在 2020/5/13 下午11:49, Daniel P. Berrangé 写道:
On Sun, May 10, 2020 at 02:14:21PM +0800, xiaoqiang zhao wrote:
add cases to test tight and non-tight for abstract address type
Signed-off-by: xiaoqiang zhao<zxq_yx_...@163.com>
---
tests/test-util-sockets.c | 83 +++++++++++++++++++++++++++++++++++++++
1 file changed, 83 insertions(+)
diff --git a/tests/test-util-sockets.c b/tests/test-util-sockets.c
index 5fd947c7bf..8042fb9276 100644
--- a/tests/test-util-sockets.c
+++ b/tests/test-util-sockets.c
@@ -227,6 +227,84 @@ static void test_socket_fd_pass_num_nocli(void)
g_free(addr.u.fd.str);
}
+static gpointer unix_server_thread_func(gpointer user_data)
+{
+ SocketAddress addr;
+ Error *err = NULL;
+ int fd = -1;
+ int connfd = -1;
+ struct sockaddr_un un;
+ socklen_t len = sizeof(un);
+ char name[] = "/tmp/unix-test.sock";
Fixed filenames are bad, as even though this is the abstract
namespace and thus safe from on-disk clashes, the abstract
namespace is still OS global. We should append both the PID
and a sequence of random bytes to get something which doesnt
clash if two copies of the unit test run concurrently.
agree.