(trafficserver) branch master updated: test_SnowflakeID: Avoid sequence number continuation (#12427)

[bneradt]

This is an automated email from the ASF dual-hosted git repository.

bneradt pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/trafficserver.git


The following commit(s) were added to refs/heads/master by this push:
     new ec956b33b6 test_SnowflakeID: Avoid sequence number continuation 
(#12427)
ec956b33b6 is described below

commit ec956b33b64d34a42141a22519bf303d530431c4
Author: Brian Neradt <brian.ner...@gmail.com>
AuthorDate: Mon Aug 11 17:41:26 2025 -0500

    test_SnowflakeID: Avoid sequence number continuation (#12427)
    
    I noticed the following test failure in a recent CI run:
    
    ```
    SnowflakeID
    
-------------------------------------------------------------------------------
    ../src/tscore/unit_tests/test_SnowflakeID.cc:94
    
...............................................................................
    
    ../src/tscore/unit_tests/test_SnowflakeID.cc:169: FAILED:
      REQUIRE( u1.sequence == 0u )
    with expansion:
      1 == 0
    ```
    
    I realized this can happen if we re-attempt to grab new snowflake IDs in
    what happens to be the same millisecond as the previous attempt. Thus
    the u1 from the new attempt will likely be in the same millisecond as
    the u2 in the previous attempt. Thus u1.squence will then be 1 instead
    of 0. This adds a 2ms sleep in the loop to avoid this situation.
    
    Note this is a timing bug in the test. The production code correctly
    assigned in this situation a sequence of 1 since the ID was in the same
    millisecond as the previous one.
---
 src/tscore/unit_tests/test_SnowflakeID.cc | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/src/tscore/unit_tests/test_SnowflakeID.cc 
b/src/tscore/unit_tests/test_SnowflakeID.cc
index 04f8e1fe02..1c259728d4 100644
--- a/src/tscore/unit_tests/test_SnowflakeID.cc
+++ b/src/tscore/unit_tests/test_SnowflakeID.cc
@@ -117,6 +117,9 @@ TEST_CASE("SnowflakeID", "[libts][SnowflakeID]")
       // Something is seriously wrong...don't infinite loop.
       FAIL("Failed to generate two snowflake IDs in the same millisecond.");
     }
+    // Sleep to ensure we generate IDs in a fresh millisecond, avoiding
+    // sequence number continuation from previous attempts.
+    std::this_thread::sleep_for(std::chrono::milliseconds(2));
     ms_since_unix_epoch_before = 
duration_cast<milliseconds>(system_clock::now().time_since_epoch()).count();
     v1                         = SnowflakeID::get_next_value();
     v2                         = SnowflakeID::get_next_value();