V2.20 Fixes an undocumented feature of IF as well as adding an optional ELSE clause and can be found here ..

http://www.themotionstore.com/leeedavison/6502/ehbasic/index.html

Lee.

Lee,

Would it be easy to port your Basic to a WDC 65C816?

Chuck

Chuck,

I am running EHBasic in native mode on my SBC-3 so I know it can be done.

I have notes on my website on the changes I made.

http://sbc.rictor.org/download/EhBASIC.zip

Daryl


EDIT: I have made most of the changes to the new version of EHBasic and was able to run a simple program on WDC65816 in native mode. I did have some trouble with the load and save functions, that's most likely my support code, not EHBasic.

Once I have it all complete, I'll update my website and post it here.

Great job Lee!!!

Thanks for continuing to develop this program!

I discovered one thing to look out for when upgrading to version 2.20.

My method of saving EHBasic programs is to scan RAM from smeml, smemh until I find the end of the program. I then do a binary save of that range to disk.

When loading a program written in v2.10 into v2.20, some of the tokenized commands are changed, due to the insertion of new tokens in the table.

what was

20 FOR x = 1 to 10
30 PRINT X , X^2

becomes

20 FOR x > 1 ELSE 10
30 PRINT X, X/2

The good side is that a conversion tool could easily be written to update the program files.

Daryl

Yes, sorry about that but it is pretty much unavoidable. I've given up on bianry LOADs and SAVEs on my systems for this very reason.


As there is only one new token, TK_ELSE - which is $AC, it should just be a case of scanning the program from start to end and incrementing every token byte of $AC or more.

Lee.

The DEEK in line 63010 needs to be to Smeml which is $79 in my case.


This can be appended to the end of any program and executed with RUN 63000

Lee.

Lee,

I understand the need for the change and it no real problem. You've already created the tool to adjust it which is great too.

I have completed the changes for V2.20 to run on SBC-3 (my earlier problem was in my supporting routines). I'll get the updated version loaded on my site soon.

thanks again Lee!

Daryl

Tokenization is a good thing to do for several reasons, but for the purposes of data import and export, its primary value is in the compression it offers. After studying LZ78 compression, you start to realize that a tokenized source file is completely equivalent to an LZ78-compressed text file with a statically predetermined dictionary.

In other words, you can get all the same benefits of tokenized source code if you just compress your raw text source (e.g., using PKZIP, for example). Perhaps, then, a binary load/save command can be added which is nothing more than a raw text load/save command piped through a post-compression or pre-decompression stage first.

I've taken a couple cracks at adding the 2.20 changes to my port of 2.10. Since my 2.10 is so heavily hacked this is a bit trickier. Probably "third time's the charm" though.

In my case, a redundant token (CATALOG, which was intended to produce an 80-column-wide directory listing while CAT produced a 40-column listing - like BASIC.SYSTEM) could be recycled. I have no room for new tokens.

I've put the details of the changes made here

Lee.