Re: Building Abstract Syntax Trees from LL(1) Grammars

"Capitaine Caverne" <no@mail.com>
23 Aug 2002 11:03:08 -0400

          From comp.compilers
Related articles
Building Abstract Syntax Trees from LL(1) Grammars bart@dynarec.com (Bart T.) (2002-08-14)
Re: Building Abstract Syntax Trees from LL(1) Grammars no@mail.com (Capitaine Caverne) (2002-08-23)
Re: Building Abstract Syntax Trees from LL(1) Grammars thp@cs.ucr.edu (2002-08-23)
| List of all articles for this month |
From: "Capitaine Caverne" <no@mail.com>
Newsgroups: comp.compilers
Date: 23 Aug 2002 11:03:08 -0400
Organization: Guest of ProXad - France
References: 02-08-056
Keywords: syntax, analysis
Posted-Date: 23 Aug 2002 11:03:08 EDT

"Bart T." <bart@dynarec.com> a écrit news:02-08-056@comp.compilers:


>
> 9+5-2:
>
> (- (+ 9 5) 2)
>
> OR
>
> (+ 9 (- 5 2))
>
> However, it isn't possible to generate the operators prior to the
> operands in my grammar. :( At least I can't think of a way. There must
> be a way, though...


Hello, I'm certainly the last person to listen in this newsgroup as I'm a
complete newbie... However in my try to implement a stack based machine
compiler I have the same problem : I can't handle operators and operands
in a way I would like (was not a problem when I implemented my own poor
parser some years ago).


The solution I've found is to use two stacks : an operand stack and an
operator stack.


For 9+3*2


Thus instead of generating (generated code is not in the source order)


PUSH VAL 3
PUSH VAL 2
MUL
PUSH 9
ADD


I do :


PUSH VAL 3 (generated code is in source order)
PUSHOP MUL
PUSH VAL 2
CALC
PUSH VAL 9
PUSHOP ADD
CALC


This is less efficient but it works... At least it seems ;-)


Not sure it answer your question, once again I'm a newbie...

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.