Project

General

Profile

Actions
Copy link

action #138590

closed

test year_2038_detection: too much detail on data source (implementation detail)

Added by dimstar 6 months ago. Updated 6 months ago.

Status:
Resolved
Priority:
High
Assignee:
favogt
Category:
Bugs in existing tests
Target version:
-
Start date:
2023-10-26
Due date:
% Done:

100%

Estimated time:
Difficulty:

Description

Observation

The y2j38test goes directly onto utmp/wtmp to identify if the system is y2k38 compatible. This is wrong insofar that those files will never be y2k38 compatible

Instead of testing utmp/wtmp directly, the code should test consumers of the data sources, such as

  • last
  • w
  • who

To fix the y2k38 issue, the data is moved to a new database (without breaking ABI of existing consumers); in recent installations, var/log/wtmp for example does not exist at all anymore, only the new location

openQA test in scenario microos-Staging:J-Staging-DVD-x86_64-microos@uefi-staging fails in
year_2038_detection

Test suite description

Reproducible

Fails since (at least) Build J.563.1

Expected result

Last good: J.562.1 (or more recent)

Further details

Always latest result in this scenario: latest

Actions
Copy link
 #1

Updated by dimstar 6 months ago

  • Subject changed from test year_2038_detection: too much detail on data source to test year_2038_detection: too much detail on data source (implementation detail)
Actions
Copy link
 #2

Updated by kukuk 6 months ago

/run/utmp will also go away with the next systemd submission.
So don't reference /run/utmp and /var/log/wtmp anymore in any test.

Actions
Copy link
 #3

Updated by favogt 6 months ago

  • Assignee set to favogt

I'll take this. FWICT who, w and last -F are the most relevant commands. Anything else?

Actions
Copy link
 #4

Updated by favogt 6 months ago

  • Status changed from New to Resolved
  • % Done changed from 0 to 100
Actions
Copy link

Also available in: Atom PDF