Mark junit test cases as skipped if no pickle step results available

[mrsheepuk]

🤔 What's changed?

This allows the junit reporter not to panic if the StopOnFirstFailure flag is used. It marks any subsequent test cases as 'skipped' (seemed like the best choice).

I kept the change as small as possible by recovering from the panics, instead of introducing a suite of "non-must' equivalents on the Storage struct. While that would have been far cleaner as code (recovering from panics feels decidedly icky), this solves the issue with relatively little impact to anything else. But if y'all think adding non-must equivalents onto the Storage struct is a better call, happy to make that change.

⚡️ What's your motivation?

Fixes #387

🏷️ What kind of change is this?

• 🐛 Bug fix (non-breaking change which fixes a defect)

📋 Checklist:

• I agree to respect and uphold the Cucumber Community Code of Conduct
• Users should know about my change
  • I have added an entry to the "Unreleased" section of the CHANGELOG, linking to this pull request.

[codecov]

Codecov Report

Attention: Patch coverage is 75.00000% with 9 lines in your changes are missing coverage. Please review.

Project coverage is 83.27%. Comparing base (153db4e) to head (e82afb8).
Report is 6 commits behind head on main.

Files	Patch %	Lines
internal/formatters/fmt_junit.go	75.00%	4 Missing and 5 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #597      +/-   ##
==========================================
+ Coverage   83.21%   83.27%   +0.06%     
==========================================
  Files          28       29       +1     
  Lines        3413     2858     -555     
==========================================
- Hits         2840     2380     -460     
+ Misses        458      357     -101     
- Partials      115      121       +6     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Johnlon]

Ant tests?

[mrsheepuk]

Ant tests?

Sorry just noticed this comment - if this approach is acceptable @Johnlon (recovering panics instead of changing the return signatures of everything), I'm happy to add tests etc in order to get this mergeable.

Will take a look shortly at what is needed.

[Johnlon]

I'm on holiday so can't remind myself about this.
your comment,"instead of changing signatures" sounds like you might be trying to make a compromise? maybe? anyway I wouldn't bother with compromising anything. just do what ever is the simple and clean thing to improve the code base and functionality.

[mrsheepuk]

I'm on holiday so can't remind myself about this. your comment,"instead of changing signatures" sounds like you might be trying to make a compromise? maybe? anyway I wouldn't bother with compromising anything. just do what ever is the simple and clean thing to improve the code base and functionality.

Hi there - sorry I was myself away on holiday when you replied @Johnlon :)

This PR was the smallest change to fix the issue - by recovering the panics that cause the issue and handling them that way. It works, but a more 'idiomatic go' solution would be to change the function signature that is panicking to return an error, and change all the places it is used to handle the error.

That would be a much, much bigger change though - hence my question about whether this approach of just catching the panics was acceptable here.

If there's no objection to this panic-recovering approach, it would be my preference to keep the change small and I'll add some tests to cover it as-is.