Re: SPARK: Coloured nodes

[Ondrej Lhotak <olhotak@sable.mcgill.ca>]

The most detailed description is in my thesis, section 4.1:
http://www.sable.mcgill.ca/publications/thesis/#olhotakMastersThesis

Briefly, the colour code that I try to use is:
Green: variable nodes, representing single location
Blue: allocation nodes, representing the objects created at an
allocation site
Red: field reference nodes, representing expressions of the form x.f,
where x is a variable node representing a variable x, and f is a field
of the object pointed to by x.
Yellow: concrete field nodes, representing locations of the form A.f,
where A is an allocation node, and f is a field. The node A.f represents
the field f of all objects allocated at A.

On rare occasions, I have used the following additional colours:
Cyan: to represent a C-style address-of (&) expression.
Purple: to represent parameterized nodes when talking about
object-sensitive analysis [Milanova et al. ISSTA 02]

Current versions of Spark should no longer refer to the colours; they
use names like VariableNode. Some of the colours are still there in the
comments.

Ondrej

On Thu, Apr 03, 2003 at 04:36:22PM +0200, Richard Stahl wrote:
> Hello,
> 
> I might have a stupid question yet I am new to the points-to analysis topic:
> 
> Could you, please, summarise the issues of colouring the nodes, its 
> meaning or, most preferably, point me to some good introductory material(s).
> 
> Thanks a lot.
> 
> Best regards,
> 
> Richard
>