This patch doesn't seem right to me.
Why doesn't "gssx_cred remote = {};" work?
On Tue, 2016-09-27 at 12:50 +0000, git repository hosting wrote:
> This is an automated email from the git hooks/post-receive script.
>
> simo pushed a commit to branch master
> in repository gssproxy.
>
> commit 0c04bee1c220812d5f55095688fb2d4256a71d43
> Author: Robbie Harwood <rharwood(a)redhat.com>
> Date: Mon Sep 19 22:01:07 2016 +0000
>
> Appease clang with memset instead of initializer
>
> Signed-off-by: Robbie Harwood <rharwood(a)redhat.com>
> Reviewed-by: Simo Sorce <simo(a)redhat.com>
> Closes 31
> ---
> proxy/src/mechglue/gpp_creds.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/proxy/src/mechglue/gpp_creds.c
> b/proxy/src/mechglue/gpp_creds.c
> index 2818dd7..31ad9d4 100644
> --- a/proxy/src/mechglue/gpp_creds.c
> +++ b/proxy/src/mechglue/gpp_creds.c
> @@ -183,9 +183,11 @@ OM_uint32 gppint_get_def_creds(OM_uint32
> *minor_status,
>
> /* Then try with remote */
> if (behavior == GPP_REMOTE_ONLY || behavior == GPP_REMOTE_FIRST)
> {
> - gssx_cred remote = {0};
> + gssx_cred remote;
> gssx_cred *premote = NULL;
>
> + memset(&remote, 0, sizeof(gssx_cred));
> +
> /* We intentionally ignore failures as finding creds is
> optional */
> maj = retrieve_remote_creds(&min, name ? name->remote :
> NULL, &remote);
> if (maj == GSS_S_COMPLETE) {
>
>