[llvm-bugs] [Bug 33694] New: False positive reported by static analyzer: "Potential memory leak"

via llvm-bugs llvm-bugs at lists.llvm.org
Wed Jul 5 13:27:30 PDT 2017


            Bug ID: 33694
           Summary: False positive reported by static analyzer:
                    "Potential memory leak"
           Product: clang
           Version: 2.7
          Hardware: Macintosh
                OS: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
          Assignee: kremenek at apple.com
          Reporter: jaf at meyersound.com
                CC: llvm-bugs at lists.llvm.org

Created attachment 18754
  --> https://bugs.llvm.org/attachment.cgi?id=18754&action=edit
Self-contained example program that reproduces the false-positive when you
compile it using scan-build

Hi all,

When I run the attached trivial example program through clang's scan-build
utility, I get a warning message that I believe is a false positive:

$ scan-build g++ bogus_memory_leak_warning.cpp 
scan-build: Using '/Users/jaf/checker-279/bin/clang' for static analysis
bogus_memory_leak_warning.cpp:26:4: warning: Potential memory leak
1 warning generated.
scan-build: 1 bug found.
scan-build: Run 'scan-view
to examine bug reports.

AFAICT the code in the trivial example program doesn't have any actual memory
leaks, but SA thinks that it does.  (If I'm wrong about that, I apologize). 
Also note that if I comment out line 25 in the attached file, the
false-positive warning goes away; however I think the presence of that line
should not make any difference regarding whether the program has a memory leak
or not.

(This example application is just a minimal self-contained example, of course;
the actual codebase where I see this sort of false positive pop up is much
larger, but this example is easier to understand and debug)


You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20170705/5ebc9974/attachment.html>

More information about the llvm-bugs mailing list