using plain code labels.

The name must start with an underscore (_), otherwise the same character restrictions apply as for regular symbols. There's no restriction on the length of the name.

Care must be taken to not use the duplicate names in the same scope when the symbol is used as a constant.

A local symbol is only looked up in it's own scope and nowhere else.

incr    inc ac
        bne _skip
        inc ac+1
_skip   rts

decr    lda ac
        bne _skip
        dec ac+1
_skip   dec ac          ;symbol reused here
        jmp incr._skip  ;this works too, but is not advised

Anonymous symbols

Anonymous symbols don't have a unique name and are always called as a single plus or minus sign. They are also called as forward (+) and backward (-) references.

When referencing them - means the first backward, -- means the second backwards and so on. It's the same for forward, but with +. In expressions it may be necessary to put them into brackets.

        ldy #4
-       ldx #0
-       txa
        cmp #3
        bcc +
        adc #44
+       sta $400,x
        bne -
        bne --

Excessive nesting or long distance references create poorly readable code. It's also very easy to copy-paste a few lines of code with these references into a code fragment already containing similar references. The result is usually a long debugging session to find out what went wrong.

These references are also useful in segments, but this can create a nice trap when segments are copied into the code with their internal references.

        bne +
        #somemakro      ;let's hope that this segment does
+       nop             ;not contain forward references...

Anonymous symbols are looked up first in the current scope, then in lower scopes until the global scope is reached.

Anonymous labels within conditionally assembled code are counted even if the code itself is not compiled and the label won't get defined. This ensures that anonymous labels are always at the same "distance" independent of the conditions in between.

Constant and re-definable symbols

Constant symbols can be created with the equal sign. These are not re-definable. Forward referencing of them is allowed as they retain the objects over compilation passes.

Symbols in front of code or certain assembler directives are created as constant symbols too. They are bound to the object following them.

Re-definable symbols can be created by the .var directive or := construct. These are also called as variables. They don't carry their content over from the previous pass therefore it's not possible to use them before their definition.

If the variable already exists in the current scope it'll get updated. If an existing variable needs to be updated in a parent scope then the ::= variable reassign operator is able to do that.

Variables can be conditionally defined using the :?= construct. If the variable was defined already then the original value is retained otherwise a new one is created with this value.

WIDTH   = 40            ;a constant
        lda #WIDTH      ;lda #$28
variabl .var 1          ;a variable
var2    := 1            ;another variable
variabl .var variabl + 1;update it verbosely
var2    += 1            ;compound assignment (add one)
var3    :?= 5           ;assign 5 if undefined

The star label

The * symbol denotes the current program counter value. When accessed it's value is the program counter at the beginning of the line. Assigning to it changes the program counter and the compiling offset.

Built-in functions

Built-in functions are pre-assigned to the symbols listed below. If you reuse these symbols in a scope for other purposes then they become inaccessible, or can perform a different function.

Built-in functions can be assigned to symbols (e.g. sinus = sin), and the new name can be used as the original function. They can even be passed as parameters to functions.

Mathematical functions

Round down. E.g. floor(-4.8) is -5.0
Round to nearest away from zero. E.g. round(4.8) is 5.0
Round up. E.g. ceil(1.1) is 2.0
Round down towards zero. E.g. trunc(-1.9) is -1
Fractional part. E.g. frac(1.1) is 0.1
Square root. E.g. sqrt(16.0) is 4.0
Cube root. E.g. cbrt(27.0) is 3.0
Common logarithm. E.g. log10(100.0) is 2.0
Natural logarithm. E.g. log(1) is 0.0
Exponential. E.g. exp(0) is 1.0
pow(<expression a>, <expression b>)
A raised to power of B. E.g. pow(2.0, 3.0) is 8.0
Sine. E.g. sin(0.0) is 0.0
Arc sine. E.g. asin(0.0) is 0.0
Hyperbolic sine. E.g. sinh(0.0) is 0.0
Cosine. E.g. cos(0.0) is 1.0
Arc cosine. E.g. acos(1.0) is 0.0
Hyperbolic cosine. E.g. cosh(0.0) is 1.0
Tangent. E.g. tan(0.0) is 0.0
Arc tangent. E.g. atan(0.0) is 0.0
Hyperbolic tangent. E.g. tanh(0.0) is 0.0
Degrees to radian. E.g. rad(0.0) is 0.0
Radian to degrees. E.g. deg(0.0) is 0.0
hypot(<expression y>, <expression x>)
Polar distance. E.g. hypot(4.0, 3.0) is 5.0
atan2(<expression y>, <expression x>)
Polar angle in −pi to +pi range. E.g. atan2(0.0, 3.0) is 0.0
Absolute value. E.g. abs(-1) is 1
Returns the sign of value as −1, 0 or 1 for negative, zero and positive. E.g. sign(-5) is -1

Byte string functions

These functions return byte strings of various lengths for signed numbers, unsigned numbers and addresses.

The naming of functions is not a coincidence and they return the bytes what the data directives with the same names normally emit.

Return a single byte string from a 8 bit unsigned (0–255) or signed number (−128–127). E.g. byte(0) is x"00" and char(-1) is x"ff"
Return a little endian byte string of 2 bytes from a 16 bit unsigned (0–65535) or signed number (−32768–32767). E.g. word(1024) is x"0004" and sint(-1) is x"ffff"
Return a little endian byte string of 3 bytes from a 24 bit unsigned (0–16777216) or signed number (−8388608–8388607). E.g. long(123456) is x"40E201" and lint(-1) is x"ffffff"
Return a little endian byte string of 4 bytes from a 32 bit unsigned (0–4294967296) or signed number (−2147483648–2147483647). E.g. dword(123456789) is x"15CD5B07" and dint(-1) is x"ffffffff"
Return a little endian byte string of 2 bytes from an address in the current program bank. E.g. addr(start) is x"0d08"
Return a little endian byte string of 2 bytes from a return address in the current program bank. E.g. rta(4096) is x"ff0f"

Other functions

Return truth for various definitions of all.
All function
all bits set or no bits at allall($f) is true
all characters non-zero or empty stringall("c") is true
all bytes non-zero or no bytesall(x"ac24") is true
all elements true or empty listall([true, true, false]) is false

Only booleans in a list are accepted with the -Wstrict-bool command line option.

Return truth for various definitions of any.
Any function
at least one bit setany(~$f) is false
at least one non-zero characterany("c") is true
at least one non-zero byteany(x"ac24") is true
at least one true elementany([true, true, false]) is true

Only booleans in a list are accepted with the -Wstrict-bool command line option.

binary(<string expression>[, <offset>[, <length>]])
Returns the binary file content as bytes.

This function reads the content of a binary file as a byte string. It also accepts optional offset and length parameters.

Binary function invocation types
Read everythingbinary(name)
Skip starting bytesbinary(name, offset)
Some bytes from offsetbinary(name, offset, length)
sid     = binary("music.sid"); read in the SID file as bytes
offs    := sid[[$7, $6]]     ; data offset (big endian)
load    := sid[[$9, $8]]     ; load address (big endian)
init    = sid[[$b, $a]]      ; init address (big endian)
play    = sid[[$d, $c]]      ; play address (big endian)

; if load address is zero then it's the first 2 bytes of data
        .if load == 0
load    := sid[offs:offs+2]  ; load address (little endian)
offs    += 2                 ; skip load address bytes

*       = load               ; set pc to load address
        .text sid[offs:]     ; dump music data
format(<string expression>[, <expression>, …])
Create string from values according to a format string.

The format function converts a list of values into a character string. The converted values are inserted in place of the % sign. Optional conversion flags and minimum field length may follow, before the conversion type character. These flags can be used:

Formatting flags
#alternate form (-$a, ~$a, -%10, ~%10, -10.)
*width/precision from list
0pad with zeros
-left adjusted (default right)
 blank when positive or minus sign
+sign even if positive
~binary and hexadecimal as bits

The following conversion types are implemented:

Formatting conversion types
cUnicode character
e Eexponential float (uppercase)
f Ffloating point (uppercase)
g Gexponential/floating point
x Xhexadecimal (uppercase)
%percent sign
        .text format("%#04x bytes left", 1000); $03e8 bytes left
Returns the number of elements.
Length of various types
bit stringlength in bitslen($034) is 12
character stringnumber of characterslen("abc") is 3
byte stringnumber of byteslen(x"abcd23") is 3
tuple, listnumber of elementslen([1, 2, 3]) is 3
dictionarynumber of elementslen({1:2, 3:4]) is 2
codenumber of elementslen(label)
random([<expression>, …])
Returns a pseudo random number.

The sequence does not change across compilations and is the same every time. Different sequences can be generated by seeding with .seed.

Random function invocation types
floating point number 0.0 <= x < 1.0random()
integer in range of 0 <= x < erandom(e)
integer in range of s <= x < erandom(s, a)
integer in range of s <= x < e, step trandom(s, a, t)
        .seed 1234      ; default is boring, seed the generator
        .byte random(256); a pseudo random byte (0–255)
        .byte random([16] x 8); 8 pseudo random bytes (0–15)
range(<expression>[, <expression>, …])
Returns a list of integers in a range, with optional stepping.
Range function invocation types
integers from 0 to e-1range(e)
integers from s to e-1range(s, a)
integers from s to e (not including e), step trange(s, a, t)
        .byte range(16) ; 0, 1, ..., 14, 15
        .char range(-5, 6); -5, -4, ..., 4, 5
mylist  = range(10, 0, -2); [10, 8, 6, 4, 2]
Returns a string representation of value.
        .warn repr(var) ; pretty print value, for debugging
Returns the size of code, structure or union in bytes.
var     .word 0, 0, 0
        ldx #size(var)  ; 6 bytes
var2    = var + 2       ; start 2 bytes later
        ldx #size(var2) ; what remains is 4 bytes
sort(<list expression>)
Returns a sorted list or tuple.

If the original list contains further lists then these must be all of the same length. In this case the order of lists is determined by comparing their elements from the start until a difference is found. The sort is stable.

; sort IRQ routines by their raster lines
sorted  = sort([(60, irq1), (50, irq2)])
lines   .byte sorted[:, 0] ; 50, 60
irqs    .addr sorted[:, 1] ; irq2, irq1



The following operators are available. Not all are defined for all types of arguments and their meaning might slightly vary depending on the type.

Unary operators
-negative +positive
!not ~invert
*convert to arguments ^decimal string

The ^ decimal string operator will be changed to mean the bank byte soon. Please update your sources to use format("%d", xxx) instead! This is done to be in line with it's use in most other assemblers.

Binary operators
+add -subtract
*multiply /divide
%modulo **raise to power
|binary or ^binary xor
&binary and <<shift left
>>shift right .member
..concat xrepeat
incontains !inexcludes

Spacing must be used for the x and in operators or else they won't be recognized as such. For example the expression [1,2]x2 should be written as [1,2]x 2 instead.

Parenthesis (( )) can be used to override operator precedence. Don't forget that they also denote indirect addressing mode for certain opcodes.

        lda #(4+2)*3

Comparison operators

Traditional comparison operators give false or true depending on the result.

The compare operator (<=>) gives −1 for less, 0 for equal and 1 for more.

Comparison operators
==equals !=not equal
<less than >=more than or equals
>more than <=less than or equals
===identical !==not identical

Bit string extraction operators

These unary operators extract 8 or 16 bits. Usually they are used to get parts of a memory address.

Bit string extraction operators
<lower byte >higher byte
<>lower word >`higher word
><lower byte swapped word `bank byte
        lda #<label     ; low byte of address
        ldy #>label     ; high byte of address
        jsr $ab1e

        ldx #<>source   ; word extraction
        ldy #<>dest
        lda #size(source)-1
        mvn #`source, #`dest; bank extraction

Please note that these prefix operators are not strongly binding like negation or inversion. Instead they apply to the whole expression to the right. This may be unexpected but is required for compatibility with old sources which expect this behaviour.

        lda #<label+10  ;This is <(label+10) and not (<label)+10

;The check below is wrong and should be written as (>start) != (>end)
        .cerror >start != >end;Effectively this is >(start != (>end))

Conditional operators

Boolean conditional operators give false or true or one of the operands as the result.

Logical and conditional operators
x || yif x is true then x otherwise y
x ^^ yif both false or true then false otherwise x || y
x && yif x is true then y otherwise x
!xif x is true then false otherwise true
c ? x : yif c is true then x otherwise y
c ?? x : yif c is true then x otherwise y (broadcasting)
x <? yif x is smaller then x otherwise y
x >? yif x is greater then x otherwise y
;Silly example for 1=>"simple", 2=>"advanced", else "normal"
        .text MODE == 1 && "simple" || MODE == 2 && "advanced" || "normal"
        .text MODE == 1 ? "simple" : MODE == 2 ? "advanced" : "normal"
;Limit result to 0 .. 8
light   .byte 0 >? range(-16, 101)/6 <? 8

Please note that these are not short circuiting operations and both sides are calculated even if thrown away later.

With the -Wstrict-bool command line option booleans are required as arguments and only the ? operator may return something else.

Address length forcing

Special addressing length forcing operators in front of an expression can be used to make sure the expected addressing mode is used. Only applicable when used directly at the mnemonic.

Address size forcing
@bto force 8 bit address
@wto force 16 bit address
@lto force 24 bit address (65816)
        lda @w $0000    ; force the use of 2 byte absolute addressing
        bne @b label    ; prevent upgrade to beq+jmp with long branches in use
        lda @w #$00     ; use 2 bytes independent of accumulator size

Compound assignment

These assignment operators are short hands for updating variables. Constants can't be changed of course.

The variables on the left must be defined beforehand by := or .var.

Compound assignment operators can modify variables defined in parent scopes as well.

Compound assignments
+=add -=subtract
*=multiply /=divide
%=modulo **=raise to power
|=binary or ^=binary xor
&=binary and ||=logical or
&&=logical and <<=shift left
>>=shift right ..=concat
<?=smaller >?=greater
x=repeat .=member
v       += 1            ; same as 'v ::= v + 1'

Slicing and indexing

Lists, character strings, byte strings and bit strings support various slicing and indexing possibilities through the [] operator.

Indexing elements with positive integers is zero based. Negative indexes are transformed to positive by adding the number of elements to them, therefore −1 is the last element. Indexing with list of integers is possible as well so [1, 2, 3][(-1, 0, 1)] is [3, 1, 2].

Slicing is an operation when parts of sequence is extracted from a start position to an end position with a step value. These parameters are separated with colons enclosed in square brackets and are all optional. Their default values are [start:maximum:step=1]. Negative start and end characters are converted to positive internally by adding the length of string to them. Negative step operates in reverse direction, non-single steps will jump over elements.

