Hi, On Wed, Apr 11, 2012 at 12:26:31PM +0800, Vijayaraghavan Murali wrote: > Any help with these two options (or slicing using CIL in general) would > be greatly appreciated.
I have really no idea what these options do and how they work. They were coded long before I joined the project and I do not actively maintain the extensions in fact, only the core of CIL. You can try to have a look in src/ext. The relevant files seem to be dataslicing.ml (which mentions that the code must be annotated with "region(i)" attributes, option --doDataSlicing), epicenter.ml (which removes functions more the n hops away from a given function, option --doepicenter) and astslicer.ml (which is not accessible as a feature for some reason, maybe it does not even compile anymore). In case you missed that in the doc, attributes are written like this: http://kerneis.github.com/cil/doc/html/cil/attributes.html#toc7 So for the doDataSlicing, it should look like (completely untested): int (__attribute__((region(1))) f)(int x, int y); /* function declaration */ struct foo { int __attribute__((region(2))) x; int y; } /* some field in a struct */ I hope someone will have a more useful answer. Do not hesitate to share your findings here for future users. Best, -- Gabriel ------------------------------------------------------------------------------ Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev _______________________________________________ CIL-users mailing list CIL-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/cil-users
