Re: Graph theory and data flow analysis w\java

Sid Ahmed Ali TOUATI <Sid-Ahmed-Ali.TOUATI@inria.fr>
16 Nov 2000 15:01:03 -0500

          From comp.compilers

Related articles
Graph theory and data flow analysis w\java david.boissy@tpx.turbopower.com (David Boissy \[TPX\]) (2000-11-14)
Re: Graph theory and data flow analysis w\java benwar@gmx.net (Ben War) (2000-11-16)
Re: Graph theory and data flow analysis w\java Sid-Ahmed-Ali.TOUATI@inria.fr (Sid Ahmed Ali TOUATI) (2000-11-16)
| List of all articles for this month |

From: Sid Ahmed Ali TOUATI <Sid-Ahmed-Ali.TOUATI@inria.fr>
Newsgroups: comp.compilers
Date: 16 Nov 2000 15:01:03 -0500
Organization: INRIA
References: 00-11-112
Keywords: Java, analysis
Posted-Date: 16 Nov 2000 15:01:03 EST

"David Boissy [TPX]" wrote:


> Group,
>
> I wanted to do some simple data-flow analysis in java and I'm looking
> for a good graph theory tool to use. I'd like it to use an adjacency
> list and not a matrix because the graphs are big.
>
> Also, I would like to perform live variable analysis on this graph
> which I assume would require backwards data-flow analysis. The dragon
> book covers the intraprocedural case. Can anybody provide me any book
> recommendations or links for the interprocedural case? The Muchnick
> book touches on this, but doesn't give any real detail.


You cannot do easily data flow analysis with java like with C or fortran.
The reason why is that the affectation "=" in java is not a copy but an
aliasing operator. There is a great amount of theoretical problems to
solve before doing any data flow analysis : the main problem is to prove
statically that there is no alias between objects, i.e. fto prove for
instance that a given array has disjoint elements.


yours,
SAAT


Post a followup to this message

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