This is quite powerful and therefore a few examples will be given here:

Positive indexing a[x]
It'll simply extracts a numbered element. It is zero based, therefore "abcd"[1] results in "b".
Negative indexing a[-x]
This extracts an element counted from the end, −1 is the last one. So "abcd"[-2] results in "c".
Cut off end a[:to]
Extracts a continuous range stopping before to. So [10,20,30,40][:-1] results in [10,20,30].
Cut off start a[from:]
Extracts a continuous range starting from from. So [10,20,30,40][-2:] results in [30,40].
Slicing a[from:to]
Extracts a continuous range starting from element from and stopping before to. The two end positions can be positive or negative indexes. So [10,20,30,40][1:-1] results in [20,30].
Everything a[:]
Giving no start or end will cover everything and therefore results in a complete copy.
Reverse a[::-1]
This gives everything in reverse, so "abcd"[::-1] is "dcba".
Stepping through a[from:to:step]
Extracts every stepth element starting from from and stopping before to. So "abcdef"[1:4:2] results in "bd". The from and to can be omitted in case it starts from the beginning or end at the end. If the step is negative then it's done in reverse.
Extract multiple elements a[list]
Extract elements based on a list. So "abcd"[[1,3]] will be "bd".

The fun start with nested lists and tuples, as these can be used to create a matrix. The examples will be given for a two dimensional matrix for easier understanding, but this also works in higher dimensions.

Extract row a[x]
Given a [(1,2),(3,4)] matrix [0] will give the first row which is (1,2)
Extract row range a[from:to]
Given a [(1,2),(3,4),(5,6),(7,8)] matrix [1:3] will give [(3,4),(5,6)]
Extract column a[x]
Given a [(1,2),(3,4)] matrix [:,0] will give the first column of all rows which is [1,3]
Extract column range a[:,from:to]
Given a [(1,2,3,4),(5,6,7,8)] matrix [:,1:3] will give [(2,3),(6,7)]

And it works for list of indexes, negative indexes, stepped ranges, reversing, etc. on all axes in too many ways to show all possibilities.

Basically it's just the indexing and slicing applied on nested constructs, where each nesting level is separated by a comma.

Compiler directives

Controlling the compile offset and program counter

Two counters are used while assembling.

The compile offset is where the data and code ends up in memory (or in image file).

The program counter is what labels get set to and what the special star label refers to.

Normally both are the same (code is compiled to the location it runs from) but it does not need to be.

*= <expression>
The compile offset is adjusted so that the program counter will match the requested address in the expression.
;Offset ;PC     ;Hex            ;Monitor        ;Source
                                                *       = $0800
.0800                                           label1
                                                        .logical $1000
.0800   1000                                    label2
                                                *       = $1200
.0a00   1200                                    label3
.0a00                                           label4
.offs <expression>
Sets the compile offset relative to the program counter.

Popular in old TASM code where this was the only way to create relocated code, otherwise it's use is not recommended as there are easier to use alternatives below.

;Offset ;PC     ;Hex            ;Monitor        ;Source
                                                *       = $1000
.1000           ea              nop                     nop
                                                        .offs 100
.1065   1001    ea              nop                     nop
.logical <expression>
Starts a relocation block
Ends a relocation block

Changes the program counter only, the compile offset is not changed. When finished all continues where it was left off before.

The naming is not logical at all for relocated code, but that's how it was named in old 6502tass.

It's used for code copied to it's proper location at runtime. Can be nested of course.

;Offset ;PC     ;Hex            ;Monitor        ;Source
                                                *       = $1000
                                                        .logical $300
.1000   0300     a9 80          lda #$80        drive   lda #$80
.1002   0302     85 00          sta $00                 sta $00
.1004   0304     4c 00 03       jmp $0300               jmp drive
.virtual [<expression>]
Starts a virtual block
Ends a virtual block

Changes the program counter to the expression (if given) and discards the result of compilation. This is useful to define structures to fixed addresses.

        .virtual $d400  ; base address
sid     .block
freq    .word ?         ; frequency
pulsew  .word ?         ; pulse width
control .byte ?         ; control
ad      .byte ?         ; attack/decay
sr      .byte ?         ; sustain/release

Or to define stack "allocated" variables on 65816.

        .virtual #1,s
p1      .addr ?         ; at #1,s
tmp     .byte ?         ; at #3,s
        lda (p1),y      ; lda ($01,s),y

Aligning data or code

Alignment is about constraining data/code placement in memory.

The processor architecture doesn't have hard constraints on instruction or data placement still pages (256 bytes) come up quite often in instruction cycle times tables. Or even in errata like the indirect JMP bug which happens only if the word of the vector is crossing such page.

Other components like video chips can only display object if placed at an address divisible by 64 for example.

For code half of an address table might be spared if it's known that all the addresses have the same high bytes. Or if all interrupt routines are on the same page then it's enough to change the low byte of the vector when selecting another one.

Now it shouldn't come as a surprise that the following directives are mainly concerned about how dividing the program counter address gives a certain remainder.

The divisor in this context is called the alignment interval and is usually a number which is a power of two. Quite often 256, so that's the default.

The remainder is called offset and is by default 0. Negative offsets are a convenience feature and are internally corrected by adding the interval to it.

An interval sized memory area is called a page. It's boundary is at it's start. If data spans more than one page it's known as a page boundary cross.

Having a non-zero offset effectively shifts the boundary of a page in memory further up or down (if negative). An interval of 256 with offset of 8 gives page boundaries of $1008, $1108 or $1208 for example.

If the alignment is not good enough some alignment directives might try to correct it by adding padding. This is by default uninitialized (skip forward) but may be a fixed byte or anything more complex similarly to what the .fill directive accepts.

When alignment is done within named structures then it's relative to the start of the structure. This means the structure layout will always be the same independent of which address it's instantiated at. Anonymous structures do not change the way the alignment works.

The -Walign command line option can be used to emit warnings on where and how much padding was necessary for alignment.

.page [<interval>[, <offset>]]
Start of page check block
End of page check block

This directive is a passive assertion and checks for a page difference or page crossing.

By default or with a negative interval parameter it verifies that the start and end directives are on the same page. This is what's needed to guard relative branches against jumping across pages:

        ldx #3
        .page           ;now this will execute
-       dex             ;in 14 cycles for sure
        bne -

With a positive size parameter it verifies that there's no page cross in the memory range between the directives. This is what's needed to guard against indexed access page cross cycle penalties:

*       = $10c0
        .page 256
table   .fill $40       ;table within the same page
        .endpage        ;different page here but no crossing

Normally a page check results in an error but the -Wno-error=page command line option can reduce it into a warning.

Once this directive reports an error it's time to rearrange the source in a way that the check passes. Or alternatively the alignment directives below can be used to avoid violating the assertion.

.align [<interval>[, <fill>[, <offset>]]]
Align the program counter to a page boundary

This directive is useful when code/data needs to be placed exactly to a page boundary. If that's not already the case sufficient padding is added until the next one is reached.

        .align $40      ;sprite bitmap (64 byte aligned)
sprite  .fill 63
        .align $400     ;screen memory (1024 byte aligned)
screen  .fill 1000
        .align $400, ?, -8;sprite pointers (last 8 bytes)
spritep .fill 8
        .align          ; page sized buffer at page boundary
sendbuf .fill 256       ;to avoid indexing penalty cycles
.alignblk [<interval>[, <fill>[, <offset>]]]
Starts alignment block.
Ends alignment block.

Often the start address is not important only avoiding the page boundary matters.

This often can be achieved without any padding at all. If padding is necessary then this directive works the same as .align including alignment within structures.

It's typically used to place tables so that absolute indexed read accesses won't suffer page crossing cycle penalties.

        .alignblk       ;avoid page cross
table   .byte 0, 1, 2, 3, 4, 5, 6, 7
        lda table,x     ;no cycles wasted on access

In case the stronger guarantee of having both the start and the end directives in the same page is required then the alignment interval needs to be given as a negative number (e.g. −256). This may be necessary for aligning code with relative branches.

If the block size varies based on its memory location then doing the alignment may become impossible.

.alignpageind <target>[, <interval>[, <fill>[, <offset>]]]
Alignment of a page block indirectly.

Using .alignblk in the middle of executable code is usually problematic as the alignment is done there as well. This directive can do the alignment padding outside of the execution flow.

        .alignpageind pageblk;add alignment padding here
wait    ldx #3
pageblk .page           ;now this will execute
-       dex             ;in 14 cycles for sure
        bne -

By default and with a negative interval it tries to avoids page differences. With positive intervals page crosses. Same as the .page assertion block.

It is assumed that the padding inserted will move the target block as if it'd be right in front of it. If this isn't the case the alignment will fail.

If the block size varies based on its memory location then doing the alignment may become impossible.

.alignind <target>[, <interval>[, <fill>[, <offset>]]]
Align the target location to a page boundary indirectly

This directive tries to align the target to a page boundary. If not already on one then sufficient padding will be added until the next one is reached.

;Align "pos" to page boundary. It must come right after "neg".
        .alignind pos
neg     .fill 8
pos     .fill 8
        .cerror (<pos) != 0, "pos should be page aligned"
        .cerror pos - neg != size(neg), "there should be no gap"

It is assumed that the padding inserted will move the target as if it'd be right in front of it. If this isn't the case the alignment will fail.

.fill <length>
Usually the .fill directive is used to reserve space but it may be useful to do alignments as well.
;replacement for a .cerror overrun check and *= combo
        .fill start_address - *
;align the vectors "block" so it ends at end_address
        .fill end_address - size(vectors) - *
vectors .logical *      ;dummy non-scoped block for size()
;screen memory is needed but if at $9xxx then take $a000 instead
        .align $400     ;next 1024 byte alignment
        .fill (* >> 12) == $9 ? ($a000 - *) : 0
screen  .fill 1000

Dumping data

Storing numeric values

Multi byte numeric data is stored in the little-endian order, which is the natural byte order for 65xx processors. Numeric ranges are enforced depending on the directives used. Signed numbers are stored as two's complement.

When using lists or tuples their content will be used one by one. Uninitialized data (?) creates holes of different sizes. Character string constants are converted using the current encoding.

Please note that multi character strings usually don't fit into 8 bits and therefore the .byte directive is not appropriate for them. Use .text instead which accepts strings of any length.

.byte <expression>[, <expression>, …]
Create bytes from 8 bit unsigned constants (0–255)
.char <expression>[, <expression>, …]
Create bytes from 8 bit signed constants (−128–127)
>1000  ff 03                             .byte 255, $03
>1002  41                                .byte "a"
>1003                                    .byte ?        ; reserve 1 byte
>1004  fd                                .char -3
;Store 4.4 signed fixed point constants
>1005  c8 34 32                          .char (-3.5, 3.25, 3.125) * 1p4
;Compact computed jumps using self modifying code
.1008  bd 0f 10  lda $1010,x             lda jumps,x
.100b  8d 0e 10  sta $100f               sta smod+1
.100e  d0 fe     bne $100e       smod    bne *
;Routines nearby (-128 to 127 bytes)
>1010  23 49                     jumps   .char (routine1, routine2)-smod-2
.word <expression>[, <expression>, …]
Create bytes from 16 bit unsigned constants (0–65535)
.sint <expression>[, <expression>, …]
Create bytes from 16 bit signed constants (−32768–32767)
>1000  42 23 55 45                       .word $2342, $4555
>1004                                    .word ?        ; reserve 2 bytes
>1006  eb fd 51 11                       .sint -533, 4433
;Store 8.8 signed fixed point constants
>100a  80 fc 40 03 20 03                 .sint (-3.5, 3.25, 3.125) * 1p8
.1010  bd 19 10  lda $1019,x             lda texts,x
.1013  bc 1a 10  ldy $101a,x             ldy texts+1,x
.1016  4c 1e ab  jmp $ab1e               jmp $ab1e
>1019  33 10 59 10               texts   .word text1, text2
.addr <expression>[, <expression>, …]
Create 16 bit address constants for addresses (in current program bank)
.rta <expression>[, <expression>, …]
Create 16 bit return address constants for addresses (in current program bank)
                                        *       = $12000
.012000  7c 03 20       jmp ($012003,x)         jmp (jumps,x)
>012003  50 20 32 03 92 15              jumps   .addr $12050, routine1, routine2
;Computed jumps by using stack (current bank)
                                        *       = $103000
.103000  bf 0c 30 10    lda $10300c,x           lda rets+1,x
.103004  48             pha                     pha
.103005  bf 0b 30 10    lda $10300b,x           lda rets,x
.103009  48             pha                     pha
.10300a  60             rts                     rts
>10300b  ff ef a1 36 f3 42              rets    .rta $10f000, routine1, routine2
.long <expression>[, <expression>, …]
Create bytes from 24 bit unsigned constants (0–16777215)
.lint <expression>[, <expression>, …]
Create bytes from 24 bit signed constants (−8388608–8388607)
>1000  56 34 12                          .long $123456
>1003                                    .long ?        ; reserve 3 bytes
>1006  eb fd ff 51 11 00                 .lint -533, 4433
;Store 8.16 signed fixed point constants
>100c  5d 8f fc 66 66 03 1e 85           .lint (-3.44, 3.4, 3.52) * 1p16
>1014  03
;Computed long jumps with jump table (65816)
.1015  bd 2a 10  lda $102a,x             lda jumps,x
.1018  8d 11 03  sta $0311               sta ind
.101b  bd 2b 10  lda $102b,x             lda jumps+1,x
.101e  8d 12 03  sta $0312               sta ind+1
.1021  bd 2c 10  lda $102c,x             lda jumps+2,x
.1024  8d 13 03  sta $0313               sta ind+2
.1027  dc 11 03  jmp [$0311]             jmp [ind]
>102a  32 03 01 92 05 02         jumps   .long routine1, routine2
.dword <expression>[, <expression>, …]
Create bytes from 32 bit unsigned constants (0–4294967295)
.dint <expression>[, <expression>, …]
Create bytes from 32 bit signed constants (−2147483648–2147483647)
>1000  78 56 34 12              .dword $12345678
>1004                           .dword ?        ; reserve 4 bytes
>1008  5d 7a 79 e7              .dint -411469219
;Store 16.16 signed fixed point constants
>100c  5d 8f fc ff 66 66 03 00  .dint (-3.44, 3.4, 3.52) * 1p16
>1014  1e 85 03 00
.text bits(<expression>[, <bit count>])
Create bytes from arbitrary precision unsigned and signed numbers.
.text bytes(<expression>[, <byte count>])
Create bytes from arbitrary precision unsigned and signed numbers.

For cases not covered by the numeric store directives above it's possible to convert numbers to byte or bit strings and store the resulting string. If the count expression of bytes() and bits() is negative then the stored number is signed otherwise unsigned.

>1000  74 65 78 74 00 00 00 00  .text bytes("text", 8);pad up to 8 bytes
>1008  f4 ff ff ff ff ff ff ff  .text bytes(-12, -8)  ;8 bytes signed
>1010  00 04 00 00 00 00        .text bits(1024, 48)  ;48 bits unsigned
>1016  f4 ff ff ff ff ff        .text bits(-12, -48)  ;48 bits signed

Storing string values

The following directives store strings of characters, bytes or bits as bytes. Small numeric constants can be mixed in to represent single byte control characters.

When using lists or tuples their content will be used one by one. Uninitialized data (?) creates byte sized holes. Character string constants are converted using the current encoding.

