> -----Original Message----- > From: Andrew Dunstan [mailto:[EMAIL PROTECTED] > Sent: Wednesday, September 01, 2004 12:31 PM > To: Dann Corbit > Cc: PostgreSQL-development > Subject: Re: [HACKERS] Also cannot build the postgresql > server under Mingw using 8.0 beta 2 > > Dann Corbit wrote: > > > Adding this to the c.h file solved most of the problems for > the libpq > > DLL: > > > > #if defined(HAVE_STRINGS_H) && !defined(_MSC_VER) > > #include <strings.h> > > #endif > > ... > > #if defined(WIN32) && defined(_MSC_VER) > > #include <winsock2.h> > > #define snprintf _snprintf > > #endif > > Since I ran configure for MINGW (which has strings.h) it creates a > > header file which is partly invalid for MS VC++ > > > We do not expect configure to create a setup for any use on Windows > other than building under MinGW.
Then why does the file \postgresql-8.0.0beta2\src\interfaces\libpq say the following: U:\postgresql-8.0.0beta2\src\interfaces\libpq>type win32.mak # Makefile for Microsoft Visual C++ 5.0 (or compat) If you really do build the interface under only one compiler that is a titanic step down from before. BTW, the missing symbol slat problem under Mingw has been isolated as the following lines from pg_os_config.h: /* * Supplement to <sys/stat.h>. */ #define lstat slat ---------------------------(end of broadcast)--------------------------- TIP 4: Don't 'kill -9' the postmaster
