Glass Z80 assembler
README: Move project source to GitHub.
Scope: Fix serialising symbols with a cycle in them.
pom: Increase version to 0.7-SNAPSHOT.


browse log
browse .tar.gz



#Glass Z80 assembler

Copyright 2013 Laurens Holst

#Project information

Glass is a cross-assembler for the Z80 processor written in Java 8. Its core principles are to be open source, cross-platform, and to provide a standard Z80 syntax infused with modern language features.

It presents a flexible language for Z80 object code generation by building an abstract syntax tree with strong scoping rules, rather than using the traditional multi-pass architecture with separate preprocessor and mnemonic translation. This allows the user to write powerful expressions, use macros as type definitions, etc. Future developments aim to bring more modern programming concepts to the Z80 assembly programming realm.

Because the binary is a jar which runs on the Java virtual machine, it can be included in a project easily without requiring the user to acquire a separate binary build for their operating system.


To run Glass from the command line, use the following command.

java -jar glass.jar [OPTION] SOURCE [OBJECT] [SYMBOL]

Source specifies the source file, object the output file, and symbol a text file which will hold a list of symbols and their addresses in the output.


  • -I include_path Include path for additional source files.

  • -L list_file File to output a listing of the assembled code to.

Note that Java 8 must be installed to run Glass. To check your Java version, invoke the java -version command.


The assembler syntax follows common style.

Lines are composed as follows:

label: mnemonic arguments ;comment

Note that the white space before the mnemonic is significant; otherwise, it will be interpreted as a label.

All identifiers are case-sensitive. Mnemonics of built-in instructions, directives, registers, flags and annotations are recognised in lowercase and uppercase, but can not be mixed case.


Labels and other identifiers follow the following grammar:

identifier = id_start_char id_char*
id_start_char = [a-z] | [A-Z] | _ | . | ? | @
id_char = id_start_char | [0-9] | ' | $

The colon after a label is optional. If a label has no colon, it can not have any leading white space, it must start at column 0.


Standard z80 instruction syntax is used:

ld a,(ix + 10)

Parentheses are used to indicate indirection.

  • Z80 instructions: adc, add, and, bit, call, ccf, cp, cpd, cpdr, cpi, cpir, cpl, daa, dec, di, djnz, ei, ex, exx, halt, im, in, inc, ind, indr, ini, inir, jp, jr, ld, ldd, lddr, ldi, ldir, neg, nop, or, otdr, otir, out, outd, outi, pop, push, res, ret, reti, retn, rl, rla, rlc, rlca, rld, rr, rra, rrc, rrca, rrd, rst, sbc, scf, set, sla, sra, srl, sub, xor

    For a complete description of the Z80 instruction set, see the official Zilog documentation:

    In addition to the documented Z80 instructions, the variations using the undocumented ixh, ixl, iyh and iyl index registers are supported, as well as the semi-documented in (c).

    For register jumps, jp (hl) etc., the parentheses are optional.

  • R800 instructions: mulub, muluw

    R800 multiplication instructions.

  • Define byte: db

    Defines a byte or a sequence of bytes.

  • Define word: dw

    Defines a word or a sequence of words.

  • Define double word: dd

    Defines a double word or a sequence of double words.

  • Define space: ds

    Defines space for a number of bytes. The first argument indicates the number of bytes, the optional second argument specifies the fill value (default 0).

    The first argument can be annotated with virtual, in which case the address counter will be incremented accordingly, but no object is actually generated in the output. If the virtual annotation is given, you can not specify a fill value.


  • Origin: org

    Changes the address location counter and sets a new origin for subsequent statements.

    org 0100H
  • Assign constant: equ

    Assigns a constant value to a symbol.

    JIFFY: equ 0FC9EH
  • Include: include

    Includes another source file. The current working directory is searched, as well as any include paths specified on the command line.

    INCLUDE "math.asm"

    Optionally you can specify a once annotation to prevent a file from being included more than once. However it is not recommended to use unless needed.

    INCLUDE ONCE "math.asm"
  • Include binary: incbin

    Includes binary data from a file. The current working directory is searched, as well as any include paths specified on the command line.

    INCBIN "image.ge5"

    Optionally you can specify a start position and length:

    INCBIN "image.ge5",7,212*128
  • Macro: macro, endm

    Defines a macro instruction, composed of all the instructions that follow until the endm directive is encountered. The definition’s arguments specify the parameters which are passed when the macro is invoked.

    ALIGN: MACRO ?boundary
           ds ?boundary - 1 - ($ + ?boundary - 1) % ?boundary
           ALIGN 100H

    All symbols defined in a macro block are local. Symbols in macro instances can be referenced by using the . operator. Symbols in macro definitions can also be referenced; the contents are assembled on address 0, effectively turning the inner symbols into offsets. This is useful for specifying structures and indexing.

    Default values for macro arguments can be specified with =:

    ALIGN: MACRO ?boundary = 100H
  • Repetition: rept, endm

    Repeats a section of code a number of times. The end of the section is marked with the endm directive. The first argument is mandatory and specifies the number of repeats. The second argument specifies a counter parameter, the third the initial value for the counter (default: 0), and the fourth argument specifies the counter increment (default: 1).