.text <expression>[, <expression>, …]
Assemble strings into 8 bit bytes.
>1000  4f 45 d5  .text "oeU"
>1003  4f 45 d5  .text 'oeU'
>1006  17 33     .text 23, $33  ; bytes
>1008  0d 0a     .text $0a0d    ; $0d, $0a, little endian!
>100a  1f        .text %00011111; more bytes
.fill <length>[, <fill>]
Reserve space (using uninitialized data), or fill with repeated bytes.
>1000            .fill $100      ;no fill, just reserve $100 bytes
>1100  00 00 00  .fill $4000, 0  ;16384 bytes of 0
>5100  55 aa 55  .fill 8000, [$55, $aa];8000 bytes of alternating $55, $aa
>7040  ff ff ff  .fill $8000 - *, $ff;fill up rest of EPROM with $ff
.shift <expression>[, <expression>, …]
Assemble strings of 7 bit bytes and mark the last byte by setting it's most significant bit.

Any byte which already has the most significant bit set will cause an error. The last byte can't be uninitialized or missing of course.

The naming comes from old TASM and is a reference to setting the high bit of alphabetic letters which results in it's uppercase version in PETSCII.

.1000  a2 00          ldx #$00                ldx #0
.1002  bd 10 10       lda $1010,x     loop    lda txt,x
.1005  08             php                     php
.1006  29 7f          and #$7f                and #$7f
.1008  20 d2 ff       jsr $ffd2               jsr $ffd2
.100b  e8             inx                     inx
.100c  28             plp                     plp
.100d  10 f3          bpl $1002               bpl loop
.100f  60             rts                     rts
>1010  53 49 4e 47 4c 45 20 53        txt     .shift "single", 32, "string"
>1018  54 52 49 4e c7
.shiftl <expression>[, <expression>, …]
Assemble strings of 7 bit bytes shifted to the left once with the last byte's least significant bit set.

Any byte which already has the most significant bit set will cause an error as this is cut off on shifting. The last byte can't be uninitialized or missing of course.

The naming is a reference to left shifting.

.1000  a2 00          ldx #$00                ldx #0
.1002  bd 0d 10       lda $100d,x     loop    lda txt,x
.1005  4a             lsr a                   lsr a
.1006  9d 00 04       sta $0400,x             sta $400,x      ;screen memory
.1009  e8             inx                     inx
.100a  90 f6          bcc $1002               bcc loop
.100c  60             rts                     rts
                                              .enc "screen"
>100d  a6 92 9c 8e 98 8a 40 a6        txt     .shiftl "single", 32, "string"
>1015  a8 a4 92 9c 8f                         .enc "none"
.null <expression>[, <expression>, …]
Same as .text, but adds a zero byte to the end. An existing zero byte is an error as it'd cause a false end marker.
.1000  a9 07          lda #$07                lda #<txt
.1002  a0 10          ldy #$10                ldy #>txt
.1004  20 1e ab       jsr $ab1e               jsr $ab1e
>1007  53 49 4e 47 4c 45 20 53        txt     .null "single", 32, "string"
>100f  54 52 49 4e 47 00
.ptext <expression>[, <expression>, …]
Same as .text, but prepend the number of bytes in front of the string (pascal style string). Therefore it can't do more than 255 bytes.
.1000  a9 1d          lda #$1d                lda #<txt
.1002  a2 10          ldx #$10                ldx #>txt
.1004  20 08 10       jsr $1008               jsr print
.1007  60             rts                     rts

.1008  85 fb          sta $fb         print   sta $fb
.100a  86 fc          stx $fc                 stx $fc
.100c  a0 00          ldy #$00                ldy #0
.100e  b1 fb          lda ($fb),y             lda ($fb),y
.1010  f0 0a          beq $101c               beq null
.1012  aa             tax                     tax
.1013  c8             iny             -       iny
.1014  b1 fb          lda ($fb),y             lda ($fb),y
.1016  20 d2 ff       jsr $ffd2               jsr $ffd2
.1019  ca             dex                     dex
.101a  d0 f7          bne $1013               bne -
.101c  60             rts             null    rts
>101d  0d 53 49 4e 47 4c 45 20        txt     .ptext "single", 32, "string"
>1025  53 54 52 49 4e 47

Text encoding

64tass supports sources written in UTF-8, UTF-16 (be/le) and RAW 8 bit encoding. To take advantage of this capability custom encodings can be defined to map Unicode characters to 8 bit values in strings. Even in plain ASCII sources it could be useful to define escape sequences for control codes.

.enc <expression>
Selects text encoding by a character string name or from an encoding object

Predefined encodings names are none and screen (screen code), anything else is user defined. All user encodings start without any character or escape definitions, add some as required. Please note that the encoding names are global.

This directive changes the text encoding after it therefore it's usually placed somewhere at the beginning of the source to make sure everything is covered.

While it is possible to juggle with multiple encodings throughout the source code using the .enc directive this is not recommended. For such use case .encode is better suited.

                                .enc "screen";screen code mode
>1000  13 03 12 05 05 0e 20 03  .text "screen codes"
>1008  0f 04 05 13
.100c  c9 15     cmp #$15       cmp #"u"    ;compare screen code
                                .enc "none" ;normal mode again
.100e  c9 55     cmp #$55       cmp #"u"    ;compare PETSCII
.encode [<expression>]
Encoding area start
Encoding area end

This directive either creates a new text encoding (if used without a parameter) or makes the one in the parameter effective within the enclosed area.

The text encoding can be assigned to a symbol in front of the directive so it can be reused whenever it's needed. This symbol can also act as a conversion function which converts a character string to a byte string using the encoding.

        .encode         ;starts anonymous local encoding scope
        .enc "titlefont";special character set
        .text "game title"
        .endencode      ;restores original encoding

vt100   .encode         ;define custom encoding
        .cdef " ~", 32
        .edef "{esc}", 27;add escape codes
        .edef "{moff}", [27, "[", "m"]
        .edef "{bold}", [27, "[", "1", "m"]

        .encode vt100   ;use custom encoding from here
        .text "{bold}bold{moff} text"
        lda #"{esc}"
        .endencode      ;restores original encoding
        cmp #vt100("{esc}");conversion when not in scope
        .enc vt100      ;select custom encoding (at start of source)
.cdef <start>, <end>, <coded> [, <start>, <end>, <coded>, …]
.cdef "<start><end>", <coded> [, "<start><end>", <coded>, …]
Assigns characters in a range to single bytes.

This is a simple single character to byte translation definition. It's useful to map a range of Unicode characters to a range of bytes. The start and end positions are Unicode character codes either by numbers or by typing them. Overlapping ranges are not allowed.

        .enc "ascii"    ;define an ascii encoding
        .cdef " ~", 32  ;identity mapping for printable
.tdef <expression>, <expression> [, <expression>, <expression>, …]
Assign single characters to byte values.

Similar to .cdef it is a single character to byte translation definition. It's easier to use when the character codes are not consecutive. Overlapping ranges with the former and itself are not allowed.

It tries to assign Unicode character codes from the first expression to byte values from the second. More than one pair of such assignments can be given.

If the byte value expression is not iterable then it will get incremented for each character definition. This allows easy assignment of randomly scattered Unicode values to a consecutive range of bytes values.

        .tdef "A", 65   ;A -> 65
        .tdef "ACX", 65 ;A -> 65, C-> 66, X -> 67
        .tdef "ACX", [65, 33, 11];A -> 65, C-> 33, X -> 11
.edef "<escapetext>", <value> [, "<escapetext>", <value>, …]
Assigns strings to byte sequences as a translated value.

When these substrings are found in a text they are replaced by bytes defined here. When strings with common prefixes are used the longest match wins. Useful for defining non-typeable control code aliases, or as a simple tokeniser.

        .edef "\n", 13  ;one byte control codes
        .edef "{clr}", 147
        .edef "{crlf}", [13, 10];two byte control code
        .edef "<nothing>", [];replace with no bytes

The example below shows how all this fits together:

petscii .namespace
common  .segment;common definitions
        .cdef " @", $20;32-64 is identical
        .tdef "[£]↑←", $5b, "┼🮌🭳", $db
        .edef "{clr}", 147, "{cr}", 13
upper   .encode;uppercase PETSCII
        .cdef "AZ", $41
        .tdef "▌▄▔▁▏▒▕🮏◤🮇├▗└┐▂┌┴┬┤▎▍🮈🮂🮃▃🭿▖▝┘▘▚🭹", $a1
        .tdef "♠🭲🭸🭷🭶🭺🭱🭴╮╰╯🭼╲╱🭽🭾●🭻♥🭰╭╳○♣🭵♦", $c1
        .tdef "◥π─│", [$df, $ff, $c0, $dd]

lower   .encode;lowercase PETSCII
        .cdef "az", $41, "AZ", $c1;the easy ranges
        .tdef "▌▄▔▁▏▒▕🮏🮙🮇├▗└┐▂┌┴┬┤▎▍🮈🮂🮃▃✓▖▝┘▘▚🭹", $a1
        .tdef "🮘🮕─│", [$df, $ff, $c0, $dd];random one to ones

                                    .encode petscii.lower
>1000  93 d4 45 58 54 20 49 4e     .text "{clr}Text in PETSCII{cr}"
>1008  20 d0 c5 d4 d3 c3 c9 c9 0d

Structured data

Structures and unions can be defined to create complex data types. The offset of fields are available by using the definition's name. The fields themselves by using the instance name.

The initialization method is very similar to macro parameters, the difference is that unset parameters always return uninitialized data (?) instead of an error.


Structures are for organizing sequential data, so the length of a structure is the sum of lengths of all items.

.struct [<name>][=<default>]][, [<name>][=<default>] …]
Begins a structure block
.ends [<result>][, <result> …]
.endstruct [<result>][, <result> …]
Ends a structure block

Structure definition, with named parameters and default values

.dstruct <name>[, <initialization values>]
.<name> [<initialization values>]
Create instance of structure with initialization values
        .struct         ;anonymous structure
x       .byte 0         ;labels are visible
y       .byte 0         ;content compiled here
        .endstruct      ;useful inside unions

nn_s    .struct col, row;named structure
x       .byte \col      ;labels are not visible
y       .byte \row      ;no content is compiled here
        .endstruct      ;it's just a definition

nn      .dstruct nn_s, 1, 2;structure instance (within label)

        lda nn.x        ;direct field access
        ldy #nn_s.x     ;get offset of field
        lda nn,y        ;and use it indirectly

nnarray .brept 4        ;4 element "array" here
        .dstruct nn_s   ;fields directly here (without a label)

        lda nnarray[0].y;access of "array" field

coords2 .bfor x2, y2 in (1,3),(4,2),(7,5)
        .dstruct nn_s, x2, y2
        .next           ;initialized "array" from list


Unions can be used for overlapping data as the compile offset and program counter remains the same on each line. Therefore the length of a union is the length of it's longest item.

.union [<name>][=<default>]][, [<name>][=<default>] …]
Begins a union block
Ends a union block

Union definition, with named parameters and default values

.dunion <name>[, <initialization values>]
.<name> [<initialization values>]
Create instance of union with initialization values
        .union          ;anonymous union
x       .byte 0         ;labels are visible
y       .word 0         ;content compiled here

nn_u    .union          ;named union
x       .byte ?         ;labels are not visible
y       .word \1        ;no content is compiled here
        .endunion       ;it's just a definition

nn      .dunion nn_u, 1 ;union instance here

        lda nn.x        ;direct field access
        ldy #nn_u.x     ;get offset of field
        lda nn,y        ;and use it indirectly

Combined use of structures and unions

The example below shows how to define structure to a binary include.

        .binary "pic.drp", 2
color   .fill 1024
screen  .fill 1024
bitmap  .fill 8000
backg   .byte ?

Anonymous structures and unions in combination with sections are useful for overlapping memory assignment. The example below shares zero page allocations for two separate parts of a bigger program. The common subroutine variables are assigned after in the zp section.

*       = $02
        .union          ;spare some memory
          .dsection zp1 ;declare zp1 section
          .dsection zp2 ;declare zp2 section
        .dsection zp    ;declare zp section


Macros can be used to reduce typing of frequently used source lines. Each invocation is a copy of the macro's content with parameter references replaced by the parameter texts.

.segment [<name>][=<default>]][, [<name>][=<default>] …]
Start of segment block
.endsegment [<result>][, <result> …]
End of segment block

Copies the code segment as it is, so symbols can be used from outside, but this also means multiple use will result in double defines unless anonymous labels are used.

.macro [<name>][=<default>]][, [<name>][=<default>] …]
Start of macro block
.endmacro [<result>][, <result> …]
End of macro block

The code is enclosed in it's own block so symbols inside are non-accessible, unless a label is prefixed at the place of use, then local labels can be accessed through that label.

#<name> [<param>][[,][<param>] …]
.<name> [<param>][[,][<param>] …]
Invoke the macro after # or . with the parameters. Normally the name of the macro is used, but it can be any expression.
.endm [<result>][, <result> …]
Closing directive of .macro and .segment for compatibility.
;A simple macro
copy    .macro
        ldx #size(\1)
lp      lda \1,x
        sta \2,x
        bpl lp

        #copy label, $500

;Use macro as an assembler directive
lohi    .macro
lo      .byte <(\@)
hi      .byte >(\@)

var     .lohi 1234, 5678

        lda var.lo,y
        ldx var.hi,y

Parameter references

The first 9 parameters can be referenced by \1\9. The entire parameter list including separators is \@.

name    .macro
        lda #\1         ;first parameter 23+1

        #name 23+1      ;call macro

Parameters can be named, and it's possible to set a default value after an equal sign which is used as a replacement when the parameter is missing.

These named parameters can be referenced by \name or \{name}. Names must match completely, if unsure use the quoted name reference syntax.

name    .macro first, b=2, , last
        lda #\first     ;first parameter
        lda #\b         ;second parameter
        lda #\3         ;third parameter
        lda #\last      ;fourth parameter

        #name 1, , 3, 4 ;call macro

Text references

In the original turbo assembler normal references are passed by value and can only appear in place of one. Text references on the other hand can appear everywhere and will work in place of e.g. quoted text or opcodes and labels. The first 9 parameters can be referenced as text by @1@9.

name    .macro
        jsr print
        .null "Hello @1!";first parameter

        #name "wth?"    ;call macro

Custom functions

Beyond the built-in functions mentioned earlier it's possible to define custom ones for frequently used calculations.

.sfunction [<name>[:<expression>][=<default>], …][*<name>,] <expression>
Defines a simple function to return the result of a parametrised expression
.function <name>[:<expression>][=<default>]], <name>[=<default>] …][, *<name>]
Defines a multi line function
.endf [<result>][, <result> …]
.endfunction [<result>][, <result> …]
End of a multi line function
#<name> [<param>][[,][<param>] …]
.<name> [<param>][[,][<param>] …]
<name> [<param>][[,][<param>] …]
Invoke a multi line function like a macro, directive or pseudo instruction

Function parameters are assigned to comma separated variable names on invocation. These variables are visible in the function scope.

Parameter values may be converted using a function whose name can be given after a colon following the variable name.

Default values may be supplied for each parameter after an equal sign. These values are calculated at function definition time only and are used when a parameter was not specified.

Extra parameters are not accepted, unless the last parameter symbol is preceded with a star, in this case these parameters are collected into a tuple.

Only those external variables and functions are available which were accessible at the place of definition, but not those at the place of invocation.

vicmem  .sfunction _font, _scr=0, ((_font >> 10) & $0f) | ((_scr >> 6) & $f0)

        lda #vicmem($2000, $0400); calculate constant
        sta $d018

If a multi line macro is used in an expression only the returned result is used. If multiple values are returned these will form a tuple.

