Re: From quads to AST

daniel@quilty.stanford.edu (Daniel Weise)
Fri, 13 Aug 1993 21:16:13 GMT

          From comp.compilers
Related articles
From quads to AST cheung@unx.sas.com (1993-08-10)
Re: From quads to AST daniel@quilty.stanford.edu (1993-08-13)
Re: From quads to AST krishna@cs.unm.edu (1993-08-18)
| List of all articles for this month |
Newsgroups: comp.compilers
From: daniel@quilty.stanford.edu (Daniel Weise)
Keywords: optimize
Organization: Computer Systems Laboratory, Stanford University
References: 93-08-049
Date: Fri, 13 Aug 1993 21:16:13 GMT

cheung@unx.sas.com (Clement Cheung) writes:
      Is it possible to reconstruct from quads the AST of a program? If yes,
      how? What other information besides the quads is needed? Is there any
      reference?


      The reason I ask is that a lot of neat applications such as construction
      of program dependence graph starts from AST rather than quads.


The construction of the PDG starts from a CFG, not an AST. A PDG
consists of an intertwining of a data dependence graph (DDG) and a
control dependence graph (CDG). Methods for constructing a CDG all
start from a CFG. The standard PDG paper (Ferrante, Ottenstein, and
Warren, TOPLAS 86?) isn't too clear on the DDG part, but it can also
be constructed from the CFG.


Going from quads to CFGs is simple. It's described in the red dragon
book.


Daniel Weise
--

Post a followup to this message

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