Kostiantyn Kostiuk <kkost...@redhat.com> writes:
> On Sat, Jul 19, 2025 at 9:27 AM Markus Armbruster <arm...@redhat.com> wrote:
>
>> Kostiantyn Kostiuk <kkost...@redhat.com> writes:
>>
>> > g_win32_error_message - translate a Win32 error code
>> > (as returned by GetLastError()) into the corresponding message.
>> >
>> > In the same time, we call error_setg_win32_internal with
>> > error codes from different Windows componets like VSS or
>> > Performance monitor that provides different codes and
>> > can't be converted with g_win32_error_message.
>>
>> Are these error codes from GetLastError()?
>>
>
> No.
> VSS functions directly return an error code.
> Section: Return value -
> https://learn.microsoft.com/en-us/windows/win32/api/vsbackup/nf-vsbackup-ivssbackupcomponents-addtosnapshotset
>
> Performance Counters API can return a system error code or a PDH error code.
> Section: Return value -
> https://learn.microsoft.com/en-us/windows/win32/api/pdh/nf-pdh-pdhopenqueryw
> System error code = GetLastError, PDH error code, something else.
>
> https://learn.microsoft.com/en-us/windows/win32/perfctrs/pdh-error-codes
> FormatMessage requires LoadLibrary(L"pdh.dll") to work properly.
The error code error_setg_win32() takes is passed to
g_win32_error_message(). Contract:
g_win32_error_message ()
gchar *
g_win32_error_message (gint error);
Translate a Win32 error code (as returned by GetLastError() or
WSAGetLastError()) into the corresponding message. The message is
either language neutral, or in the thread's language, or the user's
language, the system's language, or US English (see docs for
FormatMessage()). The returned string is in UTF-8. It should be
deallocated with g_free().
Parameters
error error code.
Returns
newly-allocated error message
https://www.manpagez.com/html/glib/glib-2.46.0/glib-Windows-Compatibility-Functions.php#g-win32-error-message
Passing error codes from sources other than GetLastError() or
WSAGetLastError() violates this contract.
Apparently, g_win32_error_message() returns NULL then. This is not
documented behavior.
Your fix relies on this undocumented behavior.
I believe we should instead fix the misuses of error_setg_win32().
[...]
