I am looking for some programs for performing empirical measurements of a new method of static analysis for resolving virtual function calls. My rough requirements are:

* The programs should be publicly available (the credibility measurements on programs that they cannot see. * The programs should be real applications, preferably not very large (say upto 20KLoC). * They should have a liberal use of virtual functions. * They should be compilable by gcc/g++. * They should have dependencies on special libraries or other programs. The build progress should be simple. Ideally, the usual sequence of `./configure' followed by 'make' should do the trick. * They should be reasonably clean programs. Some C++ features are compiled in an obscure manner by gcc/g++ and we seem to be spending much more time in fighting the IRs produced by these obscure compilations than in the actual problems at hand. Spec benchmarks have troubled us a lot on this front.

We are looking at the programs on sourceforge and github but there are no easy ways of identifying such programs. Any suggestions would be highly appreciated.

Thanks and regards,

Uday Khedker.