I found what I believe is a false positive with g++ 12 when using the "null-dereference" warning and istreambuf_iterator with any level of optimisation. And of course the "-Werror" flag is mandatory in my company process. I have searched for such similar bug report. This behavior does not show up with g++ 11.2.0 or g++ 9.4.0. Minimal example (basically the first lines of the cppreference example for istreambuf_iterator): #include <string> #include <sstream> int main() { std::istringstream in{"Hello, world"}; std::istreambuf_iterator<char> it(in), end; std::string ss{it, end}; return 0; } Compiled with: g++-12 -O -Wnull-dereference <my_file> Compiler version (default g++-12 on Ubuntu 22.04): g++ (Ubuntu 12-20220319-1ubuntu1) 12.0.1 20220319 (experimental) [master r12-7719-g8ca61ad148f] Full version: Using built-in specs. COLLECT_GCC=g++-12 COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/12/lto-wrapper OFFLOAD_TARGET_NAMES=nvptx-none:amdgcn-amdhsa OFFLOAD_TARGET_DEFAULT=1 Target: x86_64-linux-gnu Configured with: ../src/configure -v --with-pkgversion='Ubuntu 12-20220319-1ubuntu1' --with-bugurl=file:///usr/share/doc/gcc-12/README.Bugs --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --prefix=/usr --with-gcc-major-version-only --program-suffix=-12 --program-prefix=x86_64-linux-gnu- --enable-shared --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --libdir=/usr/lib --enable-nls --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --with-default-libstdcxx-abi=new --enable-gnu-unique-object --disable-vtable-verify --enable-plugin --enable-default-pie --with-system-zlib --enable-libphobos-checking=release --with-target-system-zlib=auto --enable-objc-gc=auto --enable-multiarch --disable-werror --enable-cet --with-arch-32=i686 --with-abi=m64 --with-multilib-list=m32,m64,mx32 --enable-multilib --with-tune=generic --enable-offload-targets=nvptx-none=/build/gcc-12-OcsLtf/gcc-12-12-20220319/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-12-OcsLtf/gcc-12-12-20220319/debian/tmp-gcn/usr --enable-offload-defaulted --without-cuda-driver --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu Thread model: posix Supported LTO compression algorithms: zlib zstd gcc version 12.0.1 20220319 (experimental) [master r12-7719-g8ca61ad148f] (Ubuntu 12-20220319-1ubuntu1)
The theory of the warning seems to be that if istreambuf_iterator::_M_get, called from operator* for *__beg in _M_construct, hits EOF, it clears _M_sbuf, and then ++__beg will try to refer to members of the now-null __beg._M_sbuf. At first glance, this seems like a plausible theory. Why does _M_get clear _M_sbuf? int_type _M_get() const { int_type __ret = _M_c; if (_M_sbuf && _S_is_eof(__ret) && _S_is_eof(__ret = _M_sbuf->sgetc())) _M_sbuf = 0; return __ret; }
Might be due to the changes in r253417
GCC 12.2 is being released, retargeting bugs to GCC 12.3.
(In reply to Jason Merrill from comment #1) > Why does _M_get clear _M_sbuf? Because that's what the standard implies should happen: If the end of stream is reached (streambuf_type::sgetc() returns traits::eof()), the iterator becomes equal to the end-of-stream iterator value." charT operator*() const; 1 Returns: The character obtained via the streambuf member sbuf_->sgetc(). But this means that a dereferenceable iterator can become non-dereferenceable as a result of dereferencing it ... which seems very wrong indeed. It seems better to check for EOF on operator++ and only clear _M_sbuf in that function, but we need to basically rewrite the entire class (without ABI changes) to do that.
I've implemented the suggested changes to istreamubf_iterator and also proposed them as a resolution for LWG 2366 https://wg21.link/lwg2366
GCC 12.3 is being released, retargeting bugs to GCC 12.4.
*** Bug 112421 has been marked as a duplicate of this bug. ***
*** Bug 115067 has been marked as a duplicate of this bug. ***
GCC 12.4 is being released, retargeting bugs to GCC 12.5.