Core specs behave randomly even with the same seed
Open
tests-infra
upstream fix needed
- Dominant language
- Java
- Stars
- 14.9k
- Forks
- 3.5k
- Avg merge
- 19h 14m
- Merged PRs (30d)
- 63
Description
Currently running `bin/rspec spec --seed=2992` multiple times will yield:
```
1077 examples, 0 failures
1327 examples, 0 failures
1150 examples, 0 failures
908 examples, 0 failures
```
This is happens in [unicode_trimmer_spec](https://github.com/elastic/logstash/blob/master/spec/util/unicode_trimmer_spec.rb), because of the [randomized testing spec](https://github.com/elastic/logstash/blob/master/spec/util/unicode_trimmer_spec.rb#L46-L54) that uses `stress_it` from [ruby-flores](https://github.com/jordansissel/ruby-flores/blob/master/lib/flores/rspec/stress.rb#L101-L105)
It seems to be possible to define the number of iterations in Flores so se can seed it with the rspec seed value
Contributor guide
Assessment
This issue has not been assessed yet.