amforth is written in assembler. Only a few are actually assembly words, most
are pre-compiled forth code. There are three major directories containing the
common. Each contain a number of
words that contain actual source files.
Almost every word uses its own source file with a descriptive name. These elementary
source files are collected in include file sets, called dictionary files. Depending
on the controller type, different dictionary file sets should be used. Most of the
decisions are made automatically by using the single top-level file
The assemblers used suuport a list of include directories which is used
in order. That makes it possible to have an application specific
directory that may contain the same file names as the amforth provided ones that
take precedence during the assembly process. Likewise the controller specific
directories are searched before the
Every Atmega has its own specific settings. They are based on the official include files provided by Atmel and define the important settings for the serial IO port (which port and which parameters), the interrupt vectors and some macros.
Adapting another ATmega micro controller is as easy as copy and edit an existing file from a similar type.
The last definition is a string with the device name in clear text. This string is used within the word VER.
Every build of amforth is bound to an application. There are a few sample applications, which can be used either directly (AVR Butterfly) or serve as a source for inspiration (template application).
The structure is basically always the same. First the file
preamble.inc has to be included. After that some
definitions need to done: The size of the Forth buffers,
the CPU frequency, initial terminal settings etc. As the
last step the amforth core is included.
For a comfortable development cycle the use of a build utility such as make or ant is recommended. The assembler needs a few settings and the proper order of the include directories.