Re: ncurses 5.3 backwards compatibility

[Thomas E. Dickey]

On Fri, 8 Nov 2002, Chris Zubrzycki wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> it seems that it is defined as such. here is the build output:

perhaps libtool is confusing the configure check (config.log would show
just what went wrong, if there are any messages).  You might try just
editing config.status after running the configure script to remove the
line containing BROKEN_LINKER, then run config.status again before
starting the make.

> checking if you want broken-linker support code...

what's odd is that I don't see how this shell fragment would define
BROKEN_LINKER.  Note that the AC_MSG_RESULT in your trace is empty,
and the system name certainly isn't cygwin:

AC_MSG_CHECKING(if you want broken-linker support code)
AC_ARG_ENABLE(broken_linker,
        [  --enable-broken_linker  compile with broken-linker support code],
        [with_broken_linker=$enableval],
        [with_broken_linker=$BROKEN_LINKER])
AC_MSG_RESULT($with_broken_linker)

BROKEN_LINKER=0
if test "$with_broken_linker" = yes ; then
        AC_DEFINE(BROKEN_LINKER)
        BROKEN_LINKER=1
elif test $DFT_LWR_MODEL = shared ; then
        case $cf_cv_system_name in #(vi
        cygwin*)
                AC_DEFINE(BROKEN_LINKER)
                BROKEN_LINKER=1
                CF_VERBOSE(cygwin linker is broken anyway)
                ;;
        esac
fi
AC_SUBST(BROKEN_LINKER)

-- 
T.E.Dickey <address@hidden>
http://invisible-island.net
ftp://invisible-island.net