Switched to robot framework listener version 3 because ver 2 had issues[0][1]
Issues:
when using keyword as test template and keyword had [Setup] and [Teardown] (dunno if related but adding those to examples), listener was not able to fetch the value of the variable. Feels like a scoping bug for variables with RF itself.
log_message in V2 listener interface never triggered from Log keyword from project’s resource files but it does for V3 and message is fully evaluated already at that point
So far, GHAReports did not work with Pabot but I made a small script today (that i can integrate into package) that uses ResultVisitor instead of runtime Listener and can generate github summary from just output.xml.
Feature is still undocumented because it might not contain bugs and/or breaking your reporting when using standalone ghareports.
But changes are; there’s now a cli tool added to your path called ghareports. You can pass -r output.xml to tell the tool where your robot output is. And if you want additional markdown to be generated beside just the github action summary, -m your_summary.md
-h / --help prints out available options.
And now you can run your tests with pabot and/or re-run failed tests and generate github actions summary afterwards you have up to date output.xml available.
And, here’s still one more example, what ghareports actually does:
Fixes issue where ghareports crashed when processing combined output.xml from pabot/rebot combo where start and stop timestamps where not available for the suite.
Passing/Failing & Warning sections are now collapsable to add a bit more readability.
and instead of GHAReports cli, you can run it also via python -mGHAReports as in some environments, generated executable wrappers will not work due to, say windows group policies.