QA (public) » openQA Project (public)

Motivation¶

From https://suse.slack.com/archives/C02AJ1E568M/p1705482261366879?thread_ts=1705481511.459859&cid=C02AJ1E568M

(Oliver Kurz) maybe changes have broken the dry run
(Tina Müller) yeah, line 161 is doing a openqa-cli POST request and expecting a JSON response. that won't work if it's replaced with an echo

Command-line used:

host=aquarius.suse.cz dry_run=1 exclude_no_group=false ./openqa-investigate 20047

jq (161 ./openqa-investigate): jq: parse error: Invalid numeric literal at line 1, column 11 (rc: 5 Input: >>>openqa-cli api --header User-Agent: openqa-investigate (https://github.com/os-autoinst/scripts) --host https://aquarius.suse.cz --retries=3 -X POST jobs/20047/comments text=Starting investigation for job 20047<<<)

Acceptance Criteria¶

• AC1: Dry run does not fail

Suggestions¶

• The code is POSTing a comment and then GETting it back, so we need to mock the response(s)
• Alternatively to mocking, add some input validation to the code block in question and make it do nothing
  • e.g. add an early return in sync_via_investigation_comment() echoing a made-up comment id
• Add a unit test for it (Test-Driven Development)
• "dry run" should be understood to potentially read data from external services, but never modify anything or write changes back -> Don't leave behind any traces
  • Consider documenting this somewhere