If a multi line function is used as macro, directive or pseudo instruction and there's a label in front then the returned value is assigned to it. If nothing is returned then it's used as regular label.

mva     .function value, target
        lda value
        sta target

        mva #1, label

Conditional assembly

To prevent parts of source from compiling conditional constructs can be used. This is useful when multiple slightly different versions needs to be compiled from the same source.

Anonymous labels are still recognized in the non-compiling parts even if they won't get defined. This ensures consistent relative referencing across conditionally compiled areas with such labels.

If, else if, else

.if <condition>
Compile if condition is true
.elsif <condition>
Compile if previous conditions were not met and the condition is true
Compile if previous conditions were not met
.ifne <value>
Compile if value is not zero
.ifeq <value>
Compile if value is zero
.ifpl <value>
Compile if value is greater or equal zero
.ifmi <value>
Compile if value is less than zero

The .ifne, .ifeq, .ifpl and .ifmi directives exists for compatibility only, in practice it's better to use comparison operators instead.

        .if wait==2     ;2 cycles
        .elsif wait==3  ;3 cycles
        bit $ea
        .elsif wait==4  ;4 cycles
        bit $eaea
        .else           ;else 5 cycles
        inc $2
End of conditional compilation.
.elif <condition>
Same as .elsif because it's a popular typo and it's difficult to notice.

Switch, case, default

Similar to the .if, .elsif, .else, .endif construct, but the compared value needs to be written only once in the switch statement.

.switch <expression>
Evaluate expression and remember it
.case <expression>[, <expression> …]
Compile if the previous conditions were all skipped and one of the values equals
Compile if the previous conditions were all skipped
        .switch wait
        .case 2         ;2 cycles
        .case 3         ;3 cycles
        bit $ea
        .case 4         ;4 cycles
        bit $eaea
        .default        ;else 5 cycles
        inc $2
End of .switch conditional compilation block.


Never compile.
        lda #1          ;this won't be compiled
        sta $d020
End of .comment block.


The following directives are used to repeat code or data.

The regular non-scoped variants cover most cases except when normal labels are required as those will be double defined.

Scoped variants (those starting with the letter b) create a new scope for each iteration. This allows normal labels without collision but it's a bit more resource intensive.

If the scoped variant is prefixed with a label then the list of individual scopes for each iteration will be assigned to it. This allows accessing labels within.

.for [<assignment>], [<condition>], [<assignment>]
.bfor [<assignment>], [<condition>], [<assignment>]
Assign initial value, loop while the condition is true and modify value.

First a variable is set, usually this is used for counting. This is optional, the variable may be set already before the loop.

Then the condition is checked and the enclosed lines are compiled if it's true. If there's no condition then it's an infinite loop and .break must be used to terminate it.

After an iteration the second assignment is calculated, usually it's updating the loop counter variable. This is optional as well.

        ldx #0
        lda #32
lp      .for ue := $400, ue < $800, ue += $100
        sta ue,x         ;do $400, $500, $600 and $700
        bne lp
.for <variable>[, <variable>, …] in <expression>
.bfor <variable>[, <variable>, …] in <expression>
Assign variable(s) to values in sequence one-by-one in order.

The expression is usually the range function or some sort of list.

        .for col in 0, 11, 12, 15, 1
        lda #col         ;0, 11, 12, 15 and 1
        sta $d020
End of a .for or .bfor loop block
.rept <expression>
.brept <expression>
Repeat enclosed lines the specified number of times.
        .rept 100
-       inx
        bne -

lst     .brept 100       ;each iteration into a tuple
label   jmp label        ;not a duplicate definition
        jmp lst[5].label ;use label of 6th iteration
End of a .rept or .brept block
.while <condition>
.bwhile <condition>
Repeat enclosed lines until the condition holds.

Works as expected however the scoped variant might be tricky to use as variables of the condition are usually part of the parent scope. So modifying them in the loop body should be done with compound assignments or the reassign operator (::=).

End of a .while or .bwhile loop block
Exit current repetition loop immediately.
.breakif <condition>
Exit current repetition loop immediately if the condition holds.

It's a shorthand for a .if, .break, .endif sequence.

Continue current repetition loop's next iteration.
.continueif <condition>
Continue current repetition loop's next iteration if the condition holds.

It's a shorthand for a .if, .continue, .endif sequence.

Closing directive of .for, .bfor, .rept, .brept, .while and .bwhile loop for compatibility.
Creates a special jump label that can be referenced by .goto
.goto <labelname>
Causes assembler to continue assembling from the jump label. No forward references of course, handle with care. Should only be used in classic TASM sources for creating loops.
i       .var 100
loop    .lbl
i       .var i - 1
        .ifne i
        .goto loop       ;generates 100 nops
        .endif           ;the hard way ;)

Including files

Longer sources are usually separated into multiple files for easier handling. Precomputed binary data can also be included directly without converting it into source code first.

Search path is relative to the location of current source file. If it's not found there the include search path is consulted for further possible locations.

To make your sources portable please always use forward slashes (/) as a directory separator and use lower/uppercase consistently in file names!

.include <filename>
Include source file here.
.binclude <filename>
Include source file here in it's local block. If the directive is prefixed with a label then all labels are local and are accessible through that label only, otherwise not reachable at all.

        .include "macros.asm"       ;include macros
menu    .binclude "menu.asm"        ;include in a block
        jmp menu.start
.binary <filename>[, <offset>[, <length>]]
Include raw binary data from file.

By using offset and length it's possible to break out chunks of data from a file separately, like bitmap and colors for example. Negative offsets are calculated from the end of file.

        .binary "stuffz.bin"        ;simple include, all bytes
        .binary "stuffz.bin", 2     ;skip start address
        .binary "stuffz.bin", 2, 1000;skip start address, 1000 bytes max


Scopes may contain symbols or further nested scopes. The same symbol name can be reused as long as it's in a different scope.

A symbol is looked up in the local scope first. If it's a non-local symbol then parent scopes and the global scope may be searched in addition. This means that a symbol in a parent or global scope may be shadowed.

Symbols of a named scope can be looked up using the . operator. The searched symbol stands on the right and it's looked up in the scope on the left. More than one symbol may be looked up at the same time and the result will be a list or tuple.

        lda #0
        sta vic.sprite.enable
; same as .byte colors.red, colors.green, colors.blue
ctable  .byte colors.(red, green, blue)
Start of a procedure block
End of a procedure block

If the symbol in front is not referenced anywhere then the enclosed source won't be compiled.

Symbols inside are enclosed in a scope and are accessible through the symbol of the procedure using the dot notation. This forces compilation of the whole procedure of course.

ize     .proc
cucc    nop

        jsr ize
        jmp ize.cucc

The compilation only if used behaviour of .proc eases the building of libraries from a collection of subroutines and tables where not everything is needed all the time.

Alternative dead-code reduction techniques I encountered:

Separate source files
This potentially results in a lot of small files and manually managed include directives. This is popular on external linker based systems where object files may be excluded if unused.
Conditional compilation
Few larger files with conditional compilation directives all over the place to exclude or include various parts. The source which does the include manually declares somewhere what's actually needed or not. There may be a lot of options if it's fine grained enough.
Wrap parts with macros
If a part is needed then a single macro call is placed somewhere to include that part. Much better than conditional compilation but these macro calls still need to be manually managed.
Block scoping area start
Block scoping area end

All symbols inside a block are enclosed in a scope. If the block had a symbol then local symbols are accessible through that using the dot notation.

        inc count + 1
count   ldx #0
.namespace [<expression>]
Namespace area start
Namespace area end

This directive either creates a new scope (if used without a parameter) or activates the one in the parameter.

The scope can be assigned to a symbol in front of the directive so that it can be reactivated later. This enables label definitions into the same scope in different files.

colors  .namespace
red     = 2
blue    = 6
        lda #colors.red
Weak symbol area

Any symbols defined inside can be overridden by stronger symbols in the same scope from outside. Can be nested as necessary.

This gives the possibility of giving default values for symbols which might not always exist without resorting to .ifdef/.ifndef or similar directives in other assemblers.

symbol  = 1            ;stronger symbol than the one below
symbol  = 0            ;default value if the one above does not exists
        .if symbol     ;almost like an .ifdef ;)

Other use of weak symbols might be in included libraries to change default values or replace stub functions and data structures.

If these stubs are defined using .proc/.endproc then their default implementations will not even exists in the output at all when a stronger symbol overrides them.

Multiple definition of a symbol with the same strength in the same scope is of course not allowed and it results in double definition error.

Please note that .ifdef/.ifndef directives are left out from 64tass for of technical reasons, so don't wait for them to appear anytime soon.

.with <expression>
Namespace access

This directive is similar to .namespace but it gives access to another scope's variables without leaving the current scope. May be useful to allow a short hand access in some situations.

It's advised to use the -Wshadow command line option to warn about any unexpected symbol ambiguity.


Sections can be used to collect data or code into separate memory areas without moving source code lines around. This is achieved by having separate compile offset and program counters for each defined section.

.section <name>
Starts a segment block
.send [<name>]
.endsection [<name>]
Ends a segment block

Defines a section fragment. The name at .endsection must match but it's optional.

.dsection <name>
Collect the section fragments here.

All .section fragments are compiled to the memory area allocated by the .dsection directive. Compilation happens as the code appears, this directive only assigns enough space to hold all the content in the section fragments.

The space used by section fragments is calculated from the difference of starting compile offset and the maximum compile offset reached. It is possible to manipulate the compile offset in fragments, but putting code before the start of .dsection is not allowed.

*       = $02
        .dsection zp   ;declare zero page section
        .cerror * > $30, "Too many zero page variables"

*       = $334
        .dsection bss   ;declare uninitialized variable section
        .cerror * > $400, "Too many variables"

*       = $0801
        .dsection code   ;declare code section
        .cerror * > $1000, "Program too long!"

*       = $1000
        .dsection data   ;declare data section
        .cerror * > $2000, "Data too long!"
        .section code
        .word ss, 2005
        .null $9e, format("%4d", start)
ss      .word 0

start   sei
        .section zp     ;declare some new zero page variables
p2      .addr ?         ;a pointer
        .endsection zp
        .section bss    ;new variables
buffer  .fill 10        ;temporary area
        .endsection bss

        lda (p2),y
        lda #<label
        ldy #>label
        jsr print

        .section data   ;some data
label   .null "message"
        .endsection data

        jmp error
        .section zp     ;declare some more zero page variables
p3      .addr ?         ;a pointer
        .endsection zp
        .endsection code

The compiled code will look like:

>0801    0b 08 d5 07                            .word ss, 2005
>0805    9e 32 30 36 31 00                      .null $9e, format("%4d", start)
>080b    00 00                          ss      .word 0

.080d    78                             start   sei

>0002                                   p2      .addr ?         ;a pointer
>0334                                   buffer  .fill 10        ;temporary area

.080e    b1 02                                  lda (p2),y
.0810    a9 00                                  lda #<label
.0812    a0 10                                  ldy #>label
.0814    20 1e ab                               jsr print

>1000    6d 65 73 73 61 67 65 00        label   .null "message"

.0817    4c e2 fc                               jmp error

>0004                                   p2      .addr ?         ;a pointer

Sections can form a hierarchy by nesting a .dsection into another section. The section names must only be unique within a section but can be reused otherwise. Parent section names are visible for children, siblings can be reached through parents.

In the following example the included sources don't have to know which code and data sections they use, while the bss section is shared for all banks.

;First 8K bank at the beginning, PC at $8000
*       = $0000
        .logical $8000
        .dsection bank1
        .cerror * > $a000, "Bank1 too long"

bank1   .block          ;Make all symbols local
        .section bank1
        .dsection code  ;Code and data sections in bank1
        .dsection data
        .section code   ;Pre-open code section
        .include "code.asm"; see below
        .include "iter.asm"
        .endsection code
        .endsection bank1

;Second 8K bank at $2000, PC at $8000
*       = $2000
        .logical $8000
        .dsection bank2
        .cerror * > $a000, "Bank2 too long"

bank2   .block          ;Make all symbols local
        .section bank2
        .dsection code  ;Code and data sections in bank2
        .dsection data
        .section code   ;Pre-open code section
        .include "scr.asm"
        .endsection code
        .endsection bank2

;Common data, avoid initialized variables here!
*       = $c000
        .dsection bss
        .cerror * > $d000, "Too much common data"
;------------- The following is in "code.asm"
code    sei

        .section bss   ;Common data section
buffer  .fill 10
        .endsection bss

        .section data  ;Data section (in bank1)
routine .addr print
        .endsection bss

65816 related

Select short (8 bit) or long (16 bit) accumulator immediate constants.
        lda #$4322
Select short (8 bit) or long (16 bit) index register immediate constants.
        ldx #$1000
Select automatic adjustment of immediate constant sizes based on SEP/REP instructions.
        rep #$10        ;implicit .xl
        ldx #$1000
.databank <expression>
Data bank (absolute) addressing is only used for addresses falling into this 64 KiB bank. The default is 0, which means addresses in bank zero.

When data bank is switched off only data bank indexed (,b) addresses create data bank accessing instructions.

        .databank $10   ;data bank at $10xxxx
        lda $101234     ;results in $ad, $34, $12
        .databank ?     ;no data bank
        lda $1234       ;direct page or long addressing
        lda #$1234,b    ;results in $ad, $34, $12
.dpage <expression>
Direct (zero) page addressing is only used for addresses falling into a specific 256 byte address range. The default is 0, which is the first page of bank zero.

When direct page is switched off only the direct page indexed (,d) addresses create direct page accessing instructions.

        .dpage $400     ;direct page $400-$4ff
        lda $456        ;results in $a5, $56
        .dpage ?        ;no direct page
        lda $56         ;data bank or long addressing
        lda #$56,d      ;results in $a5, $56

Controlling errors

.option allow_branch_across_page
Switches error generation on page boundary crossing during relative branch. Such a condition on 6502 adds 1 extra cycle to the execution time, which can ruin the timing of a carefully cycle counted code.
        .option allow_branch_across_page = 0
        bcc +           ;same execution time
        inx             ;needed in both cases
+       bcs +
+       .option allow_branch_across_page = 1
.error <message> [, <message>, …]
.cerror <condition>, <message> [, <message>, …]
Exit with error or conditionally exit with error
        .error "Unfinished here..."
        .cerror * > $1200, "Program too long by ", * - $1200, " bytes"
.warn <message> [, <message>, …]
.cwarn <condition>, <message> [, <message>, …]
Display a warning message always or depending on a condition
        .warn "FIXME: handle negative values too!"
        .cwarn * > $1200, "This may not work!"


.cpu <expression>
Selects CPU according to the string argument.
        .cpu "6502"     ;standard 65xx
        .cpu "65c02"    ;CMOS 65C02
        .cpu "65ce02"   ;CSG 65CE02
        .cpu "6502i"    ;NMOS 65xx
        .cpu "65816"    ;W65C816
        .cpu "65dtv02"  ;65dtv02
        .cpu "65el02"   ;65el02
        .cpu "r65c02"   ;R65C02
        .cpu "w65c02"   ;W65C02
        .cpu "4510"     ;CSG 4510
        .cpu "default"  ;cpu set on command line


Terminate assembly. Any content after this directive is ignored.
.eor <expression>
XOR output with an 8 bit value. Useful for reverse screen code text for example, or for silly encryption.
.seed <expression>
Seed the pseudo random number generator with an unsigned integer of maximum 128 bits to make the generated numbers less boring.
.var <expression>
Defines a variable identified by the label preceding, which is set to the value of expression or reference of variable.
counter .var 0                ;define, same as :=
counter .var counter + 1      ;redefine, same as += 1
.from <scope>
Defines a symbol to the value of the same symbol from another scope.

