On 18.01.2016 12:41, Huw Davies wrote:
Fixes compile failures with older versions of gnutls.
Signed-off-by: Huw Davies huw@codeweavers.com
configure.ac | 1 + dlls/bcrypt/bcrypt_main.c | 10 +++++----- 2 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/configure.ac b/configure.ac index 0766edf..0eedf43 100644 --- a/configure.ac +++ b/configure.ac @@ -1270,6 +1270,7 @@ then #include <gnutls/crypto.h>]], [[static typeof(gnutls_mac_get_key_size) *func; if (func) return 0;]])], [WINE_CHECK_SONAME(gnutls,gnutls_global_init,,[GNUTLS_CFLAGS=""],[$GNUTLS_LIBS],[[libgnutls\(-deb0\)\{0,1\}]])])], [GNUTLS_CFLAGS=""])])
- WINE_CHECK_LIB_FUNCS(gnutls_hash,[$GNUTLS_LIBS])
fi WINE_WARNING_WITH(gnutls,[test "x$ac_cv_lib_soname_gnutls" = "x"], [libgnutls ${notice_platform}development files not found, no schannel support.])
Wouldn't it be useful to warn a bit more explicit at compile-time? Otherwise packagers will have a hard time to find out if its supported by their build.
On 18 Jan 2016, at 11:56, Sebastian Lackner wrote:
On 18.01.2016 12:41, Huw Davies wrote:
Fixes compile failures with older versions of gnutls.
Signed-off-by: Huw Davies huw@codeweavers.com
configure.ac | 1 + dlls/bcrypt/bcrypt_main.c | 10 +++++----- 2 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/configure.ac b/configure.ac index 0766edf..0eedf43 100644 --- a/configure.ac +++ b/configure.ac @@ -1270,6 +1270,7 @@ then #include <gnutls/crypto.h>]], [[static typeof(gnutls_mac_get_key_size) *func; if (func) return 0;]])], [WINE_CHECK_SONAME(gnutls,gnutls_global_init,,[GNUTLS_CFLAGS=""],[$GNUTLS_LIBS],[[libgnutls\(-deb0\)\{0,1\}]])])], [GNUTLS_CFLAGS=""])])
- WINE_CHECK_LIB_FUNCS(gnutls_hash,[$GNUTLS_LIBS])
fi WINE_WARNING_WITH(gnutls,[test "x$ac_cv_lib_soname_gnutls" = "x"], [libgnutls ${notice_platform}development files not found, no schannel support.])
Wouldn't it be useful to warn a bit more explicit at compile-time? Otherwise packagers will have a hard time to find out if its supported by their build.
I figured that since this only affects fairly old versions of gnutls, that the extra bloat wouldn't be worth it.
Huw.
On 18 Jan 2016, at 12:00, Huw Davies wrote:
On 18 Jan 2016, at 11:56, Sebastian Lackner wrote:
Wouldn't it be useful to warn a bit more explicit at compile-time? Otherwise packagers will have a hard time to find out if its supported by their build.
I figured that since this only affects fairly old versions of gnutls, that the extra bloat wouldn't be worth it.
Actually, I guess it's not too bad, I'll resend with a warning.
Thanks, Huw.
-
Huw Davies -
Sebastian Lackner