Re: [patch] update M4 for INET_NTOP and INET_PTON checks.

[Simon Josefsson]

The recent inet_?to?.m4 change caused build failures in GnuTLS, and one
patch to fix it is as below.  However, what is the cause here?  Does
some autoconf version use the magic string "none needed" and other
versions "none required"?  Or can both strings happen with recent
versions of autoconf depending on some other setting?  Perhaps the
inet_?to?.m4 code needs to check for both magic strings?

Thanks,
/Simon

Brad Hards <address@hidden> writes:

> Fix:
> diff --git a/gl/m4/inet_ntop.m4 b/gl/m4/inet_ntop.m4                        
> index 85a6bb6..ea645f2 100644
> --- a/gl/m4/inet_ntop.m4
> +++ b/gl/m4/inet_ntop.m4
> @@ -18,7 +18,8 @@ AC_DEFUN([gl_INET_NTOP],
>      [AC_REPLACE_FUNCS([inet_ntop])])
>    LIBS=$gl_save_LIBS
>    INET_NTOP_LIB=
> -  if test "$ac_cv_search_inet_ntop" != "none needed"; then
> +  if test "$ac_cv_search_inet_ntop" != "none needed" &&
> +     test "$ac_cv_search_inet_ntop" != "none required"; then
>      INET_NTOP_LIB="$ac_cv_search_inet_ntop"
>    fi
>    AC_SUBST([INET_NTOP_LIB])
> diff --git a/gl/m4/inet_pton.m4 b/gl/m4/inet_pton.m4
> index f071801..068deee 100644
> --- a/gl/m4/inet_pton.m4
> +++ b/gl/m4/inet_pton.m4
> @@ -18,7 +18,8 @@ AC_DEFUN([gl_INET_PTON],
>      [AC_REPLACE_FUNCS([inet_pton])])
>    LIBS=$gl_save_LIBS
>    INET_PTON_LIB=
> -  if test "$ac_cv_search_inet_pton" != "none needed"; then
> +  if test "$ac_cv_search_inet_pton" != "none needed" &&
> +     test "$ac_cv_search_inet_pton" != "none required"; then
>      INET_PTON_LIB="$ac_cv_search_inet_pton"
>    fi
>    AC_SUBST([INET_PTON_LIB])
>
> Brad