bug-gnulib
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Compilation problems with stat-w32.c


From: Eli Zaretskii
Subject: Compilation problems with stat-w32.c
Date: Thu, 10 Sep 2020 17:22:20 +0300

I've built today's snapshot of GDB on MS-Windows, and bumped into a
problem: starting the produced binary on Windows XP fails with this
message:

     The procedure entry point GetFinalPathNameByHandleA could not be
     located in the dynamic link library KERNEL32.dll.

This happens because stat-w32.c unconditionally sets _WIN32_WINNT to
the value suitable for Windows 8.0 or later:

  #if defined _WIN32 && ! defined __CYGWIN__

  /* Ensure that <windows.h> defines FILE_ID_INFO.  */
  #if !defined _WIN32_WINNT || (_WIN32_WINNT < _WIN32_WINNT_WIN8)
  # undef _WIN32_WINNT
  # define _WIN32_WINNT _WIN32_WINNT_WIN8
  #endif

which later on causes it to expect GetFinalPathNameByHandle to be in
kernel32.dll, instead of dynamically loading it:

  #if !(_WIN32_WINNT >= _WIN32_WINNT_VISTA)
  [...]
  #else

  # define GetFileInformationByHandleExFunc GetFileInformationByHandleEx
  # define GetFinalPathNameByHandleFunc GetFinalPathNameByHandle

  #endif

The code I elided and replaced with "[...]" sets up to load the above
two APIs dynamically, and is capable of coping with these two
functions being unavailable.  But it is never used, due to the
redefinition of _WIN32_WINNT above.  That looks like a bug, or what am
I missing?

Please also note that these APIs are only needed if
_GL_WINDOWS_STAT_INODES is 2, so for lower values of
_GL_WINDOWS_STAT_INODES this redefinition is not needed at all.

In addition, VOLUME_NAME_NONE is used here, but the Windows headers
define it only when _WIN32_WINNT is greater or equal to
_WIN32_WINNT_VISTA, so for lower values I suggest to simple define it,
something like:

    #elif !defined VOLUME_NAME_NONE
    # define VOLUME_NAME_NONE 0x4
    #endif

Thanks.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]