error code 0 on successful --retry

[gondalez]

Summary

A basic working version of --retry based on work by @streetlogics, in the time I had available.

The logging could use some work and it would be nice to get feedback from the maintainers.

I figured I would put this here in case it helps someone else out and maybe it can go part-way towards a fix for #1044 :)

Details

I cherry picked the commits master...streetlogics:issue/retry-success-exit-code-0-1044 on top of master.
See #1044 (comment).

Motivation and Context

I tried using @streetlogics branch but I got an ArgumentError when a feature is retried.
This seems to have been fixed by 31968ae, that was not yet in the remote branch.

How Has This Been Tested?

I tested quickly locally and can confirm the error code is now 0 if a step succeeds on subsequent retry.

[danascheider]

Thanks @gondalez! If you can rebase this off of master I'll help figure out what's breaking in jruby.

[gondalez]

@danascheider no probs; rebase is done and the jruby build passed this time 🎉

[olleolleolle]

Golfing the code:

• an unless with != could become an if with a ==

Would that read better?

[gondalez]

@olleolleolle 👍 fixed :)

[brasmusson]

As I commented on the original PR, I think the exit code should be non-zero in --strict mode also when all scenarios are passed on rerun, that the summary should list the number of scenarios passed on rerun in the new flakey category, and that the exit code should be based on the summary (zero in --strict mode if all scenarios are passed or skipped, zero in non-strict mode if no scenarios are failed or ambiguous). I can't see that an implementation adding Configuration#total_cases is the implementation we want.

[gondalez]

@brasmusson That makes sense. I like the idea of a new category.
I'm happy to try my hand at an implementation over the coming weeks. Can you give me any helpful pointers?

[brasmusson]

Superseded by cucumber/cucumber-ruby-core#141 and #1159.

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.