http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51755
--- Comment #7 from Manuel López-Ibáñez <manu at gcc dot gnu.org> 2012-01-07 13:03:54 UTC --- (In reply to comment #6) > Thus by 'opposite' of 'complete' you mean that nothing should be *added* to > the > patch, instead something removed from it? Great. You see I meant completing Yes. But probably the patch needs to be substantially changed. And you should check with Joseph whether he is happy with the additional complexity added. > *the work*, thus committing and closing the PR. Anyway, Manuel, would you be > willing to send over a patch answeing Joseph' concerns? Otherwise I'll try to > do it myself, but probably it will take more time because I have start from > scratch on this. Thanks! Sorry, I don't have enough free time to do this. Also, I am not sure anymore that it is worth to add these checks to gcc, which is not meant to be a static analysis tool or be used to build one. GCC FEs are very limited in the kind of analysis they can make, and this seems to be on purpose. It would be better to focus on scan-build http://clang-analyzer.llvm.org/scan-build.html or build a GPL static analyzer on top of clang.
