Issue 6258 - Mitigate race condition in paged_results_test.py

[abeisemb]

The regression test dirsrvtests/tests/suites/paged_results/paged_results_test.py::test_multi_suffix_search has a race condition causing it to fail due to multiple queries potentially writing their logs out of chronological order.

This failure is mitigated by sorting the retrieved access_log_lines by their "op" value. This ensures the log lines are in chronological order, as expected by the assertions at the end of test_multi_suffix_search().

Helps fix: #6258

Reviewed by: @droideck , @progier389 (Thanks!)

[abeisemb]

Amended commit to move a helper function inside the test function for readability.

[progier389]

LGTM
FYI: rather than using indexes to extract the string you can also use directly the regular expression module. Here is an use example:

import re

str="xxxx op=123 conn=567 RESULT xxx"
result = re.search(r'(op=\d+ conn=\d+)', str)
print(result.group(1))

[vashirov]

I agree with @progier389, this should be a regex.
Minor nit: in the commit message instead of "Helps fix" please use "Fixes:", see https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/using-keywords-in-issues-and-pull-requests

[abeisemb]

Amended commit to correct the error in the message and replace the string index function with the simpler regex solution. Full test runtime for paged_results_test.py comes in at 6 minutes 46 seconds.

[progier389]

Looks good.

[abeisemb]

Added missing re import.

[droideck]

Besides that, LGTM!

[droideck]

Minor nitpick, but I think we should be as specific as possible to avoid any weird issues:

Suggested change

	access_log_lines.sort(key=lambda x: int(re.search(r"op=(.*?) RESULT", x).group(1)))
	access_log_lines.sort(key=lambda x: int(re.search(r"op=(\d+) RESULT", x).group(1)))

[abeisemb]

Good catch! The commit is amended now with that change.

[abeisemb]

Full test runtime comes in at 7 minutes 2 seconds.