This directive looks up the symbol name to be defined in the other scope for its value. Useful for shorthand definitions without repeating the name if it's unchanged.

red     .from vic.colors      ;same as red = vic.colors.red

init    .from +               ;expose these symbols publicly
play    .from +
+       .block                ;other symbols hidden in block
init    sei
Do not use these, the syntax will change in next version!

Printer control

Turn on or off source listing on part of the file.
        .proff           ;Don't put filler bytes into listing
*       = $8000
        .fill $2000, $ff ;Pre-fill ROM area
*       = $8000
        .addr reset, restore
        .text "CBM80"
reset   cld
Ignored for compatibility.

Pseudo instructions


For better code readability BCC has an alias named BLT (Branch Less Than) and BCS one named BGE (Branch Greater Equal).

        cmp #3
        blt exit        ; less than 3?

For similar reasons ASL has an alias named SHL (SHift Left) and LSR one named SHR (SHift Right). This naming however is not very common.

The implied variants LSR, ROR, ASL and ROL are a shorthand for LSR A, ROR A, ASL A and ROL A. Using the implied form is considered poor coding style.

For compatibility INA and DEA is a shorthand of INC A and DEC A. Therefore there's no implied variants like INC or DEC. The full form with the accumulator is preferred.

The longer forms of INC X, DEC X, INC Y, DEC Y, INC Z and DEC Z are available for INX, DEX, INY, DEY, INZ and DEZ. For this to work care must be taken to not reuse the x, y and z single letter register symbols for other purposes. Same goes for a of course.

Load instructions with registers are translated to transfer instructions. For example LDA X becomes TXA.

Store instructions with registers are translated to transfer instructions, but only if it involves the s or b registers. For example STX S becomes TXS.

Many illegal opcodes have aliases for compatibility as there's no standard naming convention.

Always taken branches

For writing short code there are some special pseudo instructions for always taken branches. These are automatically compiled as relative branches when the jump distance is short enough and as JMP or BRL when longer.

The names are derived from conditional branches and are: GEQ, GNE, GCC, GCS, GPL, GMI, GVC, GVS, GLT and GGE.

.0000    a9 03          lda #$03        in1     lda #3
.0002    d0 02          bne $0006               gne at          ;branch always
.0004    a9 02          lda #$02        in2     lda #2
.0006    4c 00 10       jmp $1000       at      gne $1000       ;branch farther

If the branch would skip only one byte then the opposite condition is compiled and only the first byte is emitted. This is now a never executed jump, and the relative distance byte after the opcode is the jumped over byte. If the CPU has long conditional branches (65CE02/4510) then the same method is applied to two byte skips as well.

There's a pseudo opcode called GRA for CPUs supporting BRA, which is expanded to BRL (if available) or JMP. A one byte skip will be shortened to a single byte if the CPU has a NOP immediate instruction (R65C02/W65C02).

If the branch would not skip anything at all then no code is generated.

.0009                                           geq in3         ;zero length "branch"
.0009    18             clc             in3     clc
.000a    b0             bcs                     gcc at2         ;one byte skip, as bcs
.000b    38             sec             in4     sec             ;sec is skipped!
.000c    20 0f 00       jsr $000f       at2     jsr func
.000f                                   func

Please note that expressions like Gxx *+2 or Gxx *+3 are not allowed as the compiler can't figure out if it has to create no code at all, the 1 byte variant or the 2 byte one. Therefore use normal or anonymous labels defined after the jump instruction when jumping forward!

Long branches

To avoid branch too long errors the assembler also supports long branches. It can automatically convert conditional relative branches to it's opposite and a JMP or BRL. This can be enabled on the command line using the --long-branch option.

.0000    ea             nop                     nop
.0001    b0 03          bcs $0006               bcc $1000      ;long branch (6502)
.0003    4c 00 10       jmp $1000
.0006    1f 17 03       bbr 1,$17,$000c         bbs 1,23,$1000 ;long branch (R65C02)
.0009    4c 00 10       jmp $1000
.000c    d0 04          bne $0012               beq $10000     ;long branch (65816)
.000e    5c 00 00 01    jmp $010000
.0012    30 03          bmi $0017               bpl $1000      ;long branch (65816)
.0014    82 e9 lf       brl $1000
.0017    ea             nop                     nop

Please note that forward jump expressions like Bxx *+130, Bxx *+131 and Bxx *+132 are not allowed as the compiler can't decide between a short/long branch. Of course these destinations can be used, but only with normal or anonymous labels defined after the jump instruction.

In the above example extra JMP instructions are emitted for each long branch. This is suboptimal and wasting space if there are several long branches to the same location in close proximity. Therefore the assembler might decide to reuse a JMP for more than one long branch to save space.

Original turbo assembler compatibility

How to convert source code for use with 64tass

Currently there are two options, either use TMPview by Style to convert the source file directly, or do the following:

The resulting file should then (with the restrictions below) assemble using the following command line:

64tass -C -T -a -W -i source.asm -o outfile.prg

Differences to the original turbo ass macro on the C64

64tass is nearly 100% compatible with the original Turbo Assembler, and supports most of the features of the original Turbo Assembler Macro. The remaining notable differences are listed here.


The original turbo assembler uses case sensitive labels, use the --case-sensitive command line option to enable this behaviour.

Expression evaluation

There are a few differences which can be worked around by the --tasm-compatible command line option. These are:

The original expression parser has no operator precedence, but 64tass has. That means that you will have to fix expressions using braces accordingly, for example 1+2*3 becomes (1+2)*3.

The following operators used by the original Turbo Assembler are different:

TASM Operator differences
.bitwise or, now |
:bitwise eor, now ^
!force 16 bit address, now @w

The default expression evaluation is not limited to 16 bit unsigned numbers anymore.


Macro parameters are referenced by \1\9 instead of using the pound sign.

Parameters are always copied as text into the macro and not passed by value as the original turbo assembler does, which sometimes may lead to unexpected behaviour. You may need to make use of braces around arguments and/or references to fix this.


Some versions of the original turbo assembler had bugs that are not reproduced by 64tass, you will have to fix the code instead.

In some versions labels used in the first .block are globally available. If you get a related error move the respective label out of the .block.

Command line options

Short command line options consist of - and a letter, long options start with --.

If -- is encountered then further options are not recognized and are assumed to be file names.

Options requiring file names are marked with <filename>. A single - as name means standard input or output. File name quoting is system specific.

Output options

-o <filename>, --output <filename>
Place output into <filename>. The default output filename is a.out. This option changes it.
64tass a.asm -o a.prg

This option may be used multiple times to output different sections in different formats of a single compilation.

For multiple outputs the format options and output section selection must be placed before this option. The format selection will be unchanged if no new selection was made but the output section selection and the map file must be repeated for each output. The maximum image size will be the smallest of all selected formats. Using the same name multiple times is not a good idea.

--output-append <filename>
Same as the --output option but appends instead of overwrites.

Normally output files are overwritten but in some cases it's useful to append them instead.

No output file will be written.

Useful for test compiles.

--map <file>
Specify memory map output file.

Normally the memory map is displayed on the standard output together with other messages. It's possible to write it to a file or to the standard output by using - as the file name.

--map-append <filename>
Same as the --map option but appends instead of overwrites.
Do not display or record the memory map.
--output-section <sectionname>
Specify which section to write in the output.

By default all sections go into the output file. Using this option limits the output to specific section and it's children. This is useful to split a larger program into several files.

64tass a.asm --output-section main -o main.prg \
             --output-section loader -o loader.prg
--output-exec <expression>
Sets execution address for output formats which support this.

While it's possible to enter the address as a number it's recommended to use a label instead.

-X, --long-address
Use 3 byte address/length for CBM and nonlinear output instead of 2 bytes. Also increases the size of raw output to 16 MiB.
64tass --long-address --m65816 a.asm
Generate CBM format binaries (default)

Overlapping blocks are flattened and uninitialized memory is filled up with zeros. Uninitialized memory before the first and after the last valid bytes are not saved. Up to 64 KiB or 16 MiB with long address.

Used for C64 binaries. The first 2 bytes are the little endian address of the first valid byte (load address). This is followed by the data.

64tass --cbm-prg a.asm
*       = $2000
start   rts
Example CBM format binary output
00 20load to $2000
-b, --nostart
Output raw binary data.

Overlapping blocks are flattened and uninitialized memory is filled up with zeros. Uninitialized memory before the first and after the last valid bytes are not saved. Up to 64 KiB or 16 MiB with long address.

Useful for small ROM files.

64tass --nostart a.asm
*       = $2000
Example raw output
-f, --flat
Flat address space output mode.

Overlapping blocks are flattened and uninitialized memory is filled up with zeros. Uninitialized memory after the last valid byte is not saved. Up to 4 GiB.

Useful for creating huge multi bank ROM files. See sections for an example.

-n, --nonlinear
Generate nonlinear output file.

Overlapping blocks are flattened. Blocks are saved in sorted order and uninitialized memory is skipped. Up to 64 KiB or 16 MiB with long address.

Used for linkers and downloading. Before writing each memory block the length and the memory address is saved in a little endian order. Once everything was saved a zero length block is written without an address or data. These zeros serve as an end marker.

64tass --nonlinear a.asm
*       = $1000          ;1st segment
        lda #2
*       = $2000          ;2nd segment
Example 64 KiB nonlinear output
02 00load 2 bytes
00 10to $1000
a9 02data
01 00load 1 byte
00 20to $2000
00 00load 0 bytes, end marker
Generate an Atari XEX output file.

Overlapping blocks are kept, continuing blocks are concatenated. Saving happens in the definition order without sorting, and uninitialized memory is skipped in the output. Up to 64 KiB.

Used for Atari executables. First 2 bytes of signature is written. Then before saving each memory block the words of load address and last byte address is written in little endian format.

If the --output-exec command line parameter was given then a 6 byte run block is added to the end of the output.

64tass --output-exec=start --atari-xex a.asm
*       = $2000
start   rts
Example Atari XEX format output
ff ffheader
00 20load to $2000
00 20until $2000
e0 02 e1 02run marker
00 20run address ($2000)
Generate an Apple II output file (DOS 3.3).

Overlapping blocks are flattened and uninitialized memory is filled up with zeros. Uninitialized memory before the first and after the last valid bytes are not saved. Up to 64 KiB.

Used for Apple II executables. First the load address and the data length words are written in little endian format. This is followed by the data.

64tass --apple-ii a.asm
*       = $0c00
Example of Apple II format output
00 0cload to $0c00
01 00length is 1 byte
Generate C256 Foenix PGX output file.

Overlapping blocks are flattened and uninitialized memory is filled up with zeros. Uninitialized memory before the first and after the last valid bytes are not saved. Up to 16 MiB.

Used for single segment C256 Foenix executables. After the PGX signature a four byte little endian load address is written. This is followed by the data.

64tass --c256-pgx a.asm
*       = $1000
Example PGX format output
50 47 58 01PGX signature
00 10 00 00load to $1000
Generate C256 Foenix PGZ output file.

Overlapping blocks are flattened. Blocks are saved in sorted order and uninitialized memory is skipped. Up to 16 MiB.

Used for multi segment C256 Foenix binaries. It starts with a single byte signature. Then before each memory block a three byte load address and length is written in little endian format.

If the --output-exec command line parameter was given then a 6 byte execution block is added to the end of the output.

64tass --output-exec=start --c256-pgz a.asm
*       = $1000
start   rts
Example PGZ format output
5aPGZ signature byte
00 10 00load to $1000
01 00 00length is 1 bytes
00 10 00execute at $1000
00 00 00execution marker
Use Intel HEX output file format.

Overlapping blocks are kept, data is stored in the definition order, and uninitialized areas are skipped. I8HEX up to 64 KiB, I32HEX up to 4 GiB.

Used for EPROM programming or downloading. Data bytes are written using 00 records. If the file is larger than 64 KiB then 04 records are used as needed. The output ends with a 01 record.

If the --output-exec command line parameter was given then a 05 record is added with the execution address right before the end 01 record.

64tass --intel-hex a.asm
*       = $0c00

Example Intel HEX output:

Use MOS Technology output file format. Also known as Paper Tape Format.

Overlapping blocks are kept, data is stored in the definition order, and uninitialized areas are skipped. Up to 64 KiB.

64tass --mos-hex a.asm
*       = $0c00

Example MOS Technology output:

Use Motorola S-record output file format.

Overlapping blocks are kept, data is stored in the definition order, and uninitialized memory areas are skipped. S19 up to 64 KiB, S28 up to 16 MiB and S37 up to 4 GiB.

Used for EPROM programming or downloading. First a S0 header record is written which is followed by S1, S2, or S3 data records. Then an S5 or S6 count record comes and a S9, S8 or S7 termination record.

If the --output-exec command line parameter was given then the termination record will use this address. Without this the address of the first data record is used.

64tass --s-record a.asm
*       = $0c00

Example Motorola S-record output:


Operation options

-a, --ascii
Use ASCII/Unicode text encoding instead of raw 8-bit

Normally no conversion takes place, this is for backwards compatibility with a DOS based Turbo Assembler editor, which could create PETSCII files for 6502tass. (including control characters of course)

Using this option will change the default none and screen encodings to map 'a'–'z' and 'A'–'Z' into the correct PETSCII range of $41–$5A and $C1–$DA, which is more suitable for an ASCII editor. It also adds predefined petcat style PETSCII literals to the default encodings, and enables Unicode letters in symbol names.

For writing sources in UTF-8/UTF-16 encodings this option is required!

64tass a.asm

.0000    a9 61          lda #$61        lda #"a"

>0002    31 61 41                       .text "1aA"
>0005    7b 63 6c 65 61 72 7d 74        .text "{clear}text{return}more"
>000e    65 78 74 7b 72 65 74 75
>0016    72 6e 7d 6d 6f 72 65

64tass --ascii a.asm

.0000    a9 41          lda #$41        lda #"a"
>0002    31 41 c1                       .text "1aA"
>0005    93 54 45 58 54 0d 4d 4f        .text "{clear}text{return}more"
>000e    52 45
-B, --long-branch
Automatic BXX *+5 JMP xxx. Branch too long messages are usually solved by manually rewriting them as BXX *+5 JMP xxx. 64tass can do this automatically if this option is used. BRA is of course not converted.
64tass a.asm
*       = $1000
        bcc $1233       ;error...

64tass a.asm
*       = $1000
        bcs *+5         ;opposite condition
        jmp $1233       ;as simple workaround

64tass --long-branch a.asm
*       = $1000
        bcc $1233       ;no error, automatically converted to the one above
        bcs @b $1233    ;keep this one short regardless and fail if too far
-C, --case-sensitive
Make all symbols (variables, opcodes, directives, operators, etc.) case sensitive. Otherwise everything is case insensitive by default.
64tass a.asm
label   nop
Label   nop     ;double defined...

