[Soot-list] anyone try soot (dava) on scala 2.8.x compiled code successfully?
kishen.das at gmail.com
Fri Jan 7 14:01:26 EST 2011
We have been using Soot for last one year to work with Scala's java
byte-code without any issues.
Since Scala supports closures, there will be lot of anonymous classes that
are created on the fly.
Scalap or Javap does not show these anonymous classes.
We have created our own Scala reflection tool for this that can give more
information about all these inner anonymous classes.
In any case you should be able to see all the classes in the byte-code.
On Fri, Jan 7, 2011 at 1:40 AM, Eric Bodden <
bodden at st.informatik.tu-darmstadt.de> wrote:
> Hi Craig.
> To me the names look like names of anonymous inner classes. I agree
> that javap should help de-mystify a few things. To me it looks like
> this is not really Soot's mistake: it appears that these classes are
> referenced. The question is just why the code does not fail then, if
> the classes are not provided.
> On 6 January 2011 21:40, Craig Bordelon <bord at iscp.telcordia.com> wrote:
> > Hello Eric,
> > I have no evidence that these names are actually classes at all.
> > Though, I grant that the names look like they could be classes.
> > So, it could be that the scala compiler is placing these names into the
> byte code somewhere but that they are not provided at runtime?
> > I dont know the answer, though I think i'll try javap/scalap -private on
> a few actual runtime classes to see if they show up.
> > What i do know is that the jvm is able to load all the actual runtime
> classes that i've tried from the scala distribution including some that have
> similar names.
> > So, if these names are in some of the byte code, the fact that they're
> missing at runtime, does not seem to affect the runtime class loading.
> > cheers
> > craig
> Soot-list mailing list
> Soot-list at sable.mcgill.ca
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Soot-list