Re: Line number resolving

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

Eric,

This seems like a reasonable approach. Soot should propagate the source
line numbers to tags in Jimple statements. You should then be able to
output from Soot whatever information you need about a Jimple statement
along with the source line number. If you come across problems with
this, please let us know.

Ondrej

On Mon, Jan 13, 2003 at 12:09:59PM -0000, Eric Bodden wrote:
> Hi!
> 
> I have the following problem:
> We want users to be able to highlighte a specific object in sourcecode
> and then resolve the type of that object.
> So we thought about having an internal representation of this
> sourcecode, having just one single statement per line. So with this
> precondition we should be able to get a 1 to 1 relationship between a
> sourcecode line and the appropriate bytecode line. So then we hopefully
> could use the type propagation implementation of SOOT to resolve the
> static type of the object which is located at that line.
> 
> My questions are: Do you think it's possible to solve it like this? I
> read about some inconsistency between the line numbers SOOT uses
> internally and the one which are in the bytecode table. So is it
> possible to get a 1 to 1 mapping from let's say a method call "A.foo()"
> in the sourcecode to the appropriate Jimple line ? And is it then
> possible to change the type propagation algorithm in such a way that it
> will pass you the static type of an object for this line?
> 
> I hope, you can help me.
> Thank you very much,
> Eric
> 
> --------------------------------
> Eric Bodden
> University of Kent at Canterbury
>