64tass --case-sensitive a.asm
label   nop
Label   nop     ;Ok, it's a different label...
-D <label>=<value>
Command line definition.

Same syntax is allowed as in source files. Be careful with strings, the shell might eat the quotes unless escaped.

Using hexadecimal numbers might be tricky as the shell might try to expand them as variables. Either quoting or backslash escaping might help.

In Makefiles all $ signs need to be escaped by doubling them. This needs to be done over the normal shell escaping. For example $1000 becomes \$$1000.

64tass -D ii=2 -D var=\"string\" -D FAST:=true a.asm
        lda #ii ;result: $a9, $02
FAST    :?= false ;define if undefined
-q, --quiet
Suppress messages. Disables header and summary messages.
64tass --quiet a.asm
-T, --tasm-compatible
Enable TASM compatible operators and precedence

Switches the expression evaluator into compatibility mode. This enables ., : and ! operators and disables 64tass specific extensions, disables precedence handling and forces 16 bit unsigned evaluation (see differences to original Turbo Assembler below)

-I <path>
Specify include search path

If an included source or binary file can't be found in the directory of the source file then this path is tried. More than one directories can be specified by repeating this option. If multiple matches exist the first one is used.

-M <file>, --dependencies <file>
Specify make rule output file

Writes a dependency rules suitable for make from the list of files used during compilation.

Please choose source file names which are compatible with Makefiles as there are very little escaping possibilities.

--dependencies-append <file>
Same as the --dependencies option but appends instead of overwrites.
Enable phony targets for dependencies

This is useful for automatic dependency generation to avoid missing target errors on file rename.

The following Makefile uses the rules generated by 64tass (in .dep) to achieve automatic dependency tracking:

demo: demo.asm .dep
        64tass --make-phony -M.dep $< -o $@

-include .dep

Diagnostic options

Diagnostic message switched start with a -W and can have an optional no- prefix to disable them. The options below with this prefix are enabled by default, the others are disabled.

-E <file>, --error <file>
Specify error output file

Normally compilation errors a written to the standard error output. It's possible to redirect them to a file or to the standard output by using - as the file name.

--error-append <filename>
Same as the --error option but appends instead of overwrites.
Do not output any error messages, just count them.
-w, --no-warn
Suppress warnings.

Disables warnings during compile. For fine grained diagnostic message suppression see the diagnostic options section.

64tass --no-warn a.asm
Suppress displaying of faulty source line and fault position after fault messages.

This is for cases where the fault log is automatically processed and no one ever looks at it and therefore there's no point to display the source lines.

64tass --no-caret-diag a.asm
Restrict source line and fault position display to macro expansions only.

This is for cases where the fault log is processed by an editor which also displays the compilation output somewhere. Only lines which are the result of macro processing will be output to aid debugging. Those which would just duplicate what's in the source editor window will be not.

64tass --macro-caret-diag a.asm
Enable most diagnostic warnings, except those individually disabled. Or with the no- prefix disable all except those enabled.
Make all diagnostic warnings to an error, except those individually set to a warning.
Change a diagnostic warning to an error.

For example -Werror=implied-reg makes this check an error. The -Wno-error= variant is useful with -Werror to set some to warnings.

Warns about alias opcodes.

There are several opcodes for the same task, especially for the "6502i" target. This warning helps to find where their use.

Warns when padding bytes were used for alignment.

Can be used to see where space is wasted for alignment.

Warn about alternative address modes.

Sometimes alternative addressing modes are used as the fitting one is not available. For example there's no lda direct page y so instead data bank y is used with a warning.

Warns if a branch is crossing a page.

Page crossing branches execute with a penalty cycle. This option helps to locate them easily.

Warn if symbol letter case is used inconsistently.

This option can be used to enforce letter case matching of symbols in case insensitive mode. This gives similar results to the case sensitive mode (symbols must match exactly) with the main difference of disallowing symbol name definitions differing only in case (these are reported as duplicates).

Warns for cases where immediate addressing is more likely.

It may be hard to notice if a # was missed. The code still compiles but there's a huge difference between cpx #const and cpx const. Unless the right sort of garbage was on zero page at the time of testing...

This check might have a lot of false positives if zero page locations are accessed by using small numbers, which is a popular coding style. But there are ways to reduce them.

For "known" fixed locations address(x) can be used, preferably bound to a symbol. Automatic allocation of zero page variables works too (e.g. zpstuff .byte ?). And basically everything which is a traditional "label" or derived from a label with an offset.

Warns if implied addressing is used instead of register.

Some instructions have implied aliases like asl for asl a for compatibility reasons, but this shorthand is not the preferred form.

Warns if about leading zeros.

A leading zero could be a prefix for an octal number but as octals are not supported the result will be decimal.

Warns when a long branch is used.

This option gives a warning for instructions which were modified by the long branch function. Less intrusive than disabling long branches and see where it fails.

Warn about macro call without prefix.

Such macro calls can easily be mistaken to be labels if invoked without parameters. Also it's hard to notice that an unchanged call turned into label after the definition got renamed. This warning helps to find such calls so that prefixes can be added.

Don't warn about deprecated features.

Unfortunately there were some features added previously which shouldn't have been included. This option disables warnings about their uses.

Don't warn if floating point comparisons are only approximate.

Floating point numbers have a finite precision and comparing them might give unexpected results.

For example 2.1 + 0.2 == 2.3 is true but gives a warning as the left side is actually bigger by approximately 4.44E−16.

Normally this is solved by rounding or changing the comparison values.

Warn when floating point numbers are implicitly rounded.

A lot of parameters and the data dumping directives need integers but floating point numbers are accepted as well. The style of rounding used may or may not be what you wanted.

By default floor rounding (to lower) is used and not truncate (towards zero). The reason for this is to enable calculation of fixed point integers by using floating point.

The difference is subtle and only noticable for negative numbers. The division of -300/256 is -2 which matches floor(-300/256.0) but not trunc(-300/256.0).

To get symmetric sine waves around zero trunc() needs to be used. Some other calculation might result in 126.9999997 due to inaccuracies in logarithm which would need round().

To avoid unexpected rounding this option helps to find those places where no explicit rounding was done.

Don't warn about ignored directives.
Don't warn about the jmp ($xxff) bug.

With this option it's fine that the high byte is read from the wrong address on a 6502, NMOS 6502 and 65DTV02.

        jmp (vector)
        .alignpageind vector, 256; jmp bug workaround
vector  .addr ?         ; by avoiding page cross
Don't warn about certain labels not being on left side.

You may disable this if you use labels which look like mistyped versions of implied addressing mode instructions and you don't want to put them in the first column.

This check is there to catch typos, unsupported implied instructions, or unknown aliases and not for enforcing label placement.

Ignore page assertion failures

Can be used to ignore .page assertion block failures. As a middle ground -Wno-error=page can turn the assertion to a warning only.

Don't note about common pitfalls.

There are some common mistakes, but experts and those who read this don't need extra notes about them. These are:

Use multi character strings with .byte instead of .text.
This fails because .byte enforces the 0–255 range for each value.
Using label *=*+1 style space reservations.
Warns as *= is also the compound multiply operator. The *=*+1 needs to be on a separate line without a label. A better alternatively is to use .fill 1 or .byte ?.
Negative numbers with .byte or .word
There are other directives which accept them with proper range checks like .char, .sint.
Negative numbers with lda #xxx
There's a signed variant for the immediate addressing so lda #+xx will make it work
Don't warn about source portability problems.

These cross platform development annoyances are checked for:

  • Case insensitive use of file names or use of short names.
  • Use of backslashes for path separation instead of forward slashes.
  • Use of reserved characters in file names.
  • Absolute paths
Don't warn about operator priority problems.

Not all of the unary operators are strongly binding and this may cause surprises. This warning is intended to catch mistakes like this:

        .cerror >start != >end; possibly wrong it's >(start != (>end))
        .cerror (>start) != >end; correct high byte check
Don't warn if size is larger due to negative offset

size() and len() can be used to measure a memory area. Normally there's no offset used but a positive offset may be used to reduce available space up until nothing remains.

On the other hand if a negative offset is used then more space will be available (ahead of the area) which may or may not be desired.

var     .byte ?, ?, ?
var2    = var - 2       ; start 2 bytes earlier
        ldx #size(var2) ; size is 5 bytes as it's 2 bytes ahead
Don't warn about ignored compound multiply.

Normally symbol *= ... means compound multiply of the variable in front. Unfortunately this looks the same a label *=*+x which is an old-school way to allocate space.

If the symbol was a variable defined earlier then the multiply is performed without a warning. If it's a new label definition then this warning is used to note that possibly a variable definition was missed earlier.

If the intention was really a label definition then the *= can be moved to a separate line, or in case of space allocation it could be improved to use .byte ? or .fill x.

Don't warn about memory location address space wrap around.

Applying offsets to memory locations may result in addresses which end up outside of the processors address space.

For example "tmp" is at $1000 and then it's addressed as lda tmp-$2000 then the result will be lda $f000 or lda $fff000 depending on the CPU. If this is fine then this warning can be disabled otherwise it can be made into an error by using -Werror=wrap-addr.

Don't warn for bank 0 wrap around.

Adding an offset to a bank 0 address may end up outside of bank 0. If this happens a warning is issued and the address wraps around.

The warning may be ignored using this command line parameter. Alternatively it could be turned into an error by using -Werror=wrap-bank0.

Don't warn for direct page wrap around.

Adding an offset to a direct page address may end up outside of the direct page. For a 65816 or 65EL02 an alternative addressing mode is used but on other processors if this happens a warning is issued and the address wraps around.

The warning may be ignored using this command line parameter. Alternatively it could be turned into an error by using -Werror=wrap-dpage.

Don't warn for compile offset wrap around.

While assembling the compile offset may reach the end of memory image. If this happens a warning is issued and the compile offset is set to the start of image.

The warning may be ignored using this command line parameter. Alternatively it could be turned into an error by using -Werror=wrap-mem.

The image size depends on the output format. See the Output options section above.

Don't warn for program counter bank crossing.

While assembling the program counter may reach the end of the current program bank. If this happens a warning is issued as a real CPU will not cross the bank on execution. On the other hand some addressing modes handle bank crosses so this might not be actually a problem for data.

The warning may be ignored using this command line parameter. Alternatively it could be turned into an error by using -Werror=wrap-pc.

Don't warn for program bank address calculation wrap around.

Adding an offset to a program bank address may end up outside of the current program bank. If this happens a warning is issued and the address wraps around.

The warning may be ignored using this command line parameter. Alternatively it could be turned into an error by using -Werror=wrap-pbank.

Warn about old equal operator.

The single = operator is only there for compatibility reasons and should be written as == normally.

Warn about optimizable code.

Warns on things that could be optimized, at least according to the limited analysis done. Currently it's easy to fool with these constructs:

  • Self modifying code, especially modifying immediate addressing mode instructions or branch targets
  • Using .byte $2c and similar tricks to skip instructions.
  • Using *+5 and similar tricks to skip instructions, or to loop like *-1.
  • Any other method of flow control not involving referenced labels. E.g. calculated returns.
  • Register re-mappings on 65DTV02 with SIR and SAC.

It's also rather simple and conservative, so some opportunities will be missed. Most CPUs are supported with the notable exception of 65816 and 65EL02, but this could improve in later versions.

Warn about symbol shadowing.

Checks if local variables shadow other variables of same name in upper scopes in ambiguous ways.

This is useful to detect hard to notice bugs where a new local variable takes the place of a global one by mistake.

bl      .block
a       .byte 2         ;'a' is a built-in register
x       .byte 2         ;'x' is a built-in register
        asl a           ; accumulator or the byte above?
        asl bl.x        ; not ambiguous
Warn about implicit boolean conversions.

Boolean values can be interpreted as numeric 0/1 and other types as booleans. This is convenient but may cause mistakes.

To pass this option the following constructs need improvements:

  • 1 and 0 as boolean constants. Use the slightly longer true and false.
  • Implicit non-zero checks. Write it out like .if (lbl & 1) != 0.
  • Zero checks with !. Write it out like lbl == 0.
  • Binary operators on booleans. Use the proper ||, && and ^^ operators.
  • Numeric expressions like 1 + (lbl > 3). It's better as (lbl > 3) ? 2 : 1.
Warn about unused constant symbols.

Symbols which have no references to them are likely redundant. Before removing them check if there's any conditionally compiled out code which might still need them.

The following options can be used to be more specific:

Warn about unused constants.
Warn about unused labels.
Warn about unused macros.
Warn about unused variables.

Symbols which appear in a default 64tass symbol list file and their root symbols are treated as used for exporting purposes.

Target selection on command line

These options will select the default architecture. It can be overridden by using the .cpu directive in the source.

Standard 65xx (default). For writing compatible code, no extra codes. This is the default.
64tass --m65xx a.asm
        lda $14         ;regular instructions
-c, --m65c02
CMOS 65C02. Enables extra opcodes and addressing modes specific to this CPU.
64tass --m65c02 a.asm
        stz $d020       ;65c02 instruction
CSG 65CE02. Enables extra opcodes and addressing modes specific to this CPU.
64tass --m65ce02 a.asm
-i, --m6502
NMOS 65xx. Enables extra illegal opcodes. Useful for demo coding for C64, disk drive code, etc.
64tass --m6502 a.asm
        lax $14         ;illegal instruction
-t, --m65dtv02
65DTV02. Enables extra opcodes specific to DTV.
64tass --m65dtv02 a.asm
        sac #$00
-x, --m65816
W65C816. Enables extra opcodes. Useful for SuperCPU projects.
64tass --m65816 a.asm
        lda $123456,x
-e, --m65el02
65EL02. Enables extra opcodes, useful RedPower CPU projects. Probably you'll need --nostart as well.
64tass --m65el02 a.asm
        lda #0,r
R65C02. Enables extra opcodes and addressing modes specific to this CPU.
64tass --mr65c02 a.asm
        rmb 7,$20
W65C02. Enables extra opcodes and addressing modes specific to this CPU.
64tass --mw65c02 a.asm
CSG 4510. Enables extra opcodes and addressing modes specific to this CPU. Useful for C65 projects.
64tass --m4510 a.asm

Symbol listing

-l <file>, --labels=<file>
List symbols into <file>.
64tass -l labels.txt a.asm
*       = $1000
label   jmp label

result (labels.txt):
label           = $1000

This option may be used multiple times. In this case the format and root scope options must be placed before this option. Using the same name multiple times is not a good idea.

64tass --vice-labels -l all.l --export-labels --labels-root=export -l myexport.inc source.asm

This writes symbols for VICE into all.l and symbols from scope export into myexport.inc.

Same as the --labels option but appends instead of overwrites.
Specify the scope to list labels from

This option can be used to limit the output to only a subset of labels. The parameter is an expression which must resolve to a namespace. It's usually just the name of a label in the root scope which contains the labels to be listed.

Specify the section to list labels from

This option can be used to limit the output to a section which code labels refer to.

Add prefix to labels

If defined adds a prefix to labels for some formats.

Lists labels in a 64tass readable format. (default)

List labels without any side effects. Usually for display purposes or for later include.

List labels for include in a 64tass readable format.

The difference to normal symbol listing is that 64tass assumes these symbols will be used in another source. In practice this means that any .proc/.endproc blocks appearing in the symbol file will always be compiled even if unused otherwise.

