[PATCH v2] libstdc++: Fix std::system_category().message(int) on mingw32 target

[Jonathan Wakely]

On Sat, 10 Jan 2026 at 12:24, Jonathan Wakely <jwakely.gcc@gmail.com> wrote:
>
> On Sat, 10 Jan 2026 at 12:17, Jonathan Wakely <jwakely.gcc@gmail.com> wrote:
> >
> > On Sat, 10 Jan 2026 at 10:45, Wang Jinghao <zheng.xianyuwang@gmail.com> wrote:
> > >
> > > On the mingw32 target, std::system_category().message(int) uses
> > > FormatMessage api to format error messages. When the error message contains
> > > insert sequences, it is unsafe not to use the FORMAT_MESSAGE_OGNORE_INSERTS
> > > flag, as seen at:
> > > https://devblogs.microsoft.com/oldnewthing/20071128-00/?p=24353
> > >
> > > The output of FormatMessage ends with "\r\n" and includes a Full stop character
> > > used by the current thread's UI language. Now, we will remove "\r\n" and
> > > any trailing '.' from the output in any language environment.
> > >
> > > In the testsuite for std::system_category().message(int), we first switch the
> > > thread UI language to en-US tomeet expectations in any language environment.
> >
> > Thanks, the patch looks good now.
> >
> > > libstdc++-v3/ChangeLog:
> > >
> > >         * src/c++11/system_error.cc (system_error_category) [_WIN32]:
> > >         Fix typo in __MINGW32__ macro name.
> > >         Adjust behavior on the mingw32 target.
> >
> > The two lines above relate to the error_category/system_category.cc
> > test, not the src/c++11/system_category.cc source file, right?
> >
> > >         Use FormatMessageA function instead of FormatMessage macro.
> > >         * testsuite/19_diagnostics/error_category/system_category.cc:
> > >         Adjust behavior on the mingw32 target.
> >
> > I've just noticed that there's no Signed-off-by trailer in the patch.
> > Do you have a copyright assignment filed with the FSF?
> >
> > As documented at https://gcc.gnu.org/contribute.html#legal we need
> > either a copyright assignment, or DCO sign-off.
> >
> > If you don't have a copyrigh assignment, could you please add the
> > Signed-off-by trailer if you are willing and able to agree to the
> > terms of https://gcc.gnu.org/dco.html ?
>
> Actually, don't worry, I don't think this change can be considered
> "legally significant", because the changes are small except for some
> whitespace changes and typo fixes.
>
> So I'll push this to the git repo - thanks for fixing it!

Hmm, when I test this with mingw-w64 under Wine, I get:

/home/jwakely/src/gcc/gcc/libstdc++-v3/testsuite/19_diagnostics/error_category/system_category.cc
01ec:fixme:nls:RtlSetThreadPreferredUILanguages 8, 00000001400A8122,
000000000021FE8C
/home/jwakely/src/gcc/gcc/libstdc++-v3/testsuite/19_diagnostics/error_category/system_category.cc:127:
void test03(): Assertion 'msg == "Access is denied"'
failed.
FAIL: /home/jwakely/src/gcc/gcc/libstdc++-v3/testsuite/19_diagnostics/error_category/system_category.cc

Inspecting the 'msg' string, the content is "Access denied", which is
what we have on trunk today.

Can you confirm whether it should be "Access is denied" (as in your
patch) or "Access denied"?