[Ninkazu]

Ok, as you may already know, I've been programming a scripting language....

Well, the problem lies in the fact that I've been learning from Game Scripting Mastery, and have everything working EXCEPT the variable amount operand functionality.

The big thing is, instead of Game Scripting Mastery's technique of allocating space for all operands to begin with, since they're all a fixed amount, I use a linked list to keep track of them.

In my system, operands are stored as WORDs, and are split into hi and lo bytes. The lo byte defines how many definite operands there are, and the hi byte defines how many variable operands there are (say if it's 3, you can have extra operands, but the amount has to be divisible by three - and you have to have at least one set). If the hi byte is 255, that means the operator doesn't know how many variable operands there are, because one of its operands has the capability for variable operands.

In my test source file, I use the following line to test:
CALLHOST AddWatch testFunc, ON_SCROLL, 1, 2

CALLHOST is defined with one definite operand - the API function name (in this case - AddWatch).
AddWatch is defined with two definite operands - the function to call when the watch state has been triggered, and the watch state
ON_SCROLL is the only operand with a definite amount of operands with no variable operands - and those are X and Y

This is the big problem (line 208+ in Assembler.cpp)

[Ninkazu]

Ok. All solved.

http://www.venosoft.com/venosoft/VSMC.rar

Ya, it's a full compiler. If you manage to understand the program and how to run its compiled files you can go on ahead and use it (but give me credit of course).

::EDIT::
Nope. I found something that will be a problem.
CALLHOST AddTrigger, ID, ON_SCROLL, x, y, x1, y1, x2, y2, line_of_code

Looking at this you'll see that it has two operands that have undefined variable operands. This poses a problem, since I don't know how I can go from:
CALLHOST -> AddTrigger (ID, x1, y1, x2, y2) -> state (state operands) <- AddTrigger -> line_of_code (operator with operands, quite possibly written in a way that needs to go forwards and backwards as well)

So.... I have almost no clue how to fix this, other than maybe have a linked list built as a stack to:
push callhost
add callhost's definite operands (just AddTrigger)
push AddTrigger
add AddTrigger's definite operands (ID, x1, y1, x2, y2)
add AddTrigger's variable operands (state, line_of_code)
push state
add state's definite operands (x, y)
pop state
push operator
see push callhost up to this point
pop operator
pop AddTrigger
pop callhost

then on to the next line of code Think it will work?

....

God damn it, I have to go to the animal shelter today. Fucking nag of a mom... Ok, I'll be back later to work on this some more and tell you about it, but I have to go to a play with my foreign exchange student from 6:30 to about 12:00 so no more work after I leave for that :(

[BigManJones]

Variable operands sounds like a huge pita. Why bother?

[Ninkazu]

I have it all working. Back up at the http://www.venosoft.com/venosoft/VSMC.rar link.

Oh, and I like having variable operands. That's why.

[tcaudilllg]

I think I'll enjoy using your language. Keep us posted. :)

[Aptyp]

Any particular reason why you wrote your own parser instead of using Lex/Yacc?

[Ninkazu]

This is a learning experience. I like to use things only when I really know how they work (household appliances excluded).

[LeoDraco]

[Aptyp]

Hell you don't even know how your own code works, but you use it anyway!

Besides, what exactly are you hoping to learn? How to make a tokenizer and a parser? Why, so that if one day you will have to write it, you will? Well, nobody does that anymore, everyone uses tools that were around for 30 years (yes, Lex is that old). Besides you can write a much more powerful language much more quickly using Lex/Yacc than writing your own.

[Nodtveidt]

Identify the uneducated lazy n00b in this thread, folks! :D
_________________
If you play a Microsoft CD backwards you can hear demonic voices. The scary part is that if you play it forwards it installs Windows. - wallace

[Ninkazu]

I have it all fixed and I'm going to write the VM now. If there are problems with the VM, then I'll know my compiler is wrong, but for now I'm almost certain the compiler is right now.

Oh, and I will continue to write my own code because:
1) I am a programmer
2) I want to be a better programmer
3) If you don't learn things the hard way, you'll become a lazy, crappy programmer that both André LaMothe and I hate.

So you must all:
1) Fuck off
2) Fuck off
3) Go fuck yourselves

Good. Now see 1 & 2.

[LeoDraco]

[Ninkazu]

[tcaudilllg]

It's a matter of mindset, really, and of thinking procedure. Personalities with investigative thinking procedure like to know the innards of something before they use it. If I use a library or generator of some kind to write a program, and then witness it crash in a way that I've never before seen, the first thing I suspect is the library. When you depend on others for your program accessories, without fully understanding how their systems function, you're in all practicality rolling a die.

I've heard that standardization is an important part of quick design process. But I've also heard of building a better mousetrap, however slow and painstaking a process that may be. Am I writing my own programming language? Yeah. Am I using standardized parser generator tools. Not at all. Their technology is a bit old and antiquidated, maybe a bit too specialized for my experimental purposes. Besides, the standardized lexical generation procedures were designed by INTP "mathematician" types, and their "explicit is better than implicit" mentality confuses me to no end! @_@

Maybe they confuse Ninkazu, too.

[LeoDraco]