List labels in a VICE readable format.

This format may be used to translate memory locations to something readable in VICE monitor. Therefore simple numeric constants will not show up unless converted to an address first.

VICE symbols may only contain ASCII letters, numbers and underscore. Symbols not meeting this requirement will be omitted.

There's a good chance VICE will complain about already existing labels on import. In the past an attempt was made to filter out such duplicates to eliminate these warnings. However soon it was pointed out that omitted labels are now unavailable for commands like setting breakpoints. As the latter use case is rather more important than some bogus import warnings one has to live with them.

64tass --vice-labels -l labels.l a.asm
*       = $1000
label   jmp label

result (labels.l):
al 1000 .label

For now colons are used as scope delimiter due to a VICE limitation, but this will be changed to dots in the future.

List address like symbols in a VICE readable format including numeric constants.

The normal VICE label list does not include symbols like chrout = $ffd2 or keybuff = 631 as these are numeric constants and not memory addresses.

Of course there are ways around that. For example:

chrout  = address($ffd2)
keybuff = address(631)
*       = $ffd2
chrout  .fill 3
*       = 631
keybuff .fill 10

For those who don't want to waste time on explicitly marking addresses as such there's an easy way out by using this command line option.

The tradeoff is that depending on the coding style the label list will become polluted by non-address constants to various degrees. However if one mostly uses numeric constants for addresses only this may be acceptable.

List labels for debugging.

The output will contain symbol locations and paths.

List labels in a simple label = $x fashion for interoperatibility.

Somewhat limited but much easier to parse than the normal output with all it's data types.

List labels in Mesen format

It's a text file in the following format:


  • type: the prefix set by the --labels-add-prefix command line option
  • range: a single hexadecimal number or two with a dash for a multibyte range
  • label: name of the label

If the --labels-section command line option was given then the range is relative to the section start.

If more than one type of labels need to be listed then the --labels-append command line option can be used to append them.

64tass --mesen-labels @labeloptions.txt a.asm

Option file (labeloptions.txt):
--labels-section=rom --labels-add-prefix=P --labels labels.mlb 
--labels-section=ram --labels-add-prefix=R --labels-append labels.mlb
--labels-section=save --labels-add-prefix=S --labels-append labels.mlb
--labels-section=work --labels-add-prefix=W --labels-append labels.mlb
--labels-section=registers --labels-add-prefix=G --labels-append labels.mlb

Assembly listing

-L <file>, --list=<file>
List into <file>. Dumps source code and compiled code into file. Useful for debugging, it's much easier to identify the code in memory within the source files.
; 64tass Turbo Assembler Macro V1.5x listing file
; 64tass -L list.txt a.asm
; Fri Dec  9 19:08:55 2005

;Offset ;Hex            ;Monitor        ;Source

;******  Processing input file: a.asm

.1000   a2 00           ldx #$00                ldx #0
.1002   ca              dex             loop    dex
.1003   d0 fd           bne $1002               bne loop
.1005   60              rts                     rts

;******  End of listing
Same as the --list option but appends instead of overwrites.
-m, --no-monitor
Don't put monitor code into listing. There won't be any monitor listing in the list file.
; 64tass Turbo Assembler Macro V1.5x listing file
; 64tass --no-monitor -L list.txt a.asm
; Fri Dec  9 19:11:43 2005

;Offset ;Hex            ;Source

;******  Processing input file: a.asm

.1000   a2 00                   ldx #0
.1002   ca              loop    dex
.1003   d0 fd                   bne loop
.1005   60                      rts

;******  End of listing
-s, --no-source
Don't put source code into listing. There won't be any source listing in the list file.
; 64tass Turbo Assembler Macro V1.5x listing file
; 64tass --no-source -L list.txt a.asm
; Fri Dec  9 19:13:25 2005

;Offset ;Hex            ;Monitor

;******  Processing input file: a.asm

.1000   a2 00           ldx #$00
.1002   ca              dex
.1003   d0 fd           bne $1002
.1005   60              rts

;******  End of listing
This option creates a new column for showing line numbers for easier identification of source origin. The line number is followed with an optional colon separated file number in case it comes from a different file then the previous lines.
; 64tass Turbo Assembler Macro V1.5x listing file
; 64tass --line-numbers -L list.txt a.asm
; Fri Dec  9 19:13:25 2005

;Line   ;Offset ;Hex            ;Monitor        ;Source

:1      ;******  Processing input file: a.asm

3       .1000   a2 00           ldx #$00                ldx #0
4       .1002   ca              dex             loop    dex
5       .1003   d0 fd           bne $1002               bne loop
6       .1005   60              rts                     rts

;******  End of listing
By default the listing file is using a tab size of 8 to align the disassembly. This can be changed to other more favorable values like 4. Only spaces are used if 1 is selected. Please note that this has no effect on the source code on the right hand side.
Normally the assembler tries to minimize listing output by omitting "unimportant" lines. But sometimes it's better to just list everything including comments and empty lines.
; 64tass Turbo Assembler Macro V1.5x listing file
; 64tass --verbose-list -L list.txt a.asm
; Fri Dec  9 19:13:25 2005

;Offset ;Hex            ;Monitor        ;Source

;******  Processing input file: a.asm

                                        *       = $1000

.1000   a2 00           ldx #$00                ldx #0
.1002   ca              dex             loop    dex
.1003   d0 fd           bne $1002               bne loop
.1005   60              rts                     rts

;******  End of listing

Other options

-?, --help
Give this help list. Prints help about command line options.
Give a short usage message. Prints short help about command line options.
-V, --version
Print program version

Command line from file

Command line arguments can be read from a file as well. This is useful to store common options for multiple files in one place or to overcome the argument list length limitations of some systems.

The filename needs to be prefixed with an at sign, so @argsfile reads options from argsfile. It will only work if there's not another file named @argsfile. The content is expanded in-place of @argsfile.

Stored options must be separated by white space. Single or double quotes can be used in case file names have white space in their names.

Backslash can be used to escape the character following it and it must be used to escape itself. Single and double quotes need to be escaped if needed for string quoting.

Forward slashes can be used as a portable path separation on all systems.


Faults and warnings encountered are sent to the standard error for logging. To redirect them to a file use the -E command line option. The message format is the following:

<filename>:<line>:<character>: <severity>: <message>

The faulty line will be displayed after the message with a caret pointing to the error location unless this is disabled by using --no-caret-diag option.

a.asm:3:21: error: not defined symbol 'label'
                 lda label
a.asm:3:21: note: searched in the global scope

This is helpful for macro expansions as it displays the processed line which usually looks different to the one in the original source file.

Error buried deep in included files or macros display a backtrace of files after an In file included from text where all the files and positions involved are listed down to the main file.

In file included from main.asm:3:3:
included.asm:2:11: error: not defined symbol 'test'
                #macro1 test
In file included from included.asm:2:3,
macros.asm:3:7: note: original location in an expanded macro was here
                lda test

Messages ending with [-Wxxx] are user controllable. This means that using -Wno-xxx on the command line will silence them and -Werror=xxx will turn them it into a fault. See Diagnostic options for more details.


aligned by ? bytes
alignment was necessary
approximate floating point
floating point comparisons are not exact and the numbers were close but maybe not quite
bank 0 address overflow
the calculated memory location address ended up outside of bank 0 and is now wrapped.
case ignored, value already handled
this value was already used in an earlier case so here it's ignored
compile offset overflow
compile continues at the bottom ($0000) as end of compile area was reached
constant result, possibly changeable to 'lda'
a pre-calculated value could be loaded instead as the result seems to be always the same
could be shorter by using 'xxx' instead
this shorter instruction gives the same result according to the optimizer
could be simpler by using 'xxx' instead
this instruction gives the same result but with less dependencies according to the optimizer
deprecated directive, only for TASM compatible mode
.goto and .lbl should only be used in TASM compatible mode and there are better ways to loop
deprecated equal operator, use '==' instead
single equal sign for comparisons is going away soon, update source
deprecated modulo operator, use '%' instead
double slash for modulo is going away soon, update source
deprecated not equal operator, use '!=' instead
non-standard not equal operators which will stop working in the future, update source
direct page address overflow
the calculated memory location address ended up outside of direct page and is now wrapped.
directive ignored
an assembler directive was ignored for compatibility reasons
expected ? values but got ? to unpack
the number of variables must match the number of values when unpacking
file name uses reserved character '?'
do not use \ : * ? " < > | in file names as some operating systems don't like these
immediate addressing mode suggested
numeric constant was used as an address which was likely meant as an immediate value
implicit floating point rounding
a floating point number with fractional part was used for an integer parameter
independent result, possibly changeable to 'lda'
the result does not seem to depend on the input so it could be just loaded instead
instruction 'xxx' is an alias of 'xxx'
an alternative instruction name was used
label defined instead of variable multiplication for compatibility
move the '*=' construct to a separate line or define the variable first as this construct is ambiguous
label not on left side
check if an instruction name was not mistyped and if the current CPU has it, or remove white space before label
leading zeros ignored
leading zeros in front of decimals are redundant and don't denote an octal number
long branch used
branch distance was too long so long branch was used (bxx *+5 jmp)
memory location address overflow
the calculated memory location address ended up outside of the processors address space
over the boundary by ? bytes, aligned by ? bytes
crossed boundary so alignment was necessary
please separate @b, @w or @l from label or number for future compatibility
future versions will have longer symbols after @ and so will interpret the immediately following numbers and letters as part if the symbol. Please insert a space between @b, @w or @l and the following label or number now to avoid surprises!
please use format("%d", ...) as '^' will change it's meaning
this operator will be changed to mean the bank byte later, please update your sources
possible jmp ($xxff) bug
some 6502 variants read don't increment the high byte on page cross and this may be unexpected
possibly redundant as ...
according to the optimizer this might not be needed
possibly redundant if last 'jsr' is changed to 'jmp'
tail call elimination possibility was detected
possibly redundant indexing with a constant value
the index register used seems to be constant and there's a way to eliminate indexing by a constant offset
processor program counter crossed bank
pc address had crossed into another 64 KiB program bank
program bank address overflow
the calculated memory location address ended up outside of the current program bank and is now wrapped.
symbol case mismatch '?'
the symbol is matching case insensitively but it's not all letters are exactly the same
the file's real name is not '?'
check if all characters match including their case as this is not the real name of the file
unused symbol '?'
this symbol has is not referred anywhere and therefore may be unused
use '/' as path separation '?'
backslash is not a path separator on all systems while forward slash will work independent of the host operating system
use relative path for '?'
file's path is absolute and depends on the file system layout and the source will not compile without the exact same environment


'?' expected
something is missing
? argument is missing
not enough arguments supplied
address in different program bank
this instruction is only limited to access the current bank
address not in processor address space
value larger than current CPU address space
address out of section
moving the address around is fine as long as it does not end up before the start of the section
addressing mode too complex
too much indexing or indirection for a valid address
at least one byte is needed
the expression didn't yield any bytes but it's needed here
block to long for alignment by ? bytes
impossible to align if larger than or equals alignment interval
branch crosses page by ? bytes
page crossing was on branch was detected
branch too far by ? bytes
branches have limited range and this went over by some bytes
can't calculate stable value
somehow it's impossible to calculate this expression
can't calculate this
could not get any value, is this a circular reference?
can't encode character '?' ($xx) in encoding '?'
can't translate character in this encoding as no definition was given
can't get absolute value of
not possible to calculate the absolute value of this type
can't get boolean value of
not possible to determine if this value is true or false
can't get integer value of
this value is not a number
can't get length of
this type has no length
can't get sign of
this type does not have a sign as it's not a number
can't get size of
this type has no size
closing/opening directive '?' not found
couldn't find the other half of block directive pair
at least one feature is provided, which shouldn't be there
conversion of ? '?' to ? is not possible
this type conversion can't be done
crossing of ? byte page by ? bytes
the page directive detected a page cross between start and end directives
division by zero
dividing with zero can't be done
double defined escape
escape sequence already defined in another .edef differently
double defined range
part of a character range was already defined by another .cdef and these ranges can't overlap
duplicate definition
symbol defined more than once
encoded value ? larger than 8 bit
the value for the end of character range for an encoding is too large
empty list not allowed
at least one element is required
empty range not allowed
invalid range but there must be at least one element
empty string not allowed
at least one character is required
expected exactly/at least/at most ? arguments, got ?
wrong number of function arguments used
expression syntax
syntax error
extra characters on line
there's some garbage on the end of line
floating point overflow
infinity reached during a calculation
format character expected
string ended before a format character was found
general syntax
can't do anything with this
index out of range
not enough elements in list
key not in dictionary
key not in the dictionary
label required
a label is mandatory for this directive
larger than original due to negative offset
if a negative offset is used the size gets larger than the original as this effectively adds bytes to the front.
last byte must not be gap
.shift or .shiftl needs a normal byte at the end
logarithm of non-positive number
only positive numbers have a logarithm
macro call without prefix
macro call was found without a prefix and without parameters
more than a single character
no more than a single character is allowed
more than two characters
no more than two characters are allowed
most significant bit must be clear in byte
for .shift and .shiftl only 7 bit "bytes" are valid
must be used within a loop
.break or .continue must be used within a loop
must be defined later
remote alignment must placed before the aligned label
negative number raised on fractional power
can't calculate this
no ? addressing mode for opcode 'xxx'
this addressing mode is not valid for this instruction
not a bank 0 address
value must be a bank zero address
not a data bank address
value must be a data bank address
not a direct page address
value must be a direct page address
not a key and value pair
dictionaries are built from key and value pairs separated by a colon
not a variable
only variables are changeable
not defined '?'
can't find this label at this point
not hashable
the type can't be used as a key in a dictionary
not in range -1.0 to 1.0
the function is only valid in the -1.0 to 1.0 range
not iterable
value is not a list or other iterable object
not measurable as start offset beyond size of original
the applied offset was larger than the original size. For example if size(data) is 2 then size(data + 1) is 1. However size(data + 3) makes no sense as there's no such thing as a negative size.
offset out of range
code offset too much
operands could not be broadcast together with shapes ? and ?
list length must match or must have a single element only
ptext too long by ? bytes
.ptext is limited to 255 bytes maximum
requirements not met
not all features are provided, at least one is missing
reserved symbol name '?'
do not use this symbol name
shadow definition
symbol is defined in an upper scope as well and is used ambiguously
some operation '?' of type '?' and type '?' not possible
can't do this calculation with these values
square root of negative number
can't calculate the square root of a negative number
start ? not on same ? byte page as end ?
the endpage directive detected a mismatch of page to the page directive
too large for a ? bit signed/unsigned integer
value out of range
unexpected character '?'
unexpected control or Unicode character
unknown processor '?'
unknown CPU name
unknown argument name '?'
no parameter argument known like this
unknown format character '?'
no format character known like this
use '?' instead of '?'
wrong sort of character was used. For example a left double quotation mark was used instead of a regular quotation mark.
value needs to be non-negative
only positive numbers or zero is accepted here
wrong type <?>
wrong object type used
zero raised to negative power
can't calculate this
zero value not allowed
do not use zero for example with .null

Fatal errors

can't open file
cannot open file
can't write ? file '?'
cannot write an output file
compilation was interrupted
shows the line where the interruption happened
error reading file
error while reading
file recursion
wrong nesting of .include
function recursion too deep
wrong use of nested functions
macro recursion too deep
wrong use of nested macros
option '?' doesn't allow an argument
command line option doesn't need any argument
option '?' is ambiguous
command line option abbreviation is too short
option '?' not recognized
no such command line option
option '?' requires an argument
command line option needs an argument
out of memory
won't happen ;)
section '?' for output not found
the section given on command line couldn't be found
too many passes
with a carefully crafted source file it's possible to create unresolvable situations but try to avoid this
unknown option '?'
option not known
weak recursion
excessive nesting of .weak


Original 6502tass written for DOS by Marek Matula of Taboo.

