[llvm-bugs] [Bug 28275] New: Static analyser doesn't respect #pragma GCC diagnostic ignored "-Wnonnull"

[via llvm-bugs]

https://llvm.org/bugs/show_bug.cgi?id=28275

            Bug ID: 28275
           Summary: Static analyser doesn't respect #pragma GCC diagnostic
                    ignored "-Wnonnull"
           Product: clang
           Version: 3.4
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
          Assignee: kremenek at apple.com
          Reporter: sgubmvll at mjt.me.uk
                CC: llvm-bugs at lists.llvm.org
    Classification: Unclassified

I have a method that should not be passed a null pointer under normal
operation, so I have declared it like so (in C):

uint16_t CRC_xmodem(const uint8_t* data, const uint16_t length)
__attribute__((nonnull));

but in unit testing I want to pass it a null pointer to ensure it accepts it
gracefully. To avoid a compiler warning, I do this like so (in C++, using
catch.hpp if that's relevant):

TEST_CASE("CRC_xmodem null pointer produces zero, "[crc]") {
#pragma GCC diagnostic ignored "-Wnonnull"
    CHECK( 0 == CRC_xmodem(NULL, 0) );
#pragma GCC diagnostic pop
}

This is sufficient to suppress the compile-time warnings, but scan-build still
produces a warning:-

snip/crc/crc_unittest.cpp:34:21: warning: Null pointer passed as an argument to
a 'nonnull' parameter
    CHECK( 0 == CRC_xmodem(NULL, 0) );
                ^          ~~~~

Currently I'm suppressing the warning using #ifndef __clang_analyzer__ but it
would be tidier if the clang analyser respected the GCC pragma.

-- 
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/20160623/78b8937b/attachment.html>