On Wed, May 21, 2008 at 10:07:25AM +1000, Bojan Smojver wrote: > On Tue, 2008-05-20 at 20:04 -0400, Bob Rossi wrote: > > > I noticed you changed everything but the AC_CHECK_LIB, do you believe > > that is unnecessary, or are you thinking about it? > > Not required. We set LDFLAGS specifically for the test (hence the > APR_ADDTO(LDFLAGS, [$pgsql_LDFLAGS])) and then we restore back the > original after the test. Something like this: > > old_libs="$LIBS" > old_cppflags="$CPPFLAGS" > old_ldflags="$LDFLAGS" > > [... test ...] > > LIBS="$old_libs" > CPPFLAGS="$old_cppflags" > LDFLAGS="$old_ldflags" > > So, we should be all good now, as long as the latest build/dbd.m4 works > for you on all platforms.
OK, the problem is, if you don't add the libraries to the AC_CHECK_LIB, you end up with this, configure:36502: gcc-sjlj -o conftest.exe -g -gstabs -Wall -D_LARGEFILE64_SOURCE -I.../POSTGR~1/POSTGR~1.1/include -L...POSTGR~1/POSTGR~1.1/lib -lpgport -lm -lSecur32 -lws2_32 -lshfolder conftest.c -lpq >&5 which still failed. That's because as you can see the -lpq went at the end which is incorrect. If you add the libs to the AC_CHECK_LIB command, all the other libraries go after the -lpq. Here is some documentation on AC_CHECK_LIB from the autoconf manual, If linking with library results in unresolved symbols that would be resolved by linking with additional libraries, give those libraries as the other-libraries argument, separated by spaces: e.g., -lXt -lX11. Otherwise, this macro fails to detect that library is present, because linking the test program always fails with unresolved symbols. The other-libraries argument should be limited to cases where it is desirable to test for one library in the presence of another that is not already in LIBS. I hope this is helpful. Thank you for your tireless effort! Bob Rossi
