Replies: 4 comments 9 replies
-
|
The release notes task has no dependencies on any other tasks. It populates the variable using the Azure DevOps REST API based on parameters picked up from the Azure DevOps runtime parameters e.g. the build ID etc. So I am unclear as to why it is not at least populating some of the variables for you. Remember some will be empty if the REST API returns no data for a request e.g. the PR array is empty if the build is not associated with a PR. Could you run your build again but with diagnostics on, either
Have a look at the detailed log, this will show all the calls that have been made and maybe explain the issues. if you upload the log file here I will have a look too |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for your quick response! I am still very new to this so it's entirely possible I'm missing something really obvious. The build source I used for this run had one related work item, a bug fix. The resulting entry in the wiki looks like this:
|
Beta Was this translation helpful? Give feedback.
-
|
I can see you are passing a parameter into the task of You can see in the log it is scanning build but none have this tag. This means the task will just look for WI, CS, PR etc directly associated with this build run.
So the can is done against a build
and the task finds the following.
It finds a CS that probably triggered the build, but there were no WI etc. directly associated This means you should see the build information being expanded and the commit loop The bottom line is it working as I would expect. But, you need to decide how you want to use this task, it is so flexible it is easy to get confused. I would at least as a start
You might find the conference session video I did on this subject help |
Beta Was this translation helpful? Give feedback.
-
|
I believe I'm getting closer. I notice that its attempting to compare this build to previous builds in this same pipeline. All this pipeline does is take the output from the developer's pipeline and deploy it into a couple of environments. The dev pipeline consists of PR builds and release candidates. I think I need to be comparing the release in my pipeline to the builds in the developer pipeline. Either that or there needs to be a history of builds in this new one before it generates meaningful data but I feel like that's not right. I know that's not really a question, just thinking out loud and hoping you can interpret my thoughts into something meaningful. :) Some days I'm just pounding on my keyboard like one of the monkeys trying to reproduce the works of Shakespeare. |
Beta Was this translation helpful? Give feedback.
-
|
My task can only compare a run against previous runs of the same build definition. Maybe you need to generate the release notes in the developers build, maybe publishing them to a wiki when they merge to the trunk. The video I mentioned above explains how we use the task |
Beta Was this translation helpful? Give feedback.
-
|
I reworked the pipeline so that it is a deployment instead of just a build. It runs fine and creates the deploy in the environment as it should. The issue is still with creating the valid markdown files. Not sure why it's saying there no lastMergCommit. Any insight on this? I imagine I'm missing something simple but I don't know what it might be. I admit I'm far over my head here. This is an YAML pipeline, not a "classic" 2021-06-08T23:16:33.3791360Z ##[section]Starting: Generate Release Notes for Wiki |
Beta Was this translation helpful? Give feedback.
-
|
The environment Changes and Workitems show legit entries so I kno the pipeline is seeing stuff. Just not getting reflected in the morkdown for some reason. |
Beta Was this translation helpful? Give feedback.
-
|
Is this a rerun of a pipeline without any code changes? If it is then there will be no commits or work items associated with it. The release notes are based on changes. If you want to see more detail of the logic used in the task, set a build variable |
Beta Was this translation helpful? Give feedback.
-
|
I found another issue which was that none of the previous deploy builds in the environment were tagged "release" I was thinking that was referring to the previous pipeline builds. I manually tagged one as release and I can see that it got recognized. (I had high hopes this was the 'silver bullet') There is a comparison that runs for the 8 commits but nothing appears to match and I don't understand why. 2021-06-09T18:38:30.9733420Z ##[debug]Evaluating condition for step: 'Generate Release Notes for Wiki'
|
Beta Was this translation helpful? Give feedback.
-
|
More clarification here. This is being run in a deployment environment, but the build pipeline is not. I don't know if that matters. |
Beta Was this translation helpful? Give feedback.
-
|
The logs show that the comparison is being done These commits can be added to the release notes using the loop in the general form (adding what ever properties you are interested in within the loop. Remember There has been a scan of all PRs in the repo to see if any match the commits associated with the run, none were found to match. The |
Beta Was this translation helpful? Give feedback.
-
|
If you want to see the raw results to the API call used to get the associated CS and PR you can run Also look at the command line tester to explore what the task can do outside a build pipeline |
Beta Was this translation helpful? Give feedback.
-
|
We did finally get this going by running it against the build pipeline. Thanks for the hand holding! |
Beta Was this translation helpful? Give feedback.
-
I am attempting to use this to create wiki notes for several pipelines/stages. Everything runs fine but nothing in the template gets populated.
At the start of the build pipeline I prompt for resources and pick the pipeline buld I want to deploy to the environment. I use the $(resources.pipeline.mobile.runID) since that identifies the pipeline build I want to use to get my artifacts and begin the deploy process.
I pass that on to the template that downloads the artifact and then publish it:
` jobs:
displayName: Download the artifacts
parameters:
artifactPipeline: $(resources.pipeline.mobile.pipelineName)
artifactRunId: $(resources.pipeline.mobile.runID)`
After that the pipeline goes to the next stage, downloads the published artifact and goes through the build process. During the download the build gets assigned a new number and uses that to compare to previous builds. Does it also try to use it as the build number to match work items to? If so it will always fail.
I feel like I must be missing something that it needs in order to complete the process.
FYI..
The task looks like this (this was a standard template in the doc):
`task: XplatGenerateReleaseNotes@3
displayName: 'Generate Release Notes for Wiki'
inputs:
useAgentToken: true
stopOnRedeploy: true
stopOnError: false
ReplaceFile: true
CheckStage: true
tags: release
outputfile: '$(System.DefaultWorkingDirectory)/$(MOBILE_CI_NAME)/release-notes.md'
localPath: '$(System.DefaultWorkingDirectory)/$(MOBILE_CI_NAME)'
templateLocation: InLine
inlinetemplate: |
## Build {{buildDetails.buildNumber}}
* Branch: {{buildDetails.sourceBranch}}
* Tags: {{buildDetails.tags}}
* Completed: {{buildDetails.finishTime}}
* Previous Build: {{compareBuildDetails.buildNumber}}
Beta Was this translation helpful? Give feedback.
All reactions