Since some change during last 3 months, regrtest.py on 2.7 branch prints "${test_file_name} skipped" message twice for each testfilename,firstlydirectlyaftergiven{test_file_name}, firstly directly after given testfilename,firstlydirectlyaftergiven{test_file_name}, and secondly after the following test. Example: 0:00:06 [ 17/401] test_abc 0:00:06 [ 18/401] test_abstract_numbers 0:00:06 [ 19/401] test_aepack test_aepack skipped -- No module named aetypes 0:00:06 [ 20/401] test_aifc -- test_aepack skipped 0:00:06 [ 21/401] test_al test_al skipped -- No module named al 0:00:06 [ 22/401] test_anydbm -- test_al skipped 0:00:08 [ 23/401] test_applesingle test_applesingle skipped -- No module named MacOS 0:00:08 [ 24/401] test_argparse -- test_applesingle skipped 0:00:14 [ 25/401] test_array 0:00:15 [ 26/401] test_ascii_formatd 0:00:15 [ 27/401] test_ast
The output is correct. First, regrtest says that the test is running. Then the test is skipped and a message is logged. Finally, the test result is logged: the test was skipped. Using -j0, the output is less surprising ;-)
This bug in output actually occurs in sequential mode. Apparently problem can be reproduced in 3.*, but only if Tools/scripts/run_tests.py is changed to run tests sequentially. (That problem reported as issue 31098.) Example from 3.7: 0:00:42 load avg: 1.89 [ 11/405] test_cprofile 0:00:42 load avg: 1.89 [ 12/405] test_winsound test_winsound skipped -- Use of the 'audio' resource not enabled 0:00:43 load avg: 1.89 [ 13/405] test_pickletools -- test_winsound skipped (resource denied) 0:00:46 load avg: 1.82 [ 14/405] test_select 0:00:58 load avg: 1.54 [ 15/405] test_bisect