I'd like to try adding a 'BYE' command to the basic which causes the basic interpreter to quit and exit back to the operating system.

Any tips on how to implement this ?

_________________
http://www.finitron.ca

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

I've just been glancing at the source code, I don't know for sure, but adding a command should be reasonably straight forward. All of this is supposition, I don't know.

The first step is defining a TOKEN value for it. This is done around line 296, you can see all of the token values defined. It starts with TK_END, and a value of $80, then rest build incrementally from there.

Couple of issues that I don't know of the top of my head.

1) is there even any room for another token, I assume a token must start with the high bit set, and there are a lot of tokens here, so there may well not even be room for another token. I would run the assembly listing and then look for the value of TK_MIDS. If it's less than $FF, then you should have room. If not, then you'll need to remove something.

2) I don't know if the order of the tokens matter. I sense not, but I'm not sure. He has them broken up in to primary commands, secondary commands, operators, and functions. It appears to be simply organization, rather than architecture. In theory, to be consistent, you would insert TK_BYE after TK_NMI, and update the definition for TK_TAB. But I'm not sure if you MUST do this, since you're defining a primary command.

Moving on...

Next, you need to define the code that is invoked when the command is executed. These are identified by LAB_XXX labels. In your case, LAB_BYE would be appropriate. (Mind, LOTS of things are identified by LAB_XXX, but within that set are the actual command runtime routines.). Since your code is just an escape, it should be a straight forward JMP to some known place, or a BRK, or whatever, so you don't have to know anything about the actual BASIC runtime here. Just make sure you cold start BASIC if you try to come back, or you'd likely be in trouble.

You'd probably, again for consistency, put your code just before LAB_MAX, which comes after all of the interrupt code and runtime for LAB_NMI.

Next...

Add your runtime routine to the command table, at LAB_CTBL. This list is in TOKEN order, so, again, you would put your entry just after LAB_NMI, and it would look something like:

Continuing...

Add the command to the command text table. Since the name of your command is BYE, you would put this in the B section, identified by TAB_ASCB. These lists are ALPHABETICAL, but BYE is going to be at the end anyway, so add your entry after LBB_BITTST. It will look like:

Then...

Add the entry for the LIST command decode table at LBB_KEYT. This is in TOKEN order, so your entry will be after NMI, again, and will look like:


And that should be it.

Good luck, let us know if I guessed right. I only looked at this thing for a couple minutes.

I added the SYS command to my EhBASIC's to allow you to go back to the Monitor. These are my notes from the change log:



I don't recall many details, but I think there are two groups - Primary commands and secondary. This is a primary command. I was able to figure it all out from the comments in the code.

I hope this helps!

_________________
Please visit my website -> https://sbc.rictor.org/

See the 5th post down. --> viewtopic.php?f=5&t=1341

Thanks for the tips, I think I've got it working.

The command exits EhBASIC and returns to my OS, bit sometimes the OS reboots as a result.
The OS is still buggy...

_________________
http://www.finitron.ca

All --

Adding to this thread a bit, how would you recommend dealing with optional parameters? For example, my Commodore IEC extensions can have an optional device number for the disk drives. Device 8 is assumed unless otherwise specified. Here's the code:



LAB_SCGB automatically returns Syntax Error if the comma isn't found. So, is there a better routine buried within EhBASIC that deals with this?

Also, a few disk commands require both numeric and string parameters -- any recommendations on parsing those?

Thanks!

_________________
Rich Cini
Build Master and maintainer of the Altair32 Emulation project
http://cini.classiccmp.org
http://altair32.classiccmp.org

Before calling LAB_SCGB you could check for end of statement with LAB_GBYT. LAB_GBYT returns Z=1 if end of statement (null or ":") and you can skip the call to LAB_SCGB with a simple BEQ.

Generally LAB_PRINT is a good place to check how to implement parsing as it works with any mixture of variables but of course also with none.

_________________
6502 sources on GitHub: https://github.com/Klaus2m5

Great, Klaus, thank you. Just made the change and it works fine.

I will check out _PRINT as well.

I'm also adding an "LLIST" command which will echo to the printer.

_________________
Rich Cini
Build Master and maintainer of the Altair32 Emulation project
http://cini.classiccmp.org
http://altair32.classiccmp.org

All --

I've made some good progress in the last week or two. The boards I had made work fine so I'm going to do a larger quantity order.

I'm presently working on adding more commands to EhBasic...LLIST is the command I'm working on right now. It works...mostly...but I've discovered an anomaly that I'm sure relates to me not fully understanding how EhBasic dispatches commands.

While LIST will work with no command line parameters, LLIST doesn't. But, if I "LLIST 10-100" it works fine. So, even though I'm saving the registers, something is getting messed up.

Here's the code:


What am I missing here? I figured that saving the registers would do it, but it doesn't. A fresh set of eyes would be helpful and much appreciated Thanks!

_________________
Rich Cini
Build Master and maintainer of the Altair32 Emulation project
http://cini.classiccmp.org
http://altair32.classiccmp.org

Maybe you should save the processor status as well (PHP PLP).

_________________
6502 sources on GitHub: https://github.com/Klaus2m5

Yes, that was it. Works perfectly now. I guess I have to be extra-careful in saving the processor context.

Thanks Klaus!

_________________
Rich Cini
Build Master and maintainer of the Altair32 Emulation project
http://cini.classiccmp.org
http://altair32.classiccmp.org

Hi Everyone, sorry for posting in an old thread but I'm having difficulty figuring this out myself and need some advice.
I figured out most of what I needed to do with the tables / token etc. though, I originally added the vector to LAB_FTBL and had no success at all.
After reading through this thread, I moved the vector to LAB_CTBL and changed the token placement to match, now, I can get it to run without a syntax error so I know I'm on the right track but adding arguments always results in a syntax error. Out of desperation, I even tried this:
.word LAB_PRINT-1 ; PRINT <-----------------------The original "PRINT" entry
.word LAB_CONT-1 ; CONT
.word LAB_LIST-1 ; LIST
.word LAB_CLEAR-1 ; CLEAR
.word LAB_NEW-1 ; NEW
.word LAB_WDTH-1 ; WIDTH new command
.word LAB_GET-1 ; GET new command
.word LAB_SWAP-1 ; SWAP new command
.word LAB_BITSET-1 ; BITSET new command
.word LAB_BITCLR-1 ; BITCLR new command
.word LAB_IRQ-1 ; IRQ new command
.word LAB_NMI-1 ; NMI new command
.word LAB_PRINT-1 ; TESTFUNC <----------------------My new function, pointed to LAB_PRINT
it's in the same table as PRINT so I figure it should essentially work as PRINT but it still gives a syntax error if I add arguments, ie:
testfunc <------Works
testfunc "hello" <------Syntax Error

It's as if there should be some sort of pre-process like in the case of a 'function' but there's no such table.
I basically need to create a function which is provided a string so modelling it on Print seems to be a logical approach.
Can anyone please suggest what I have missed here?