    REPT 10, ?counter, 0, 2
    ld bc,(table + ?counter)
    REPT 3
    add hl,bc

    All symbols defined in a repeat block are local. If a repeat is labeled, the inner repeat scopes can be accessed by index, e.g.: mylist.0.

  • Indefinite repetition: irp, endm

    Repeats a section of code for each of the arguments specified. The end of the section is marked with the endm directive. The first argument is mandatory and specifies the parameter the current repetition’s value is passed to. The remaining arguments are passed one by one as the section is repeated.

    IRP ?value, 1, 2, 4, 8, 16, 32, 64, 128
    or ?value

    All symbols defined in a indefinite repeat block are local. If a repeat is labeled, the inner repeat scopes can be accessed by index, e.g.: mylist.0.

  • Procedure: proc, endp

    Defines a section of code as a procedure. Currently mostly serves to establish a local scope.

    shift5: PROC
            ld b,5
            jp shiftl.loop
    shiftl: PROC
            ld b,1
    loop:   add a,a
            djnz loop

    All symbols defined in a procedure block are local. Symbols in inner scopes can be referenced by using the . operator.

  • Condition: if, else, endif

    Conditionally assembles a section of code, or an optional alternative section. The end of the section is either marked with endif, or with else in which case an alternative will follow up to the endif. The argument is evaluated as an integer, and if the result is nonzero (true) the first section is assembled, and if the result is zero (false) the alternative is assembled if one is provided.

    PAD: MACRO ?address
         IF $ > ?address
             ERROR "Padding address exceeded."
             ds ?address - $
  • Source file end: end

    Indicates the end of the current source file. Specifying this is optional, and usually omitted. Any content on lines beyond this directive will not be parsed.

  • Error: error

    Generates an error and aborts the compilation. Optionally a message can be specified.

    ERROR "Limit exceeded."
  • Warning: warning

    Generates a warning. Optionally a message can be specified.

    WARNING "Nearly out of space."
  • Section: section

    Defines a section of code or data that will be assembled inside the space of a ds statement. This allows you to have nonadjacent code or data sections and group them into separate regions, such as ROM and RAM pages. The mandatory argument references the DS statement that is the target of the section.

        org 4000H
    ROM_PAGE1: ds 4000H
    ROM_PAGE2: ds 4000H
    RAM: ds VIRTUAL 3000H
        ld (value),a
    value: db 0


  • Decimal: 127
  • Hexadecimal: 0FC9EH, #FC9E, $FC9E or 0xFC9E
  • Binary: 10110001B or %10110001
  • Octal: 377O
  • Character: 'c'
  • String: "abc"

Character literals can contain the ' character by repeating it as '', and string literals can contain the " character by repeating it as "".

Character and string literals support the following escape sequences:

  • \0 (NUL)
  • \a (bell)
  • \t (tab)
  • \n (line feed)
  • \f (form feed)
  • \r (carriage return)
  • \e (escape)
  • \" (double quotation mark)
  • \' (single quotation mark)
  • \\ (backslash)

Numeric escape sequences are not supported. In stead, you can insert them using the comma operator: "abc", 0FFH, "def".

The character set used to read files is ISO-8859-1, this maps the file’s bytes 1:1 to the Unicode code points used internally so the object code output matches the input file bytes verbatim.

The assembler uses 32-bit integer math internally. When a 8-bit or 16-bit value is generated, the excessive bits are usually truncated. Except for addresses, used in jumps, calls and indirect loads, they generate an error. Index and relative jump offsets are also checked to be in their allowed range.


  • Member: .
  • Positive: +a
  • Negative: -a
  • Complement: ~a
  • Not: !a
  • Multiply: a * b
  • Divide: a / b
  • Modulo: a % b
  • Add: a + b
  • Subtract: a - b
  • Shift left: a << b
  • Shift right: a >> b
  • Shift right unsigned: a >>> b
  • Less than: a < b
  • Less or equal: a <= b
  • Greater than: a > b
  • Greater or equal: a >= b
  • Equal: a = b
  • Not equal: a != b
  • Bitwise and: a & b
  • Bitwise xor: a ^ b
  • Bitwise or: a | b
  • Logical and: a && b
  • Logical or: a || b
  • Ternary if: a ? b : c
  • Annotation: a b
  • Sequence: a, b
  • Group: ()
  • Index: []

Logical operators use integers to represent true / false values. 0 means false, any other value means true. They return -1 for true values.

Logical and / or apply short-circuit evaluation and evaluate to the last evaluated value, so they can also be used similar to a ternary operator.

Expressions can span multiple lines when they’re incomplete at the line ends.

Operator precedence:

  1. .
  2. unary + - ~ !
  3. * / %
  4. + -
  5. << >> >>>
  6. < <= > >=
  7. = !=
  8. &
  9. ^
  10. |
  11. &&
  12. ||
  13. ?:
  14. ,

#Development information

Glass is free and open source software. If you want to contribute to the project you are very welcome to. Please contact me at any one of the places mentioned in the project information section.

You are also free to re-use code for your own projects, provided you abide by the license terms.

Glass is written in Java 8. To check your Java version, invoke the java -version command. The project can be built using Maven by invoking the following command on the command line:

mvn verify

The jar binary will be output to the target directory.