[grilo] tests: really fix common test failure in GrlNet

[Victor Toso <victortoso src gnome org>]

commit 19126d7cf8c9b37bb5f35c63dd2b32e7e7ea0e25
Author: Victor Toso <me victortoso com>
Date:   Wed Aug 29 14:50:18 2018 +0200

    tests: really fix common test failure in GrlNet
    
    Based on intention of commit a176b9a03d04d3f1, we should be going up
    to 1s of threshold to avoid failure such as:
    
        ERROR:../tests/lib-net.c:114:test_net_wc_throttling_cb: assertion
        failed (received_time >= op->expected_time):
        (118905943655 >= 118905946568)
    
    This has been tested over 500 times, with different overloads in the
    machine, no failure. It still should be able to find valid regressions
    on the code as the most important one would be stress test and
    big-delay, the later having a 5 seconds timeout to guarantee that the
    network request has not been made.
    
    Related: https://bugzilla.gnome.org/show_bug.cgi?id=774578
    Signed-off-by: Victor Toso <victortoso gnome org>

 tests/lib-net.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
---
diff --git a/tests/lib-net.c b/tests/lib-net.c
index 565847f..61858b5 100644
--- a/tests/lib-net.c
+++ b/tests/lib-net.c
@@ -138,7 +138,7 @@ throttling_operation_new (Fixture *f, guint delay_in_sec)
   /* Some THRESOLD seems necessary as the test time happens before the actual
    * GrlNetWc computation. This test is more about throttling working than
    * precision. */
-  op->expected_time -= THRESHOLD;
+  op->expected_time -= G_USEC_PER_SEC * THRESHOLD;
 
   id++;
   f->num_operations++;