Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-33520

test main.table_value_constr fails on armhf with output missing extra 'r_total_time_ms' field

Details

    • Bug
    • Status: Open (View Workflow)
    • Minor
    • Resolution: Unresolved
    • 10.11
    • 10.11
    • Tests
    • None

    Description

      This is a minor issue and seems to only affect armhf, so filing with 'trivial' priority.

      I saw this test fail already earlier in MDEV-30411, which led to fixing the test `explain_json_format_partitions`. This Jira is about test `table_value_constr` that still fails with:

      main.table_value_constr                  w1 [ fail ]
              Test ended at 2024-02-22 05:19:56
       
      CURRENT_TEST: main.table_value_constr
      --- /<<PKGBUILDDIR>>/mysql-test/main/table_value_constr.result	2024-02-01 17:36:14.000000000 +0000
      +++ /<<PKGBUILDDIR>>/mysql-test/main/table_value_constr.reject	2024-02-22 05:19:56.815765117 +0000
      @@ -1790,9 +1790,6 @@
       values (1,2),(3,4);
       ANALYZE
       {
      -  "query_optimization": {
      -    "r_total_time_ms": "REPLACED"
      -  },
         "query_block": {
           "union_result": {
             "table_name": "<union1,2>",
       
      mysqltest: Result length mismatch
      

      Full log at https://launchpadlibrarian.net/715550344/buildlog_ubuntu-noble-armhf.mariadb_1%3A10.11.7-2~bpo24.04.1~1708576752.b53aa569b70+bugfix.1063738.revert.c432c9ef_BUILDING.txt.gz

      Attachments

        Issue Links

          Activity

            The test could be making an incorrect assumption about time tracking. There are 4 places where such output can be generated. Here is one of them:

            Filesort_tracker::print_json_members(Json_writer *writer)

              if (get_r_loops() && time_tracker.timed)
              {
                writer->add_member("r_total_time_ms").
                        add_double(time_tracker.get_time_ms());
              }
            

            Perhaps this part of the test was run faster than the timer would have been updated?

            marko Marko Mäkelä added a comment - The test could be making an incorrect assumption about time tracking. There are 4 places where such output can be generated. Here is one of them: Filesort_tracker::print_json_members(Json_writer *writer) if (get_r_loops() && time_tracker.timed) { writer->add_member( "r_total_time_ms" ). add_double(time_tracker.get_time_ms()); } Perhaps this part of the test was run faster than the timer would have been updated?
            otto Otto Kekäläinen added a comment - This was now passing in https://buildd.debian.org/status/fetch.php?pkg=mariadb&arch=armel&ver=1%3A11.4.5-2%7Eexp1&stamp=1740282021&raw=0 main.table_value_constr w2 [ pass ] 181

            People

              psergei Sergei Petrunia
              otto Otto Kekäläinen
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

                Created:
                Updated:

                Git Integration

                  Error rendering 'com.xiplink.jira.git.jira_git_plugin:git-issue-webpanel'. Please contact your Jira administrators.