On Fri, 2008-05-16 at 06:26 -0400, Bob Rossi wrote: > configure:40279: gcc -o conftest -g -Wall -pthread -DLINUX=2 -D_REENTRANT > -D_GNU_SOURCE conftest.c -lcrypt >&5 > configure:40285: $? = 0 > configure:40313: result: -lcrypt
So, the test for crypt actually succeeds. Maybe the whole thing is caused by the fact that we check for crypt after we did DBD stuff. Can you tell me if the attached patch works for you? PS. The patch is against 1.3.x, but it should apply against 1.2.x as well, with some fuzz. -- Bojan
Index: configure.in =================================================================== --- configure.in (revision 655406) +++ configure.in (working copy) @@ -142,6 +142,21 @@ ]) AC_SUBST(APR_ICONV_DIR) +AC_SEARCH_LIBS(crypt, crypt ufc) +AC_MSG_CHECKING(if system crypt() function is threadsafe) +if test "x$apu_crypt_threadsafe" = "x1"; then + AC_DEFINE(APU_CRYPT_THREADSAFE, 1, [Define if the system crypt() function is threadsafe]) + msg="yes" +else + msg="no" +fi +AC_MSG_RESULT([$msg]) + +AC_CHECK_FUNCS(crypt_r, [ crypt_r="1" ], [ crypt_r="0" ]) +if test "$crypt_r" = "1"; then + APU_CHECK_CRYPT_R_STYLE +fi + dnl Find LDAP library dnl Determine what DBM backend type to use. dnl Find Expat @@ -159,21 +174,6 @@ APU_FIND_EXPAT APU_FIND_ICONV -AC_SEARCH_LIBS(crypt, crypt ufc) -AC_MSG_CHECKING(if system crypt() function is threadsafe) -if test "x$apu_crypt_threadsafe" = "x1"; then - AC_DEFINE(APU_CRYPT_THREADSAFE, 1, [Define if the system crypt() function is threadsafe]) - msg="yes" -else - msg="no" -fi -AC_MSG_RESULT([$msg]) - -AC_CHECK_FUNCS(crypt_r, [ crypt_r="1" ], [ crypt_r="0" ]) -if test "$crypt_r" = "1"; then - APU_CHECK_CRYPT_R_STYLE -fi - so_ext=$APR_SO_EXT lib_target=$APR_LIB_TARGET AC_SUBST(so_ext)