[COMMITTERS] pgsql: Fix macro-redefinition warning on MSVC.

2017-09-03 Thread Tom Lane
Fix macro-redefinition warning on MSVC. In commit 9d6b160d7, I tweaked pg_config.h.win32 to use "#define HAVE_LONG_LONG_INT_64 1" rather than defining it as empty, for consistency with what happens in an autoconf'd build. But Solution.pm injects another definition of that macro into ecpg_config.h,

[COMMITTERS] pgsql: Fix macro-redefinition warning on MSVC.

2017-09-03 Thread Tom Lane
Fix macro-redefinition warning on MSVC. In commit 9d6b160d7, I tweaked pg_config.h.win32 to use "#define HAVE_LONG_LONG_INT_64 1" rather than defining it as empty, for consistency with what happens in an autoconf'd build. But Solution.pm injects another definition of that macro into ecpg_config.h,

[COMMITTERS] pgsql: Fix macro-redefinition warning on MSVC.

2017-09-03 Thread Tom Lane
Fix macro-redefinition warning on MSVC. In commit 9d6b160d7, I tweaked pg_config.h.win32 to use "#define HAVE_LONG_LONG_INT_64 1" rather than defining it as empty, for consistency with what happens in an autoconf'd build. But Solution.pm injects another definition of that macro into ecpg_config.h,

[COMMITTERS] pgsql: Fix macro-redefinition warning on MSVC.

2017-09-03 Thread Tom Lane
Fix macro-redefinition warning on MSVC. In commit 9d6b160d7, I tweaked pg_config.h.win32 to use "#define HAVE_LONG_LONG_INT_64 1" rather than defining it as empty, for consistency with what happens in an autoconf'd build. But Solution.pm injects another definition of that macro into ecpg_config.h,

[COMMITTERS] pgsql: Fix macro-redefinition warning on MSVC.

2017-09-03 Thread Tom Lane
Fix macro-redefinition warning on MSVC. In commit 9d6b160d7, I tweaked pg_config.h.win32 to use "#define HAVE_LONG_LONG_INT_64 1" rather than defining it as empty, for consistency with what happens in an autoconf'd build. But Solution.pm injects another definition of that macro into ecpg_config.h,

[COMMITTERS] pgsql: Fix macro-redefinition warning on MSVC.

2017-09-03 Thread Tom Lane
Fix macro-redefinition warning on MSVC. In commit 9d6b160d7, I tweaked pg_config.h.win32 to use "#define HAVE_LONG_LONG_INT_64 1" rather than defining it as empty, for consistency with what happens in an autoconf'd build. But Solution.pm injects another definition of that macro into ecpg_config.h,

[COMMITTERS] pgsql: Fix macro-redefinition warning on MSVC.

2017-09-03 Thread Tom Lane
Fix macro-redefinition warning on MSVC. In commit 9d6b160d7, I tweaked pg_config.h.win32 to use "#define HAVE_LONG_LONG_INT_64 1" rather than defining it as empty, for consistency with what happens in an autoconf'd build. But Solution.pm injects another definition of that macro into ecpg_config.h,

[COMMITTERS] pgsql: Suppress compiler warnings in dshash.c.

2017-09-03 Thread Tom Lane
Suppress compiler warnings in dshash.c. Some compilers complain, not unreasonably, about left-shifting an int32 "1" and then assigning the result to an int64. In practice I sure hope that this data structure never gets large enough that an overflow would actually occur; but let's cast the constan