'smbc_new_context() calls smbc_setOptionUseCCache(context, True) if LIBSMBCLIENT_NO_CCACHE environemnt variable is not defined. I suppose it is not defined in your system, but can you please confirm it?
smbclient --kerberos --use-ccache doesn't work as it forces kerberos to be used, however gvfs uses smbc_setOptionFallbackAfterKerberos (op_backend->smb_context, 1), which I suppose should fallback to ccache as well. It would be nice to confirm that downgrade of samba/gvfs package solves the problem, to figure out where is the regression.
In theory, we could break something by the commit a0aec329, but it is 4 years old.
I am not sure how exactly is ccache authentication handled, but first iteration uses smbc_setOptionFallbackAfterKerberos (smb_context, op_backend->user != NULL). I wonder if it helps to specify the username in the URI, e.g. gio mount smb://username@server/share?'
Upstream comment
'smbc_new_context() calls smbc_setOptionU seCCache( context, True) if LIBSMBCLIENT_ NO_CCACHE environemnt variable is not defined. I suppose it is not defined in your system, but can you please confirm it?
smbclient --kerberos --use-ccache doesn't work as it forces kerberos to be used, however gvfs uses smbc_setOptionF allbackAfterKer beros (op_backend- >smb_context, 1), which I suppose should fallback to ccache as well. It would be nice to confirm that downgrade of samba/gvfs package solves the problem, to figure out where is the regression.
In theory, we could break something by the commit a0aec329, but it is 4 years old.
I am not sure how exactly is ccache authentication handled, but first iteration uses smbc_setOptionF allbackAfterKer beros (smb_context, op_backend->user != NULL). I wonder if it helps to specify the username in the URI, e.g. gio mount smb://username@ server/ share?'