It was ported to ANSI C by BigFoot/Breeze. This is when it's name changed to 64tass.

Soci/Singular reworked the code over the years to the point that practically nothing was left from original at this point.

Improved TASS compatibility, PETSCII codes by Groepaz.

Additional code: my_getopt command-line argument parser by Benjamin Sittler, avl tree code by Franck Bui-Huu, ternary tree code by Daniel Berlin, snprintf Alain Magloire, Amiga OS4 support files by Janne Peräaho.

Pierre Zero helped to uncover a lot of faults by fuzzing. Also there were a lot of discussions with oziphantom about the need of various features.

Main developer and maintainer: soci at c64.rulez.org

Default translation and escape sequences

Raw 8-bit source

By default raw 8-bit encoding is used and nothing is translated or escaped. This mode is for compiling sources which are already PETSCII.

The none encoding for raw 8-bit

Does no translation at all, no translation table, no escape sequences.

The screen encoding for raw 8-bit

The following translation table applies, no escape sequences.

Built-in PETSCII to PETSCII screen code translation table
InputByte InputByte
00–1F80–9F 20–3F20–3F
40–5F00–1F 60–7F40–5F
80–9F80–9F A0–BF60–7F
C0–FE40–7E FF5E

Unicode and ASCII source

Unicode encoding is used when the -a option is given on the command line.

The none encoding for Unicode

This is a Unicode to PETSCII mapping, including escape sequences for control codes.

Built-in Unicode to PETSCII translation table
GlyphUnicodeByte GlyphUnicodeByte
 –@U+0020–U+004020–40 A–ZU+0041–U+005AC1–DA
[U+005B5B ]U+005D5D
a–zU+0061–U+007A41–5A £U+00A35C
πU+03C0FF U+21905F
U+21915E U+2500C0
U+2502DD U+250CB0
U+2510AE U+2514AD
U+2518BD U+251CAB
U+2524B3 U+252CB2
U+2534B1 U+253CDB
U+256DD5 U+256EC9
U+256FCB U+2570CA
U+2571CE U+2572CD
U+2573D6 U+2581A4
U+2582AF U+2583B9
U+2584A2 U+258CA1
U+258DB5 U+258EB4
U+258FA5 U+2592A6
U+2594A3 U+2595A7
U+2596BB U+2597AC
U+2598BE U+259ABF
U+259DBC U+25CBD7
U+25CFD1 U+25E4A9
U+25E5DF U+2660C1
U+2663D8 U+2665D3
U+2666DA U+2713BA
🭰U+1FB70D4 🭱U+1FB71C7
🭲U+1FB72C2 🭳U+1FB73DD
🭴U+1FB74C8 🭵U+1FB75D9
🭶U+1FB76C5 🭷U+1FB77C4
🭸U+1FB78C3 🭹U+1FB79C0
🭺U+1FB7AC6 🭻U+1FB7BD2
🮂U+1FB82B7 🮃U+1FB83B8
🮇U+1FB87AA 🮈U+1FB88B6
🮕U+1FB95FF 🮘U+1FB98DF
Built-in PETSCII escape sequences
EscapeByte EscapeByte EscapeByte
{bell}07 {black}90 {blk}90
{blue}1F {blu}1F {brn}95
{brown}95 {cbm-*}DF {cbm-+}A6
{cbm--}DC {cbm-0}30 {cbm-1}81
{cbm-2}95 {cbm-3}96 {cbm-4}97
{cbm-5}98 {cbm-6}99 {cbm-7}9A
{cbm-8}9B {cbm-9}29 {cbm-@}A4
{cbm-^}DE {cbm-a}B0 {cbm-b}BF
{cbm-c}BC {cbm-d}AC {cbm-e}B1
{cbm-f}BB {cbm-g}A5 {cbm-h}B4
{cbm-i}A2 {cbm-j}B5 {cbm-k}A1
{cbm-l}B6 {cbm-m}A7 {cbm-n}AA
{cbm-o}B9 {cbm-pound}A8 {cbm-p}AF
{cbm-q}AB {cbm-r}B2 {cbm-s}AE
{cbm-t}A3 {cbm-up arrow}DE {cbm-u}B8
{cbm-v}BE {cbm-w}B3 {cbm-x}BD
{cbm-y}B7 {cbm-z}AD {clear}93
{clr}93 {control-0}92 {control-1}90
{control-2}05 {control-3}1C {control-4}9F
{control-5}9C {control-6}1E {control-7}1F
{control-8}9E {control-9}12 {control-:}1B
{control-;}1D {control-=}1F {control-@}00
{control-a}01 {control-b}02 {control-c}03
{control-d}04 {control-e}05 {control-f}06
{control-g}07 {control-h}08 {control-i}09
{control-j}0A {control-k}0B {control-left arrow}06
{control-l}0C {control-m}0D {control-n}0E
{control-o}0F {control-pound}1C {control-p}10
{control-q}11 {control-r}12 {control-s}13
{control-t}14 {control-up arrow}1E {control-u}15
{control-v}16 {control-w}17 {control-x}18
{control-y}19 {control-z}1A {cr}0D
{cyan}9F {cyn}9F {delete}14
{del}14 {dish}08 {down}11
{ensh}09 {esc}1B {f10}82
{f11}84 {f12}8F {f1}85
{f2}89 {f3}86 {f4}8A
{f5}87 {f6}8B {f7}88
{f8}8C {f9}80 {gray1}97
{gray2}98 {gray3}9B {green}1E
{grey1}97 {grey2}98 {grey3}9B
{grn}1E {gry1}97 {gry2}98
{gry3}9B {help}84 {home}13
{insert}94 {inst}94 {lblu}9A
{left arrow}5F {left}9D {lf}0A
{lgrn}99 {lower case}0E {lred}96
{lt blue}9A {lt green}99 {lt red}96
{orange}81 {orng}81 {pi}FF
{pound}5C {purple}9C {pur}9C
{red}1C {return}0D {reverse off}92
{reverse on}12 {rght}1D {right}1D
{run}83 {rvof}92 {rvon}12
{rvs off}92 {rvs on}12 {shift return}8D
{shift-*}C0 {shift-+}DB {shift-,}3C
{shift--}DD {shift-.}3E {shift-/}3F
{shift-0}30 {shift-1}21 {shift-2}22
{shift-3}23 {shift-4}24 {shift-5}25
{shift-6}26 {shift-7}27 {shift-8}28
{shift-9}29 {shift-:}5B {shift-;}5D
{shift-@}BA {shift-^}DE {shift-a}C1
{shift-b}C2 {shift-c}C3 {shift-d}C4
{shift-e}C5 {shift-f}C6 {shift-g}C7
{shift-h}C8 {shift-i}C9 {shift-j}CA
{shift-k}CB {shift-l}CC {shift-m}CD
{shift-n}CE {shift-o}CF {shift-pound}A9
{shift-p}D0 {shift-q}D1 {shift-r}D2
{shift-space}A0 {shift-s}D3 {shift-t}D4
{shift-up arrow}DE {shift-u}D5 {shift-v}D6
{shift-w}D7 {shift-x}D8 {shift-y}D9
{shift-z}DA {space}20 {sret}8D
{stop}03 {swlc}0E {swuc}8E
{tab}09 {up arrow}5E {up/lo lock off}09
{up/lo lock on}08 {upper case}8E {up}91
{white}05 {wht}05 {yellow}9E

The screen encoding for Unicode

This is a Unicode to PETSCII screen code mapping, including escape sequences for control code screen codes.

Built-in Unicode to PETSCII screen code translation table
GlyphUnicodeTranslated GlyphUnicodeTranslated
 –?U+0020–U+003F20–3F @U+004000
A–ZU+0041–U+005A41–5A [U+005B1B
]U+005D1D a–zU+0061–U+007A01–1A
£U+00A31C πU+03C05E
U+21901F U+21911E
U+250040 U+25025D
U+250C70 U+25106E
U+25146D U+25187D
U+251C6B U+252473
U+252C72 U+253471
U+253C5B U+256D55
U+256E49 U+256F4B
U+25704A U+25714E
U+25724D U+257356
U+258164 U+25826F
U+258379 U+258462
U+258C61 U+258D75
U+258E74 U+258F65
U+259266 U+259463
U+259567 U+25967B
U+25976C U+25987E
U+259A7F U+259D7C
U+25CB57 U+25CF51
U+25E469 U+25E55F
U+266041 U+266358
U+266553 U+26665A
U+27137A 🭰U+1FB7054
🭱U+1FB7147 🭲U+1FB7242
🭳U+1FB735D 🭴U+1FB7448
🭵U+1FB7559 🭶U+1FB7645
🭷U+1FB7744 🭸U+1FB7843
🭹U+1FB7940 🭺U+1FB7A46
🭻U+1FB7B52 🭼U+1FB7C4C
🭽U+1FB7D4F 🭾U+1FB7E50
🭿U+1FB7F7A 🮂U+1FB8277
🮃U+1FB8378 🮇U+1FB876A
🮈U+1FB8876 🮌U+1FB8C5C
🮏U+1FB8F68 🮕U+1FB955E
🮘U+1FB985F 🮙U+1FB9969
Built-in PETSCII screen code escape sequences
EscapeByte EscapeByte EscapeByte
{cbm-*}5F {cbm-+}66 {cbm--}5C
{cbm-0}30 {cbm-9}29 {cbm-@}64
{cbm-^}5E {cbm-a}70 {cbm-b}7F
{cbm-c}7C {cbm-d}6C {cbm-e}71
{cbm-f}7B {cbm-g}65 {cbm-h}74
{cbm-i}62 {cbm-j}75 {cbm-k}61
{cbm-l}76 {cbm-m}67 {cbm-n}6A
{cbm-o}79 {cbm-pound}68 {cbm-p}6F
{cbm-q}6B {cbm-r}72 {cbm-s}6E
{cbm-t}63 {cbm-up arrow}5E {cbm-u}78
{cbm-v}7E {cbm-w}73 {cbm-x}7D
{cbm-y}77 {cbm-z}6D {left arrow}1F
{pi}5E {pound}1C {shift-*}40
{shift-+}5B {shift-,}3C {shift--}5D
{shift-.}3E {shift-/}3F {shift-0}30
{shift-1}21 {shift-2}22 {shift-3}23
{shift-4}24 {shift-5}25 {shift-6}26
{shift-7}27 {shift-8}28 {shift-9}29
{shift-:}1B {shift-;}1D {shift-@}7A
{shift-^}5E {shift-a}41 {shift-b}42
{shift-c}43 {shift-d}44 {shift-e}45
{shift-f}46 {shift-g}47 {shift-h}48
{shift-i}49 {shift-j}4A {shift-k}4B
{shift-l}4C {shift-m}4D {shift-n}4E
{shift-o}4F {shift-pound}69 {shift-p}50
{shift-q}51 {shift-r}52 {shift-space}60
{shift-s}53 {shift-t}54 {shift-up arrow}5E
{shift-u}55 {shift-v}56 {shift-w}57
{shift-x}58 {shift-y}59 {shift-z}5A
{space}20 {up arrow}1E


Standard 6502 opcodes

The standard 6502 opcodes
ADC61 65 69 6D 71 75 79 7D AND21 25 29 2D 31 35 39 3D
ASL06 0A 0E 16 1E BCC90
BIT24 2C BMI30
DEY88 EOR41 45 49 4D 51 55 59 5D
JSR20 LDAA1 A5 A9 AD B1 B5 B9 BD
LSR46 4A 4E 56 5E NOPEA
ORA01 05 09 0D 11 15 19 1D PHA48
PLP28 ROL26 2A 2E 36 3E
ROR66 6A 6E 76 7E RTI40
RTS60 SBCE1 E5 E9 ED F1 F5 F9 FD
SEI78 STA81 85 8D 91 95 99 9D
STX86 8E 96 STY84 8C 94
Aliases, pseudo instructions
BLT90 CPAC1 C5 C9 CD D1 D5 D9 DD
GLT4C 90 GMI30 4C
GVC4C 50 GVS4C 70
ROR6A SHL06 0A 0E 16 1E
SHR46 4A 4E 56 5E

6502 illegal opcodes

This processor is a standard 6502 with the NMOS illegal opcodes.

Additional opcodes
LDSBB NOP04 0C 14 1C 80
RLA23 27 2F 33 37 3B 3F RRA63 67 6F 73 77 7B 7F
SAX83 87 8F 97 SBXCB
SLO03 07 0F 13 17 1B 1F SRE43 47 4F 53 57 5B 5F
Additional aliases

65DTV02 opcodes

This processor is an enhanced version of standard 6502 with some illegal opcodes.

Additionally to 6502 illegal opcodes
Additional pseudo instruction
GRA12 4C    
These illegal opcodes are not valid
LDSBB NOP04 0C 14 1C 80
These aliases are not valid

Standard 65C02 opcodes

This processor is an enhanced version of standard 6502.

Additional opcodes
BIT34 3C 89 BRA80
STA92 STZ64 74 9C 9E
TRB14 1C TSB04 0C
Additional aliases and pseudo instructions
CLR64 74 9C 9E CPAD2

R65C02 opcodes

This processor is an enhanced version of standard 65C02.

Please note that the bit number is not part of the instruction name (like rmb7 $20). Instead it's the first element of coma separated parameters (e.g. rmb 7,$20).

Additional opcodes
NOP44 54 82 DC RMB07 17 27 37 47 57 67 77
SMB87 97 A7 B7 C7 D7 E7 F7

W65C02 opcodes

This processor is an enhanced version of R65C02.

Additional opcodes
Additional aliases

W65816 opcodes

This processor is an enhanced version of 65C02.

Additional opcodes
ADC63 67 6F 73 77 7F AND23 27 2F 33 37 3F
COP02 EOR43 47 4F 53 57 5F
ORA03 07 0F 13 17 1F PEAF4
SEPE2 STA83 87 8F 93 97 9F
Additional aliases

65EL02 opcodes

This processor is an enhanced version of standard 65C02.

Additional opcodes
ADC63 67 73 77 AND23 27 33 37
CMPC3 C7 D3 D7 DIV4F 5F 6F 7F
ENT22 EOR43 47 53 57
ORA03 07 13 17 PEAF4
SEPE2 STA83 87 93 97
Additional aliases

65CE02 opcodes

This processor is an enhanced version of R65C02.

Additional opcodes
ASR43 44 54 ASWCB
JSR22 23 LDAE2
Additional aliases
This alias is not valid
CLR64 74 9C 9E      

CSG 4510 opcodes

This processor is an enhanced version of 65CE02.

Additional opcodes
Additional aliases