Re: [PATCH v2 2/9] tests: fix test-io-channel-command on win32

[Thomas Huth]

On 29/01/2023 19.24, marcandre.lur...@redhat.com wrote:
From: Marc-André Lureau <marcandre.lur...@redhat.com>
socat "PIPE:"" on Windows are named pipes, not fifo path names.

Fixes: commit 68406d10859 ("tests/unit: cleanups for test-io-channel-command")
Signed-off-by: Marc-André Lureau <marcandre.lur...@redhat.com>
---
  tests/unit/test-io-channel-command.c | 6 ++++++
  1 file changed, 6 insertions(+)

diff --git a/tests/unit/test-io-channel-command.c 
b/tests/unit/test-io-channel-command.c
index 096224962c..e76ef2daaa 100644
--- a/tests/unit/test-io-channel-command.c
+++ b/tests/unit/test-io-channel-command.c
@@ -31,8 +31,12 @@ static char *socat = NULL;
  
  static void test_io_channel_command_fifo(bool async)
  {
+#ifdef WIN32
+    const gchar *fifo = TEST_FIFO;

Question from a Windows ignorant: Won't this cause a race condition in case 
someone is trying to run tests in parallel (i.e. shouldn't there be a random 
part in the name)? Or are these named pipes local to the current process?

 Thomas

+#else
      g_autofree gchar *tmpdir = g_dir_make_tmp("qemu-test-io-channel.XXXXXX", 
NULL);
      g_autofree gchar *fifo = g_build_filename(tmpdir, TEST_FIFO, NULL);
+#endif
      g_autofree gchar *srcargs = g_strdup_printf("%s - PIPE:%s,wronly", socat, 
fifo);
      g_autofree gchar *dstargs = g_strdup_printf("%s PIPE:%s,rdonly -", socat, 
fifo);
      g_auto(GStrv) srcargv = g_strsplit(srcargs, " ", -1);
@@ -57,7 +61,9 @@ static void test_io_channel_command_fifo(bool async)
      object_unref(OBJECT(src));
      object_unref(OBJECT(dst));
  
+#ifndef WIN32
      g_rmdir(tmpdir);
+#endif
  }