|
|
Subscribe / Log in / New account

Filesystems, testing, and stable trees

Filesystems, testing, and stable trees

Posted Jun 1, 2022 13:44 UTC (Wed) by metan (subscriber, #74107)
In reply to: Filesystems, testing, and stable trees by Paf
Parent article: Filesystems, testing, and stable trees

In LTP we do maintain all of that as a static data. For example all that is needed to set minimal kernel version required for the functionality is to set the 'min_kver' variable in a test, the rest of the magic happens in the test library. Generally it takes a lot of work to properly track all the metadata such as kernel commits that fixes regression the test is testing etc., however all this efforts actually pays for itself in the long term. After we managed to tag significant amount of tests with different kinds of metadata the number of emails and issues asking for a help with test failures dropped significantly.

Also I would say that this is more about setting the right guidelines and culture rather than being a technical problem. Once implementation and processes are in place all new tests include the right metadata from a start and everything grows into the right direction effortlessly.


to post comments


Copyright © 2025, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds