https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98001
--- Comment #3 from rguenther at suse dot de <rguenther at suse dot de> --- On Thu, 26 Nov 2020, redi at gcc dot gnu.org wrote: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98001 > > Jonathan Wakely <redi at gcc dot gnu.org> changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > Assignee|unassigned at gcc dot gnu.org |redi at gcc dot > gnu.org > Status|NEW |ASSIGNED > > --- Comment #2 from Jonathan Wakely <redi at gcc dot gnu.org> --- > I don't think we actually need fflush(__file) to set errno in order to verify > that the library doesn't set it to zero. > > So this should be fine: > > --- a/libstdc++-v3/testsuite/ext/stdio_filebuf/char/79820.cc > +++ b/libstdc++-v3/testsuite/ext/stdio_filebuf/char/79820.cc > @@ -26,10 +26,10 @@ void > test01() > { > FILE* f = std::fopen("79820.txt", "w"); > - std::fclose(f); > errno = 127; > __gnu_cxx::stdio_filebuf<char> b(f, std::ios::out, BUFSIZ); > VERIFY(errno == 127); // PR libstdc++/79820 > + std::fclose(f); > } Is the DTOR of stdio_filebuf fine with the file being closed under it? For a fix in our GCC 7 tree I wrapped it in its own scope to ensure the DTOR is called before fclose > int > >
