[PATCH] Set _WIN32_WINNT in common.m4 configure check

[PATCH] Set _WIN32_WINNT in common.m4 configure check

[Tom Tromey]

GCC recently added support for the Windows thread model, enabling
libstdc++ to support Windows natively.  However, this supporrt
requires a version of Windows later than the minimum version that is
supported by GDB.

PR build/29966 points out that the GDB configure test for std::thread
does not work in this situation, because _WIN32_WINNT is not defined
in test program, and so <thread> seems to be fine.

This patch is an attempt to fix the problem, by using the same setting
for _WIN32_WINNT at configure time as is used at build time.

I don't have access to one of the older systems so I don't think I can
truly test this.  I did do a mingw cross build, though.  I'm going to
ask the bug reporter to test it.

Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=29966
---
 gdb/configure            | 12 +++++++++++-
 gdbserver/configure      | 12 +++++++++++-
 gdbsupport/common-defs.h |  4 +++-
 gdbsupport/common.m4     | 13 ++++++++++++-
 gdbsupport/configure     | 12 +++++++++++-
 5 files changed, 48 insertions(+), 5 deletions(-)

diff --git a/gdb/configure b/gdb/configure
index 0591d187f0a..0455af120f3 100755
--- a/gdb/configure
+++ b/gdb/configure
@@ -24969,7 +24969,17 @@ if ${gdb_cv_cxx_std_thread+:} false; then :
 else
   cat confdefs.h - <<_ACEOF >conftest.$ac_ext
 /* end confdefs.h.  */
-#include <thread>
+  #if defined (__MINGW32__) || defined (__CYGWIN__)
+    # ifdef _WIN32_WINNT
+    #  if _WIN32_WINNT < 0x0501
+    #   undef _WIN32_WINNT
+    #   define _WIN32_WINNT 0x0501
+    #  endif
+    # else
+    #  define _WIN32_WINNT 0x0501
+    # endif
+    #endif	/* __MINGW32__ || __CYGWIN__ */
+    #include <thread>
     void callback() { }
 int
 main ()
diff --git a/gdbserver/configure b/gdbserver/configure
index e5e0cb14d03..e807c60c08a 100755
--- a/gdbserver/configure
+++ b/gdbserver/configure
@@ -7951,7 +7951,17 @@ if ${gdb_cv_cxx_std_thread+:} false; then :
 else
   cat confdefs.h - <<_ACEOF >conftest.$ac_ext
 /* end confdefs.h.  */
-#include <thread>
+  #if defined (__MINGW32__) || defined (__CYGWIN__)
+    # ifdef _WIN32_WINNT
+    #  if _WIN32_WINNT < 0x0501
+    #   undef _WIN32_WINNT
+    #   define _WIN32_WINNT 0x0501
+    #  endif
+    # else
+    #  define _WIN32_WINNT 0x0501
+    # endif
+    #endif	/* __MINGW32__ || __CYGWIN__ */
+    #include <thread>
     void callback() { }
 int
 main ()
diff --git a/gdbsupport/common-defs.h b/gdbsupport/common-defs.h
index bcebd7187d9..5e529f6febf 100644
--- a/gdbsupport/common-defs.h
+++ b/gdbsupport/common-defs.h
@@ -70,7 +70,9 @@
 
 /* We don't support Windows versions before XP, so we define
    _WIN32_WINNT correspondingly to ensure the Windows API headers
-   expose the required symbols.  */
+   expose the required symbols.
+
+   NOTE: this must be kept in sync with common.m4.  */
 #if defined (__MINGW32__) || defined (__CYGWIN__)
 # ifdef _WIN32_WINNT
 #  if _WIN32_WINNT < 0x0501
diff --git a/gdbsupport/common.m4 b/gdbsupport/common.m4
index c6b9a25b744..3909ec81ccb 100644
--- a/gdbsupport/common.m4
+++ b/gdbsupport/common.m4
@@ -100,7 +100,18 @@ AC_DEFUN([GDB_AC_COMMON], [
   AC_CACHE_CHECK([for std::thread],
 		 gdb_cv_cxx_std_thread,
 		 [AC_COMPILE_IFELSE([AC_LANG_PROGRAM(
-  [[#include <thread>
+  dnl NOTE: this must be kept in sync with common-defs.h.
+  [[#if defined (__MINGW32__) || defined (__CYGWIN__)
+    # ifdef _WIN32_WINNT
+    #  if _WIN32_WINNT < 0x0501
+    #   undef _WIN32_WINNT
+    #   define _WIN32_WINNT 0x0501
+    #  endif
+    # else
+    #  define _WIN32_WINNT 0x0501
+    # endif
+    #endif	/* __MINGW32__ || __CYGWIN__ */
+    #include <thread>
     void callback() { }]],
   [[std::thread t(callback);]])],
 				gdb_cv_cxx_std_thread=yes,
diff --git a/gdbsupport/configure b/gdbsupport/configure
index a0e4a3f94a4..9433ac41468 100755
--- a/gdbsupport/configure
+++ b/gdbsupport/configure
@@ -8967,7 +8967,17 @@ if ${gdb_cv_cxx_std_thread+:} false; then :
 else
   cat confdefs.h - <<_ACEOF >conftest.$ac_ext
 /* end confdefs.h.  */
-#include <thread>
+  #if defined (__MINGW32__) || defined (__CYGWIN__)
+    # ifdef _WIN32_WINNT
+    #  if _WIN32_WINNT < 0x0501
+    #   undef _WIN32_WINNT
+    #   define _WIN32_WINNT 0x0501
+    #  endif
+    # else
+    #  define _WIN32_WINNT 0x0501
+    # endif
+    #endif	/* __MINGW32__ || __CYGWIN__ */
+    #include <thread>
     void callback() { }
 int
 main ()
-- 
2.38.1

