[llvm-dev] Query about Clang Static Analyzer behaviour

[David Blaikie via llvm-dev]

It's a testing tool for the analyzer:
https://clang.llvm.org/docs/analyzer/developer-docs/DebugChecks.html#exprinspection-checks

On Tue, Jan 19, 2021 at 12:46 AM Deep Majumder via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
>
> Hi everyone,
> The following is a code from a test in the Clang Static Analyzer (clang/test/Analysis/pointer-to-member.cpp): Why are the last two statements expected to produce a warning?
> struct B {
> int f;
> };
> struct L1 : public B { };
> struct R1 : public B { };
> struct M : public L1, R1 { };
> struct L2 : public M { };
> struct R2 : public M { };
> struct D2 : public L2, R2 { };
>
> void diamond() {
> M m;
>
> static_cast<L1 *>(&m)->f = 7;
> static_cast<R1 *>(&m)->f = 16;
>
> int L1::* pl1 = &B::f;
> int M::* pm_via_l1 = pl1;
>
> int R1::* pr1 = &B::f;
> int M::* pm_via_r1 = pr1;
>
> clang_analyzer_eval(m.*(pm_via_l1) == 7); // expected-warning {{TRUE}}
> clang_analyzer_eval(m.*(pm_via_r1) == 16); // expected-warning {{TRUE}}
> }
> Warm Regards,
> Deep
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev