================ @@ -1,4 +1,9 @@ -// RUN: %clang_analyze_cc1 -analyzer-checker=core,debug.ExprInspection -verify %s -Wno-undefined-bool-conversion +// RUN: %clang_analyze_cc1 \ +// RUN: -analyzer-checker=core,debug.ExprInspection,unix.Malloc \ +// RUN: -verify %s \ +// RUN: -Wno-undefined-bool-conversion +// unix.Malloc is necessary to model __builtin_alloca, +// which could trigger an "unexpected region" bug in StackAddrEscapeChecker. ---------------- NagyDonat wrote:
Malloc / etc. modeling should not be a dependency of `StackAddrEscapeChecker` (and in particular the `unix.Malloc` part of `MallocChecker` cannot be a strong dependency of anything, because it can generate reports, while only modeling checkers may be dependencies of other checkers). I don't think that we __need__ a separate testcase to see that `StackAddrEscapeChecker` does not crash on `__builtin_alloca` when `unix.Malloc` is not enabled -- the reason for this crash was that `unix.Malloc` introduced an unusual region type (`AllocaRegion`) and without that modeling the return value of `__builtin_alloca` is just a generic conjured region which should be already covered by the many other testcases. However, I'm not opposed to adding that testcase if you would like to see it. https://github.com/llvm/llvm-project/pull/109655 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
