Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

Azure DevOps pipeline 'view changes' show wrong number of changes

Is anyone faced with such problem as wrong number in "view xxx changes" in build? I dunno why, but it show not only commits between build, but everything else! For example I had two commits between releases, but it show me 300+ changes. My git showed below. I expect 'View changes' show me only commits between r-2.8 and r-2.9.

enter image description here

enter image description here

enter image description here

enter image description here

like image 723
Vladimir Titkov Avatar asked Nov 26 '22 23:11

Vladimir Titkov


1 Answers

The following list describes which changesets are shown in the run details for different pipeline run types:

  • For "Individual CI" runs
    • Show all changesets in the entire branch
  • For "PR automated" runs
    • Show diff between PR source branch and PR dest branch
  • For "Manually triggered" runs
    • XXX Determined not true: Show the run branch's diff between current changeset and previous merge into that branch

This info is based on walking through many YAML pipeline runs. Corrections/refinements to this answer would be helpful.

This is news to me as of today (and I have an audit meeting in an hour, so this will be interesting).

I think that this means that CI runs are not useful for the purpose of audit, change management, release notes, etc.

If the above info holds, then we'll have to sort out how to have useful info in pipeline run history. May be that CI is turned off and YAML conditionals take care of what can be deployed where, with PR triggers and manual runs as the possible initiators.

Update:

A deployment from today showed all changesets for the branch rather than diffs, so we need to rule that one out.

Update - 2021-09-08:

The Environments area under Pipelines seems to be accurate every time. It's a different approach and is a few clicks away from the pipeline run details screen, but it does happen automatically if you're using the YAML "deployment" type of job.

like image 182
GaTechThomas Avatar answered Dec 08 '22 00:12

GaTechThomas