Fasmw 1.73
I want do this. Parse on fasm text.
Load text file in TxtBuffer and find all defs.
My text look like this:


My question how do fast\more efective searching in X and Z A and B ?
Because if i have thousand #defs and thousand values (like X Z )its slow down my program.

I dont know what do good.
Search all #defs and put in another buffer.
Or always search in X A then search in text #def A ?

Don't update dynamically. Instead only update part that changes. Loading a new file will still incur a delay, but active file updates will be negligible. The fasm source itself is a good example on how to store the symbols.

As soon as you use . as a delimiter you’re stuck with assembler stage to parse this stuff. As soon as you use # as the first character of certain lines you’re stuck with switching to another tool.

Dimon Soft please say you algorithm and your proposition.
More concrete to my problem.

FASM uses # as a special character. I really doubt there’s a way to override it to be ignored while leaving it possible to use it in macros (which might be useful for the rest of parsing).

Breaking an identifier into parts on dots is hardly possible with preprocessor, so the only way might be to convert it to string data in a virtual block and then walking through it somehow converting the string parts to an index into another virtual block that would store actual values. Doesn’t look promising as well.

I write my parser program.
And write parser on asm.
I mean this.

I want own script language.

Ah, sorry. Is it going to be limited to the features you’ve shown? Just plain value lists and indexing into them? Or we’re talking about complex scripting language in the future?

I changed my script logic.
And now I don't need defs and x z.
No need search and parse defs and x z.
Now my parser more simple, faster and small.