Re: [PATCH] Set _WIN32_WINNT in common.m4 configure check

[Tom Tromey]

>>>>> "Tom" == Tom Tromey via Gdb-patches <gdb-patches@sourceware.org> writes:

Tom> I don't have access to one of the older systems so I don't think I can
Tom> truly test this.  I did do a mingw cross build, though.  I'm going to
Tom> ask the bug reporter to test it.

BTW, if this works, I think we should consider it for GDB 13.

Tom

Re: [PATCH] Set _WIN32_WINNT in common.m4 configure check

[Eli Zaretskii]

> Cc: Tom Tromey <tromey@adacore.com>
> Date: Mon,  9 Jan 2023 07:47:16 -0700
> From: Tom Tromey via Gdb-patches <gdb-patches@sourceware.org>
> 
> GCC recently added support for the Windows thread model, enabling
> libstdc++ to support Windows natively.  However, this supporrt
> requires a version of Windows later than the minimum version that is
> supported by GDB.
> 
> PR build/29966 points out that the GDB configure test for std::thread
> does not work in this situation, because _WIN32_WINNT is not defined
> in test program, and so <thread> seems to be fine.
> 
> This patch is an attempt to fix the problem, by using the same setting
> for _WIN32_WINNT at configure time as is used at build time.
> 
> I don't have access to one of the older systems so I don't think I can
> truly test this.  I did do a mingw cross build, though.  I'm going to
> ask the bug reporter to test it.

The patch LGTM, but of course actually testing on an old system would
be the best.

Thanks.

Re: [PATCH] Set _WIN32_WINNT in common.m4 configure check

[Tom Tromey]

>>>>> "Tom" == Tom Tromey via Gdb-patches <gdb-patches@sourceware.org> writes:

Tom> GCC recently added support for the Windows thread model, enabling
Tom> libstdc++ to support Windows natively.  However, this supporrt
Tom> requires a version of Windows later than the minimum version that is
Tom> supported by GDB.

Tom> PR build/29966 points out that the GDB configure test for std::thread
Tom> does not work in this situation, because _WIN32_WINNT is not defined
Tom> in test program, and so <thread> seems to be fine.

Tom> This patch is an attempt to fix the problem, by using the same setting
Tom> for _WIN32_WINNT at configure time as is used at build time.

Tom> I don't have access to one of the older systems so I don't think I can
Tom> truly test this.  I did do a mingw cross build, though.  I'm going to
Tom> ask the bug reporter to test it.

There turns out to also be some small problem in libstdc++ so this patch
can't fully be confirmed.  However it also seems like an improvement, so
I am going to check it in.

Joel, I tend to think we should consider this for GDB 13.  Without it,
GDB 13 won't build with a newer GCC on some older Windows systems.

Tom

Re: [PATCH] Set _WIN32_WINNT in common.m4 configure check

[Joel Brobecker]

Hi Tom,

> Tom> GCC recently added support for the Windows thread model, enabling
> Tom> libstdc++ to support Windows natively.  However, this supporrt
> Tom> requires a version of Windows later than the minimum version that is
> Tom> supported by GDB.
> 
> Tom> PR build/29966 points out that the GDB configure test for std::thread
> Tom> does not work in this situation, because _WIN32_WINNT is not defined
> Tom> in test program, and so <thread> seems to be fine.
> 
> Tom> This patch is an attempt to fix the problem, by using the same setting
> Tom> for _WIN32_WINNT at configure time as is used at build time.
> 
> Tom> I don't have access to one of the older systems so I don't think I can
> Tom> truly test this.  I did do a mingw cross build, though.  I'm going to
> Tom> ask the bug reporter to test it.
> 
> There turns out to also be some small problem in libstdc++ so this patch
> can't fully be confirmed.  However it also seems like an improvement, so
> I am going to check it in.
> 
> Joel, I tend to think we should consider this for GDB 13.  Without it,
> GDB 13 won't build with a newer GCC on some older Windows systems.

Thanks for the heads up, Tom. Sounds good to me!

-- 
Joel