Download

Quick search

Rate this software

• Currently 0/5 Stars.
• 1
• 2
• 3
• 4
• 5

No. Votes

0

Linux

Windows

Mac

Mobile

Drivers

Scripts - DHTML

Web Developer Blog

Scripts and Applications

amforth

Details

amforth is a compact forth interpreter and compiler. amforth project is implemented in assembly language and forth itself. It is not a cross-compiler.
amforth implements an ans94 stylish forth with a few missing words and a few words from the extension word sets.
amforth is designed to run on atmel atmega processors with built-in RAM, flash and EEPROM. It is a pure 16 bit forth. It uses the indirect threaded model. It works interactivly on the MC via the serial interface (usart0). Use of double cell operands is dropped completly. They may be added later, if needed.
Flash structure
+------------------ 0000 --------------+
| Interrupt vectors |
| |
+------------------ 0026 --------------+
| initialization code |
| interrupt routines |
| "colon" dictionary |
| |
|LATEST: |
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
+------------- nrww section -----------+
| inner interpreter (docol, next) |
| "primitive" dictionary |
| Link to "colon" dictionary |
+----------- flashend -----------------+
The flash contains both the inner interpreter as well as the whole dictionary. New words are compiled directly into flash. The dictionary is splited into two parts. One contains all assembler (primitive) words including NEXT and DOCOL. This part is located at the end of the flash area.
The other part contains words coded as colon definitions. This comes down to the beginning of the flash. The reason for this split is that the spm instruction is used to compile new definitions directly into flash.
This instruction can only be called from an nrww area of the flash (booloader area). To use this operation, most of the core routines of the forth systems needs to be located in the nrww section.
RAM
+--------------- 0000 -----------------+
| 0-31: register set |
|31-59: io space |
+--------------------------------------+
|60-84: usart buffer |
|84.. : TIB, PAD, HLD buffers |
| HEAP: start of variable space |
| |
| ^ |
| | |
| data stack, growing downward |
+---------- ramsize - $80 -------------|
| ^ |
| | |
| return stack, growing downward |
+---------- ram size ------------------+
TIB is currently 80 bytes long, PAD and HLD 16 bytes.
What's New in This Release:
· core: new defining words code and end-code. code starts a new dictionary header with the XT set to the data field. The 2nd one appends the jmp NEXT call into the dictionary.
· core: removed the pre-assembled case / endcase words. Added them as forth library.
· core: new words -jtag (turns off JTAG at runtime) and -wdt (turns off watch dog timer at runtime. They need to be implemented as primitives due to timing requirements.
· core: quit: Keep base when handling an exception.

0 comments

Add to

 Del.icio.us   Digg It   Furl   YahooMyWeb   Blinklist

amforth Version History

Related Software