Source files

Source files consist of valid tokens and comments

COMMENTS

comments start with a semicolon (;) and end at the end of the line Comments are fully ignored by the assembler.

TOKENS

tokens are a string of non-whitespace non-semicolon characters in a line They can be surrounded by any number of whitespace and can be followed by a comment in the same line, but no line can contain two tokens.

Tokens must be of 3 types:

• directive
• label
• instruction

labels*

Labels are aliases for addresses in memory. They automatically mark the next address where they are defined, but they do not take any space in the resulting binary. Label tokens must end with a colon (:)

directives

directives are commands for the assembler. They might or might not take space in the binary.
directive tokens must start with a dot (.)
Valid commands:

• define
• ifdef
• printc
• print
• include
• endif
• ifndef
• ifbeq
• org
• data
• pad

The assembler is not case-sensitive with directive names

define <name> <value>

Define a constant with name and value.

ifdef <constant>

Conditional compilation until the matching .endif if the constant is defined.
Node: does not need @ in front of the name!

printc <constant>

Log the name and value of a constant.
Note: does not need @ in front of the constant name!

print <line>

Logs the rest of the line

include <file>

Include another file in the source.
Basically automated copy-paste.
Searches the file in CWD.

endif

Close a block of conditional compilation

ifndef <constant>

Inverted ifdef

ifbeq <first> <second>

Conditional compilation: first >= second
Both can be numbers!

org <number>

Set PC to number

data <entry> [...]

Can have many entries of these 3 types:

• byte: <number> (raises an error if it can't fit in a single byte)
• word: w:<number> (gets encoded in $LLHH big-endian format)
• string: "<str>" (must not include whitespace) The binary data from these entries will be concatenated an added in the result

pad <to> <with>

Pad remaining space to to with byte with

instructions

instructions are actual 6502 instructions

They must consist of a valid instruction mnemonic (3 letters), followed by an additional whitespace and argument
(i.e they must end after the mnemonic, or have a space and an argument)

See this page for a list of valid mnemonics, arguments and their combinations.

Differences to that page:

• this assembler uses star (*) to indicate zeropage (and indexed zeropage) addressing
• instructions with relative addressing modes take memory addresses instead of 8 bit addresses and the offsets are calculated by the assembler

The assembler is not case-sensitive with instruction mnemonics or operands.

Numbers

Numbers in directives and operands can take 4 main forms:

• labels addresses: &<labelname>
• defined constants: @<constant>
• numbers in hex: $<number>
• numbers in decimal: <number>

The assembler is not case sensitive with hex numbers (a-f, A-F), but IS case sensitive with constant and label names.

Both constant and label names have an upper limit set in map.h

All types can have 2 additional modifiers:

• < to extract low byte (lower 8 bits)
• > to extract high byte (higher 8 bits)

When an operand or a directive takes a 16 bit value, it will get stored as the 6502's format (high endian, $LLHH).

Labels can also be forward-referenced in 2 cases:

• instruction operand
• .data entry