https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79588
--- Comment #2 from prathamesh3492 at gcc dot gnu.org --- Created attachment 40771 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=40771&action=edit untested patch Hi Jakub, IIUC, the patch punts if param_pos < args->length(), ie, the default value has been passed. I suppose it won't warn for case when an argument passed by caller would equal default argument ? for eg: int obj; bool foo (int *__restrict q, int *__restrict p = &obj) { return false; } int bar () { return foo(&obj) == 0; } param_pos for p is 1 which equals number of arguments passed and so warn_for_restrict won't get called ? I have attached untested patch that tries to warn for above case by appending default arguments to args vector. Does it look OK ? Thanks, Prathamesh
