a80 is an assembler written in D for the
Intel 8080 (and, by extension, the Zilog Z80) CPU.
a80 is developed on OpenBSD but
should work on any system that D targets.
a80 is not an exact clone of any pre-existing CP/M assember, nor
does it want to be. The differences will be explained in this
document.
a80 also is quite conscious about its design practice and
implementation. It is written to be the subject of a series of
blog posts in
which we attempt to demystify the building of programming tools,
and as such very intentionally does not use some very obvious data
structures. And it may make some seemingly peculiar design choices.
My goal is to have written a real assembler for a real CPU that you
can still purchase today (in the form of the Z80) that true
beginners can come to understand.
After the blog series, if we want to turn this into a clone of an existing CP/M assembler, I'm all for it.
Bug reports are welcome at any time.
a80 file.asm
The output will be file.com.
A line of assembly takes the following form:
[label:] [op [arg1[, arg2]]] [; comment]
An example of an assembly program can be found in fib.asm.
a80 only understands Intel 8080 opcodes.
The CP/M EQU directive is supported however you cannot use other
labels as a value nor can you use expressions.
All op and arg must be lowercase, though labels may include capital letters.
Numbers may be in decimal or hex.
Hex numbers must end with an h. If a hex number begins with
a-f, it must be prefixed with 0. This is not too dissimilar
compared to other CP/M assemblers.
The DB pseudo-op is available. Strings can be written within
single quotes. Non-numeric, non-comma, non-semicolon, and
non-single quote single characters will be interpreted to ASCII.
If you want a comma you need to use db 2c or put it in a string.
Both the semicolon and the single quote are illegal in strings and
cannot be written as a single character. However, they can still be
input by using their hex code: db 3b will input a semicolon
and db 27 will input a single quote. This is a limitation of the
parser and is welcome to be fixed.
There is none. That is to keep things simple.
The equ pseudo-op does allow for $ and very simple $+number
expressions only. No spaces in expressions. Valid expression
operands are +, -, *, /, and %.
ISC license. See LICENSE for more information.